Fourier transform spectrometer

Abstract

A Fourier transform spectrometer including an interferometer is provided. In one embodiment, an FT spectrometer includes an excitation light source, a beam splitter adapted to separate a metrology signal from the excitation light source and direct the metrology signal through an interferometer to a metrology detector. The FT spectrometer is additionally adapted to receive a spectral signal from the sample and pass the spectral signal through the interferometer to the spectrometer detector. The invention provides a method for correcting the movement of an interferometer reflector. In addition, a total spectroscopic method is provided that includes a spectral signal that includes Raman scattering, fluorescence, and near-infrared absorption components.

Description

[0001] Cross-reference to related applications

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 439,230, filed January 16, 2023, U.S. Provisional Application No. 63 / 463,597, filed May 3, 2023, U.S. Provisional Patent Application No. 63 / 529,612, filed July 28, 2013, and U.S. Provisional Patent Application No. 63 / 601,080, filed November 20, 2023, each of which is incorporated herein by reference as if fully set forth herein. Technical Field

[0003] The present disclosure relates to Fourier transform (FT) spectrometers. In particular, the present disclosure relates to FT spectrometers including an interferometer configured as a very small device with many applications, including wearable devices. Background Art

[0004] A Fourier transform spectrometer including an interferometer is provided. Summary of the Invention

[0005] A Fourier transform spectrometer including an interferometer is provided. In one embodiment, the FT spectrometer includes an excitation light source, a beam splitter adapted to separate a metrology signal from the excitation light source, and direct the metrology signal to a metrology detector via the interferometer. The FT spectrometer is further adapted to receive a spectral signal from a sample and transmit the spectral signal to the spectrometer detector via the interferometer. A method for correcting the motion of the interferometer mirror is provided. Additionally, a total spectral method is provided that includes a spectral signal comprising Raman scattering, fluorescence, and near-infrared absorption components.

[0006] The foregoing and other aspects, features, details, utilities, and advantages of the present invention will be apparent from reading the following description and claims, and from examining the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 An embodiment of a Fourier transform (FT) Raman spectrometer is shown.

[0008] Figure 2 Another example embodiment of an FT spectrometer is shown.

[0009] Figure 3 Examples are shown of possible imperfections in the signal generated within an interferometer due to imperfections in mirror movement and / or changes in environmental conditions, such as changes in temperature or light conditions.

[0010] Figure 4Furthermore, problems that may be introduced in an FT spectrometer comprising an interferometer due to mirror motion imperfections and / or variations in environmental conditions are detailed.

[0011] Figure 5 is a graph illustrating an example of a method for correcting for imperfect motor motion, thermal expansion or contraction of components within the spectrometer (eg, due to changes in environmental conditions), and / or changes in laser or other light source wavelength.

[0012] Figure 6 To illustrate the use of reference Figure 5 Graph showing an example of a corrected FT signal resulting from the described correction method.

[0013] Figures 7A to 7C A further example of an implementation for correcting for imperfect mirror motion in a small scalable FT spectrometer is shown.

[0014] Figure 8 A pair of graphs showing more details of an example of a scalable FT spectrometer with imperfect mirror motion are shown.

[0015] Figure 9 Graphs showing additional details of an example of a scalable FT spectrometer with imperfect mirror motion are shown.

[0016] Figure 10 An example of a scalable FT spectrometer with imperfect mirror motion corrected by implementation of a corrective post-processing method is shown.

[0017] Figures 11A to 11C A graph illustrating the effect of imperfect mirror motion and its impact on the FT spectrum calibration is shown.

[0018] Figure 12 An example of calibration using a scalable FT spectrometer is shown.

[0019] Figure 13 This is an example of combining multiple spectral signals into a total spectral signal.

[0020] Figure 14 is another example of an FT spectrometer adapted to generate a total spectral signal as a combination of at least Raman, fluorescence and NIR absorption spectra.

[0021] Figures 15A to 15D An example of an implementation of VCSEL excitation to generate spatial information about a sample is shown.

[0022] Figure 16 is another example of an implementation of an FT spectrometer comprising a relatively large area excitation pattern adapted to produce a spatially averaged and low power density on the sample.

[0023] Figure 17 Spectra obtained using a microlens array, a ball lens, and a convex lens to generate Raman spectra are shown.

[0024] Figure 18 is an example of an implementation of a scalable FT spectrometer for rejecting interference from ambient light conditions.

[0025] Figure 19 Another example implementation of a scalable FT spectrometer for interference rejection is shown.

[0026] Figure 20 is another example of an implementation of a device for operation at long wavelengths above 900 nm.

[0027] Figure 21 Two example implementations of actuators that may be used to move components of an FT spectrometer, such as a moving mirror of an interferometer of an FT spectrometer, are shown.

[0028] Figures 22A to 22B Three example embodiments of actuators that may be used to move components of an FT spectrometer, such as a moving mirror of an interferometer of an FT spectrometer, are shown.

[0029] Figure 23A and 23B Two examples of actuators that can be used to move components of an FT spectrometer, such as a moving mirror of an interferometer of an FT spectrometer, are shown. DETAILED DESCRIPTION

[0030] The following description of the present invention is provided as an implementation teaching of the present invention with the best currently known embodiment of the present invention. For this reason, those skilled in the relevant art will recognize and understand that many changes can be made to the different aspects of the present invention described herein while still obtaining the beneficial results of the present invention. It is also apparent that some of the expected benefits of the present invention can be obtained by selecting some of the features of the present invention without utilizing other features. Therefore, those skilled in the art will recognize that many modifications and adaptations of the present invention are possible and, in some cases, even desirable and are part of the present invention. Therefore, the following description is provided as an illustration of the principles of the present invention, rather than limiting it.

[0031] As used throughout, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to "a" component may include two or more such components unless the context clearly dictates otherwise. Additionally, the words "proximal" and "distal" are used to describe items or portions of items that are located closer to and farther from, respectively, a user or operator (such as a surgeon). Thus, for example, the tip or free end of a device may be referred to as the distal end, while the generally opposite end or handle may be referred to as the proximal end.

[0032] All directional references (e.g., upper, lower, upward, downward, left, right, leftward, rightward, top, bottom, above, below, vertical, horizontal, clockwise, and counterclockwise) are used for identification purposes only to assist the reader in understanding the present invention and do not constitute limitations on the position, orientation, or use of the present invention in particular. References to joining (e.g., attaching, coupling, connecting, etc.) are to be interpreted broadly and may include intermediate members between the connection of elements and relative movement between elements. Likewise, references to joining do not necessarily infer that two elements are directly connected and in fixed relation to each other.

[0033] Ranges may be expressed herein as from "about" one particular value, and / or to "about" another particular value. When such ranges are expressed, another aspect includes from the one particular value and / or to the other particular value. Similarly, when a value is expressed as an approximation by using the antecedent "about," it is understood that the particular value forms another aspect. It is further understood that each endpoint in a range is meaningful both relative to the other endpoint and independently of the other endpoint.

[0034] As used herein, the terms "optional" or "optionally" mean that the subsequently described event or circumstance may or may not occur, and that the description includes instances where said event or circumstance occurs and instances where it does not.

[0035] As used herein, the term "substantially" may be applied to modify any quantitative representation that could permissibly vary without resulting in a change in the basic function to which it is related.

[0036] Various embodiments are provided in which various concepts, components, and techniques are described that can be combined in further embodiments. The intended embodiments are not limited to the various embodiments provided herein, but include variations of the concepts, components, and techniques provided. Figure 3-19 The process shown and described in D can be used with FT spectrometers such as but not limited to Figure 1 and 2 In addition, components such as those shown in Figures 15-17, 19D, 21A-21C, 22A and 22B may also be used in any spectrometer, such as but not limited to, Figure 1 and 2FT spectrometer shown.

