WO2023091709A2 - Compensation en ligne de la dérive de réponse instrumentale dans des spectromètres miniaturisés - Google Patents

Compensation en ligne de la dérive de réponse instrumentale dans des spectromètres miniaturisés Download PDF

Info

Publication number
WO2023091709A2
WO2023091709A2 PCT/US2022/050469 US2022050469W WO2023091709A2 WO 2023091709 A2 WO2023091709 A2 WO 2023091709A2 US 2022050469 W US2022050469 W US 2022050469W WO 2023091709 A2 WO2023091709 A2 WO 2023091709A2
Authority
WO
WIPO (PCT)
Prior art keywords
spectrometer
temperature
detector
correction
correction matrix
Prior art date
Application number
PCT/US2022/050469
Other languages
English (en)
Other versions
WO2023091709A3 (fr
Inventor
Bassem Mortada
Yasser M. Sabry
Samir Abozyd
Original Assignee
Si-Ware Systems
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US17/989,317 external-priority patent/US20230152216A1/en
Application filed by Si-Ware Systems filed Critical Si-Ware Systems
Publication of WO2023091709A2 publication Critical patent/WO2023091709A2/fr
Publication of WO2023091709A3 publication Critical patent/WO2023091709A3/fr

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
    • G01J3/02—Details
    • G01J3/0286—Constructional arrangements for compensating for fluctuations caused by temperature, humidity or pressure, or using cooling or temperature stabilization of parts of the device; Controlling the atmosphere inside a spectrometer, e.g. vacuum
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
    • G01J3/02—Details
    • G01J3/0256—Compact construction
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
    • G01J3/02—Details
    • G01J3/0275—Details making use of sensor-related data, e.g. for identification of sensor parts or optical elements
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G01N21/27—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands using photo-electric detection ; circuits for computing concentration
    • G01N21/274—Calibration, base line adjustment, drift correction
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2201/00—Features of devices classified in G01N21/00
    • G01N2201/12—Circuits of general importance; Signal processing
    • G01N2201/121—Correction signals
    • G01N2201/1211—Correction signals for temperature

