A spectrometer based on optical waveguide leaky mode spectroscopy and a spectral reconstruction method thereof

Abstract

The application discloses a spectrometer based on light waveguide leakage mode light splitting and a spectrum reconstruction method thereof. The spectrometer comprises an image sensor and at least one light waveguide, the surface of the image sensor is coated with a substrate layer, the first end of the light waveguide is connected with a light source as an entrance, and the second end of the light waveguide is adsorbed on the surface of the substrate layer after being heated and drawn. The spectrum reconstruction method comprises the following steps: calibration: the photocurrent value of the sensor unit at each wavelength is saved as a spectral response function; collecting photocurrent: collecting the photocurrent value of the to-be-detected light in any sensor unit; constructing an equation set and solving: obtaining k actual spectral functions of the to-be-detected light, solving the corresponding photocurrent equation and composing a linear equation set, and solving to obtain the reconstructed actual spectral function of the to-be-detected light. The application has the advantages of small physical volume, multi-channel detection and accurate detection result.

Description

Technical Field

[0001] This invention belongs to the field of spectrometer technology, and specifically relates to a spectrometer based on optical waveguide leakage mode spectroscopy and its spectral reconstruction method. Background Technology

[0002] Spectroscopy is one of the most powerful and widely used tools in scientific and industrial research. Currently, spectrometers capable of providing ultra-high resolution and wide spectral range often have bulky optics, moving parts, and long working distances. However, as the application areas of spectral analysis continue to expand, the need to reduce the physical size, cost, and power consumption of spectrometers is gradually increasing, even taking precedence over the need for high performance.

[0003] Currently, there are roughly four technologies for miniaturizing spectrometers.

[0004] The first technology involves the miniaturization of traditional spectrometer components. Most traditional spectrometers use diffraction gratings for spectral dispersion. Advances in micro and nanofabrication technologies have provided opportunities for shrinking traditional spectrometer components; currently, miniature spatial dispersive spectrometers developed using electrochemical etching techniques have emerged. However, as the size of spectrometer components shrinks, the light propagation distance also decreases. For a given grating and detector array, resolution is proportional to the system's optical path length; therefore, as the device becomes more compact, the spectral resolution decreases.

[0005] The second technique utilizes narrowband filter arrays for spectral dispersion, with each filter mounted on its own detector. Although the system size is still limited by the detector and filter sizes, compared to traditional spectrometers, micro-spectroscopy based on filter arrays is no longer constrained by path length. The advantages of narrowband filters are significant, but their fabrication process is complex, involving multiple deposition processes. This limits the number of channels in the filter array, thus limiting the spectrometer's resolution.

[0006] The third technology is the Fourier transform spectrometer. A Fourier transform spectrometer centers on an interferometer. The interference of light changes over time and is received by a detector. The time-varying data received by the detector is then subjected to a Fourier transform to reconstruct the spectrum of the light being measured. Compared to spectrophotometers, Fourier transform spectrometers require fewer detectors and yield data with a higher signal-to-noise ratio. Traditional Fourier transform spectrometers require moving plane mirrors to alter the interference of light. However, with technological advancements, Fourier transform spectrometers based on integrated optical waveguides and without moving mirrors have emerged. These spectrometers are based on Mach-Zehnder interferometers (MZIs) and utilize the phase differences of light in multiple channels to generate interference. While Fourier transform spectrometers require fewer detectors, their performance depends on the maximum optical path difference between different channels of the interferometer, which limits the extent to which the physical size of the spectrometer can be reduced.

[0007] The fourth technique is computational spectrum reconstruction. It relies on computation to reconstruct the spectrum of incident light from encoded pre-calibration information detected by the detector. Currently, a nanowire spectrometer has been developed, using a single nanowire as a spatial encoder. Incident light passes through the nanowire and illuminates an array of detectors. By utilizing the intensity information received from different detectors and pre-calibrated data, a system of linear equations is solved by a computer to reconstruct the incident light spectrum with relatively high accuracy. This spectrometer is one of the most successfully miniaturized spectrometers in the world, but it is also limited by the physical size of the detectors. Theoretically, it requires more detectors, i.e., more equations, to ensure the accuracy of the spectrum reconstruction. Summary of the Invention

[0008] The purpose of this invention is to disclose a spectrometer based on optical waveguide leaky mode spectroscopy and its spectral reconstruction method, so as to overcome the problem that the performance of existing spectrometers will be reduced when the physical scale is reduced.

