Terahertz spectroscopy detection method and system

Abstract

This paper discloses a terahertz spectral detection method and system. The terahertz spectral detection method includes: modulating an incident terahertz wave signal using a terahertz wave modulation device to generate multiple terahertz wave modulation signals; wherein any two terahertz wave modulation signals have completely different or partially identical spectra; detecting the intensity of the multiple terahertz wave modulation signals using a terahertz wave detector; inputting the spectral data of the multiple terahertz wave modulation signals and the detected intensity data into a spectral reconstruction model for spectral reconstruction calculation to obtain the reconstructed spectrum of the incident terahertz wave. The terahertz spectral detection method presented in this paper has a fast detection speed and a simple and small detection system structure.

Description

Technical Field

[0001] This application relates to the field of electromagnetic spectrum detection technology, and in particular to a terahertz spectrum detection method and system. Background Technology

[0002] Terahertz spectral detection technology aims to reveal the spectral information contained in broadband terahertz waves, and has important application value in fields such as hyperspectral imaging in the terahertz band, high-bandwidth communication, and non-destructive testing.

[0003] In related technologies, terahertz spectral detection is typically achieved based on time-domain or spatial-domain coherent detection methods. Time-domain coherent detection relies on femtosecond optical excitation to generate terahertz waves, which are then focused onto a photoconductive antenna or electro-optic crystal via a complex optical path system to achieve time-domain signal detection. The spectral information is then obtained through Fourier transform. This detection method usually requires the construction of optical delay lines and signal acquisition through scanning, limiting the spectral detection rate. Furthermore, the systems are often complex, bulky, and expensive, hindering their widespread application.

[0004] The spatial coherent detection method splits the light under test into two beams, transmits them separately in space, and then combines them to cause interference. The optical path length of one beam is adjusted while the interference signal is measured simultaneously, and the spectrum information is obtained by Fourier transform. This method also suffers from slow scanning measurement speed, system complexity, and large size. Summary of the Invention

[0005] This application provides a terahertz spectral detection method, including:

[0006] Multiple terahertz wave modulation signals are generated by modulating the incident terahertz wave signal using a terahertz wave modulation device; wherein any two terahertz wave modulation signals have completely different or partially identical spectra.

[0007] The intensity of the multiple terahertz wave modulation signals is detected by a terahertz wave detector;

[0008] The spectral data of the multiple terahertz wave modulation signals and the detected intensity data are input into the spectral reconstruction model to perform spectral reconstruction calculations, thereby obtaining the reconstructed spectrum of the incident terahertz wave.

[0009] This application provides a terahertz spectral detection system, including:

[0010] A terahertz wave modulation device is configured to modulate an incident terahertz wave signal to generate multiple terahertz wave modulation signals; wherein any two terahertz wave modulation signals have completely different or partially identical spectra.

[0011] A terahertz wave detector, configured to detect the intensity of the plurality of terahertz wave modulation signals;

[0012] The calculation module is configured to input the spectral data of the multiple terahertz wave modulation signals and the detected intensity data into the spectral reconstruction model to perform spectral reconstruction calculations and obtain the reconstructed spectrum of the incident terahertz wave.

[0013] The terahertz spectral detection method and system provided in this application utilize a terahertz wave modulation device to modulate an incident terahertz wave signal, generating multiple terahertz wave modulation signals. The spectra of any two terahertz wave modulation signals are either completely different or partially identical. The intensity of the multiple terahertz wave modulation signals is detected by a terahertz wave detector. The spectral data of the multiple terahertz wave modulation signals and the detected intensity data are input into a spectral reconstruction model for spectral reconstruction calculation to obtain the reconstructed spectrum of the incident terahertz wave. The terahertz spectral detection method provided in this application has a fast detection speed, a simple detection system structure, and a small size.

[0014] After reading and understanding the accompanying diagrams and detailed descriptions, the other aspects can be understood. Attached Figure Description

[0015] The accompanying drawings are used to provide an understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0016] Figure 1 This is a flowchart of a terahertz spectral detection method according to an embodiment of this application;

[0017] Figure 2 This is a structural diagram of a terahertz spectral detection system according to an embodiment of this application;

[0018] Figure 3 This is a structural diagram of a terahertz spectral detection system according to Example 1 of this application;

[0019] Figure 4 This is a schematic diagram of the artificial microstructure array and terahertz transmission spectrum of the terahertz wave control device in Example 1 of this application;

[0020] Figure 5 A flowchart of a method for terahertz wave detection using the terahertz spectral detection system of Example 1 of this application;

[0021] Figure 6 This is a structural diagram of a terahertz spectral detection system according to Example 2 of this application;

[0022] Figure 7 This is a schematic diagram comparing the original spectrum and the reconstructed spectrum of a terahertz wave to be measured, provided as an embodiment of this application. Detailed Implementation

[0023] This application describes several embodiments, but these descriptions are exemplary and not restrictive, and it will be apparent to those skilled in the art that many more embodiments and implementations are possible within the scope of the embodiments described herein. Although many possible combinations of features are shown in the drawings and discussed in the detailed description, many other combinations of the disclosed features are also possible. Unless specifically limited, any feature or element of any embodiment may be used in combination with, or may replace, any feature or element of any other embodiment.