Definitions

  • a spectrometer measures a single-beam spectrum (e.g., a power spectral density (PSD)).
  • PSD power spectral density
  • the intensity of the single-beam spectrum is proportional to the power of the radiation reaching the detector.
  • absorbance of a sample is its fingerprint, which is used in spectral processing operations to enable material identification, along with quantitative and qualitative analysis.
  • absorbance is not measured in one step. Instead, it is extracted from the measured spectrum in presence of a sample with respect to a background (BG) spectrum measurement with no sample in the light path, to account for the instrumental transfer function.
  • BG background
  • the background measurement is acquired, in the case of transmittance, by measuring the spectrum of a light beam without placing any sample in the light pass. While in diffuse reflection, a reference material with 95% or higher reflectance and approximately a flat response across the spectral range of interest, is used to obtain a background measurement.
  • This methodology enables the compensation of instrumental response on condition of having the same instrumental drift between the background measurement and sample measurement.
  • instrumental response may change with environmental conditions, and as a result, satisfying this criterion may be difficult. For example, background measurements may need to be taken more frequently or even before each sample measurement. This consumes time and effort and sometimes is not possible.
  • a spectrometer including a light modulator optically coupled to receive input light and to produce modulated light based on the input light, a detector optically coupled to detect the modulated light in presence of a sample and to produce an output based on the modulated light, a sensor configured to obtain a value of a condition contributing to instrumental drifts in the spectrometer, a memory configured to store a correction matrix associating a plurality of sets of correction parameters with sensor values, and a processor coupled to the detector and configured to extract a set of correction parameters of the plurality of sets of correction parameters from the correction matrix based on the value.
  • the processor can further be configured to process the output of the detector based on the set of correction parameters to produce a corrected spectrum of the sample.
  • FIG. 1 is a diagram illustrating an example of a spectrometer according to some aspects
  • FIG. 2 is a diagram illustrating another example of a spectrometer according to some aspect.
  • FIG. 3 is a diagram illustrating another example of a spectrometer according to some aspects.
  • FIG. 4 is a diagram illustrating an example of a photodiode, which may be used in a photodetector of a spectrometer according to some aspects.
  • FIG. 5A is a graph illustrating ideal responsivity for photodiodes with constant quantum efficiency according to some aspects.
  • FIG. 5B is a graph illustrating typical responsivities of silicon (Si) and InGaAs photodiodes according to some aspects.
  • FIG. 6 is a graph illustrating the shift in cut-off wavelength with temperature for a semiconductor photodiode according to some aspects.
  • FIG. 7 is a graph illustrating the spectral temperature coefficient of the responsivity of a photodiode according to some aspects.
  • FIG. 8 is a graph illustrating the shift in the cut-off wavelength with temperature for a semiconductor photoconductive detector according to some aspects.
  • FIG. 9 is a diagram illustrating an example of a spectrometer configured for online compensation of instrumental drift according to some aspects.
  • FIG. 10 is a diagram illustrating another example of a spectrometer configured for on-line compensation of instrumental drift according to some aspects.
  • FIGs. 11 A and 1 IB are diagrams illustrating the relation between temperature and wavelength according to some aspects.
  • FIG. 12 is a flow chart illustrating a process for on-line compensation of temperature drifts in a spectrometer according to some aspects.
  • FIG. 13 is a diagram illustrating an example of a spectrometer employing the detector as a temperature sensor to be used for the correction of thermal drift according to some aspects.
  • FIG. 14 is a diagram illustrating another spectrometer employing the detector as a temperature sensor to be used for the correction of thermal drift according to some aspects.
  • FIG. 15 is a diagram illustrating an example of a spectrometer employing a dedicated temperature sensor to be used for the correction of thermal drift according to some aspects.
  • FIG. 16 is a diagram illustrating another example of a spectrometer employing a dedicated temperature sensor to be used for the correction of thermal drift according to some aspects.
  • FIG. 17 is a diagram illustrating another example of a spectrometer employing a dedicated temperature sensor to be used for the correction of thermal drift according to some aspects.
  • FIGs. 18A and 18B are diagrams illustrating another example of of a spectrometer employing the detector as a temperature sensor to be used for the correction of thermal drift according to some aspects.
  • FIG. 19 is a diagram illustrating a thermal drift correction process using a third order polynomial fitting according to some aspects.
  • FIG. 20 is a graph illustrating an original response corresponding to the thermal drift taken from a background measurement along with a corrected response according to some aspects.
  • FIG. 21 is another graph illustrating an original response corresponding to the thermal drift taken from a background measurement along with a corrected response according to some aspects.
  • FIG. 22 is another graph illustrating an original response corresponding to the thermal drift taken from a background measurement along with a corrected response according to some aspects.
  • FIG. 23 is a graph illustrating a combination of the two-decomposition corrections shown in FIGs. 21 and 22 at two different values of temperature change according to some aspects.
  • FIG. 24 is a diagram illustrating an example of a spectrometer including a plurality of sensors according to some aspects.
  • FIG. 25 is a flow chart illustrating a process for generating a correction matrix for on-line compensation of temperature drifts according to some aspects.
  • FIG. 26 is a diagram illustrating an example of a system configured to sweep the temperature of a spectrometer to generate a correction matrix according to some aspects.
  • FIG. 27 is a diagram illustrating another example of a system configured to sweep the temperature of a spectrometer to generate a correction matrix according to some aspects.
  • FIGs. 28A and 28B are diagrams illustrating another example of a system configured to sweep the temperature of a spectrometer to generate a correction matrix according to some aspects.
  • FIG. 29 is a diagram illustrating an example flow for generating the correction matrix including coefficients of fitted functions as the correction parameters, according to some aspects.
  • FIG. 30 is a graph illustrating an example of a third-order polynomial fitting for the thermal drift at the 1850 nm wavelength point of an NIR spectrometer according to some aspects.
  • FIG. 31 is a diagram illustrating an example of a system configured to generate a global correction matrix according to some aspects.
  • FIG. 32 is a diagram illustrating another example of a system configured to generate a global correction matrix according to some aspects.
  • FIG. 33 is a graph illustrating the thermal drift of a typical InGaAs PIN detector responsivity according to some aspects.
  • FIGs. 34A and 34B are graphs illustrating an AirPLS method used to separate the baseline first order drift from the dip higher order drift according to some aspects.
  • FIG. 35 is a graph illustrating the common (average) correction matrix for the dip drift per degree Celsius according to some aspects.
  • FIG. 36 is a graph illustrating a mechanism for finding the detector feature position using the slope of the detector drift according to some aspects.
  • FIG. 37 is a graph illustrating a detector dip feature position according to some aspects.
  • FIGs. 38A and 38B are graphs illustrating the extraction of detector higher-order thermal drift caused by cut-off shift with the temperature at different temperatures according to some aspects.
  • FIG. 39 is a diagram illustrating another example of a system configured to generate a global correction matrix according to some aspects.
  • a spectrometer may include a light modulator (e.g., an FT-IR interferometer, Fabry-Perot interferometer, diffraction grating, etc.), a detector, and a processor.
  • the processor may be configured to apply a correction matrix to an output of the detector to produce a corrected spectrum of a sample under test.
  • the processor may be configured to extract a set of correction parameters from the correction matrix and apply the set of correction parameters to the output of the detector (e.g., a measured spectrum of a sample) to compensate for instrumental drifts in the spectrometer.
  • the spectrometer may further include a sensor configured to obtain a value (e.g., a real-time value during operation of the spectrometer) of a condition contributing to instrumental drifts in the spectrometer.
  • the processor may then extract the set of correction parameters based on the value.
  • the value may represent an environmental condition (e.g., thermal drift or humidity drift) or other condition (e.g., light source inefficiency or optical misalignment).
  • the correction matrix may be generated during production of the spectrometer or may be based on a global correction matrix fitted to the spectrometer.
  • temperature sensing can be performed with the aid of a separate temperature sensor, a temperature sensor embedded in the control/detection electronics, or the photodetector may be exploited to sense the temperature based on the relation between the detector bandgap and the temperature. Temperature sensing can be performed during generation of the correction matrix or during application of the correction matrix to an output of the detector. In some examples, a heat source may be used to raise the temperature during the correction matrix generation process.
  • the heat source can be an external heat source, such as an oven or temperature chamber, or self-heating of the spectrometer device itself, which can be due to the light source of the spectrometer or the heat generated by the electronic components, such as current driver low-dropout regulators (LDOs), analog-to-digital converters (ADCs), or processor chips.
  • LDOs current driver low-dropout regulators
  • ADCs analog-to-digital converters
  • a feedback system uses the temperature reading (e.g., temperature value) to correct the measured spectrum using a set of correction parameters (e.g., a pre-calibrated correction function) based on the measured temperature.
  • the pre-calibrated correction function can be generated from a direct fitting algorithm, a decomposition algorithm, or the drift physical governing equations can be fed to a machine learning algorithm along with spectrometry raw data to create an adaptive correction model for the correction matrix.
  • the correction matrix can adapt automatically to each spectrometer unit. Updated correction matrices taking into account the aging effect over time can be extracted and fed back to spectrometer units through a cloud-based interface.
  • FIG. 1 is a diagram illustrating an example of a spectrometer 100 according to some aspects.
  • the spectrometer 100 may be, for example, a Fourier Transform infrared (FTIR) spectrometer that exploits light interference and Fourier transform to calculate the spectral content of an infrared light beam.
  • FTIR Fourier Transform infrared
  • the spectrometer 100 is a Michelson FTIR interferometer.
  • FTIR spectrometers measure a single-beam spectrum (power spectral density (PSD)), where the intensity of the single-beam spectrum is proportional to the power of the radiation reaching the detector.
  • PSD power spectral density
  • the background spectrum i.e., the single-beam spectrum in absence of a sample
  • the absorbance of the sample 112 may be calculated from the transmittance, reflectance, or trans-reflectance of the sample 112, the former being illustrated.
  • the absorbance of the sample 112 may be calculated as the ratio of the spectrum of transmitted light, reflected light, or trans -reflected light from the sample to the background spectrum.
  • the FT-IR spectrometer 100 includes a fixed mirror 106, a moveable mirror 108, a beam splitter 104, and a detector 114 (e.g., a photodetector).
  • a light source 102 associated with the spectrometer 100 is configured to emit an input beam and to direct the input beam towards the beam splitter 104.
  • the light source 102 may include, for example, a laser source, one or more wideband thermal radiation sources, or a quantum source with an array of light emitting devices that cover the wavelength range of interest.
  • the beam splitter 104 is configured to split the input beam into two beams. One beam is reflected off of the fixed mirror 106 back towards the beam splitter 104, while the other beam is reflected off of the moveable mirror 108 back towards the beam splitter 104.
  • the moveable mirror 108 may be coupled to an actuator 110 to displace the movable mirror 108 to the desired position for reflection of the beam. An optical path length difference (OPD) is then created between the reflected beams that is substantially equal to twice the mirror 108 displacement.
  • the actuator 110 may include a micro-electro-mechanical systems (MEMS) actuator, a thermal actuator, or other type of actuator.
  • MEMS micro-electro-mechanical systems
  • the reflected beams interfere at the beam splitter 104 to produce an output light beam, allowing the temporal coherence of the light to be measured at each different Optical Path Difference (OPD) offered by the moveable mirror 108.
  • OPD Optical Path Difference
  • the signal corresponding to the output light beam may be detected and measured by the detector 114 at many discrete positions of the moveable mirror 108 to produce an interferogram.
  • the detector 114 may include a detector array or a single pixel detector.
  • the interferogram data verses the OPD may then be input to a processor (not shown, for simplicity).
  • the spectrum may then be retrieved, for example, using a Fourier transform carried out by the processor.
  • the spectrometer 100 may be implemented as a MEMS interferometer (e.g., a MEMS chip).
  • the MEMS chip may be attached to a printed circuit board (PCB) that may include, for example, one or more processors, memory devices, buses, and/or other components.
  • PCB printed circuit board
  • MEMS refers to an actuator, a sensor, or the integration of sensors, actuators and electronics on a common silicon substrate through microfabrication technology to build a functional system.
  • Microelectronics are typically fabricated using an integrated circuit (IC) process, while the micromechanical components are fabricated using compatible micromachining processes that selectively etch away parts of the silicon wafer or add new structural layers to form the mechanical and electromechanical components.
  • IC integrated circuit
  • MEMS element is a micro-optical component having a dielectric or metallized surface working in a reflection or refraction mode.
  • Other examples of MEMS elements include actuators, detector grooves and fiber grooves.