[0009] The first objective of this invention is achieved by the following means:

[0010] A spectrometer based on optical waveguide leakage mode spectroscopy includes an image sensor and at least one optical waveguide. The surface of the image sensor is coated with a substrate layer. The first end of the optical waveguide serves as a light inlet connected to a light source. The second end of the optical waveguide is heated, drawn into wire, and then adsorbed onto the surface of the substrate layer.

[0011] The principle of this invention is as follows: Before measuring the spectrum of a specific incident light, the entire system is calibrated using monochromatic light sources of different wavelengths with known intensity to determine the response of each pixel on the image sensor to light of different wavelengths. During the actual measurement, the light to be measured is passed through an optical waveguide, guided to the substrate layer, and then introduced into the image sensor. Based on the pre-calibrated response function, a system of linear equations is solved to reconstruct the target spectrum.

[0012] Preferably, the refractive index of the substrate layer is close to the effective refractive index of light propagating in the optical waveguide. In an optical waveguide with a diameter close to or smaller than the wavelength of the light guide, a certain proportion of power is guided out of the waveguide as an evanescent wave. The longer the wavelength, the greater the evanescent power, and the lower the effective refractive index of the corresponding guided mode within the optical waveguide. In this case, if the refractive index of the substrate layer is close to the effective refractive index within the optical waveguide, light will leak into the substrate layer, thus filtering out longer wavelengths of light.

[0013] Preferably, the substrate layer is made by using magnesium fluoride or calcium fluoride. The substrate is deposited on the surface of the image sensor at a high speed, and the thickness of the substrate layer is 150nm-300nm.

[0014] Preferably, the first end of the optical waveguide is cut flat and then connected to the FC connector.

[0015] Preferably, the second end of the optical waveguide is drawn to the micrometer scale by flame heating, and the second end of the optical waveguide is fixed to the substrate by van der Waals forces and electrostatic attraction. The optical waveguide, drawn to the micrometer scale, is very flexible and small, while the substrate is relatively smooth; these two elements can be firmly bonded together under the action of van der Waals forces and electrostatic attraction.

[0016] Preferably, the image sensor is a CMOS image sensor, which undergoes the following treatment before the substrate is deposited: first, the protective glass of the CMOS image sensor chip is cut off and removed; then, it is immersed in acetone solution to wash away the glass debris attached to the surface; and finally, it is air-dried.

[0017] The second objective of this invention is achieved by the following means:

[0018] A spectral reconstruction method for a spectrometer based on optical waveguide leaky mode spectroscopy is characterized by defining k sensor units in the image sensor that can receive leaked light, including:

[0019] Calibration: K different wavelengths of light are introduced into the optical waveguide of the spectrometer based on optical waveguide leakage mode spectroscopy, and the photocurrent value of the sensor unit at each wavelength is saved as a spectral response function;

[0020] Photocurrent acquisition: The light to be measured is introduced into the image sensor through different optical waveguides to obtain the photocurrent value in any sensor unit;

[0021] Construct and solve a system of equations: Obtain the actual spectral function of the light to be measured by using the photocurrent value and the spectral response function obtained by calibration. Combine the actual spectral function with the Gaussian function of the corresponding sensor unit to obtain the photocurrent equation of a single sensor unit. Then, combine the photocurrent equations of k sensor units into a linear system of equations and solve to obtain the reconstructed actual spectral function of the light to be measured.

[0022] Preferably, in the calibration step, the optical power output of the adjustable light source is confirmed with an optical power meter before the light is introduced, and the output light intensity is adjusted to ensure that the power is approximately the same at each wavelength.

[0023] Preferably, in the steps of constructing and solving the system of equations, a loss function is also introduced, and a linear regression algorithm is used to gradually minimize the loss function during the solution process.

[0024] Preferably, an additional regularization coefficient is introduced into the loss function, and the optimal solution is obtained through repeated iterations during the solution process to improve the accuracy of the results.