[0024] This application includes and contemplates combinations of features and elements known to those skilled in the art. The embodiments, features, and elements disclosed in this application may also be combined with any conventional features or elements to form a unique inventive scheme as defined by the appended claims. Any feature or element of any embodiment may also be combined with features or elements from other inventive schemes to form another unique inventive scheme as defined by the appended claims. Therefore, it should be understood that any feature shown and / or discussed in this application may be implemented individually or in any suitable combination. Therefore, the embodiments are not limited except by the limitations imposed by the appended claims and their equivalents. Furthermore, various modifications and changes may be made within the scope of the appended claims.

[0025] This application provides a terahertz spectral detection method. For example... Figure 1 As shown, a terahertz spectral detection method includes:

[0026] Step S10: Modulate the incident terahertz wave signal using a terahertz wave modulation device to generate multiple terahertz wave modulation signals; wherein, the spectra of any two terahertz wave modulation signals are completely different or partially the same.

[0027] Step S20: Detect the intensity of the plurality of terahertz wave modulation signals using a terahertz wave detector;

[0028] Step S30: Input the spectral data of the multiple terahertz wave modulation signals and the detected intensity data into the spectral reconstruction model to perform spectral reconstruction calculations and obtain the reconstructed spectrum of the incident terahertz wave.

[0029] The terahertz spectral detection method provided in this application utilizes a terahertz wave modulation device to modulate an incident terahertz wave signal, generating multiple terahertz wave modulation signals. The spectra of any two terahertz wave modulation signals are either completely different or partially identical. The intensity of the multiple terahertz wave modulation signals is detected by a terahertz wave detector. The spectral data of the multiple terahertz wave modulation signals and the detected intensity data are input into a spectral reconstruction model for spectral reconstruction calculation, yielding the reconstructed spectrum of the incident terahertz wave. The terahertz spectral detection method provided in this application does not require strictly separating the incident terahertz wave to be measured into subdivided frequency bands. It only requires using a broadband detector to detect the intensity of multiple terahertz modulation waves carrying different spectral characteristics, and the spectral information can be calculated and reconstructed using the spectral reconstruction model. The spectral measurement speed of the terahertz spectral detection method provided in this application is mainly determined by the detector response speed, and it does not require the optical phase scanning device necessary in time-domain measurement methods, thus resulting in a fast detection speed.

[0030] In some exemplary embodiments, the terahertz wave modulation device has different spectral transfer functions at different locations in the spatial domain or at different times in the temporal domain.

[0031] In some exemplary embodiments, the terahertz wave control device includes an array type or a single-point type.

[0032] In some exemplary embodiments, the array-type terahertz wave modulation device includes multiple sets of artificial microstructure arrays, which are arranged in spatial regions, and any two sets of artificial microstructure arrays have different spectral transfer functions.

[0033] In some exemplary embodiments, the array-type terahertz wave manipulation device can be fabricated using high-transmittance dielectrics (such as high-resistivity silicon, sapphire, etc.) in the terahertz band through micro-nano fabrication processes.

[0034] In some exemplary embodiments, the shape of the microstructure unit includes any of the following: a circular hole, a cylinder, a square hole, and a square prism. In other embodiments, the shape of the microstructure unit may also be other shapes.

[0035] In some exemplary embodiments, the artificial microstructure array includes multiple microstructure units, the shape and size of which affect the spectral transfer function. By setting microstructure units of different shapes and / or sizes for different microstructure unit arrays, the spectra of terahertz wave modulation signals passing through different artificial microstructure arrays can be made different.

[0036] In some exemplary embodiments, the single-point terahertz wave modulation device can change the transmission or reflection spectrum through external modulation means; wherein, the external modulation means include: electric field, light field, heat, etc. The single-point terahertz wave modulation device can be realized by two-dimensional materials such as graphene and molybdenum disulfide.

[0037] In some exemplary embodiments, before detecting the intensity of the plurality of terahertz wave modulated signals by a terahertz wave detector, the method further includes:

[0038] The terahertz wave modulation signal is focused onto the terahertz wave detector through a terahertz wave coupling device.

[0039] In some exemplary embodiments, the terahertz wave coupling device includes any one or more combinations of the following: a terahertz lens, a parabolic mirror, and a silicon lens. In other embodiments, the terahertz wave coupling device may also be other devices with terahertz wave focusing capabilities.

[0040] In some exemplary embodiments, the terahertz wave detector includes either an array type or a single-point type.

[0041] In some exemplary embodiments, the array-type terahertz wave detector includes multiple terahertz wave detection units, which are arranged in spatial regions.