  • the MEMS interferometer may be fabricated using a Deep Reactive Ion Etching (DRIE) process on a Silicon On Insulator (SOI) wafer in order to produce the micro-optical components and other MEMS elements that are able to process free-space optical beams propagating parallel to the SOI substrate.
  • DRIE Deep Reactive Ion Etching
  • SOI Silicon On Insulator
  • the electro-mechanical designs may be printed on masks and the masks may be used to pattern the design over the silicon or SOI wafer by photolithography. The patterns may then be etched (e.g., by DRIE) using batch processes, and the resulting chips (e.g., MEMS chip) may be diced and packaged (e.g., attached to the PCB).
  • the beam splitter 104 may be a silicon/air interface beam splitter (e.g., a half-plane beam splitter) positioned at an angle (e.g., 45 degrees) from the input beam.
  • the input beam may then be split into two beams LI and L2, where LI propagates in air towards the moveable mirror 108 and L2 propagates in silicon towards the fixed mirror 106.
  • LI originates from the partial reflection of the input beam from the half-plane beam splitter 104, and thus has a reflection angle equal to the beam incidence angle.
  • L2 originates from the partial transmission of the input beam through the half-plane beam splitter 104 and propagates in silicon at an angle determined by Snell’s Law.
  • the fixed and moveable mirrors 106 and 108 are metallic mirrors, where selective metallization (e.g., using a shadow mask during a metallization step) is used to protect the beam splitter 104.
  • the mirrors 106 and 108 are vertical Bragg mirrors that can be realized using, for example, DRIE.
  • the MEMS actuator 110 may be an electrostatic actuator formed of a comb drive and spring. For example, by applying a voltage to the comb drive, a potential difference results across the actuator 110, which induces a capacitance therein, causing a driving force to be generated as well as a restoring force from the spring, thereby causing a displacement of moveable mirror 108 to the desired position for reflection of the beam back towards the beam splitter 104.
  • FIG. 2 is a diagram illustrating another example of a spectrometer 200 according to some aspects.
  • the spectrometer 200 may be, for example, a Fabry-Perot spectrometer that includes a fixed mirror 206, a moveable mirror 208, and a detector 212 (e.g., a photodetector).
  • a light source 202 associated with the spectrometer 200 is configured to emit an input beam and to direct the input beam towards the fixed mirror 206.
  • the light source 202 may include, for example, a laser source, one or more wideband thermal radiation sources, or a quantum source with an array of light emitting devices that cover the wavelength range of interest.
  • a single-beam spectrum (power spectral density (PSD)) may be obtained using the Fabry-Perot spectrometer 200 shown in FIG. 2, where the intensity of the single-beam spectrum is proportional to the power of the radiation reaching the detector 212.
  • the background spectrum i.e., the single-beam spectrum in absence of a sample
  • the single-beam spectrum of light transmitted or reflected from the sample 204, the former being illustrated, may then be measured.
  • FIG. 3 is a diagram illustrating another example of a spectrometer 300 according to some aspects.
  • the spectrometer 300 may be, for example, a diffraction grating spectrometer that includes a diffraction grating 306 and a detector 308 (e.g., a photodetector).
  • a light source 302 associated with the spectrometer 300 is configured to emit an input beam and to direct the input beam towards the diffraction grating 306.
  • the light source 302 may include, for example, a laser source, one or more wideband thermal radiation sources, or a quantum source with an array of light emitting devices that cover the wavelength range of interest.
  • d is the periodicity of the grating
  • 0 m is the angle of diffracted beam
  • m is the order of diffraction.
  • the detector 308 may be a multi-pixel detector, as shown in FIG. 3, to detect the different intensity of light on every point on the detector 308 and convert that to an image that can then be processed to produce the light spectrum.
  • a single detector can be used.
  • a movable mirror or slit may be needed to direct each wavelength separately to the detector.
  • the resulting light spectrum produced by the diffraction grating spectrometer 300 corresponds to a power spectral density (PSD), where the intensity of the spectrum is proportional to the power of the radiation reaching the detector 308 at each point.
  • PSD power spectral density
  • the background spectrum i.e., the spectrum in absence of a sample
  • the spectrum of light transmitted or reflected from the sample 304, the former being illustrated, may then be measured.
  • Temperature variations can cause a change in instrument response including interferometer, optics, and photodetector response, which can be challenging to overcome even with post-processing of data. Stabilizing the temperature is one solution; however temperature stabilization may not be possible, especially in miniaturized spectrometers, as this affects size, power consumption, and possibility of sealing.
  • the photodetector is considered as the main source of thermal drifts in spectrometers, such as the spectrometers shown in FIGs. 1-3.
  • FIG. 4 is a diagram illustrating an example of a photodiode 400, which may be used in a photodetector of a spectrometer according to some aspects.
  • the photodiode 400 shown in FIG. 4 is a PIN photodiode that includes five main layers.
  • the first layer is the substrate 404 that carries the photodiode's other layers and acts as the N region of the PIN photodiode 400.
  • the second layer is a buffer layer 406 that prevents the migration of defects and impurities from the substrate 404 to the active layer.
  • the buffer layer 406 also serves as a matching layer to decrease the lattice mismatch between the active layer and the substrate 404.
  • the third layer is the intrinsic I active layer 408 where light is converted to electric current.
  • the fourth layer is the cap layer 410/412 that includes the P region.
  • the cap layer 410/412 should be built from a high bandgap material relative to the active layer 408 to prevent the surface recombination of photo-generated carriers. For example, if the cap layer 410/412 has a small bandgap, photons may be able to generate carriers in the cap layer 410/412; hence, these photons will be absorbed and will not reach the active layer 408.
  • these photo-generated carriers on the cap layer 410/412 may be lost as they may recombine quickly due to the high doping in the cap layer 410/412.
  • the last layer is a dielectric layer 414/416 that serves as a passivation, protective, and antireflection coating layer. Contacts 402 and 418 may be present on either side of the PIN photodiode 400.
  • Quantum efficiency of a photodiode is defined as the number of generated electrons due to electron-photon interaction, divided by the number of incident photons on the photodiode.
  • Responsivity .S’ of the photodiode is defined as the photodiode output electric power divided by the incident light radiant power. Therefore, quantum efficiency and responsivity are linked by the following equation: he 1 1240 where e is the electron charge, c is the speed of light, z is the wavelength, and h is the plank’s constant. Therefore, if r] is constant, the responsivity should be linear with wavelength.
  • FIG. 5A is a graph illustrating ideal responsivity for photodiodes with constant quantum efficiency rj.
  • FIG. 5B is a graph illustrating typical responsivities of silicon (Si) and InGaAs photodiodes.
  • FIG. 5B shows the deviation of responsivity from the linear behaviour for typical semiconductor photodetectors shown in FIG. 5A.
  • the responsivity decreases sharply to reach zero at the longer wavelengths when the photon energy is smaller than the bandgap, while the left side of each responsivity curve also deviates from the expected linear behaviour.
  • the responsivity r] may also be defined as the total photon flux divided by the total electrons flux as follows:
  • the generated electrons can be calculated from the following equation: where R is the reflection of photons from the top surface of the photo-diode, a c and d c are the absorption coefficient and the thickness of the cap layer respectively, a a and d a are the absorption coefficient and the thickness of the active layer respectively and is the fraction of absorbed photons in the active layer that leads to generation of useful electron flux. From the previous two equations,
  • Equation (6) has a contribution to the behaviour of the photodiode.
  • a typical AR coating may be able to decrease reflection to less than 0.5%.
  • every AR coating has a bandwidth (BW). Outside the BW, reflection starts to increase, hence, quantum efficiency decreases. Thus, one reason behind the responsivity deviation from the linear behaviour is the non-constant reflection along the whole spectrum.
  • the term (1 — e a a d a ⁇ describes the absorption that happens in the active layer. In some examples, this term may be made as big as possible (close to 1) since the absorption of photons in the active layer is the process that leads to electrons generation.
  • the absorption coefficient for some semiconductor materials used in the manufacturing of photodiodes may sharply increase when the photon energy exceeds the band-gap energy, especially the direct band-gap materials. This high absorption is due to the interband transition, which is the electron transition between valence and conduction bands. Indeed, electro-photon generation is an interband transition from valence to conduction band. Therefore, the responsivity of the detector approaches zero for high wavelengths when photon energy is less than the band gap.
  • the absorption coefficient for a direct band gap E9 transition is related to the energy gap by the following relation: where m is a material dependent parameter and it is equal to Vi in many materials, while the indirect band transition absorption coefficient is given by: a oc (hv — Eg nd ⁇ h Q ) 2 (8)
  • the Urbach tail is the representation of the non-abrupt decay of absorption beyond the cut-off wavelength. It is due to statistical broadening of the band gap energy. For example, band gap energy is not a sharp constant in every place of the material, instead, it has some sort of statistical distribution. This is due to the defects in the crystalline structure. Moreover, the change of the composition of alloy materials from place to place may cause this broadening. For example, alloy materials are deposited using epitaxial growth, which may have some variations that lead to this broadening.
  • Thermal absorption also leads to some absorption beyond the cut-off wavelength. However, its effect is weak compared to the Urbach tail. Thermal absorption can happen due to interband transition, exciton absorption below bandgap, or phonon absorption. By combining the aforementioned equations of absorption in direct and indirect band gap transitions, it can be deduced that absorption is increasing for the wavelengths shorter than the cut-off. This means that quantum efficiency should increase and the responsivity should be higher than expected by the linear model. However, the opposite happens. As can be seen in FIG. 5B, the InGaAs curve shows that the responsivity reaches a maximum value then decays faster than expected. Moreover, below a specific wavelength (about 0.97 pm), it decays exponentially.
  • the first decay is related to the term in the quantum efficiency equation. This term describes how many of the generated electrons will lead to useful electric current. For example, not all generated electrons may participate in the electric current, because some of them recombine before they reach the electrodes, and decreases for shorter wavelengths because of hot carriers’ effect.
  • electrons When electrons have higher energy than the bandgap, they can reach a higher position in the conduction band. However, this position is not stable; hence, they lose this excess energy gradually as phonons to reach the bandgap energy. This leads to a faster recombination rate of these hot electrons due to phonons assisted recombination. Furthermore, these hot electrons suffer from a higher rate of surface traps capturing, because their high energy enables them to hit the surface on a higher rate.
  • the last term in the quantum efficiency equation e acdc is the reason behind the exponential decay in responsivity that starts at the wavelength about 0.97 pm in the InGaAs responsivity curve in FIG. 5B.
  • This term describes the absorption in the cap layer, where a c is the absorption coefficient of the cap layer and d c is its thickness.
  • the cap layer 412 in the InGaAs photodiode shown in FIG. 4 is made of InP, a material with a cut-off wavelength of 0.97 pm (this cut-off wavelength changes depending on the cap layer martial).
  • a c Before the cut-off, a c is negligible and the cap layer is transparent, but after reaching the cut-off, a c rockets up and much of the light intensity is absorbed. Absorption in the cap layer does not create electric current, because the cap layer is highly p doped, hence, the generated electrons recombine very fast. This absorption may be decreased by decreasing the thickness d c of the cap layer; however, decreasing it lower than a certain value affects the reliability of the photodetector.
  • FIG. 6 is a graph illustrating the shift in cut-off wavelength with temperature for a semiconductor photodiode according to some aspects. As shown in FIG. 6, the cut-off wavelength of photodetectors shifts to higher wavelengths, as a result of the reduction in band gap energy with temperature.
  • FIG. 7 is a graph illustrating the spectral temperature coefficient of the responsivity of a photodiode according to some aspects.
  • the temperature coefficient versus wavelength can be divided into two regions.
  • the strong positive increase at the right side (right region) of the graph is due to the previously explained shift in cut-off to higher wavelengths with increasing temperature.
  • the left region, which has a negative temperature coefficient, is due to the increase of absorption in the cap layer.
  • the cap layer cut-off is located in the middle of the spectrum.
  • FIG. 8 is a graph illustrating the shift in the cut-off wavelength with temperature for a semiconductor photoconductive detector (e.g., an MIR PbSe photoconductive photodetector) according to some aspects. Since the bandgap of this material increases with temperature, the cutoff wavelength decreases with temperature.
  • a semiconductor photoconductive detector e.g., an MIR PbSe photoconductive photodetector
  • a spectrometer may be configured with a correction matrix including a plurality of sets of correction parameters (e.g., correction functions).
  • the correction matrix may be generated during assembly line production of the spectrometer or may be based on a global (e.g., average) correction matrix applicable to a plurality of spectrometers.
  • FIG. 9 is a diagram illustrating an example of a spectrometer 900 configured for on-line compensation of instrumental drift according to some aspects.
  • the spectrometer 900 includes a light modulator 906, a detector 910, and a processor 914.
  • the light modulator 906 may use a spectroscopic technique, including, but not limited to, direct absorption spectroscopy, indirect absorption spectroscopy, such as photo-acoustic spectroscopy, photo-thermal spectroscopy, or Raman spectroscopy.