[0037] Figure 1 An embodiment of a Fourier transform (FT) Raman spectrometer is shown. The spectrometer includes a spectroscopic laser adapted to provide an excitation light signal to excite a sample and generate a spectral signal (e.g., Raman scattering, fluorescence, and / or near-infrared (NIR) absorption) from the sample. The spectroscopic laser (e.g., a Raman laser) is shown to direct the excitation light signal toward the sample using a first dichroic beam splitter, which sends a relatively large percentage of the light beam to the sample. For example, in one embodiment, less than 5% of the laser excitation light signal leaks through the dichroic beam splitter. The first dichroic beam splitter also allows longer wavelengths, such as spectral signals (e.g., Raman scattering, fluorescence, and / or NIR absorption), to pass through an interferometer. A second laser, referred to as a metrology laser, generates a metrology light signal that is directed into the interferometer. The metrology laser serves two purposes. First, the metrology laser is used to schedule the sampling of the spectral signal (e.g., Raman scattering) and ensure that the sampling is greater than twice the frequency of the Raman signal. This is known as the Nyquist condition. Secondly, the principle of FT spectroscopy is that a moving mirror within an interferometer will produce an interference pattern whose horizontal axis is the distance between the two mirrors. The accuracy of this distance determines the accuracy of the spectrum produced by the FT algorithm. A quasi-monochromatic source like a laser will produce a sinusoidal pattern that has a maximum at the precise wavelength of the laser and is therefore a standard for the distance between the mirrors. A third source that can be used by FT spectrometers is a broadband source, such as "white light" or other broadband source, to produce a large signal when the mirrors are exactly the same distance apart. This is in Figure 1 Shown in FIG is an LED source that generates a broadband signal.

[0038] For example, in Figure 1 In the embodiment shown, the interferometer may include a Michelson design in which an approximately 50% to approximately 50% beam splitter is placed between a moving mirror and a fixed mirror. A source signal (spectral signal, metrological light signal, and / or broadband signal) is introduced into the beam splitter. The beam splitter sends a portion of the source signal to the fixed mirror and another portion of the source signal to the moving mirror. This produces an interference pattern determined by the distance between the mirrors and the wavelength of the source. The interference pattern (interferogram) is projected onto a transducer (e.g., a photodetector) to produce an electrical signal to record the interference pattern. For example, in Figure 1 In the example shown, three detectors are shown for three sources.

[0039] In this embodiment, the spectrometer includes three detectors that provide corresponding detection signals. The LED detector produces an LED detection signal that includes a sharp peak at the zero path difference (ZPD). This can be used by the FT algorithm, which integrates the interferogram from 0 to infinity and defines the signal at 0. FT theory states that a signal from a broad source will produce a large interferometer signal at the ZPD.

[0040] The metrological detector generates a metrological detection signal that is used to determine the optical path difference (OPD), which can be used to correct for inaccuracies in the mechanical drive that moves the mirror. FT theory states that a sharp (monochromatic) signal will produce a continuous sinusoidal interference pattern.

[0041] A spectral detector (eg, a Raman detector) generates a spectral detection signal (eg, a Raman detection signal) that is a combination of multiple sharp bands, and thus generates an interferogram that is a combination of multiple sinusoidal signals at different frequencies.

[0042] When FT algorithms are applied to these signals (intensity vs. path difference), they produce a spectrum (intensity vs. frequency). The FT algorithms in various implementations may be executed by a controller or other processor device and may include a Fast Fourier Transform (FFT) or a Discrete Fourier Transform (DFT).

[0043] The calculation and correction algorithms are executed via a controller. In one embodiment, the controller is a component of the FT spectrometer. In another embodiment, the controller is remotely connected to the FT spectrometer, such as wirelessly (e.g., Bluetooth, WiFi, or other wireless connection) or wiredly (e.g., USB or Lightning connection). For example, the wearable device can be adapted to wirelessly connect to an external processor or controller, such as a smartphone, tablet, laptop, personal computer, or other computing device. For example, in one embodiment, the FT spectrometer is adapted to send interferograms (metrology and spectra) to an Android device via Bluetooth. In this embodiment, there would be one controller on the interferometer to form the waveform to control the motor, and there would be a second wireless "controller" to perform the Fourier transform and correction / calibration method.

[0044] Figure 2 Another example embodiment of an FT spectrometer is shown. In this embodiment, the FT spectrometer includes a spectral laser source, but eliminates Figure 1 The broadband light source and metrological light source shown in the FT spectrometer are shown. Figure 2 In the FT spectrometer shown, Figure 1 The functions of the broadband light source and metrology light source shown are replaced by the spectral light source.

[0045] The FT spectrometer also includes a reflector and a filter (e.g., a neutral density filter). The reflector and neutral density filter are adapted to isolate, attenuate, and forward the small-amplitude optical signal from the spectral laser that leaks into the interferometer via the dichroic beam splitter. This signal is used as a metrology signal from the spectral laser source. For example, in this embodiment, the FT spectrometer is adapted to operate without the need for a secondary metrology laser. The FT spectrometer also generates a metrology signal at the frequency of the spectral laser source, and this can be used to calibrate the spectral signal.

[0046] In one embodiment, a reflector can be used without a filter to reflect the leakage portion of the spectral light signal toward the interferometer as a metrological light signal, where the signal does not need to be attenuated, or the reflector is adapted to attenuate the leakage portion of the spectral light signal. For example, in one embodiment, the surface of the "reflector" can be modified to reduce reflectivity. For example, the "reflector" can include a glass plate, a bead-blasted plate, a poorly reflective metal surface, or even a slight misalignment of the optics or detector. Additionally, a neutral density filter can be placed at or near the detector. In yet another example, the beam shaping optics can defocus the light on the detector.

[0047] exist Figure 2 In the embodiment of the FT spectrometer shown, the FT spectrometer is adapted to use the leaked attenuated portion of the spectral laser signal as a metrology laser. The laser signal generated by the spectral laser source is directed to a first beam splitter. The first beam splitter primarily reflects all of the spectral laser signal toward the sample. The spectral laser signal is focused onto the sample via one or more beam shaping optical devices (such as lenses). However, a small portion of the spectral laser signal will pass through the first beam splitter toward the mirror and neutral density filter, as shown in the dotted pattern. The mirror and neutral density filter are used to attenuate the portion of the spectral light signal that leaks through the neutral density filter through the first dichroic beam splitter, and reflect the attenuated light signal back to the first beam splitter via the mirror. The first dichroic beam splitter then reflects the attenuated leaked portion of the spectral light signal into the interferometer. In the interferometer, the interferometer beam splitter divides the attenuated light signal between the fixed mirror and the movable mirror of the interferometer. The split attenuated light signal is then passed to the third beam splitter.

[0048] In some implementations, a compensator is used, such as a beam splitter whose surface is coated to create light splitting. In such examples, one arm of the interferometer has a longer optical distance due to the refractive index of the beam splitter material. In other words, the path length of one arm is n(refractive index) × the thickness of the beam splitter. Because the refractive index also depends on wavelength, the distance difference between the two arms varies with wavelength. In this example, the compensator can be made of the same material as the beam splitter.