[0025] Compared with existing technologies, this invention has the following advantages: It utilizes optical waveguides for beam splitting, breaking away from the traditional spectrometer's grating-based method, and is not constrained by optical path length, resulting in a more compact structure and significantly reduced spatial volume. Each part of the optical waveguide can be considered a bandpass filter, similar to the principle of a spectrometer using a narrowband filter array. However, compared to a spectrometer using a narrowband filter array, this invention only requires thinning the ends of the optical waveguide to a certain extent, greatly reducing manufacturing difficulty and cost. In ordinary filter arrays, each filter only utilizes a single wavelength of the incident light, while this invention fully utilizes various components of the incident light from infrared to visible light, improving the utilization efficiency of the incident light. Where computing power allows, this invention can achieve multi-channel processing and can calculate and reconstruct the spectra of multiple incident beams. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the spectrometer based on optical waveguide leakage mode spectroscopy in the embodiment.

[0027] Figure 2 for Figure 1 Enlarged view of point A in the image.

[0028] Figure 3 This is a schematic diagram of the multi-channel processing state of a spectrometer based on optical waveguide leaky mode spectroscopy, as an example.

[0029] In the figure, 1 is the image sensor; 2 is the optical waveguide; 2a is the first end; 2b is the second end; 3 is the substrate; and 4 is the FC connector. Detailed Implementation

[0030] The technical solution of the present invention will be further described in detail below with reference to specific embodiments. It should be understood that the following embodiments are merely illustrative and explanatory of the present invention, and should not be construed as limiting the scope of protection of the present invention. All technologies implemented based on the above content of the present invention are covered within the scope of protection intended by the present invention.

[0031] Figure 1 The overall structure of an embodiment of the present invention is shown. As can be seen from the figure, the spectrometer based on optical waveguide leaky mode spectroscopy of the present invention mainly consists of an image sensor 1, an optical waveguide 2, a substrate layer 3, and an FC connector 4. The optical waveguide 2 is made of optical fiber, and the image sensor 1 is a CMOS image sensor 1.

[0032] Specifically, the CMOS image sensor 1 first undergoes the following processing: A hot air gun is set to 350°C, and air is blown onto its pins until the solder melts. The chip is then removed with tweezers. After removing the protective glass using a diamond wire cutter, the chip is immersed in an acetone solution to wash away any glass debris adhering to its surface, and then air-dried. Since acetone is used to clean the glass debris because it is non-conductive and does not react with the gold wires on the chip, the purer the acetone concentration, the better. In this example, the acetone concentration should exceed 99.5%.

[0033] Combination Figure 2 The surface of the air-dried CMOS image sensor 1 is then coated with... A layer of magnesium fluoride or calcium fluoride with a thickness of approximately 200 nm is deposited as a substrate layer at a specific evaporation rate. It should be noted that the evaporation rate is not limited to a specific speed. It can also be in The thickness of the substrate layer 3 is not limited to 200nm; it can be selected within the range of 150nm-300nm as needed. Since the effective refractive index of light propagating in the optical waveguide 2 is similar to that of the substrate, it will leak into the substrate. Therefore, any colorless and transparent material with a refractive index similar to the effective refractive index of the light wave can be used as the substrate of this patent. In addition to magnesium fluoride, calcium fluoride and other materials can also be used. The use of colorless and transparent materials is to reduce the influence of the color of the crystal itself.

[0034] The coating layers at both ends of the optical fiber are removed using wire strippers. The first end, 2a, is cut flat or ground smooth before being connected to FC connector 4 to reduce input loss. FC connector 4 is a commonly used end-face fiber optic connector used for repeatedly connecting and disconnecting optical fibers. Making the other end of the fiber into FC connector 4 facilitates the coupling of the light under test and can also be used to create other types of interfaces. The second end, 2b, is softened by flame heating and then drawn to the micrometer level, specifically about 1 micrometer. It is then placed on the surface of a CMOS image sensor array and connected using van der Waals forces and electrostatic attraction. The connection is generally stable without strong external forces. Since the second end, 2b, only needs to be drawn to about 1 micrometer, there are no strict limitations on the flame temperature and heating time. The heating source is a blowtorch fueled by butane, with a temperature of approximately 800℃. Heating for a few seconds softens the fiber, after which it is drawn apart. According to experimental results, the spectrometer in the example has a good response in the visible light range of 550nm-780nm.