  • the light modulator 906 may include a diffraction element, a Michelson interferometer, a Fabry-Perot cavity, a spatial light modulator, or a birefringent device.
  • the light modulator may include a MEMS interference device.
  • the detector 910 may correspond to, for example, to a photodetector.
  • a light source 902 may further be coupled to or incorporated within the spectrometer 900 (the former being illustrated).
  • the light source 902 may include, for example, a laser source, one or more wideband thermal radiation sources, or a quantum source with an array of light emitting devices that cover the wavelength range of interest.
  • the processor 914 may include a single processing device or a plurality of processing devices. Such a processing device may be a microprocessor, micro-controller, digital signal processor, microcomputer, central processing unit, field programmable gate array, programmable logic device, logic circuitry, analog circuitry, digital circuitry, and/or any device that manipulates signals (analog and/or digital) based on hard coding of the circuitry and/or operational instructions.
  • the processor 914 may be coupled to a memory, such as memory 920.
  • the memory 920 may be a single memory device, a plurality of memory devices, and/or embedded circuitry of the processor 914.
  • Such a memory device may be a read-only memory, random access memory, volatile memory, non-volatile memory, static memory, dynamic memory, flash memory, cache memory, and/or any device that stores digital information, including instructions (e.g., code) that may be executed by the processor 914.
  • the memory 920 may further store a correction matrix 922.
  • the correction matrix 922 may include a plurality of sets of correction parameters (e.g., correction functions) that may be utilized to provide on-line compensation of spectrum measurements.
  • the correction matrix 922 may provide on-line compensation due to variations in environmental conditions, including thermal (temperature) drift and humidity drift, and due to other sources of instrumental drift, such as light source inefficiency and optical misalignment.
  • the correction matrix 922 may include a plurality of correction matrices, each configured to provide on-line compensation due to a different condition.
  • the correction matrix 922 may include a temperature drift correction matrix and a humidity drift correction matrix.
  • the correction matrix 922 may include a temperature drift correction matrix, a humidity drift correction matrix, and a light source inefficiency correction matrix. Any combination of correction matrices is within the scope of the present disclosure.
  • the correction matrix 922 may be locally generated for the spectrometer 900 or based on a cloud-based (global) correction matrix applicable to a plurality of spectrometers.
  • the correction matrix 922 may be continuously or periodically updated via a machine learning (ML) or artificial intelligence (Al) engine, which may include a local AI/ML engine and/or a cloud-based AI/ML engine.
  • ML machine learning
  • Al artificial intelligence
  • the spectrometer 900 may further include other circuitry, such as control circuitry configured to control the light modulator 906 and light source(s) 902, powering circuitry configured to provide power to the various components of the spectrometer 900, communication circuitry (e.g., Wi-Fi, Bluetooth, cellular, etc.) configured to enable wireless communication with the spectrometer 900 (e.g., for cloud-based Al applications), and/or other suitable circuitry.
  • control circuitry configured to control the light modulator 906 and light source(s) 902
  • powering circuitry configured to provide power to the various components of the spectrometer 900
  • communication circuitry e.g., Wi-Fi, Bluetooth, cellular, etc.
  • wireless communication with the spectrometer 900 e.g., for cloud-based Al applications
  • the spectrometer 900 may further include one or more sensor(s) 916, each configured to obtain a respective value 918 of a condition contributing to instrumental drifts in the spectrometer 900.
  • the sensor 916 may be a separate sensor (e.g., separate from the light modulator 906, detector 910 and processor 914), a sensor embedded in or coupled to the light modulator 906, control/detection electronics, light source, or other component of the spectrometer 900, or may be realized as an algorithm that may be executed by the processor 914 to sense the value 918 based on a relation between the detector output and the condition being sensed. Examples of values include, but are not limited to, temperature values, humidity values, light source efficiency values, and optical misalignment values.
  • the light source 902 may be configured to generate input light 904 and to direct the input light to the light modulator 906.
  • the input light 904 may be directed to the light modulator 906 via an optical fiber and/or one or more optical elements (e.g., mirrors and/or focusing lenses).
  • the light modulator 906 may be configured to produce modulated light 908 based on the input light 904 and to direct the modulated light 908 towards the detector 910 to produce an output 912.
  • the modulated light 908 may be directed to the detector 910 via an optical fiber and/or one or more optical elements (e.g., mirrors and/or focusing lenses).
  • the modulated light 908 may correspond to interference beams produced over time with an OPD between beams.
  • the output 912 of the detector 910 may then correspond to an interferogram.
  • the modulated light 908 may correspond to diffracted light across a plurality of wavelengths.
  • the output 912 of the detector 910 may then correspond to an image representing the light intensity at each wavelength point on the detector 910.
  • the output 912 of the detector 910 may then be input to the processor 914 to produce a sample spectrum of a sample under test (not shown).
  • the processor 914 may apply a Fourier transform of the interferogram data produced by the detector and the OPD in the light modulator 906 to produce the spectrum.
  • the processor 914 may further be configured to correct the sample spectrum based on instrumental response drifts in the spectrometer 900 to produce a corrected spectrum 926.
  • the processor 914 may be configured to receive from the one or more sensor(s) 916 respective values 918 of one or more conditions measured by the sensor(s) 916 contributing to the instrumental drifts.
  • the processor 914 may further be configured to access the memory 920 to retrieve a respective set of correction parameters 924 from the correction matrix 922 based on each received value 918. The processor 914 may then be configured to apply the set(s) of correction parameters 924 to the sample spectrum to produce the corrected spectrum 926.
  • FIG. 10 is a diagram illustrating another example of a spectrometer 1000 configured for on-line compensation of instrumental drift according to some aspects.
  • the spectrometer 1000 includes a spectral sensor 1002 and a digital signal processing (DSP) unit 1018.
  • the spectral sensor 1002 includes a light source 1004 (e.g., a wideband light source), a light modulator chip 1008 (e.g., a light modulator), a detector 1012 (e.g., a photodetector), and an electronics interface 1016.
  • a light source 1004 e.g., a wideband light source
  • a light modulator chip 1008 e.g., a light modulator
  • detector 1012 e.g., a photodetector
  • the light modulator 1008 may use a spectroscopic technique, including, but not limited to, direct absorption spectroscopy, indirect absorption spectroscopy, such as photo-acoustic spectroscopy, photo-thermal spectroscopy, or Raman spectroscopy.
  • the light modulator 906 may include a diffraction element, a Michelson interferometer, a Fabry-Perot cavity, a spatial light modulator, or a birefringent device.
  • the light modulator may include a MEMS interference device.
  • the electronics interface 1016 may include the control circuitry to control the light modulator 1008 and light source 1004.
  • the electronics interface 1016 may be configured to provide a control signal 1040 to the light modulation chip 1008 to control operation of the light modulator.
  • the electronics interface 1016 provides an electrical connection to the DSP unit 1018.
  • the spectral sensor 1002 may further include one or more environmental conditions sensor(s) 1026, each configured to obtain a respective value 1028 of a condition contributing to instrumental drifts in the spectral sensor 1002.
  • the environmental conditions sensor 1026 may be a separate sensor (e.g., separate from the light source 1004, the light modulator 1008, and the detector 1012 or may be a sensor embedded in or coupled to the light modulator 1008, electronics interface 1016, light source 1004, or other component of the spectral sensor 1002.
  • values 1028 include, but are not limited to, temperature values, humidity values, light source efficiency values, and optical misalignment values.
  • the DSP unit 1018 includes an optical signal processing unit 1020, an artificial intelligence (AI)/machine learning (ML) engine 1030, and a memory 1032.
  • the optical signal processing unit 1020 may include a single processing device or a plurality of processing devices.
  • Such a processing device may be a microprocessor, micro-controller, digital signal processor, microcomputer, central processing unit, field programmable gate array, programmable logic device, logic circuitry, analog circuitry, digital circuitry, and/or any device that manipulates signals (analog and/or digital) based on hard coding of the circuitry and/or operational instructions.
  • the memory 1032 may be a single memory device or a plurality of memory devices.
  • Such a memory device may be a readonly memory, random access memory, volatile memory, non-volatile memory, static memory, dynamic memory, flash memory, cache memory, and/or any device that stores digital information, including instructions (e.g., code) that may be executed by the optical signal processing unit 1020.
  • instructions e.g., code
  • the memory 1032 may further store a correction matrix 1034.
  • the correction matrix 1034 may include a look up table of sets of correction parameters (e.g., correction functions) verses environmental conditions (or other conditions).
  • the sets of correction parameters in the correction matrix 1034 may provide on-line compensation due to variations in environmental conditions, including thermal (temperature) drift and humidity drift, and due to other sources of instrumental drift, such as light source inefficiency and optical misalignment.
  • the correction matrix 1034 may include a plurality of correction matrices, each configured to provide on-line compensation due to a different condition.
  • the correction matrix 1034 may be generated and/or updated using the AI/ML engine 1030.
  • the AI/ML engine 1030 may fully reside in hardware and/or software on the spectrometer 1000 or may be implemented using a cloud-based AI/ML engine.
  • a local AI/ML engine 1030 residing in hardware and/or software on the spectrometer may be in communication with (e.g., wireless communication) the cloud-based AI/ML engine.
  • the cloud-based AI/ML engine may provide access to one or more cloud-based correction matrices that may be downloaded into the memory 1032 and adapted (e.g., updated) for the spectrometer 1000 by the AI/ML engine 1030.
  • the AI/ML engine 1030 may use the output (e.g., values 1028) obtained by one or more environmental sensors 1026 at each sample measurement, along with the sample spectrum (e.g., power spectral density (PSD)) of each sample measurement to update the correction matrix 1034.
  • the sensor value e.g., an environment conditions measurement 1024
  • the environment conditions measurement 1024, together with the PSD 1022 may be input to the AI/ML engine 1030 to update the correction matrix 1034.
  • the DSP unit 1018 may include circuitry configured to execute the AI/ML engine 1030 (e.g., software or instructions for performing ML functions).
  • the DSP unit 1018 may include dedicated ML circuitry (e.g., one or more application specific integrated circuits (ASICs)) configured to perform one or more functions of the AI/ML engine 1030.
  • ASICs application specific integrated circuits
  • the light source 1004 may be configured to generate input light 1006 and to direct the input light to the light modulator 1008.
  • the input light 1006 may be directed to the light modulator 1008 via an optical fiber and/or one or more optical elements (e.g., mirrors and/or focusing lenses).
  • the light modulator 1008 may be configured to produce modulated light 1010 based on the input light 1006 and to direct the modulated light 1010 towards the detector 1012 to produce an output 1014.
  • the modulated light 1010 may be directed to the detector 1012 via an optical fiber and/or one or more optical elements (e.g., mirrors and/or focusing lenses).
  • the modulated light 1010 may correspond to interference beams produced over time with an OPD between beams.
  • the output 1014 of the detector 1012 may then correspond to an interferogram.
  • the modulated light 1010 may correspond to diffracted light across a plurality of wavelengths.
  • the output 1014 of the detector 1012 may then correspond to an image representing the light intensity at each wavelength point on the detector 1012.
  • the output 1014 of the detector 1012 may then be output from the electronics interface 1016 to the DSP unit 1018.
  • the optical signal processing unit 1020 within the DSP unit 1018 may process the output 1014 of the detector 1012 to produce a sample spectrum (e.g., a PSD) of a sample under test (not shown).
  • the optical signal processing unit 1020 may transform the interferogram 1014 provided by the detector 1012 to a PSD using a Fast Fourier transform (FFT).
  • FFT Fast Fourier transform
  • the optical signal processing unit 1020 may further be configured to correct the sample spectrum based on instrumental response drifts in the spectral sensor 1002 to produce a corrected spectrum (e.g., corrected optical PSD) 1038.
  • the optical signal processing unit 1020 may be configured to receive from the AI/ML engine 1030 a set of correction parameters (e.g., a correction signal) 1036 to be applied to the PSD to produce the corrected PSD 1038.
  • the AI/ML engine 1030 may receive from one or more environmental conditions sensor(s) 1026 and/or from the optical signal processing unit 1020, respective values 1028 and/or 1024 of one or more conditions measured by the sensor(s) 1026 and/or optical signal processing unit 1020 contributing to the instrumental drifts.
  • the AI/ML engine 1030 may further be configured to access the memory 1032 to retrieve a respective set of correction parameters 1036 from the correction matrix 1034 based on each received value 1028/1024.
  • the optical signal processing unit 1020 may then be configured to apply the set(s) of correction parameters 1036 to the sample spectrum (PSD) to produce the corrected spectrum (corrected PSD) 1038.
  • the correction matrix may be a temperature correction matrix that is based on the wavelength cut-off of the detector (e.g., detector 1012).
  • the cut-off of photodetectors responsivity shifts with temperature.
  • the shift takes place due to heat-assisted electron generation.
  • PSD power spectral density
  • FIGs. 11 A and 1 IB are diagrams illustrating the relation between temperature and wavelength according to some aspects.
  • FIG. 11A illustrates the PSD of the spectrometer at two different temperatures and the corresponding extracted 10% wavelengths. The relation may be described by the equation:
  • the slope A is nearly constant for the same type of detector, while C can be changing from spectrometer unit to spectrometer unit.
  • Higher order polynomials can be used as well to increase the accuracy if needed.
  • any criteria rather than the 10% wavelength can be used to define a threshold for the calculations of the shift.
  • the detector cut-off may be utilized in generating the correction matrix by extracting the change in temperature T from the change in the cut-off wavelength.
  • the temperature T may be measured using a dedicated temperature sensor. In the former case, in order to discern the temperature, the parameter C or other higher order polynomial fitting parameters can be calibrated for each spectrometer unit using a single calibration measurement at room temperature.
  • FIG. 12 is a flow chart illustrating a process for on-line compensation of temperature drifts in a spectrometer according to some aspects. As described below, some or all illustrated features may be omitted in a particular implementation within the scope of the present disclosure, and some illustrated features may not be required for implementation of all examples. In some examples, the process may be performed by a processor, processing unit, or processing system of a spectrometer, or by any suitable means for carrying out the described functions.
  • the processor may obtain a PSD based on an output of a detector of a spectrometer.
  • the processor may obtain a temperature sensor reading.
  • the temperature sensor reading may be a temperature value provided by a dedicated temperature sensor within the spectrometer.
  • the temperature sensor reading (indicated as a change in temperature) may be extracted by the processor based on a change in the cut-off wavelength of the spectrometer.
  • the processor may enter the temperature value into an algorithm executed by the processor for correcting the PSD.
  • the processor may select a set of correction parameters from a stored correction matrix based on the temperature value.
  • the selected set of correction parameters may indicate a correction function to be applied to the PSD.
  • the processor may calculate a drift percentage (e.g., a percentage of drift of the PSD) from the correction parameters.
  • the processor may divide the original PSD obtained at block 1202 by the calculated drift percentage to obtain a corrected PSD.
  • FIG. 13 is a diagram illustrating an example of a spectrometer 1300 employing the detector as a temperature sensor to be used for the correction of thermal drift according to some aspects.
  • the spectrometer 1300 includes a light source 1302 (e.g., a wideband light source), a light modulator 1306, a detector 1310, an electronics interface 1314, a processor 1316 and a memory 1320.
  • the light modulator 1306 may use a spectroscopic technique, including, but not limited to, direct absorption spectroscopy, indirect absorption spectroscopy, such as photo- acoustic spectroscopy, photo-thermal spectroscopy, or Raman spectroscopy.
  • the light modulator 1306 may include a diffraction element, a Michelson interferometer, a Fabry-Perot cavity, a spatial light modulator, or a birefringent device.
  • the light modulator may include a MEMS interference device.
  • the detector 1310 may correspond to, for example, to a photodetector.
  • the memory 1320 may further store a correction matrix 1324.
  • the correction matrix 1324 may include a look up table of sets of correction parameters (e.g., correction functions) verses temperature (T). [0107] Thermal drift in the spectrometer 1300 may occur, for example, due to heat excitation raising the temperature of the detector 1310 during operation.
  • the heat source can be the environmental conditions or heat generated internally in the spectrometer 1300 during operation.
  • a PSD can be generated by the processor 1316 from the detector output 1312 versus wavelength.
  • the light modulator 1306 may be an interferometer and the detector output may be the interferogram signal versus the interferometer mirror motion (OPD).
  • OPD interferometer mirror motion
  • the PSD may be obtained.
  • An analog-to-digital converter (ADC) may be used to digitize the detector output.
  • the processor 1316 can extract the cut-off wavelength 1318 from the PSD curve and feed the cut-off wavelength 1318 to the memory 1320. Based on the stored relation between the temperature and the cut-off wavelength, which can be a simple linear relation as mentioned before, the processor 1316 can extract the temperature 1322 and feed the temperature 1322 to the correction matrix 1324 stored in the memory 1320. Based on the temperature 1322, the processor 1316 can select the set of correction parameters 1326 from the correction matrix 1324 (e.g., from a look-up table), which is fed back to the processor 1316. The processor 1316 can then use the previously obtained PSD and the set of correction parameters 1326 to obtain a new PSD 1328 that is corrected from the temperature drift.
  • the processor 1316 can select the set of correction parameters 1326 from the correction matrix 1324 (e.g., from a look-up table), which is fed back to the processor 1316.
  • the processor 1316 can then use the previously obtained PSD and the set of correction parameters 1326 to obtain a new PSD 1328 that is corrected from the temperature drift
  • FIG. 14 is a diagram illustrating another spectrometer 1400 employing the detector as a temperature sensor to be used for the correction of thermal drift according to some aspects.
  • the spectrometer 1400 includes a light source 1402 (e.g., a wideband light source), a light modulator 1406, a detector 1410, an electronics interface 1414, a processor 1416 and a memory 1420.
  • the light modulator 1406 may use a spectroscopic technique, including, but not limited to, direct absorption spectroscopy, indirect absorption spectroscopy, such as photo- acoustic spectroscopy, photo-thermal spectroscopy, or Raman spectroscopy.
  • the light modulator 1406 may include a diffraction element, a Michelson interferometer, a Fabry-Perot cavity, a spatial light modulator, or a birefringent device.
  • the light modulator may include a MEMS interference device.
  • the detector 1410 may correspond to, for example, to a photodetector.
  • the memory 1420 may further store a correction matrix 1424.
  • the correction matrix 1424 may include a look up table of sets of correction parameters (e.g., correction functions) verses temperature (T).
  • a PSD 1418 can be generated by the processor 1416 from the detector output 1412 versus wavelength.
  • the temperature can be extracted from PSD measurement 1418 with the aid of a PLS method or other similar method.
  • the temperature 1422 can be calculated from any PSD 1418 for the same sample without the need for calculating the detector-cut-off wavelength.
  • the processor 1416 can select the set of correction parameters 1426 from the correction matrix 1424 (e.g., from a look-up table), which is fed back to the processor 1416.
  • the processor 1416 can then use the previously obtained PSD and the set of correction parameters 1426 to obtain a new PSD 1428 that is corrected from the temperature drift.
  • FTIR spectrometers may include a temperature sensor to monitor the temperature of the spectrometer and shut it down if the temperature exceeds a specified value. Temperature sensor readings can be also used to extract the temperature of one of the spectrometer components, where:
  • the spectrometer 1500 includes a light source 1502 (e.g., a wideband light source), a light modulator 1506, a detector 1510, an electronics interface 1514, a processor 1516 and a memory 1520.
  • the light modulator 1506 may use a spectroscopic technique, including, but not limited to, direct absorption spectroscopy, indirect absorption spectroscopy, such as photo- acoustic spectroscopy, photo-thermal spectroscopy, or Raman spectroscopy.
  • the light modulator 1506 may include a diffraction element, a Michelson interferometer, a Fabry-Perot cavity, a spatial light modulator, or a birefringent device.
  • the light modulator may include a MEMS interference device.
  • the detector 1510 may correspond to, for example, to a photodetector.
  • the memory 1520 may further store a correction matrix 1524.
  • the correction matrix 1524 may include a look up table of sets of correction parameters (e.g., correction functions) verses temperature (T).
  • a PSD can be generated by the processor 1516 from the detector output 1512 versus wavelength.
  • a temperature sensor 1518 is configured to measure a temperature value 1530 (e.g., change in temperature value) of the detector 1510 (or other component of the spectrometer 1500). The measured temperature value 1530 (T se nsor) may be fed to the memory 1520.
  • the processor 1516 can extract the detector temperature 1522 and feed the detector temperature 1522 to the correction matrix 1524 stored in the memory 1520. Based on the detector temperature 1522, the processor 1516 can select the set of correction parameters 1526 from the correction matrix 1524 (e.g., from a look-up table), which is fed back to the processor 1516. The processor 1516 can then use the previously obtained PSD and the set of correction parameters 1526 to obtain a new PSD 1528 that is corrected from the temperature drift.
  • calibration of the temperature sensor 1518 can be carried out by a few measurements to enable the electronics and spectrometer to heat up, while monitoring the temperature of the spectrometer 1500 using a calibrated temperature sensor (not shown) and comparing it to the reading of the temperature sensor 1518.
  • the temperature sensor 1518 may be a thermistor or a thermocouple to measure the spectrometer temperature.
  • these type of temperature sensors typically utilize a specific electronic interface that consumes power and needs a hosting package that consumes area too. Therefore, this configuration may be costly, but does provide the flexibility to choose the location of the temperature sensor in the optimal location minimizing the error between the actual temperature of the component of interest and the predicted temperature from the sensor reading.
  • FIG. 16 is a diagram illustrating another example of a spectrometer 1600 employing a dedicated temperature sensor to be used for the correction of thermal drift according to some aspects.
  • the spectrometer 1600 includes a light source 1602 (e.g., a wideband light source), a light modulator 1606, a detector 1610, an electronics interface 1614 (e.g., an ASIC chip), a processor 1618, and a memory 1622.
  • the light modulator 1606 may use a spectroscopic technique, including, but not limited to, direct absorption spectroscopy, indirect absorption spectroscopy, such as photo-acoustic spectroscopy, photo-thermal spectroscopy, or Raman spectroscopy.
  • the light modulator 1606 may include a diffraction element, a Michelson interferometer, a Fabry- Perot cavity, a spatial light modulator, or a birefringent device.
  • the light modulator may include a MEMS interference device.
  • the detector 1610 may correspond to, for example, to a photodetector.
  • the memory 1622 may further store a correction matrix 1626.
  • the correction matrix 1626 may include a look up table of sets of correction parameters (e.g., correction functions) verses temperature (T).
  • a PSD can be generated by the processor 1618 from the detector output 1612 versus wavelength.
  • a temperature sensor 1620 configured to measure a temperature value 1628 (e.g., change in temperature value) of the spectrometer 1600 is not a separate chip, but rather a sensor that is embedded in the electronics interface 1614.
  • the electronics interface 1614 is a central chip including an ADC 1616 for converting the output 1612 of the detector 1610 to a digital signal for input to the processor 1618 and the actuation circuitry 1634.
  • the temperature sensor 1620 may be embedded in a processing unit chip including the processor 1618 or a separate chip providing the actuation signal 1636.
  • the measured temperature value 1628 may be fed to the memory 1622.
  • the processor 1618 can extract the detector temperature 1624 and feed the detector temperature 1624 to the correction matrix 1626 stored in the memory 1622.
  • the processor 1618 can select the set of correction parameters 1630 from the correction matrix 1626 (e.g., from a look-up table), which is fed back to the processor 1618.
  • the processor 1618 can then use the previously obtained PSD and the set of correction parameters 1630 to obtain a new PSD 1632 that is corrected from the temperature drift.
  • FIG. 17 is a diagram illustrating another example of a spectrometer 1700 employing a dedicated temperature sensor to be used for the correction of thermal drift according to some aspects.
  • the spectrometer 1700 includes a spectral sensor 1702 including a light source 1704 (e.g., a wideband light source), a light modulator 1708 (e.g., a silicon (Si) light modulation chip), a detector 1716 and an electronics interface 1720.
  • a light source 1704 e.g., a wideband light source
  • a light modulator 1708 e.g., a silicon (Si) light modulation chip
  • the light modulator 1708 includes an optical MEMS device 1712 (e.g., an FT-IR MEMS-based interferometer) and light guides (e.g., optical fibers) 1710a and 1710b for guiding input light 1706 from the light source 1704 into the optical MEMS device 1712 and guiding modulated light 1714 (e.g., interference beams over time with OPDs therebetween) from the optical MEMS device 1712 to the detector 1716.
  • the light modulator 1708 may further be actuated by a control signal 1726 (e.g., actuation signal) supplied by the electronics interface 1720.
  • the detector 1716 may correspond to, for example, to a photodetector.
  • the output 1718 of the detector 1716 (e.g., an interferogram) may be provided to the electronics interface 1720, which may provide the interferogram 1718 (e.g., a digital signal representing the interferogram) to a processor (not shown).
  • a temperature sensor 1722 configured to measure a temperature value 1724 (e.g., change in temperature value) of the spectrometer 1700 is fabricated on the light modulation chip 1708.
  • the temperature sensor 1722 may be a piezoresistive temperature sensor fabricated on the same MEMS chip used in the light modulation element.
  • the piezoresistive temperature sensor 1722 can be made of a double-layer cantilever beam, where each layer has a different thermal expansion coefficient. When the temperature changes, the cantilever beam will bend up or down based on the sign of the temperature change. This bending will cause strain.
  • the resistance of the piezoresistive element will change and, hence, the temperature can be measured.
  • the piezoresistive element can be made on the crystalline silicon of the silicon wafer of the light modulation chip 1708 using, for example, ion implantation of p-type impurities. In some examples, four piezoresistive elements can be used to form a Wheatstone bridge for better resistance change sensing.
  • the temperature sensor 1722 By placing the temperature sensor 1722 on the light modulation chip 1708, the resulting temperature value 1722 reflects the actual temperature value of the light modulator 1708.
  • fabrication of a piezoresistive temperature sensor is compatible with light modulation element fabrication, as it requires a few simple surface micromachining steps.