[0049] In this embodiment, the metering signal path for the attenuated optical signal uses Figure 2The dashed path in FIG. The last beam splitter in the system reflects this short laser wavelength and passes the longer wavelength spectral signal. The laser metrology signal is then focused onto a metrology detector (e.g., a photodiode) via one or more beam shaping optics (such as lenses). For example, in this particular embodiment, the metrology detector comprises the photodiode depicted on the left side of the interferometer. Similarly, the longer wavelength spectral signal (shown by the solid gray path) is passed by the third and final beam splitter toward the longpass filter and is then focused onto the spectral detector via one or more beam shaping optics (such as lenses). In this embodiment, the spectral detector comprises a second photodiode depicted on the top of the interferometer.

[0050] A controller is provided to perform one or more calculations or operations of the FT spectrometer. For example, in one embodiment, one or more calculations and corrections are performed by the controller. In one embodiment, the controller is a component of the FT spectrometer. In another embodiment, the controller is remotely connected to the FT spectrometer, such as wirelessly (e.g., Bluetooth, WiFi, or other or wireless connection) or wiredly (e.g., USB or Lightning connection). For example, the wearable device can be adapted to wirelessly connect to an external processor or controller, such as a smartphone, tablet, laptop, personal computer, or other computing device. For example, in one embodiment, the FT spectrometer is adapted to send interferograms (metrology and spectra) to an Android device via Bluetooth. In this embodiment, there is one controller on the interferometer to form the waveform to control the motor, and there will be a second wireless "controller" to perform the Fourier transform and correction / calibration method.

[0051] Figure 3 Examples are shown of possible imperfections in the signal generated within an interferometer due to imperfections in mirror movement and / or changes in environmental conditions, such as changes in temperature or light conditions.

[0052] In one embodiment of an FT spectrometer, for example, a method is provided for correcting a signal, such as for imperfections in a mirror and / or changes in environmental conditions (such as light conditions). In this embodiment, for example, the method of correcting an optical signal for imperfections in mirror motion allows for perfect mirror motion to be produced from motor-driven mirror motion that may be imperfect and low-cost compared to current methods such as air bearings. This also enables scalable spectrometers to utilize small, low-cost components that would otherwise produce imperfect motion.

[0053] although Figure 1 and 2 An interferometer is shown for an FT spectrometer comprising a single fixed mirror and a single moving mirror, but other implementations are also contemplated. A wishbone design interferometer comprising two moving mirrors and no fixed mirror may also be used. For example, in this design, Figure 1 and 2 The fixed mirrors of the interferometer shown can be replaced by movable mirrors. Similarly, Figure 1 and 2 The spectrometer may include a fixed mirror and a plurality of movable mirrors.

[0054] For example, Figure 3 An example of an interferometer such as a Michelson interferometer including a first fixed mirror and a second movable mirror is shown. The interferometer receives an input optical signal such as a reference Figure 1 and 2 as described, and generating an interferogram output signal which is directed to a detector.

[0055] exist Figure 3 In the top graph shown on the right, the mechanical path differences can be shown as an irregular non-sinusoidal pattern as shown. Changes in the movement of the second movable mirror or changes in components, such as due to temperature fluctuations, can provide an irregular signal.

[0056] exist Figure 3 In the bottom graph shown on the right, a uniform sinusoidal pattern corresponding to a known light source (e.g., a metrology laser source, a spectral laser source, a broadband light source) can be used to map multiple data points from the irregular non-sinusoidal pattern shown in the top graph to correct for changes in mirror movement and / or environmental conditions. Figure 3 In Figure 1, the top graph shows the mechanical path difference observed from an imperfect motor. The optical path difference is the true position. In other words, the optical path difference is the actual position of the motor as determined by the sine wave.

[0057] Figure 4 Furthermore, problems that may be introduced in an FT spectrometer comprising an interferometer due to mirror motion imperfections and / or variations in environmental conditions are detailed.

[0058] Figure 4 Illustrate the problem of motion of an imperfect mirror. Figure 4 The spectrum generated by a 520nm VCSEL metrology laser and the signal generated by a 977nm VCSEL laser are shown. Both the 520nm metrology laser and the 977nm spectral signal should theoretically be single lines, as they originate from quasi-monochromatic sources. The wide frequency range surrounding each of these signals stems from the imperfect mirror motion generated by a moving mirror, such as that used by a voice coil as a motor in one embodiment of an FT spectrometer. Voice coils, such as those from small headphone speakers, represent one possibility for creating scalable miniaturized FT spectroscopy systems.

[0059] Figure 4A second aspect shown is that VCSELs (Vertical Cavity Surface Emitting Lasers) can be used as one or more sources for FT spectroscopy. VCSELs represent low-cost, small-scale laser sources that can have the linewidth and stability required for spectroscopy sources (e.g., Raman sources) or metrology lasers.

[0060] The broadening of the signal is due to the convolution of the true interferogram, as if the mirror motion were perfect, and the imperfect motion of the mirror.

[0061] Figure 5 A graph illustrating an example of a method for correcting for imperfect motor motion, thermal expansion or contraction of spectrometer internal components (e.g., due to changes in environmental conditions), and / or changes in the wavelength of a laser or other light source. Correction can be particularly useful in scalable FT spectrometers with imperfect mirror motion, such as due to small, relatively inexpensive components. In this embodiment, post-processed data (e.g., metrology data) is digitally mapped to an ideal sine wave calculated from the frequency of a source signal (e.g., a metrology laser source).

[0062] Figure 5 Shown from Figure 5 A graph of a typical metrology signal for an imperfect mirror motion in an FT spectrometer on top. Ideally, the metrology signal should be something like Figure 5 The perfect sinusoidal signal is shown at the bottom of Figure 3 and 4 As discussed, imperfect signals can be corrected by post-processing the inspection data by obtaining evenly spaced data points from a perfect sine wave, which is formed by counting the number of peaks in the metrology interferogram and forming a sine wave with the same number of peaks. These evenly spaced points are specified in a model sinusoidal signal and used to map the imperfect signal from the motion of the imperfect mirror. This method corrects the metrology signal and aligns it to match the calculated sinusoidal signal.

[0063] The mapping is also applied to the spectral signal, thereby correcting it for imperfect mirror motion. This embodiment is useful for creating small, scalable FT spectrometers that may be subject to imperfect mirror motion. It will also correct for changes in the interferometer that may occur due to environmental changes. An example would be temperature changes that cause mechanical expansion of the interferometer material. A temperature increase can increase the distance between the fixed and moving mirrors. This will change the interferogram in unpredictable ways. This embodiment will similarly correct for these changes. For example, if the scalable FT spectrometer is used as a wearable device, it will be affected by these temperature changes due to body temperature and ambient temperature.

[0064] As mentioned above Figure 3 As described, an example method for correcting for mirror motion defects is as follows:

[0065] Counting the number of peaks in a metering signal;

[0066] Use the number of peaks to form the average frequency of the peaks in the interferogram;

[0067] Use this frequency to form a theoretical interference pattern with a perfect sinusoidal pattern;

[0068] Locate peaks and valleys in the acquired and theoretical interferograms;

[0069] Adding or removing these distances from the acquired interferogram to make it match the theoretical interferogram; and

[0070] • Use the same distance to map the spectral interferogram to the theoretical interferogram, thereby correcting it as well.

[0071] For example, in one embodiment, the metrology signal is sampled at 500 kHz as the mirror moves back and forth. This yields 1 million data points during the mirror's forward motion. The received data is intensity and appears to be time-varying. However, because the mirror is moving, it can be converted to distance rather than time. Because the motor is imperfect, the distance cannot be accurately known.