[0035] With the above structure, the second end 2b of the optical fiber is very small compared to the entire image sensor 1. This means that when light passes through a single optical fiber, a large portion of the sensor units on the image sensor 1 do not participate in the spectral reconstruction of the incident light in that fiber. Therefore, these idle sensor units can be used for the spectral reconstruction of another beam of light, thereby achieving multi-channel processing. Figure 3As shown, four optical fibers are mounted on the image sensor 1 array, and they are all far apart from each other, so they will not have a significant impact on their respective spectral reconstruction processes.

[0036] A multi-channel fiber optic computational spectrometer performs the same calibration and reconstruction process for each channel as a single-channel fiber optic computational spectrometer, and, given sufficient computing power, can simultaneously calculate the spectra of multiple beams. Furthermore, in... Figure 3 In the four-channel system shown, a large area of ​​the image sensor 1 is still unused. This means that as long as an appropriate FC interface is selected and there is no conflict at the fiber optic front end, the system can accommodate more channels and process information from more optical paths simultaneously.

[0037] The usage process of this invention and the principle of reconstructing the spectrum of the light under test are as follows:

[0038] This example uses an adjustable light source. If there are k sensor units capable of receiving leaked light, then k different wavelengths of light need to be introduced during calibration. The specific wavelengths are...

[0039]

[0040] In the above formula:

[0041] λ m The m-th wavelength that needs to be calibrated;

[0042] λ max The maximum wavelength that the spectrometer can respond to;

[0043] λ min The smallest wavelength that the spectrometer can respond to.

[0044] Before introducing the light, use an optical power meter to confirm the output optical power of the adjustable light source. Adjust the output light intensity to ensure that the power is approximately the same at each wavelength. Then, introduce monochromatic light of different wavelengths from the first end 2a of the optical fiber and save the photocurrent value of the sensor unit at each wavelength as the spectral response function R(λ).

[0045] The process of spectral reconstruction is as follows:

[0046] Let the spectral function of the light to be measured be S(λ). For any sensor unit that can receive the leaked light, its photocurrent value I can be expressed as:

[0047]

[0048] In the above formula:

[0049] I j : The photocurrent of the j-th sensor unit;

[0050] S(λ): Spectral function of the light to be measured;

[0051] R j (λ): The spectral response function obtained from calibration, where the subscript j indicates that it is the spectral response of the j-th sensor.

[0052] The actual spectral function can be approximated as a Gaussian function with the same number of sensor units. The combination is:

[0053]

[0054] In the above formula:

[0055] α i :for The coefficient;

[0056] With λ i The Gaussian function with the center wavelength (i-th wavelength, calibrated) is expressed as follows: σ is the variance of the Gaussian function, which is related to the full width at half maximum (FWHM) of the basis functions.

[0057] The specific full width at half maximum (FWHM) (variance σ) of the basis functions is determined by the actual number of sensor units and the wavelength range. After converting the spectral function into a combination of basis functions based on the above formula, the photocurrent equation for a single sensor becomes:

[0058]

[0059] The equations for k different sensor units can form a system of linear equations. The problem of spectral reconstruction is transformed into solving a system of linear equations. The matrix form of the transformed system of linear equations is as follows:

[0060] Aα=c (4)

[0061] In the above formula:

[0062] A: A K*k matrix, where each element is an integral of the Gaussian function and the spectral response function. For example:

[0063] α: Coefficient matrix, specifically in the form [α1, α2, ..., α k ] T ;

[0064] c: Photocurrent matrix, specifically in the form [I1,I2,…,I…] k ] T

[0065] The loss function for this linear equation is:

[0066]

[0067] That is, the square of the Euclidean norm of matrix Aα-c.

[0068] The specific solution process is to use the linear regression algorithm to gradually change the values ​​of each element of α to minimize the loss function. The minimized α is the coefficient of the Gaussian function in the spectrum to be measured. Substituting it into equation (2) will reconstruct the spectrum to be measured.

[0069] Due to the influence of high-frequency noise, the result obtained by minimizing the loss function in equation (5) may be inaccurate. Therefore, an additional regularization coefficient γ needs to be introduced into the loss function, and the loss function becomes:

[0070]

[0071] The solution process becomes: first, fix the regularization coefficient γ, and then solve for the coefficient matrix α that minimizes the loss function. γ Then, the value of the regularization coefficient γ is updated using this matrix, and the iteration is repeated to obtain the optimal solution.