[0042] In some exemplary embodiments, when the terahertz wave modulation device is an array type and the terahertz wave detector is an array type, the artificial microstructure array of the array type terahertz wave modulation device and the terahertz wave detection unit of the array type terahertz wave detector are arranged in a one-to-one correspondence. Multiple sets of artificial microstructure arrays of the array type terahertz wave modulation device simultaneously transmit terahertz wave modulation signals, and multiple terahertz wave detection units of the array type terahertz wave detector simultaneously receive terahertz wave modulation signals.

[0043] In some exemplary embodiments, when the terahertz wave modulation device is a single-point type and the terahertz wave detector is a single-point type, the single-point terahertz wave modulation device sequentially sends multiple terahertz wave modulation signals, and the single-point terahertz wave detector sequentially receives multiple terahertz wave modulation signals.

[0044] In some exemplary embodiments, the spectral reconstruction model is constructed based on machine learning methods, and the spectral reconstruction model is I = H * S;

[0045] I is the intensity of the terahertz wave signal, H is the spectral transfer function, and S is the spectrum of the terahertz wave signal; where I, H, and S are all wavelength-dependent.

[0046] In some exemplary embodiments, the spectral reconstruction model is built based on a deep learning method, and the spectral reconstruction model includes a neural network.

[0047] In some exemplary embodiments, the spectral transfer function is obtained by machine learning of the response of a terahertz wave signal with a known spectrum.

[0048] In some exemplary embodiments, the neural network parameters of the spectral reconstruction model are obtained by deep learning of the response of a terahertz wave signal with a known spectrum.

[0049] In some exemplary embodiments, obtaining the spectral transfer function by performing machine learning on the response of a terahertz wave signal with a known spectrum includes:

[0050] The training terahertz wave signal is acquired multiple times. After each acquisition, the incident training terahertz wave signal is modulated using a terahertz wave modulation device to generate multiple training terahertz wave modulation signals. The spectra of any two training terahertz wave modulation signals are completely different or partially the same, and the spectra of any two acquired training terahertz wave signals are completely different or partially the same. The intensity of the multiple training terahertz wave modulation signals is detected by a terahertz wave detector.

[0051] Using the spectral data of all incident training terahertz wave signals and the intensity data of all training terahertz wave modulated signals detected by the terahertz wave detector as the training set, a machine learning process is performed to obtain the spectral transfer function.

[0052] In some exemplary embodiments, the step of obtaining the neural network parameters of the spectral reconstruction model by deep learning the response of a terahertz wave signal with a known spectrum includes:

[0053] The training terahertz wave signal is acquired multiple times. After each acquisition, the incident training terahertz wave signal is modulated using a terahertz wave modulation device to generate multiple training terahertz wave modulation signals. The spectra of any two training terahertz wave modulation signals are completely different or partially the same, and the spectra of any two acquired training terahertz wave signals are completely different or partially the same. The intensity of the multiple training terahertz wave modulation signals is detected by a terahertz wave detector.

[0054] Using the spectral data of all incident training terahertz wave signals and the intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as the training set, a deep learning process is performed to obtain the neural network parameters of the spectral reconstruction model.

[0055] In some exemplary embodiments, the training terahertz wave signal is acquired multiple times, including:

[0056] During the i-th acquisition, a raw terahertz wave signal is generated using a broadband terahertz light source. This raw terahertz wave signal is then filtered using a terahertz band spectral filter to obtain the terahertz wave signal for the i-th training iteration. Different terahertz band spectral filters have different spectral transfer functions. By combining a broadband terahertz light source with different filters, a large sample set of terahertz signals with different spectra can be established, and the output spectra of the filters can be measured in advance.

[0057] In some exemplary embodiments, the broadband terahertz light source includes: a terahertz photoconductive antenna, or a continuously tunable terahertz quantum cascade laser, etc.

[0058] In some exemplary embodiments, the terahertz band spectral filter is a terahertz band spectral filter based on artificial microstructures.

[0059] In some exemplary embodiments, the terahertz band spectral filter is a terahertz band spectral filter based on artificial microstructures, fabricated using micro-nano fabrication processes. The terahertz band spectral filter is either an artificial metasurface device based on all-dielectric materials or a metamaterial device based on metallic microstructures.

[0060] In some exemplary embodiments, the terahertz band spectral filter has significant frequency-dependent transmission characteristics, enabling the transmission of spectra within a specific frequency band.

[0061] This application provides a terahertz spectral detection system. For example... Figure 2 As shown, a terahertz spectral detection system includes:

[0062] The terahertz wave modulation device 10 is configured to modulate the incident terahertz wave signal to generate multiple terahertz wave modulation signals; wherein any two terahertz wave modulation signals have completely different or partially identical spectra.