[0072] The calibration is done by measuring the peaks and valleys that the metrology laser produces optically. By summing the peaks in the imperfect metrology signal and forming a uniform sinusoidal pattern, we can calculate the distance from the imperfect metrology signal to the perfect waveform of our calculated sinusoidal pattern. The calibration can then be applied to the spectral signal, thereby correcting it as well.

[0073] The calculations and corrections are performed via a controller. In one embodiment, the controller is a component of the FT spectrometer. In another embodiment, the controller is remotely connected to the FT spectrometer, such as wirelessly (e.g., Bluetooth, WiFi, or other wireless connection) or wiredly (e.g., USB or Lightning connection). For example, the wearable device can be adapted to wirelessly connect to an external processor or controller, such as a smartphone, tablet, laptop, personal computer, or other computing device. For example, in one embodiment, the FT spectrometer is adapted to send interferograms (metrology and spectra) to an Android device via Bluetooth. In this embodiment, there is one controller on the interferometer to form the waveform to control the motor, and there will be a second wireless "controller" to perform the Fourier transform and correction / calibration method.

[0074] Figure 6 To illustrate the use of reference Figure 5 Graph showing an example of a corrected FT signal resulting from the described correction method.

[0075] Figure 4The two signals shown, the metrological and spectral signals, should be sharp lines produced by a quasi-monochromatic laser. However, Figure 4 The uncorrected signal is shown to be a very broad signal, such as that due to the imperfect mirror motion produced by the voice coil. Figure 5 When the post-processing method described is applied to this data, the results are Figure 6 Shown in. Figure 4 The broad spectral features shown are folded into Figure 6 Corrected sharp lines are shown.

[0076] In this example, the x-axis shows the data points. For an interferogram 100,000 points long, determined by the ADC sampling rate and the speed of the mirror movement, the x-axis would run from 1 to 100,001. In the calibration method, everything can be done in data points until the data is calibrated.

[0077] Figures 7A to 7C A further example of an implementation for correcting for imperfect mirror motion in a small scalable FT spectrometer is shown.

[0078] Figures 7A to 7C This graph shows a data set from an FT spectroscopy experiment used to measure the spectrum of an LED spectral source. In this example, the mirror motion is controlled using a voice coil. A sawtooth voltage waveform is applied to the voice coil to produce linear forward and backward motion. Figures 7A to 7C This figure shows the imperfections when the motor slows down and stops to change its direction of motion. It also illustrates an imperfect metrology signal. The intensity variations are caused by the imperfect motion leading to poor alignment of the interferometer. The distance variations between the metrology signal peaks are due to the poor linear motion of the voice coil driving the moving mirror.

[0079] In this example, Figure 7A Actual data from an interferometer is shown. Frequency distortion is evident near the point where the mirror changes direction (shown at the area highlighted by the circle), and variations in the metrology signal are evident throughout the interferogram. Figure 7B A small portion of the metrology signal is shown, and what should be a sine wave is an irregular sawtooth wave due to undersampling of the interferogram. Figure 7C A theoretical sine wave is shown for comparison. In one embodiment, the portion of the data corresponding to where the mirror changes direction (highlighted area) may be discarded from the data used to correct for the metrology and spectral signals for mirror motion and / or environmental changes due to its increased noise level.

[0080] Figure 8 Included are a pair of graphs showing more details of an example of a scalable FT spectrometer with imperfect mirror motion.

[0081] Figure 8The graph shown details Figures 7A to 7C The data shown. Figures 7A to 7C This data subset illustrates the imperfect mirror motion in this voice coil-driven spectrometer. The solid line is calculated based on the total number of peaks in the metrological (dashed) interferogram. In this example, a portion of the rising edge of the sawtooth mirror drive waveform is shown. The total number of peaks in the metrological interferogram is 2077, and there are 1,000,000 data points. This frequency (2077 / 1,000,000) is used to form a theoretically perfect sine wave, shown as the dashed line. Figure 8 This illustrates the error in the theoretical signal in a subset of the data (i.e., how far away is the measured signal from the calculated signal). In this small subset of the interferogram, you can see that the mirror is moving faster, producing a sine wave with a higher frequency than the calculated sine wave. These changes in the mirror's motion speed result in a broadening of the spectrum, such as Figure 4 shown.

[0082] Figure 8 Also shown in the application Figure 5 Following the described method, the metrology data follows the exact frequency of the calculated interferogram. Figure 5 The method described can be used to correct both the frequency of the data and the intensity of the data.

[0083] Figure 9 Graphs showing additional details of an example of a scalable FT spectrometer with imperfect mirror motion are shown.

[0084] Figure 9 FT results illustrating poor mirror motion in a small scalable FT spectrometer. Figure 9 Show Figures 7A to 7C FT of the data presented in . Figure 4 As shown, due to the large variability of frequency in the interference pattern, the metrology signal is distributed over a large range. In this example, different LEDs are used as separate simulated spectral signals set at the simulated sample position to simulate the spectral signal. Interference patterns of two simulated spectral signals from the LED source are also collected and labeled as LED1 and LED2. They are distorted due to poor reflector movement, but the distribution in frequency space is smaller than that of the metrology signal. This is because the LED output forms a small burst signal around the ZPD and is not affected by much distortion during the longer reflector movement captured by the metrology laser. We calculated that the LED simulated spectral signal distortion is approximately 6% based on the component specification full width at half maximum (FWHM). In this particular example, the Marktech Optoelectronics multi-chip emitter product number MTMD6788594SMT6 is used as the LED source for the simulated spectral signal source.

[0085] Figure 10An example of a scalable FT spectrometer with imperfect mirror motion corrected by implementation of a corrective post-processing method is shown.

[0086] Figure 10 illustrate Figure 5 The results of the implementation of the correction process outlined in . Figure 9 This same data is shown before correction. Figure 9 We show that poor mirror motion in a small, scalable FT spectrometer can lead to spurious data when applying the FT algorithm. Figure 10 Shown from Figure 5 The algorithm corrects the data over the entire distance of the mirror movement, resulting in a suitably sharp line for the metrology laser, and also corrects the line shape and FWHM of the two LED sources representing the simulated spectral source.

[0087] Figures 11A to 11C A graph illustrating the effect of imperfect mirror motion and its impact on the FT spectrum calibration is shown.

[0088] An alternative to FT spectroscopy is to use dispersive optical elements such as diffraction gratings. Gratings differ significantly from FT spectrometers in the way they produce spectra. Gratings disperse light into a pattern that is linear in wavelength. For spectrometers where X, Y are paired as wavelength and intensity, gratings work well, although they still need to be calibrated against a standard to map the spectrum in pixel space, as with multi-channel detectors with pixel elements. For spectrometers where X, Y are paired as frequency (cm -1 ) and intensity paired Raman-like spectra, where cm -1 Since the wavelength is a well-known unit of frequency, the linear pattern of the grating in wavelength can cause calibration problems. For example, a Raman spectrum collected using a 633nm laser will not match a spectrum collected using a 532nm laser when plotted in wavelength. Raman spectroscopy requires X, Y as wavenumbers (cm -1 ) and intensity to produce the same energy difference spectrum regardless of the laser excitation wavelength. Unlike a simple calibration that maps the number of pixels from a multichannel detector to the wavelength in a grating system, the relationship between wavelength and wavenumber is an inverse relationship, where cm -1=1 / cm and cm is equivalent to wavelength. This means that a simple (linear) mapping between pixels and wavenumbers is impossible for a grating spectrometer. The recognized method is ASTM E 1840, which uses a grating spectrometer to calibrate Raman spectra. This ASTM method requires that the peaks in the uncalibrated pixel space be matched with peaks provided by ASTM as acceptable Raman standard peaks. The 1 / wavelength relationship requires a high-order least squares fit to find a suitable mapping, rather than a linear mapping. This method also requires the frequent use of at least one chemical standard to ensure proper calibration, and ASTM standards are typically toxic and often have volatility associated with them. Implementations of methods for correcting for mirror motion and other environmental changes can eliminate these problems and can produce improved calibrations for each spectrum.