[0072] The regularization coefficient γ is updated using generalized cross-validation (GCV), and its loss function is:

[0073]

[0074] In the above formula:

[0075] I: A matrix whose diagonal elements are all 1s and the rest are 0s;

[0076] The definitions of the remaining parameters are the same as those in the aforementioned formula.

[0077] The regularization coefficient can be obtained by minimizing the loss function using univariate linear regression analysis.

[0078] The above reconstruction method can accurately obtain the spectrum of the light to be measured, has good applicability, and is suitable for widespread use.

[0079] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A spectrometer based on optical waveguide leakage mode spectroscopy, characterized in that, The device includes an image sensor (1) and at least one optical waveguide (2). The surface of the image sensor (1) is coated with a substrate layer (3). The first end (2a) of the optical waveguide (2) is connected to the light source as a light inlet. The second end (2b) of the optical waveguide (2) is drawn to the micron level by flame heating and then adsorbed on the surface of the substrate layer (3). The refractive index of the substrate layer (3) is similar to the effective refractive index of light propagating in the optical waveguide (2). The projection of the second end (2b) of the optical waveguide (2) on the surface of the image sensor (1) is much smaller than the surface of the image sensor (1). The surface of the image sensor (1) can accommodate multiple optical waveguides (2) to achieve multi-channel processing.

2. The spectrometer based on optical waveguide leakage mode spectroscopy according to claim 1, characterized in that, The substrate layer (3) is obtained by vapor deposition of magnesium fluoride or calcium fluoride on the surface of the image sensor (1) at a rate of 0.3 Å / s-1 Å / s, and the thickness of the substrate layer (3) is 150 nm-300 nm.

3. A spectrometer based on optical waveguide leakage mode spectroscopy according to claim 1 or 2, characterized in that, The first end (2a) of the optical waveguide (2) is cut flat and then connected to the FC connector (4).

4. A spectrometer based on optical waveguide leakage mode spectroscopy according to claim 1 or 2, characterized in that, The second end (2b) of the optical waveguide (2) is fixed to the substrate (3) by van der Waals force and electrostatic attraction.

5. A spectrometer based on optical waveguide leakage mode spectroscopy according to claim 1 or 2, characterized in that, The image sensor (1) is a COMS image sensor (1). Before the substrate layer (3) is deposited, the COMS image sensor (1) undergoes the following treatment: first, the protective glass of the chip of the COMS image sensor (1) is cut off and removed, then it is soaked in acetone solution to wash away the glass debris attached to the surface, and finally it is air-dried.

6. A spectral reconstruction method applied to a spectrometer based on optical waveguide leaky mode spectroscopy as described in any one of claims 1-5, characterized in that, The image sensor (1) is defined to have k sensor units capable of receiving leaked light. The method includes: Calibration: K different wavelengths of light are introduced into the optical waveguide (2) of the spectrometer based on optical waveguide leakage mode spectroscopy, and the photocurrent value of the sensor unit at each wavelength is saved as a spectral response function; Photocurrent acquisition: The light to be measured is introduced into the image sensor (1) through different optical waveguides (2) to obtain the photocurrent value in any sensor unit; Construct and solve a system of equations: Obtain the actual spectral function of the light to be measured by using the photocurrent value and the spectral response function obtained by calibration. Combine the actual spectral function with the Gaussian function of the corresponding sensor unit to obtain the photocurrent equation of a single sensor unit. Then, combine the photocurrent equations of k sensor units into a linear system of equations and solve to obtain the reconstructed actual spectral function of the light to be measured.

7. The spectral reconstruction method for a spectrometer based on optical waveguide leaky mode spectroscopy according to claim 6, characterized in that, During the calibration process, before introducing the light, use an optical power meter to confirm the output optical power of the adjustable light source and adjust the output light intensity to ensure that the power is the same at each wavelength.

8. A spectral reconstruction method for a spectrometer based on optical waveguide leaky mode spectroscopy according to claim 6 or 7, characterized in that, In the process of constructing and solving the system of equations, a loss function is introduced, and a linear regression algorithm is used to gradually minimize the loss function during the solution process.

9. The spectral reconstruction method for a spectrometer based on optical waveguide leaky mode spectroscopy according to claim 8, characterized in that, An additional regularization coefficient is introduced into the loss function, and the optimal solution is obtained through repeated iterations during the solution process to improve the accuracy of the results.

Images