[0063] Terahertz wave detector 20 is configured to detect the intensity of the plurality of terahertz wave modulation signals;

[0064] The calculation module 30 is configured to input the spectral data of the multiple terahertz wave modulation signals and the detected intensity data into the spectral reconstruction model to perform spectral reconstruction calculations and obtain the reconstructed spectrum of the incident terahertz wave.

[0065] The terahertz spectral detection system provided in this application includes a terahertz wave modulation device, a terahertz wave detector, and a computing module. The terahertz wave modulation device modulates the incident terahertz wave signal to generate multiple terahertz wave modulation signals; wherein, the spectra of any two terahertz wave modulation signals are completely different or partially the same; the terahertz wave detector detects the intensity of the multiple terahertz wave modulation signals; the computing module inputs the spectral data of the multiple terahertz wave modulation signals and the detected intensity data into a spectral reconstruction model to perform spectral reconstruction calculations, obtaining the reconstructed spectrum of the incident terahertz wave. The terahertz spectral detection system provided in this application can include only a terahertz wave modulation device, a terahertz wave detector, and a computing module. It has a simple and compact structure and can be used to manufacture small-volume, portable terahertz spectrometers with real-time detection capabilities, making it highly practical. The spectral measurement speed is mainly determined by the detector response speed, and it does not require the optical phase scanning device necessary in time-domain measurement methods, thus resulting in a fast detection speed.

[0066] In some exemplary embodiments, the terahertz wave modulation device has different spectral transfer functions at different locations in the spatial domain or at different times in the temporal domain.

[0067] In some exemplary embodiments, the terahertz wave control device includes an array type or a single-point type.

[0068] In some exemplary embodiments, the array-type terahertz wave modulation device includes multiple sets of artificial microstructure arrays, which are arranged in spatial regions, and any two sets of artificial microstructure arrays have different spectral transfer functions.

[0069] In some exemplary embodiments, the array-type terahertz wave manipulation device can be fabricated using high-transmittance dielectrics (such as high-resistivity silicon, sapphire, etc.) in the terahertz band through micro-nano fabrication processes.

[0070] In some exemplary embodiments, the shape of the microstructure unit includes any of the following: a circular hole, a cylinder, a square hole, and a square prism. In other embodiments, the shape of the microstructure unit may also be other shapes.

[0071] In some exemplary embodiments, the artificial microstructure array includes multiple microstructure units, the shape and size of which affect the spectral transfer function. By setting microstructure units of different shapes and / or sizes for different microstructure unit arrays, the spectra of terahertz wave modulation signals passing through different artificial microstructure arrays can be made different.

[0072] In some exemplary embodiments, the single-point terahertz wave modulation device can change the transmission or reflection spectrum through external modulation means; wherein, the external modulation means include: electric field, light field, heat, etc. The single-point terahertz wave modulation device can be realized by two-dimensional materials such as graphene and molybdenum disulfide.

[0073] In some exemplary embodiments, the terahertz spectral detection system further includes a terahertz wave coupling device;

[0074] The terahertz wave coupling device is configured to converge the terahertz wave modulation signal onto the terahertz wave detector.

[0075] In some exemplary embodiments, the terahertz wave coupling device includes any one or more combinations of the following: a terahertz lens, a parabolic mirror, and a silicon lens. In other embodiments, the terahertz wave coupling device may also be other devices with terahertz wave focusing capabilities.

[0076] In some exemplary embodiments, the terahertz wave detector includes either an array type or a single-point type.

[0077] In some exemplary embodiments, the array-type terahertz wave detector includes multiple terahertz wave detection units, which are arranged in spatial regions.

[0078] In some exemplary embodiments, when the terahertz wave modulation device is an array type and the terahertz wave detector is an array type, the artificial microstructure array of the array type terahertz wave modulation device and the terahertz wave detection unit of the array type terahertz wave detector are arranged in a one-to-one correspondence. Multiple sets of artificial microstructure arrays of the array type terahertz wave modulation device simultaneously transmit terahertz wave modulation signals, and multiple terahertz wave detection units of the array type terahertz wave detector simultaneously receive terahertz wave modulation signals.

[0079] In some exemplary embodiments, when the terahertz wave modulation device is a single-point type and the terahertz wave detector is a single-point type, the single-point terahertz wave modulation device sequentially sends multiple terahertz wave modulation signals, and the single-point terahertz wave detector sequentially receives multiple terahertz wave modulation signals.

[0080] In some exemplary embodiments, the spectral reconstruction model is constructed based on machine learning methods, and the spectral reconstruction model is I = H * S;

[0081] I is the intensity of the terahertz wave signal, H is the spectral transfer function, and S is the spectrum of the terahertz wave signal; where I, H, and S are all wavelength-dependent.

[0082] In some exemplary embodiments, the spectral reconstruction model is built based on a deep learning method, and the spectral reconstruction model includes a neural network.

[0083] In some exemplary embodiments, the spectral transfer function is obtained by machine learning of the response of a terahertz wave signal with a known spectrum.