[0089] Calibration within and between spectrometers is often important for applications where acquired spectra are correlated with a standard library. Complex and often poor calibration of grating spectrometers results in poor matches with the standard library. Another use for spectral data is as a large dataset for forming machine learning models. In these machine learning tools, sufficient calibration is again crucial to form meaningful, robust models.

[0090] Yet another complication with grating spectrometers with multi-channel detectors is intensity calibration. As with the mapping of pixels to wavenumbers, consistency in the intensity at each wavenumber is important for matching a library to a peak or a standard intensity for a machine learning model. A grating spectrometer disperses light into a pattern of wavelengths and intensities orthogonal to the direction of the light. A multi-channel detector is placed along this orthogonal axis to form a digital spectrum. The problem is that the dispersed light must be focused along this axis, and the optics used for focusing must be significantly larger than the grating, or if small focusing optics are used, a pattern with vignetting will be formed. Large optics limit the small size that can be achieved using this type of system, and vignetting creates an inconsistent intensity distribution on the multi-channel detector. This is an advantage of FT spectrometers that have a single detector and are not affected by the size requirements of lenses that focus onto a multi-channel array detector.

[0091] Figures 11A to 11C Shown are data from interferograms generated by multiple sawtooth waves that resulted in 15 interferograms consisting of forward and reverse mirror motions. Figure 11A Corrected but uncalibrated data along multiple data points are shown. Figure 11B Show Figure 11A The data shown corresponds to 15 separate (some of which at least partially overlap) laser lines. Figure 11C Shown from 1100 to 900 cm -1 In the area of ​​uncalibrated data Figure 11AZoomed-in view of the data region shown. The separation of the metrology laser signal can be seen due to the difference in speed between the motor moving in the forward and reverse directions. It can also be seen that even within these two subsets, there is variation between scans. This is an artifact caused by poor mirror motion in a small, scalable FT spectrometer. This is an artifact of the application Figure 5 This presents a new calibration problem. Different mirror velocities within the waveform in the forward or reverse direction and even greater velocity differences between forward and reverse result in different frequencies for the laser and Raman signals in the FT spectrum. Figure 5 The post-processing described can correct Figure 11A The data shown is analyzed and the FT algorithm is applied to each interferogram generated by the forward and reverse directions of multiple sawtooth patterns. Each of these data subsets from the sawtooth will have its own unique laser line frequency and Raman spectrum. Because Raman scattering is inelastic, or in other words, energy loss, each subset can be calibrated to achieve improved calibration, and multiple subsets from the sawtooth pattern can be accurately averaged.

[0092] Figure 12 An example of calibration using a scalable FT spectrometer is shown.

[0093] exist Figure 12 The FT spectra shown on the graph on the left are uncalibrated data from a metrological detector (top) and a spectral detector (bottom). These were acquired using a 633 nm laser with the spectrometer configured at Figure 2 In this implementation, the Raman laser and the metrology laser are the same source. The sample is toluene. Toluene is characterized by a -1 The large peak at 15798 cm is shown in the upper left corner (reported in ASTM E 1840). -1 The laser line at 28980 cm appears at an arbitrary data point in the FT spectrum. Unlike grating spectrometers, which require nonlinear mapping, the mapping in this case is linear with a linear coefficient of 1.834409. This is calculated as 28980 / 157898 = 1.834409. When this mapping coefficient is applied to the data from the spectral detector, the toluene peak is found at 14794.4 cm -1 Since the Raman spectrum is the energy lost by laser energy, the energy of the toluene peak is 15798-14794.4=1003.6cm -1 .

[0094] This implementation is similar to Figure 5 The described implementation combination pairs normalized intensity with calibration wavenumber position, which is useful for robust and scalable FT spectrometers that can be used in conjunction with libraries or machine learning models.

[0095] Figure 13 This is an example of combining multiple spectral signals.

[0096] Spectroscopy often involves solving problems using only one technique, or by applying multiple techniques independently. For example, a book on analytical spectroscopy will have chapters dedicated to different types of spectroscopy. Our implementation of scalable FT spectroscopy enables one method to simultaneously perform three different spectroscopic methods. Depending on the sample, laser excitation can produce Raman scattering alone, or a combination of Raman scattering and fluorescence from the sample or impurities within it. Both Raman scattering and fluorescence can be considered spectral information. In addition to Raman scattering and fluorescence, a third technique provides key information about the sample. Light can be absorbed by the sample. In fact, if fluorescence is present, it is due to absorption from the laser excitation and emission of a fluorescence signal. The fluorescence signal is broad and often represents a continuous background across the entire Raman spectrum. In addition to electronic absorption of light, there is also absorption due to molecular vibrations within the sample. The fundamental wavelength of light that causes molecular vibrations due to absorption is between 2,500 and 50,000 nm. These wavelengths are outside the emission range of fluorescence or Raman scattering. This constitutes a technique known as IR absorption spectroscopy, or when combined with FT, it becomes FTIR. In addition to the fundamental vibrations of molecules, there are also overtones ranging from 700 nm to 2500 nm. The spectroscopic technique that measures these overtones is called NIR absorption spectroscopy. Raman spectroscopy is commonly performed in this region of the spectrum. Judicious selection of laser excitation wavelengths that produce fluorescence and fall within this wavelength range allows for simultaneous observation of Raman scattering, fluorescence, and NIR absorbance. These three sources of information from a single FT spectrometer enhance its spectra for use in building machine learning models and accurate correlation with existing libraries.

[0097] Figure 13 The left side of the diagram is a conceptualization of what we call the "total spectrum," or the combination of Raman scattering, fluorescence, and NIR absorption. The top is the Raman scattering of the sample, the second from the top is the fluorescence, the third from the top is the NIR absorbance of the sample, and the bottom is the combined "total spectrum" signal that will be observed. By observing the rate of change of the signal across the spectrum, these three signals can be extracted. This technique is similar to Sequential Excitation Raman Difference Spectroscopy (SERD). If the fluorescence spectrum is standard and in a library, it can be compared to the results of SERDS to generate the NIR absorption spectrum. While this is possible, having a fluorescence library is not standard.

[0098] In one implementation of a spectrometer, the total detected spectrum can be used, without separating it into Raman scattering, fluorescence, and NIR absorbance, to build a heuristic model around the desired properties of the sample. For example, measurements can be taken from a wearable device or a small device to collect the total spectrum and build a model around properties such as nutrition or hydration levels.

[0099] For example, in one embodiment, the total spectrum method may include the following:

[0100] Exciting the sample using at least one light source, such as a laser light source, to produce a total spectral signal including Raman scattering, fluorescence, and absorption components. In another embodiment, the at least one light source may include a laser light source for producing Raman scattering and fluorescence, and a relatively weak broadband source for producing the absorption component.

[0101] Obtaining a total spectral signal including at least a Raman spectral component, a fluorescence spectral component, and a NIR absorption component, thereby forming an integrated total spectral signal;

[0102] comparing the total spectral signal to a library of discrete total spectral signals and corresponding corresponding materials and / or comparing the total spectral signal to a standard intensity or absorbance to generate a quantitative response; and

[0103] • Identifying at least one of a corresponding qualitative or quantitative response based on the total spectral signal.