  • Each photodetector has a shunt resistance. Shunt resistance decreases with temperature increase. If this resistance is calculated and calibrated with temperature values, it can be used to calculate the temperature. However, it may not be practical to measure the shunt resistance of the detector while being in operation. It is also known that the thermal noise of the detector is proportional to shunt resistance. Therefore, calculating the noise level can be used as a calibration for shunt resistance, and hence, a calibration for temperature. In order to calculate the root mean square (RMS) noise level, and consequently extracting the corresponding temperature, a dark measurement is obtained by the photodetector.
  • RMS root mean square
  • FIGs. 18A and 18B are diagrams illustrating another example of of a spectrometer 1800 employing the detector as a temperature sensor to be used for the correction of thermal drift according to some aspects.
  • the spectrometer 1800 includes a light source 1806 (e.g., a wideband light source), a light modulator 1808, a detector 1810, an electronics interface 1822, and a DSP Unit 1826 (e.g., a processor).
  • the relation between dark measurements and temperature may be calibrated using a temperature sensor 1830 embedded in or coupled to the spectrometer 1800.
  • a temperature reading 1814 is obtained by the temperature sensor 1830.
  • the RMS noise level may be calculated (e.g., by the processor/DSP unit 1826) and both the RMS noise level and corresponding temperature reading may be stored in a matrix 1816.
  • the matrix 1816 may then be fed into a regression module 1818 to extract the relation 1820 between the RMS noise level and temperature.
  • the relation 1820 may be stored, for example, within a memory of the spectrometer 1800 to be used during on-line compensation of illuminated spectral measurements.
  • a spectral sensor 1802 including a light modulation module 1832 including, for example, at least the light modulator 1808 and the detector 1810 may be configured to obtain a dark interferogram 1812 that is provided by the electronics interface 1822 to the DSP unit 1826.
  • the dark interferogram 1812 may be obtained, for example, immediately prior to obtaining an illuminated interferogram 1824 of a sample under test.
  • the DSP unit 1826 can extract the temperature 1814 and feed the temperature 1814 to a correction matrix (not shown) to select a set of correction parameters therefrom.
  • the DSP unit 1826 can then use the illuminated interferogram 1824 (e.g., the PSD obtained from the illuminated interferogram) and the set of correction parameters to obtain a new PSD 1828 that is corrected from the temperature drift.
  • the application of the set of correction parameters may include three steps.
  • the first step includes measuring the spectrometer temperature using any of the aforementioned mechanisms.
  • the temperature value can be delivered as a feedback to tailor the correction matrix that corresponds to the spectrometer temperature.
  • the tailored correction matrix is applied to the original PSD to produce the corrected PSD.
  • the stored correction matrix can be retrieved for any temperature measurement and the temperature drift can then be calculated at all wavelengths for the measured temperature.
  • the thermal drift is additive in absorbance and multiplicative in reflection, transmission and PSD domains. Therefore, the correction can be applied on the PSD itself, and then transformed into the needed response to save processing time.
  • the correction can be done according to the following equation: where S dpd is the spectrum of drift per degree Celsius.
  • FIG. 19 is a diagram illustrating a thermal drift correction process using a third order polynomial fitting according to some aspects.
  • a set of correction parameters e.g., a correction function
  • a compensation unit 1908 e.g., a processor
  • the stored correction matrix can be retrieved and tailored for correction according to the following equation: where a and b are equal to fitting parameters that can be set to unity if not needed. However, they can be used as a tuning factor for better correction to account for the variations from spectrometer unit to spectrometer unit. As the drift as a function of temperature may change from unit to unit, thus, a and b can be used to account for a nonlinear temperature-drift relation. Moreover, the position of the dip changes across units, but it can be determined using a few measurements, then a shift (l sftl ) in the position of the correction matrix can be performed to produce a global correction matrix. As can be seen in FIG.
  • the global correction matrix was applied on 130 miniaturized near infrared (NIR) spectrometers and the dip drift decreased to less than 15% of its value in most of them.
  • NIR near infrared
  • the detector cut-off shift feature is not linear through the temperature range. Therefore, the temperature range can be divided into regions and a piecewise correction function can be used to cover the whole temperature range and account for the change in detector cut-off shift feature in each region.
  • the correction function can be retrieved according to the temperature value measured by one of the aforementioned mechanisms. For example, temperatures from 20 degrees to 25 degrees can have a specific correction function. Then, correction can be applied according to the following equation:
  • a shift (A sh2 ) in the position of the correction matrix is needed as in the case of dip correction.
  • a and b may be used to improve correction, as in the case of dip correction. This correction was applied on 102 units and cut-off drift decreased to less than 15% of its value in most of the miniaturized spectrometers.
  • FIG. 23 is a graph illustrating a combination of the two-decomposition corrections shown in FIGs. 21 and 22 at two different values of AT. As can be seen in FIG. 23, the higher order drifts were corrected and the only remaining drift component is a first order baseline drift that can be removed by simple pre-processing before chemometrics analysis.
  • Background measurements may enable the instrumental response to be removed, thus allowing only the sample response to be measured. Background measurements typically have approximately the same instrumental response as the sample measurements, though as discussed herein, the instrumental response may also change with environmental conditions. In some examples, a background measurement may be taken every certain period of time, e.g., one hour. However, some applications require using the same background measurement for a longer period of time. Hence, a generalized background measurement may be acquired by measuring all the instrumental response drift causes, making models for them and compensating these drifts to restore the same instrumental response of the background measurement. [0134]
  • the main conditions contributing to instrumental response change include, as discussed above, temperature variations, humidity variations, light source efficiency and optical misalignment. For example, humidity can cause some absorption features and photometric errors. Optical misalignment can cause baseline drifts especially in the shorter wavelengths. In some examples, models (e.g., correction matrices) for these other conditions can be generated in the same way used for temperature drift compensation discussed above.
  • FIG. 24 is a diagram illustrating an example of a spectrometer 2400 including a plurality of sensors according to some aspects.
  • the spectrometer 2400 includes a spectral sensor 2402, an electronics interface 2416, a processor 2418, a plurality of sensors 2420, 2422, and 2424, and a memory 2426.
  • the plurality of sensors include different sensors 2420, 2422, and 2424 for temperature, humidity, and source consumption and optical alignment (e.g., light source efficiency and misalignment), respectively.
  • the memory 2426 may store a correction matrix 2432 including, for example, a lookup table of correction parameters mapped to sensor readings.
  • the memory 2426 may further store a factory-saved background measurement 2428 of a stored background measurement.
  • the stored background measurement 2428 may be modified to include current variations with respect to stored background conditions.
  • the spectral sensor 2402 includes a light source 2404 (e.g., a wideband light source), a light modulator 2408, and a detector 2412.
  • the light source 2404 may be configured to direct input light 2406 to the light modulator 2408.
  • the light modulator 2408 may then produce modulated light 2410 (e.g., interference beams over time with OPDs therebetween) based on the input light 2406 and may further direct the modulated light 2410 to the detector 2412.
  • the light modulator 2408 may further be actuated by an actuation signal 2438 provided by an electronics interface 2416.
  • the detector 2412 may correspond to, for example, to a photodetector.
  • the output 2414 of the detector 2412 (e.g., an interferogram) may be provided to the electronics interface 2416, which may provide the interferogram 2414 (e.g., a digital signal representing the interferogram) to the processor 2418.
  • the processor 2418 may process the output 244 of the detector 2412 to produce a sample spectrum (e.g., a PSD) of a sample under test (not shown). For example, the processor 2418 may transform the interferogram 2414 provided by the detector 2412 to a PSD using a Fast Fourier transform (FFT). The processor 2418 may further be configured to correct the sample spectrum based on instrumental response drifts in the spectral sensor 2402 and the factory stored background 2428 to produce a corrected spectrum (e.g., corrected optical PSD) 2436. For example, one or more of the sensors 2420, 2422, and/or 2424 may obtain a respective sensor reading (e.g., value), which may be fed to a sensor readings reception algorithm 2430 in the memory 2426.
  • a respective sensor reading e.g., value
  • the processor 2418 may be configured to execute the sensor readings reception algorithm 2430 to feed the sensor reading(s) to the correction matrix 2432 stored in the memory 2426. Based on the sensor reading(s), the processor 2418 can select the set(s) of correction parameters 2434 from the correction matrix 2432 (e.g., from a look-up table), which is fed back to the processor 2418. The processor 2418 can then use the previously obtained PSD, the factory-saved background measurement 2428, and the set(s) of correction parameters 2434 to obtain a new PSD 2436 that is corrected from the instrumental drift(s).
  • Creating the correction matrix involves not only the total measured intensity at the detector output, but also the power spectral density (PSD), at each wavelength. Since the drift may change from spectrometer unit to spectrometer unit, a unique correction matrix can be created for each unit.
  • the correction matrix can be calculated by measuring the instrumental response while sweeping across temperature. The responses can then be stored and processed to obtain a correction function (e.g., indicated by the correction matrix) that relates the thermal drift of the PSD at different wavelengths across temperatures.
  • FIG. 25 is a flow chart illustrating a process for generating a correction matrix for on-line compensation of temperature drifts according to some aspects. As described below, some or all illustrated features may be omitted in a particular implementation within the scope of the present disclosure, and some illustrated features may not be required for implementation of all examples. In some examples, the process may be performed by a processor, processing unit, or processing system of a spectrometer, or by any suitable means for carrying out the described functions.
  • the processor may take a spectrometer measurement (e.g., a PSD) of a sample.
  • the spectrometer temperature may be measured (e.g., by a dedicated temperature sensor within the spectrometer).
  • the temperature value may be assigned to the spectrometer measurement.
  • the processor may determine whether the temperature range of interest is covered. If the temperature range of interest has not been covered (N branch of block 2508), at block 2510, the spectrometer may be heated and the process repeated to assign a respective temperature value across the range of interest to each spectrometer measurement. Once the temperature range of interest has been covered (Y branch of block 2508), at block 2512, the processor may extract the relation between drift and temperature, and at block 2514, store the extracted relation as the correction matrix.
  • FIG. 26 is a diagram illustrating an example of a system 2600 configured to sweep the temperature of a spectrometer 2602 to generate a correction matrix according to some aspects.
  • the spectrometer 2602 includes a light source 2606, a light modulator 2608, and a detector 2610.
  • the spectrometer 2602 may be inserted into an oven or a temperature chamber 2604 having a controller 2622 configured to control the spectrometer temperature by controlling the chamber temperature.
  • a temperature sensor 2624 in the chamber 2604 may be configured to measure a temperature (e.g., obtain a temperature reading 2614) of the chamber 2604/spectrometer 2602 for each spectrometer measurement 2612 (e.g., PSD verses wavelength at that temperature). Both the spectrometer measurement 2612 and the corresponding temperature reading 2614 may be stored in a matrix 2616. To cover all possible temperature conditions, the matrix 2616 may then be fed into a regression module 2618 to extract the relation 2620 between the spectrometer measurement 2612 and temperature 2614. The relation 2620 may be stored, for example, as the correction matrix within a memory of the spectrometer 2602 to be used during on-line compensation.
  • FIG. 27 is a diagram illustrating another example of a system 2700 configured to sweep the temperature of a spectrometer 2702 to generate a correction matrix according to some aspects.
  • the spectrometer 2702 includes a light source 2706, a light modulator 2708, and a detector 2710.
  • a Peltier element in a thermos -electric cooler or heater (TEC) 2704 may be attached to the detector 2710 or placed in close proximity thereto.
  • TEC thermos -electric cooler or heater
  • a temperature sensor 2722 attached to or embedded in the spectrometer 2702 may be configured to measure a temperature (e.g., obtain a temperature reading 2714) of the spectrometer 2702 for each spectrometer measurement 2712 (e.g., PSD verses wavelength at that temperature). Both the spectrometer measurement 2712 and the corresponding temperature reading 2714 may be stored in a matrix 2716. To cover all possible temperature conditions, the matrix 2716 may then be fed into a regression module 2718 to extract the relation 2720 between the spectrometer measurement 2712 and temperature 2714. The relation 2720 may be stored, for example, as the correction matrix within a memory of the spectrometer 2702 to be used during on-line compensation. [0144] FIGs.
  • FIGs. 28A and 28B are diagrams illustrating another example of a system 2800 configured to sweep the temperature of a spectrometer 2802 to generate a correction matrix according to some aspects.
  • the heat energy emerging from electronics e.g., electronic components 2806 producing heat
  • a light source 2804 in the spectrometer 2802 can be exploited to raise the temperature of temperature sensitive elements 2810 in the spectrometer 2802, such as the photodetector or the light modulation/MEMS chip. Heating can further be accelerated by using a thermally insulating enclosure 2812 to prevent heat from escaping outside the spectrometer 2802.