[0084] In some exemplary embodiments, the neural network parameters of the spectral reconstruction model are obtained by deep learning of the response of a terahertz wave signal with a known spectrum.

[0085] In some exemplary embodiments, the terahertz wave modulation device is further configured to acquire training terahertz wave signals multiple times. After each acquisition of a training terahertz wave signal, the incident training terahertz wave signal is modulated to generate multiple training terahertz wave modulation signals. The spectra of any two training terahertz wave modulation signals are completely different or partially the same, and the spectra of any two acquired training terahertz wave signals are completely different or partially the same.

[0086] The terahertz wave detector is also configured to detect the intensity of the plurality of training terahertz wave modulation signals;

[0087] The computing module is further configured to use the spectral data of all incident training terahertz wave signals and the intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as a training set to perform a machine learning process to obtain the spectral transfer function.

[0088] In some exemplary embodiments, the terahertz wave modulation device is further configured to acquire training terahertz wave signals multiple times. After each acquisition of a training terahertz wave signal, the incident training terahertz wave signal is modulated to generate multiple training terahertz wave modulation signals. The spectra of any two training terahertz wave modulation signals are completely different or partially the same, and the spectra of any two acquired training terahertz wave signals are completely different or partially the same.

[0089] The terahertz wave detector is also configured to detect the intensity of the plurality of training terahertz wave modulation signals;

[0090] The computing module is further configured to use the spectral data of all incident training terahertz wave signals and the intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as a training set to perform a deep learning process to obtain the neural network parameters of the spectral reconstruction model.

[0091] In some exemplary embodiments, the terahertz spectral detection system further includes: a broadband terahertz light source and multiple terahertz band spectral filters;

[0092] Broadband terahertz light, configured to generate raw terahertz wave signals;

[0093] The i-th terahertz band spectral filter is configured to filter the original terahertz wave signal to obtain the i-th training terahertz wave signal.

[0094] Different terahertz band spectral filters have different spectral transfer functions.

[0095] In some exemplary embodiments, the broadband terahertz light source includes: a terahertz photoconductive antenna, or a continuously tunable terahertz quantum cascade laser, etc.

[0096] In some exemplary embodiments, the terahertz band spectral filter is a terahertz band spectral filter based on artificial microstructures.

[0097] In some exemplary embodiments, the terahertz band spectral filter is a terahertz band spectral filter based on artificial microstructures, fabricated using micro-nano fabrication processes. The terahertz band spectral filter is either an artificial metasurface device based on all-dielectric materials or a metamaterial device based on metallic microstructures.

[0098] In some exemplary embodiments, the terahertz band spectral filter has significant frequency-dependent transmission characteristics, enabling the transmission of spectra within a specific frequency band.

[0099] The following example illustrates how a terahertz spectral detection system works.

[0100] Example 1

[0101] like Figure 3 As shown, this application provides a terahertz spectral detection system. Figure 3 As shown, a terahertz spectral detection system includes: a terahertz light source 1, a terahertz wave modulation device 2, a terahertz wave coupling device 3, a terahertz wave detector 4, and a data processing device 5.

[0102] Terahertz light sources are the starting components of the terahertz optical path of interest in terahertz spectral detection systems. In the machine learning stage, they can be light sources that directly emit terahertz waves, while in the stage of detecting unknown spectra, they can be light sources to be tested that have terahertz waves passing through them and whose spectral information is unknown.

[0103] The terahertz wave modulation device is placed behind the terahertz light source in the optical path to modulate the passing terahertz waves. The terahertz wave modulation device can be fabricated using high-transmittance dielectric materials in the terahertz band, such as high-resistivity silicon or sapphire, through micro-nano fabrication processes. In this example, the terahertz wave modulation device comprises a 6×6=36 artificial microstructure array, numbered m (m=1, 2, 3…36).

[0104] Figure 4 The diagram shows a partial array of artificial microstructures, where array 1 is the first numbered artificial microstructure array, array 2 is the second numbered artificial microstructure array, array 3 is the third numbered artificial microstructure array, and so on. Each artificial microstructure array consists of 25 microstructure units, and the shape of the microstructure units can be circular holes, cylinders, square holes, square pillars, etc.

[0105] Figure 4 The image also shows the terahertz wave transmission spectra corresponding to the artificial microstructure arrays numbered 1-3. The shapes and / or sizes of the microstructure units in the different artificial microstructure arrays vary, therefore their terahertz transmission spectra are also different. m (λ), the spectral transfer function is a function of wavelength λ.

[0106] The terahertz wave coupling device is located behind the terahertz wave manipulation device in the optical path, and it focuses the terahertz waves passing through the terahertz wave coupling device onto the terahertz wave detector. In this example, a terahertz focusing lens is used as the terahertz wave coupling device.