[0104] In another embodiment, a training set of total spectral signals is acquired and run through a machine learning algorithm to build a model for comparing detected total spectral signals to identify materials or quantify responses.

[0105] Figure 14 is another example of an FT spectrometer adapted to generate a total spectral signal as a combination of at least Raman, fluorescence and NIR absorption spectra.

[0106] In this example, the total spectral signal is used to determine a measure of the hydration level of tissue in the body. For example, a high hydration level may indicate edema, a serious indicator of cardiovascular disease. A low hydration level may indicate dehydration, a serious condition that can lead to death, several diseases, and excessive exercise and heat.

[0107] Figure 14Shown are the total spectroscopic spectra of chicken tissue in its natural hydration state (bottom spectrum) and after dehydration (top spectrum). These spectra were collected using an FT spectrometer operated under a 1064nm excitation source. At this wavelength, animal tissue produces Raman scattering, fluorescence, and NIR absorbance. Although NIR absorbance originates from many components of the tissue, the largest absorption component comes from water. Water is strongly absorbed in this region of the spectrum. The Raman spectrum of water is very weak, making this combination orthogonal in terms of information content. Orthogonal in this context is used to refer to two methods that produce different information.

[0108] Figures 15A to 15D An example of an implementation of VCSEL excitation to generate spatial information about a sample is shown.

[0109] One of the characteristics of dispersive spectrometers is the requirement that the signal pass through an aperture that determines the spectral resolution of the spectrometer. This requirement does not exist for FT spectrometers, and is referred to in the literature as the Jacquinot advantage. In one implementation, an FT spectrometer can use a VCSEL array spectral light source. In this example, the VCSEL array spectral light source enables two implementations of a scalable FT spectrometer. First, the array can be used to generate excitation over a large area, which, through the Jacquinot advantage, will allow it all to enter the interferometer. This means that the power density on the sample can be very small compared to dispersive systems, which require a small focal spot at the sample to pass through their aperture to produce a spectrum. Second, the elements of the VCSEL array can be individually addressed to illuminate different points of the sample.

[0110] When sampling human tissue in vivo, it is crucial to reduce the power density at the sample through this implementation of a laser or VCSEL array. The light intensity allowed on human tissue is defined by the Maximum Permissible Exposure (MPE). To develop regulatory-acceptable FT spectrometers for wearable or general in vivo spectroscopic analysis, the MPE dictates that laser powers can be prohibitively low for focused beams in dispersive spectrometers. The Jaquino advantage, coupled with a laser or VCSEL array, enables high laser intensities due to the large-area excitation.

[0111] The ability to individually address the spectral light source elements of the VCSEL array enables analysis of the spatial distribution of the sample. This will also enable averaging of large sample areas at lower power densities. Figures 15A to 15D The implementation of sampling the molecular distribution in the sample is shown in FIG. Figure 15A The VCSEL array spectrum light source is shown with all elements turned on. In this case, this implementation will produce a spectrum with low power density and material spatial distribution average ( Figure 15D Shown in FIG. 1 is spectrum A). Figure 15BA VCSEL array spectrum light source is shown where only half of its lasers are turned on and the resulting spectrum is a subset of the average. Figure 15D This spectrum, shown as spectrum B, is different from Figure 15A The spectrum shown in is indicated by the highlighted rectangle surrounding the 1000 wavenumber region. Figure 15C The VSEL array spectrum light source is shown, where the other half of the array is turned on, and the resulting spectrum is Figure 15D is shown as spectrum C, which is different from the spectrum corresponding to Figure 15A The spectroscopic spectrum A corresponds to Figure 15B Spectrum B of the array. This illustrates how the spatial distribution of a sample can be determined. The example of turning on half the array at a time is just one implementation of a laser or VCSEL array spectral light source; it is possible to turn on only one element at a time until all elements are turned on, but only one at a time. Regardless, the spatial pattern of the VCSEL array spectral light source can be correlated with the resulting spectrum to produce a spatial map of the sample. An example application would be the distribution of components in a pharmaceutical mixture. Another example application would be the distribution of illicit materials in a sample, where matching individual components can lead to more accurate identification from a spectral library rather than an average of the sample.

[0112] Figure 16 is another example of an implementation of an FT spectrometer comprising a relatively large area excitation pattern adapted to produce a spatially averaged and low power density on the sample.

[0113] As reference Figures 15A to 15C As mentioned above, the amount of laser power that can be used for in vivo spectroscopy is strictly regulated by the MPE. Figure 16 This example shows how to achieve low power density and average power over a large sample area using a VCSEL-type laser or other type of laser. In this example, a comparison of three possible sampling configurations is shown: a microlens array, a ball lens, and a lens that focuses the laser onto the sample from a distance. If a ball lens or lens focuses a 100 mW laser onto the sample, the power density from, for example, a 50 micron focus will be 50 W / mm 2 This situation will also result in a small spatial average of a 50 μm spot on the sample. Using the same 100 mW laser to implement a 10 × 10 mm lens array will produce 0.001 W / mm 2 This means a 50,000-fold reduction in power density, which increases the probability of not exceeding the MPE by a factor of 50,000. This will also improve sample averaging. An application described is in vivo tissue sampling, where the sample is very complex and has spatial variations greater than a 50-micron spot.

[0114] A microlens array (or other lens array) can be used with a spectral light source array such as a VCSEL array spectral light source, where one or more individual lens elements correspond to one or more individual light source elements, or can be used with a single light source element such as a single light source element that passes through a beam expander to reach the lens array.

[0115] Figure 17 is an example of an implementation of a microlens array for generating Raman spectra.

[0116] Figure 17 Show Figure 16 The data for the three sampling configurations discussed in [1] are shown. The sample was citric acid, a poor Raman scatterer, and the laser power was 30 mW at 633 nm. All of these spectra were collected using an FT Raman spectrometer. The importance of these spectra is that they are nearly equal in signal, but as Figure 17 The power density of the microlens array is calculated to be 1 in 50,000.

[0117] Figure 18 is an example of an implementation of a scalable FT spectrometer for rejecting interference from ambient light conditions.

[0118] Interference in Raman spectroscopy is the contribution of non-Raman scattering to the spectrum. These can be caused by competing light sources such as room lights or solar radiation. Dispersive multi-channel spectrometers collected during periods when room lights are operating at 50 or 60 Hz are averaged into the Raman signal. Similarly, solar radiation can fluctuate naturally at low frequencies due to cloud cover and natural fluctuations, and at higher frequencies due to turbulence, dust, and changes in atmospheric temperature. Dispersive spectrometers, which average the light within the spectrometer's spectral window, cannot distinguish between interfering sources. FT spectroscopy detects optical interference and converts it into frequency components.

[0119] In one implementation of an FT spectrometer, the FT spectrometer is adapted to differentiate the light sources detected by the FT spectrometer. This is valuable for rejecting ambient and solar interference.

[0120] Figure 18 The spectra produced by an FT spectrometer over a large frequency range are shown. This demonstrates an implementation of an FT spectrometer for rejecting ambient room light at frequencies much lower than optical frequencies. Low- and high-frequency variations in solar radiation will also be rejected. For example, the ability to distinguish spectral signals from interfering ambient light conditions is useful for wearable devices worn under varying ambient light conditions. Another example would be an FT spectrometer operating under varying ambient light conditions to detect and identify materials.

[0121] Figure 19 Another example implementation of a scalable FT spectrometer for interference rejection is shown.