  • a temperature sensor 2824 within the enclosure 2812 and/or attached to or embedded in the spectrometer 2802 may be configured to measure a temperature (e.g., obtain a temperature reading 2816) of the spectrometer 2802 for each spectrometer measurement 2814 (e.g., PSD verses wavelength at that temperature). Both the spectrometer measurement 2814 and the corresponding temperature reading 2816 may be stored in a matrix 2818. To cover all possible temperature conditions, the matrix 2818 may then be fed into a regression module 2820 to extract the relation 2822 between the spectrometer measurement 2814 and temperature 2816. The relation 2822 may be stored, for example, as the correction matrix within a memory of the spectrometer 2802 to be used during on-line compensation.
  • the correction matrix can include coefficients of fitted functions that relate the PSD change of each wavelength/wavenumber point with temperature. These fitted functions can be polynomial, exponential, or others, depending on the thermal drift characteristics of the instrument.
  • FIG. 29 is a diagram illustrating an example process flow for generating the correction matrix including coefficients of fitted functions as the correction parameters, according to some aspects.
  • a control module 2902 e.g., a controller
  • the control module 2902 may be configured to control a scan time or frequency of capturing spectrometer measurements and corresponding temperature values.
  • control module 2902 may be configured to perform continuous or periodic multiple scans of the spectrometer 2904.
  • the spectrometer measurements e.g., PSD
  • temperature (T) 2906 may be output by the spectrometer 2904 and provided to a post processing unit 2908 configured to calculate the correction matrix using an N lh order polynomial fitting.
  • the correction matrix 2912 may then be output by the post processing unit 2908 and stored in a memory (not shown) of the spectrometer 2904 by a memory controller 2910.
  • the correction matrix 2912 may include the drift fitting coefficients for each wavelength point as a function of temperature (e.g., based on a polynomial fitting relation).
  • FIG. 30 is a graph illustrating an example of a third-order polynomial fitting for the thermal drift at the 1850 nm wavelength point of an NIR spectrometer according to some aspects.
  • the minimum temperature of -10 degrees C is taken as a reference.
  • a general (global) correction matrix can be calculated for the common or mean behaviour of the spectrometer units and production line testing of each spectrometer unit may focus on the variation between spectrometer units that can be extracted in a relatively short time.
  • the general correction matrix and the features that vary from spectrometer unit to spectrometer unit may be extracted with a machine learning algorithm, which may use less extensive and less time-consuming testing to extract the non-common features that are unique to each unit.
  • FIG. 31 is a diagram illustrating an example of a system 3100 configured to generate a global correction matrix according to some aspects.
  • each of a plurality of spectrometer units 3102a, 3102b, ..., 3102N may obtain spectrometer measurement/temperature data 3108 (e.g., PSD verses temperature) and transmit the spectrometer measurement/temperature data 3108 over a network (e.g., via cloud 3106) to an AI/ML engine 3104 (e.g., a neural network).
  • the AI/ML engine 3104 may be configured to fit a respective correction matrix for each spectrometer unit 3102a, 3102b, . .
  • the AI/ML engine 3104 can further be configured to modify the global correction matrix for each spectrometer unit 3102a, 3102b, . . ., 3102N using minimal spectrometer measurement/temperature data 3108 from each spectrometer unit 3102a, 3102b, . . . , 3102N to produce respective updated correction matrices 3112 for each spectrometer unit 3102a, 3102b, . . ., 3102N.
  • FIG. 32 is a diagram illustrating another example of a system 3200 configured to generate a global correction matrix according to some aspects.
  • each of a plurality of spectrometer units 3202a, 3202b, ..., 3202N may obtain spectrometer measurement/temperature data 3206 (e.g., PSD verses temperature) and transmit the spectrometer measurement/temperature data 3206 over a network (e.g., via cloud 3205) to an AI/ML engine (e.g., a neural network), represented by boxes 3208 and 3210.
  • a network e.g., via cloud 3205
  • AI/ML engine e.g., a neural network
  • the AI/ML engine at box 3208 may be configured to fit a respective correction matrix for each spectrometer unit 3202a, 3202b, ..., 3202N and produce an average correction matrix (e.g., a global correction matrix) based on the respective correction matrices.
  • the AI/ML engine at box 3210 can further be configured to adaptively scale the average correction matrix using two spectrometer measurements at two different temperatures for each spectrometer unit 3202a, 3202b, ..., 3202N and compare the drift at each wavelength point with the expected drift from the average correction matrix. Then, the ratio between the actual and expected drifts can be used to scale the average correction matrix to produce a respective updated correction matrix 3212 for each of the spectrometer units 3202a, 3202b, ...
  • an average correction matrix can be calculated for each production lot to account for the differences in drifts from lot to lot, hence, decreasing the error in correction. In some examples, this may decrease the testing time per unit as fewer measurements may be needed to spot the drift variations for each spectrometer unit.
  • the thermal drift of spectrometers includes different superimposed components. Some of these components are approximately constant for all units, for example the detector drift features, while others vary from one spectrometer to another, for example the MEMS drift features in miniaturized spectrometers. It is possible to decompose the drift components using wavelet transform, polynomial fitting, morphological weighted penalized least squares, adaptive iterative reweight penalized least squares, automatic iterative moving average, asymmetric least squares or any another decomposition method. The example below is based on an adaptive iteratively reweighted penalized least squares (AirPLS) algorithm.
  • AirPLS adaptive iteratively reweighted penalized least squares
  • the PLS algorithm has a roughness term:
  • the PLS algorithm has a fidelity term: where x is the vector of the measured spectrum, z is the fitted vector, and w is the fidelity weight.
  • Q F + IR
  • I a user defined parameter that controls the roughness of the fitted curve.
  • a traditional PLS algorithm solves the last equation to get the fitted curve z and by changing I it can be used as a smoothing method. However, if W can be tuned to detect the maximas and prevent the fitted curve from following them, PLS can be used as a baseline extraction method.
  • AirPLS is the simplest and most effective in that AirPLS uses an iterative algorithm to adaptively change the value of W in the following equation: where W is defined according to the following iteration equation [0155]
  • W is defined according to the following iteration equation [0155]
  • a correction matrix for only the high order global features can be built. After applying this correction matrix and removing the global features, only a linear baseline drift may be left in the response. Linear baseline drift can be easily corrected using any simple pre-processing before chemometrics analysis of the spectral data.
  • FIG. 33 illustrates the thermal drift of a typical InGaAs PIN detector responsivity.
  • This feature represents a reduction in responsivity with the temperature at a specific position.
  • This dip feature represents a second-order drift that is not easily recognized as a baseline drift in chemometrics analysis.
  • the AirPLS method can also be exploited to extract a correction matrix for this feature.
  • %S should be 100% corresponding to the 100% line. Then, denoising of S can be performed by smoothing of the spectrum using the AirPLS method. After that, the drifted 100% line can be converted to absorbance according to:
  • FIG. 34A and 34B illustrate the AirPLS method, as shown in FIG. 34A and 34B, to separate the baseline first order drift from the dip higher order drift.
  • the dotted lines represent the first-order baseline drift in the absorbance domain, while the solid line represents the original drift.
  • FIG. 34B illustrates the result of subtracting the baseline first- order drift from the original drift and represents the dip feature of the detector near the responsivity knee around 1850 nm.
  • the drift function per degree can then be calculated according to:
  • FIG. 35 is a graph illustrating the common (average) correction matrix for the dip drift per degree Celsius according to some aspects.
  • the position of the detector feature may change from spectrometer unit to spectrometer unit. Therefore, the position of the detector dip feature can be detected and then the correction function can be repositioned based on that. This can be performed using two measurements at two different temperatures of a standard reference sample, such as using a 95% diffuse reflection standard in case of reflection- mode spectrometers.
  • the position can be defined by applying the dip feature extraction algorithm on two or more measurements. Then, the slope of the dip feature can be calculated and the wavelength that corresponds to the value of zero slope in a range of wavelengths (for example 1750 nm to 1900 nm) can correspond to the feature minima, as shown in FIG. 36.
  • FIG. 36 is a graph illustrating a mechanism for finding the detector feature position using the slope of the detector drift according to some aspects.
  • the detector dip feature position may be also determined by subtracting the measurement with higher temperature from the background.
  • the feature minimum is at the minimum value of the resulted spectrum, as shown in FIG. 37. This position value can then be used to shift the correction file to the correct position. In some examples, this process may be performed only one time for each lot, as the position of the detector dip feature does not change dramatically in the same lot.
  • FIGs. 38 A and 38B are graphs illustrating the extraction of detector higher-order thermal drift caused by cut-off shift with the temperature at different temperatures according to some aspects.
  • the dotted lines represent the cut-off drift in the reflectance domain after removing the features that change from spectrometer unit to spectrometer unit, while the solid line is the original cut-off drift.
  • FIG. 38B the same dotted line as in FIG. 38 A is shown; however, FIG. 38B illustrates the feature of the cut-off drift and how its slope changes and its position shifts at different temperatures.
  • a correction matrix can be extracted for multiple temperature points with a certain temperature change/step (for example 5 degrees). This algorithm can be repeated on certain number of different spectrometer units (for example 10 units) and the average correction matrix can be saved.
  • FIG. 39 is a diagram illustrating another example of a system 3900 configured to generate a global correction matrix according to some aspects.
  • an AI/ML engine e.g., a neural network
  • ML guided machine learning
  • the ML algorithm can be trained by feeding an enough amount of spectrometry data linked to temperature values (e.g., PDS, T data) 3908.
  • the data 3908 can be collected from the production line or from a plurality of spectrometer units 3902a, 3902b, ..., 3902N by connecting the units 3902a, 3902b, ..., 3902N to the cloud 3906 and sending their temperature states and the measured power spectral density in the presence of a reference sample, the latter being illustrated in FIG. 39.
  • the ordinary performance tests, whether production line or calibration at spectrometers, which are done on any spectrometer may be valuable for the ML training process.
  • This data may be stored in an online cloud for future usage.
  • the ML algorithm training can be enhanced and take less time if provided with the governing equations 3910 of thermal drift in the spectrometer.
  • the ML training process is only looking for the fitting parameters in the equations with the least provided data.
  • a feedback to the cloud 3906 can be optionally used to improve the on-line compensation for instrumental drifts at the user side (e.g., during operation of the spectrometer units 3902a, 3902b, ..., 3902N).
  • instrumental response may change with the aging of the spectrometer, which may decrease the performance of thermal drift correction algorithm.
  • each spectrometer unit 3902a, 3902b, ..., 3902N may be recalibrated.
  • the AI/ML engine 3904 can be configured to use a few measurements with corresponding temperatures of each spectrometer unit to provide respective updated correction matrices 3912 for each spectrometer unit 3902a, 3902b, ..., 3902N.
  • the word “exemplary” is used to mean “serving as an example, instance, or illustration.” Any implementation or aspect described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects of the disclosure. Likewise, the term “aspects” does not require that all aspects of the disclosure include the discussed feature, advantage or mode of operation.
  • the term “coupled” is used herein to refer to the direct or indirect coupling between two objects. For example, if object A physically touches object B, and object B touches object C, then objects A and C may still be considered coupled to one another — even if they do not directly physically touch each other. For instance, a first object may be coupled to a second object even though the first object is never directly physically in contact with the second object.
  • circuit and “circuitry” are used broadly, and intended to include both hardware implementations of electrical devices and conductors that, when connected and configured, enable the performance of the functions described in the present disclosure, without limitation as to the type of electronic circuits, as well as software implementations of information and instructions that, when executed by a processor, enable the performance of the functions described in the present disclosure.
  • One or more of the components, steps, features and/or functions illustrated in FIGs. 1-39 may be rearranged and/or combined into a single component, step, feature or function or embodied in several components, steps, or functions. Additional elements, components, steps, and/or functions may also be added without departing from novel features disclosed herein.
  • the apparatus, devices, and/or components illustrated in FIGs. 1-39 may be configured to perform one or more of the methods, features, or steps described herein.
  • the novel algorithms described herein may also be efficiently implemented in software and/or embedded in hardware.
  • “at least one of: a, b, or c” is intended to cover: a; b; c; a and b; a and c; b and c; and a, b and c.
  • All structural and functional equivalents to the elements of the various aspects described throughout this disclosure that are known or later come to be known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed by the claims.
  • nothing disclosed herein is intended to be dedicated to the public regardless of whether such disclosure is explicitly recited in the claims. No claim element is to be construed under the provisions of 35 U.S.C. ⁇ 112(f) unless the element is expressly recited using the phrase “means for” or, in the case of a method claim, the element is recited using the phrase “step for.”