[0107] The terahertz wave detector is located behind the terahertz wave coupling device in the optical path and is used to detect the intensity of terahertz waves. In this example, the terahertz wave detector is an area array detector, whose detection surface can be spatially divided into 6×6=36 regions. Each region can independently output a light intensity signal, denoted as Im (m=1, 2, 3…36). Furthermore, the detection regions on the terahertz wave detector correspond one-to-one with the artificial microstructure array of the terahertz wave control device. That is, the terahertz wave controlled by artificial microstructure array 1 on the terahertz wave control device will illuminate detection region 1 on the terahertz wave detector, and the measured light intensity is I1; the terahertz wave controlled by artificial microstructure array 2 on the terahertz wave control device will illuminate detection region 2 on the terahertz wave detector, and the measured light intensity is I2; and so on.

[0108] The data processing device is electrically connected to the terahertz wave detector, reads the light intensity information output by the terahertz wave detector, performs terahertz spectrum reconstruction calculations, and outputs the reconstructed terahertz spectrum.

[0109] Figure 5 The diagram shows a flowchart of a method for terahertz wave detection using the terahertz spectral detection system provided in Example 1 of this invention. Figure 5 As shown, the terahertz spectral detection method may include the following steps:

[0110] Step 101: Fabricate a terahertz wave control device using micro-nano fabrication technology. The device contains an array of 36 artificial microstructures, numbered m (m = 1, 2, 3...36).

[0111] Each array consists of 25 periodically arranged subwavelength microstructure units. These microstructure units are different from each other in shape and / or size. Therefore, the terahertz transmission spectrum of each microstructure array is also different.

[0112] Step 102: Prepare N types of terahertz band spectral filters based on artificial microstructures using micro-nano fabrication technology, numbered as n (n = 1, 2, 3... N).

[0113] These filters can be artificial metasurface devices based on all-dielectric materials or metamaterial devices based on metallic microstructures; and for these filters, it is not necessary to focus on their blocking effect on terahertz waves of a specific frequency, but only to make them have obvious frequency-dependent transmission characteristics.

[0114] Step 103: Measure the transmission spectra of these filters using a terahertz time-domain spectroscopy system, denoted as f. n .

[0115] Step 104: Emit terahertz waves using a broadband terahertz light source, and filter the terahertz waves one by one using terahertz band spectral filters. The filtered terahertz wave spectrum is then S. n =S0*f n Where S0 is the spectrum of the terahertz wave emitted by the broadband terahertz source;

[0116] Broadband terahertz light sources can be terahertz photoconductive antennas or continuously tunable terahertz quantum cascade lasers, etc.

[0117] Step 105: The filtered terahertz wave is transmitted into the array-type terahertz wave detector via the terahertz wave modulation device and the terahertz wave coupling device. The light intensity received by the 36 regions of the array-type terahertz wave detector corresponds one-to-one with the 36 artificial microstructure arrays on the terahertz wave modulation device. When using the nth filter, the light intensity detected in the mth detection region is recorded as I. mn ;

[0118] It should be noted that if the light intensity on the plane of the terahertz wave detector exhibits a background distribution pattern (such as a Gaussian curve), then I here... mn This refers to the result after removing the background distribution through normalization.

[0119] Step 106: Design a spectral reconstruction model, using each set of spectral signals (S) n ) and intensity signal (I mn Using this as the training set, the spectral transfer matrix H is learned. m (Spectral transfer function);

[0120] I mn=H m *S n

[0121] Steps 102 to 106 describe the process of performing machine learning, and the final result of the learning is to establish a spectral reconstruction model.

[0122] Step 107: The terahertz spectrum to be measured is denoted as X(λ). It is then passed through the terahertz wave modulation device and the terahertz wave coupling device and incident on the terahertz detector. The light intensity received by different regions of the terahertz wave detector corresponds one-to-one with the different artificial microstructure arrays on the terahertz wave modulation device, denoted as I. m .

[0123] Step 108: Using the trained spectral reconstruction model, based on I m The reconstructed terahertz spectrum X'(λ) is obtained: I m =H m *X′(λ).

[0124] In the above formula, f, S, H, X, etc. are all functions of the terahertz wavelength.

[0125] Example 2

[0126] like Figure 6 As shown, this application provides a terahertz spectral detection system. Figure 6 As shown, a terahertz spectral detection system includes: a terahertz light source 1, a terahertz wave modulation device 2, a terahertz wave coupling device 3, a terahertz wave detector 4, and a data processing device 5.

[0127] Terahertz light sources are the starting components of the terahertz optical path of interest in terahertz spectral detection systems. In the machine learning stage, they can be light sources that directly emit terahertz waves, while in the stage of detecting unknown spectra, they can be light sources to be tested that have terahertz waves passing through them and whose spectral information is unknown.