[0122] Figure 19 Spectrum 19A shown shows the full FT spectrum of toluene collected with 30 mW of 633 nm light. The frequency axis is determined in part by the rate at which the mirror is moved. Spectrum 19A contains three important regions. First, at low frequencies, room light interference is observed. Second, in the region of 10 kHz to 20 kHz, a Raman signal is observed. Third, at high frequencies (above about 80 kHz), the noise decreases. The first region of enhancement in spectrum 19B is Figure 18 The implementation discussed in

[19] is to separate the indoor light flicker caused by AC voltage from the Raman signal. Spectrum 19C enhances the Raman spectrum frequency and noise reduction. Figure 19 C has led to another innovation in using a lock-in amplifier to collect frequency-modulated signals from a modulated laser source. It has been found that for a 633 nm laser and a Si photodiode, this increases the signal-to-noise ratio by approximately 2. Since one of the noise sources is thermal noise from dark current, and this is low frequency, by modulating the laser at a high frequency and detecting only these higher frequencies with a lock-in amplifier, the noise is rejected, thereby improving the signal-to-noise ratio.

[0123] Figure 19B 2 is a block diagram illustrating an example embodiment of a lock-in amplifier that may be used in an FT spectrometer to collect a frequency modulated signal from a modulated laser source.

[0124] An effective way to recover small signals masked by large ambient noise is to use so-called phase-sensitive detectors or lock-in amplifiers. Phase-sensitive detectors achieve narrowband amplification by reducing the noise content that falls outside the bandwidth of interest.

[0125] When the noise is essentially white, its amplitude level can be significantly reduced by limiting the detection bandwidth (which includes the modulation frequency at which the signal appears while excluding the frequency at which the noise appears). Phase-sensitive detection enables extremely narrow bandwidth detection (0.001Hz is typical). A typical application scenario is with electrical transducers, where the noise amplitude is in the millivolt region and the signal falls in the nanovolt region.

[0126] FT Raman spectroscopy is challenging because the optical signal generated is low, about 0.3nA, for a laser power of 500mW. This is almost equal to the dark current of a silicon sensor for visible wavelengths and about an order of magnitude smaller than the dark current of an InGaAs sensor for NIR wavelengths. Other noise sources in the system are the Johnson noise of the sensor load resistor and the noise generated by the preamplifier operational amplifier. The amplitudes of all these noise sources are similar to or greater than the amplitude of the desired signal, and they are broadband, so they will appear in the interference pattern regardless of the frequency of the interference pattern controlled by the mirror speed. The total noise power is proportional to the detector bandwidth, which implies that as long as the signal does not have a corresponding attenuation, the narrowing of the bandwidth will reduce the power density in the signal. Using a phase-sensitive detector, also known as a lock-in amplifier, allows the system to reduce the effective bandwidth to a few tenths of a hertz while maintaining the full power of the signal of interest. It does this by modulating the signal source (laser) with a controlled frequency and phase and then multiplying the received signal by the reference used to modulate the source. This result is low-pass filtered to remove the modulation frequency and then passed to the data acquisition part of the system. This allows achieving high SNR even when the broadband noise in the system is larger than the desired interferogram signal.

[0127] Figure 20 is another example of an implementation of a device for operation at long wavelengths above 900 nm.

[0128] Skin pigmentation typically has two sources: ethnicity and sun exposure. One implementation of our scalable FT spectrometer is for in vivo measurements through the skin to monitor a person's health. Figure 20 The absorption of chromophores (pigments) is shown as a function of wavelength. As expected, it is very high at dangerous UV wavelengths and drops at harmless longer wavelengths. Laser excitation and spectra with wavelengths longer than about 800 nm will reduce signal loss in people of color and reduce signal changes due to sun-related pigmentation.

[0129] Figure 21 Two example implementations of actuators that can be used to move components of an FT spectrometer, such as a moving mirror of an interferometer of an FT spectrometer, are shown. Figure 21 A moving magnet voice coil motor embodiment that can be used as an actuator in an FT spectrometer is shown. In this embodiment, a magnet is disposed between a set of coils within a conductive housing (eg, copper).

[0130] to produce a scalable FT spectrometer.

[0131] Shape memory alloy (SMA) materials will contract in length by up to 7% when heated above their transition temperature, while providing a significant motive force. The materials are available in a variety of transition temperatures ranging from -20°C to 100°C, with 70°C and 90°C being typical values. This heating can be provided by an external source or, more commonly, by Joule heating caused by passing an electric current through a wire.

[0132] The actuator SMA material can be in wire or ribbon form and can be made up of multiple elements that are mechanically parallel. Using smaller, thinner elements in parallel has the benefit of heating and cooling faster, and therefore providing faster motion.

[0133] Figure 22A A bi-directionally driven shape memory alloy (SMA) rotary actuator is shown.

[0134] The left and right SMA wires are alternately heated by passing current through them. This allows for symmetric speed in both directions, rather than waiting for the wires to cool by dissipating heat to the ambient air. Mechanical amplification is provided by placing the SMA element's anchor point closer to the pivot than the mirror. For typical applications, the Y to X ratio can be anywhere from 2 to 20. This allows for compact designs with SMA lengths as small as 1 cm.

[0135] Figure 22B A one-way driven shape memory alloy (SMA) rotary actuator is shown.

[0136] The SMA wire on the right is heated by passing an electric current through it, providing clockwise motion of the mirror. When the wire cools, counterclockwise motion is provided, which typically takes twice as long as heating. Because the speed varies greatly depending on direction, this design is best suited for capturing interferograms in only one direction.

[0137] Figure 22C A shape memory alloy (SMA) linear actuator is shown.

[0138] The wires are heated by an applied current to provide motion toward the device, while the return motion is provided by the spring constant of the flexure bearing. A symmetrical version can also be created by pulling a second set of wires in the opposite direction to the one shown and by alternating the current between the wire pairs. Linear actuators do not provide any mechanical amplification that would result in larger dimensions. For an application requiring 0.6 mm of motion, the SMA element would need to be 6 cm long.

[0139] Piezoelectric actuators

[0140] Piezoelectric actuators offer high force, high speed and good control with strokes of 40 to 120 μm. The PowerHap series from TDK / Epcos is representative of mechanically amplified piezoelectric actuators that could potentially be used to provide mirror motion in interferometers.

[0141] Mechanical amplification is provided by placing the thrust point of the piezoelectric actuator closer to the pivot than the mirror.For typical applications, the Y to X ratio can be anywhere from 5 to 20.

[0142] Figure 22A A square-shaped amplified piezoelectric actuator is shown.

[0143] Based on a 12.7mm square actuator, the preload will be 2N and the mechanical magnification ratio of Y to X will be 15:1 to provide 0.6mm of mirror travel and 40um of actuator stroke.

[0144] The drive required for full range of motion is 120V, very low current, and a small negative voltage will be applied to overcome the hysteresis inherent in the piezoelectric ceramic element.

[0145] Figure 22B A rectangular amplified piezoelectric actuator is shown.

[0146] Based on a 60mm x 5mm rectangular actuator, the preload would be 15N and the mechanical magnification in Y to X would be 5:1, providing a mirror travel of 0.6mm and an actuator stroke of 120µm. This design is larger but would have the benefit of being much more mechanically rigid to perturbations of the rotating arm and mirror.

[0147] A 120V drive is necessary, and very low current is required for full range of motion, and a small negative voltage will be applied to overcome the hysteresis inherent in the piezoelectric ceramic element.