[0128] The terahertz wave modulation device is an electrically controlled, adjustable terahertz wave modulator (belonging to the category of single-point modulators). By changing the applied voltage, it can switch between M different operating states, and in each operating state, the terahertz wave modulator's transfer function exhibits different spectral characteristics. Terahertz wave modulators with this property can be realized using two-dimensional materials such as graphene and molybdenum disulfide. The spectral transfer function H of the terahertz wave modulation device varies under different operating states. m (λ) are different.

[0129] The terahertz wave coupling device is located behind the terahertz wave control device in the optical path, and it focuses the terahertz waves that have passed through the terahertz wave coupling device onto the terahertz wave detector.

[0130] The terahertz wave detector is located behind the terahertz wave coupling device in the optical path and is used to detect the intensity of terahertz waves. In this example, the terahertz wave detector is a single-point detector. By continuously adjusting the operating state of the terahertz wave modulator, a set of corresponding light intensity values ​​can be detected on the terahertz wave detector, denoted as I. m .

[0131] The data processing device is electrically connected to the terahertz wave detector, reads the light intensity information output by the terahertz wave detector, performs terahertz spectrum reconstruction calculations, and outputs the reconstructed terahertz spectrum.

[0132] Figure 7 This is a schematic diagram comparing the original spectrum and the reconstructed spectrum of a terahertz wave to be measured, provided in an embodiment of this application. The left image shows the original spectrum of the terahertz wave to be measured. For comparison purposes, the original spectrum can be obtained by using a terahertz time-domain spectroscopy method. The right image shows the terahertz spectrum obtained by measurement and reconstruction using the terahertz spectral detection method provided in this embodiment of the application.

[0133] The method for terahertz wave detection using the terahertz spectroscopy detection system in this example may include the following steps:

[0134] The terahertz wave to be measured, after being modulated by a terahertz wave modulation device, is focused onto a terahertz wave detector by a terahertz wave coupling device. Depending on the state of the terahertz wave modulator, the terahertz wave detector detects and records the corresponding light intensity value. Through external electrical control, the terahertz wave modulator can switch between M different operating states, each corresponding to a unique transmission spectrum. By continuously adjusting the operating state of the terahertz wave modulator, the terahertz wave detector can detect and record a set of corresponding light intensity values, denoted as I. m By using a pre-established spectral reconstruction model, it is possible to obtain from I m The reconstructed terahertz spectrum was calculated.

[0135] In this example, the spectral reconstruction model can be obtained through machine learning. By combining broadband terahertz sources with different filters, a large sample set of terahertz sources with different spectra can be established, and their output spectra can be measured in advance. Under illumination from terahertz sources with different spectra, the M operating states of the terahertz wave modulator are traversed, and the output data of the corresponding terahertz wave detector are recorded to form a training set. Through training on a large amount of training data, the spectral transfer function of the optimized spectral reconstruction model is obtained.

[0136] Those skilled in the art will understand that the functional modules / units in the apparatus disclosed above can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or a microprocessor, or as hardware, or as integrated circuits, such as application-specific integrated circuits (ASICs).

Claims

1. A terahertz spectral detection method, comprising: Multiple terahertz wave modulation signals are generated by modulating the incident terahertz wave signal using a terahertz wave modulation device; wherein any two terahertz wave modulation signals have completely different or partially identical spectra. The intensity of the multiple terahertz wave modulation signals is detected by a terahertz wave detector; The spectral data of the multiple terahertz wave modulation signals and the detected intensity data are input into the trained spectral reconstruction model to perform spectral reconstruction calculations and obtain the reconstructed spectrum of the incident terahertz wave. The spectral reconstruction model is constructed based on machine learning or deep learning methods. The trained spectral reconstruction model is obtained by performing a machine learning or deep learning process on the spectral reconstruction model using a large amount of incident spectral data of training terahertz wave signals and intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as a training set.

2. The method as described in claim 1, characterized in that: The spectral reconstruction model is constructed based on machine learning methods, and the spectral reconstruction model is I = H * S; I is the intensity of the terahertz wave signal, H is the spectral transfer function, and S is the spectrum of the terahertz wave signal; where I, H, and S are all wavelength-dependent.

3. The method as described in claim 2, characterized in that: The spectral transfer function of the trained spectral reconstruction model is obtained by machine learning from the response of a terahertz wave signal with a known spectrum.

4. The method as described in claim 3, characterized in that: The step of obtaining the spectral transfer function of the trained spectral reconstruction model by performing machine learning on the response of a terahertz wave signal with a known spectrum includes: The training terahertz wave signal is acquired multiple times. After each acquisition, the incident training terahertz wave signal is modulated using a terahertz wave modulation device to generate multiple training terahertz wave modulation signals. The spectra of any two training terahertz wave modulation signals are completely different or partially the same, and the spectra of any two acquired training terahertz wave signals are completely different or partially the same. The intensity of the multiple training terahertz wave modulation signals is detected by a terahertz wave detector. Using the spectral data of all incident training terahertz wave signals and the intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as the training set, a machine learning process is performed to obtain the spectral transfer function of the trained spectral reconstruction model.

5. The method as described in claim 1, characterized in that: The spectral reconstruction model is built based on deep learning methods and includes a neural network. The neural network parameters of the trained spectral reconstruction model are obtained in the following way: The training terahertz wave signal is acquired multiple times. After each acquisition, the incident training terahertz wave signal is modulated using a terahertz wave modulation device to generate multiple training terahertz wave modulation signals. The spectra of any two training terahertz wave modulation signals are completely different or partially the same, and the spectra of any two acquired training terahertz wave signals are completely different or partially the same. The intensity of the multiple training terahertz wave modulation signals is detected by a terahertz wave detector. Using the spectral data of all incident training terahertz wave signals and the intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as the training set, a deep learning process is performed to obtain the neural network parameters of the trained spectral reconstruction model.

6. The method as described in claim 1, characterized in that: The terahertz wave control device has different spectral transfer functions at different locations in the spatial domain or at different times in the temporal domain.

7. The method according to claim 6, wherein: The terahertz wave control device includes an array type or a single-point type; The array-type terahertz wave control device includes multiple sets of artificial microstructure arrays, which are arranged in spatial regions, and any two sets of artificial microstructure arrays have different spectral transfer functions. The single-point terahertz wave modulation device can change the transmission or reflection spectrum through external modulation means; wherein, the external modulation means include: electric field, light field or heat.

8. The method as described in claim 1, characterized in that: Before detecting the intensity of the plurality of terahertz wave modulated signals using a terahertz wave detector, the method further includes: The terahertz wave modulation signal is focused onto the terahertz wave detector through a terahertz wave coupling device.

9. The method as described in claim 7, characterized in that: When the terahertz wave control device is an array, the terahertz wave detector is also an array; the array-type terahertz wave detector includes multiple terahertz wave detection units, which are arranged in spatial regions. The artificial microstructure array of the array-type terahertz wave modulation device is configured in a one-to-one correspondence with the terahertz wave detection unit of the array-type terahertz wave detector. Multiple artificial microstructure arrays of the array-type terahertz wave modulation device simultaneously transmit terahertz wave modulation signals, and multiple terahertz wave detection units of the array-type terahertz wave detector simultaneously receive terahertz wave modulation signals.

10. The method as described in claim 7, characterized in that: When the terahertz wave control device is a single-point type, the terahertz wave detector is also a single-point type; The single-point terahertz wave modulation device sequentially sends multiple terahertz wave modulation signals, and the single-point terahertz wave detector sequentially receives multiple terahertz wave modulation signals.

11. A terahertz spectral detection system, comprising: A terahertz wave modulation device is configured to modulate an incident terahertz wave signal to generate multiple terahertz wave modulation signals; wherein any two terahertz wave modulation signals have completely different or partially identical spectra. A terahertz wave detector, configured to detect the intensity of the plurality of terahertz wave modulation signals; The calculation module is configured to input the spectral data of the multiple terahertz wave modulation signals and the detected intensity data into the trained spectral reconstruction model to perform spectral reconstruction calculations and obtain the reconstructed spectrum of the incident terahertz wave. The spectral reconstruction model is constructed based on machine learning or deep learning methods. The trained spectral reconstruction model is obtained by using a large number of incident spectral data of training terahertz wave signals and the intensity data of all training terahertz wave modulation signals detected by the terahertz wave detector as a training set, and performing a machine learning or deep learning process on the spectral reconstruction model.

12. The system as described in claim 11, characterized in that: The terahertz wave control device has different spectral transfer functions at different locations in the spatial domain or at different times in the temporal domain.

13. The system as described in claim 11, characterized in that: The terahertz wave control device includes an array type or a single-point type; The array-type terahertz wave control device includes multiple sets of artificial microstructure arrays, which are arranged in spatial regions, and any two sets of artificial microstructure arrays have different spectral transfer functions. The single-point terahertz wave modulation device can change the transmission or reflection spectrum through external modulation means; wherein, the external modulation means include: electric field, light field or heat.

14. The system as described in claim 13, characterized in that: When the terahertz wave control device is an array, the terahertz wave detector is also an array; the array-type terahertz wave detector includes multiple terahertz wave detection units, which are arranged in spatial regions. The artificial microstructure array of the array-type terahertz wave modulation device is configured in a one-to-one correspondence with the terahertz wave detection unit of the array-type terahertz wave detector. Multiple artificial microstructure arrays of the array-type terahertz wave modulation device simultaneously transmit terahertz wave modulation signals, and multiple terahertz wave detection units of the array-type terahertz wave detector simultaneously receive terahertz wave modulation signals.

15. The system as described in claim 13, characterized in that: When the terahertz wave control device is a single-point type, the terahertz wave detector is also a single-point type; The single-point terahertz wave modulation device sequentially sends multiple terahertz wave modulation signals, and the single-point terahertz wave detector sequentially receives multiple terahertz wave modulation signals.

Images