[0148] In order to produce high quality data from a Michelson interferometer, it is helpful to move one of the mirrors in a smooth, continuous manner. One way to provide such motion is to use a linear or rotary voice coil motor (VCM) comprising one or more coil windings, a support structure and a permanent magnet. The motor can be configured to either move the coil or the magnet. Since the movement of the VCM in the interferometer is relatively slow (<20Hz), there is little efficiency loss by using mechanical damping. However, such damping should be achieved in a way that does not cause any jolts in the motion. To achieve this, we recommend using a highly conductive metal layer that will generate a back EMF that will dampen the motion. Damping can be achieved by making the coil out of copper or another highly conductive material that will generate a back EMF force whenever the coil moves. Another means of achieving this damping is to use a ferrofluid inside the VCM, which can be used in combination with eddy current damping. In particular, a high viscosity ferrofluid will provide damping while also providing smooth motion. These are described in the following sections. Figure 23A and 23B It is illustrated as a moving magnet VCM and a moving coil VCM.

Claims

1. A Fourier transform (FT) spectrometer comprising: an excitation light source, adapted to provide an excitation light signal; Spectrometer beam splitters suitable for: receiving the excitation light signal from the excitation light source, separating the excitation light signal into a first spectral excitation portion and a second spectral measurement portion of the excitation light signal, directing the first spectral excitation portion toward a sample and directing the second metrological portion of the excitation light signal toward a metrological mirror, wherein the metrology mirror is adapted to reflect the second metrology portion of the excitation light signal, and the spectrometer beam splitter is adapted to reflect the second metrology portion of the excitation light signal received from the metrology mirror and pass a spectral signal received from the sample; An interferometer comprising: an interferometer beam splitter adapted to receive the reflected metered portion of the excitation light signal and a corresponding one of the spectral signal and split the reflected metered portion of the excitation light signal and the corresponding one of the spectral signal between a first mirror and a second mirror, and direct the reflected signals from the first and second mirrors from the interferometer; a first metering signal detector; a second spectral signal detector; and A second spectrometer beam splitter is adapted to direct the second metrology portion of the excitation signal toward the first metrology signal detector and to direct the spectral signal to the second spectral signal detector. 2 . The FT spectrometer according to claim 1 , wherein the metrology mirror is adapted to attenuate the second metrology portion of the excitation light signal.

3. The FT spectrometer of claim 1, wherein an optical filter is adapted to attenuate the second metering portion of the excitation light signal. 4 . The FT spectrometer according to claim 3 , wherein the optical filter is arranged between the metrology mirror and the spectrometer beam splitter, or between the second spectrometer beam splitter and the first metrology signal detector. 5 . The FT spectrometer of claim 1 , wherein the controller is adapted to perform a Fourier transform on the metrological output of the first metrological detector and on the spectral signal output of the second spectral detector.

6. The FT spectrometer of claim 1 , wherein the controller is adapted to: determining the number of peaks in the metrology signal interferogram; determining an average frequency of peaks in the metrology signal interferogram based on the number of peaks; generating a theoretical interference pattern having a sinusoidal pattern based on the frequencies; locating a plurality of data points in a metrology signal interferogram and in the theoretical interferogram; as well as A plurality of distances corresponding to the plurality of data points in the metrology signal interferogram are adjusted to match the theoretical interferogram. 7 . The FT spectrometer according to claim 6 , wherein the plurality of adjusted distances corresponding to the plurality of data points of the metrology signal interferogram are used to map a spectral interferogram to the theoretical interferogram to correct the spectral interferogram.

8. The FT spectrometer according to claim 1, wherein the spectrometer is adapted to collect a total spectral signal comprising a Raman scattering component, a fluorescence component, and a near infrared (NIR) absorption component.

9. The FT spectrometer of claim 8, wherein the total spectrum is compared to a library or a machine learning model.

10. A method for providing a Fourier transform (FT) spectrum, comprising: providing an excitation light signal; Splitting the excitation light signal into a first spectral excitation portion and a second metering portion; directing the first spectral excitation portion of the excitation light signal toward a sample; directing the second metered portion of the excitation light signal toward an interferometer beam splitter; splitting the second metered portion of the excitation light between a first interferometer mirror and a second interferometer mirror; moving at least one of the first and second interferometer mirrors to generate an interference pattern between the separated portions of the second metered portion of the excitation light signal; directing a reflected portion of the metrology portion of the excitation light signal from the first interferometer mirror and the second interferometer mirror toward a first metrology detector; splitting the spectral signal between the first interferometer mirror and the second interferometer mirror; moving at least one of the first and second interferometer mirrors to generate an interference pattern between the separated portions of the spectral signal; as well as A reflected portion of the spectral signal is directed from the first interferometer mirror and the second interferometer mirror toward a second spectral detector. The method of claim 10 , wherein the metrology mirror is adapted to attenuate the second metrology portion of the excitation light signal.

12. The method of claim 10, wherein an optical filter is adapted to attenuate the second metered portion of the excitation light signal. 13 . The FT spectrometer according to claim 12 , wherein the optical filter is arranged between the metrology mirror and the spectrometer beam splitter, or between the second spectrometer beam splitter and the first metrology signal detector.

14. The method of claim 10, wherein the controller is adapted to perform a Fourier transform on the metrological output of the first metrological detector and on the spectral signal output of the second spectral detector.

15. The method of claim 10, wherein the controller is adapted to: determining the number of peaks in the metrology signal interferogram; determining an average frequency of peaks in the metrology signal interferogram based on the number of peaks; generating a theoretical interference pattern having a sinusoidal pattern based on the frequencies; locating a plurality of data points in a metrology signal interferogram and in the theoretical interferogram; as well as adjusting a plurality of distances corresponding to the plurality of data points in the metrology signal interferogram to match the theoretical interferogram; as well as 16. The method of claim 15, wherein the plurality of adjusted distances corresponding to the plurality of data points of the metrology signal interferogram are used to map a spectral interferogram to the theoretical interferogram to correct the spectral interferogram.

17. The method of claim 10, wherein the spectrometer is adapted to collect a total spectral signal comprising a Raman scattering component, a fluorescence component, and a near infrared (NIR) absorption component.

18. The method of claim 8, wherein the total spectrum is compared to a library or a machine learning model.

19. A method of calibrating a signal from an interferometer of an FT spectrometer, comprising: determining the number of peaks in the metrology signal interferogram; determining an average frequency of peaks in the metrology signal interferogram based on the number of peaks; generating a theoretical interference pattern having a sinusoidal pattern based on the frequencies; locating a plurality of data points in a metrology signal interferogram and in the theoretical interferogram; as well as adjusting a plurality of distances corresponding to the plurality of data points in the metrology signal interferogram to match the theoretical interferogram; as well as 20. The FT spectrometer of claim 6, wherein the plurality of adjusted distances corresponding to the plurality of data points of the metrology signal interferogram are used to map a spectral interferogram to the theoretical interferogram to correct the spectral interferogram.

21. A total spectral method comprising: Exciting the sample using at least one light source, such as a laser light source, to produce a total spectral signal including Raman scattering, fluorescence, and absorption components. In another embodiment, the at least one light source may include a laser light source for producing Raman scattering and fluorescence and a relatively weak broadband source for producing the absorption component. Obtaining a total spectral signal including at least a Raman spectral component, a fluorescence spectral component, and a NIR absorption component, thereby forming an integrated total spectral signal; comparing the total spectral signal to a library of discrete total spectral signals and corresponding corresponding materials and / or comparing the total spectral signal to a standard intensity or absorbance to generate a quantitative response; as well as • identifying at least one of the corresponding qualitative or quantitative response based on the total spectral signal.

22. The method of claim 21, wherein a training set of total spectral signals is collected and run through a machine learning algorithm to build a model for comparing detected total spectral signals to identify materials or quantify responses.

Images