Landscapes

  • Physics & Mathematics (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)
  • Electron Tubes For Measurement (AREA)
  • Spectrometry And Color Measurement (AREA)

Abstract

Des aspects concernent la compensation en ligne de dérives instrumentales dans des spectromètres miniaturisés en raison de variations de conditions environnementales et d'autres sources de dérive instrumentale. Le spectromètre peut comprendre un modulateur de lumière, un détecteur et un processeur. Le spectromètre peut en outre comprendre un capteur configuré pour obtenir une valeur d'une condition contribuant à des dérives instrumentales dans le spectromètre. Le processeur peut être configuré pour extraire un ensemble de paramètres de correction d'une matrice de correction associant une pluralité d'ensembles de paramètres de correction à des valeurs de capteur sur la base de la valeur et pour appliquer l'ensemble de paramètres de correction à une sortie du détecteur en vue de produire un spectre corrigé d'un échantillon testé. La matrice de correction peut être générée pour le spectromètre ou peut être basée sur une matrice de correction globale adaptée au spectromètre.
PCT/US2022/050469 2021-11-18 2022-11-18 Compensation en ligne de la dérive de réponse instrumentale dans des spectromètres miniaturisés WO2023091709A2 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US202163281042P 2021-11-18 2021-11-18
US63/281,042 2021-11-18
US17/989,317 2022-11-17
US17/989,317 US20230152216A1 (en) 2021-11-18 2022-11-17 On-line compensation of instrumental response drift in miniaturized spectrometers

Publications (2)

Publication Number Publication Date
WO2023091709A2 true WO2023091709A2 (fr) 2023-05-25
WO2023091709A3 WO2023091709A3 (fr) 2023-07-27

Family

ID=84942998

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2022/050469 WO2023091709A2 (fr) 2021-11-18 2022-11-18 Compensation en ligne de la dérive de réponse instrumentale dans des spectromètres miniaturisés

Country Status (1)

Country Link
WO (1) WO2023091709A2 (fr)

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS61213650A (ja) * 1985-03-19 1986-09-22 Chino Works Ltd 光学的測定装置
US6864978B1 (en) * 1999-07-22 2005-03-08 Sensys Medical, Inc. Method of characterizing spectrometer instruments and providing calibration models to compensate for instrument variation
DE102008054056A1 (de) * 2008-10-31 2010-05-06 Carl Zeiss Microimaging Gmbh Spektrometrische Anordnung und Verfahren zum Ermitteln eines Temperaturwerts für einen Detektor eines Spektrometers
DE102014013848B4 (de) * 2014-09-24 2016-08-04 Insion Gmbh Mikrospektrometer, Mikrospektrometersystem und Kalibrationsverfahren
DE102018205400A1 (de) * 2018-04-10 2019-10-10 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Spektralanalysesystem, mobiles gerät mit einem spektralanalysesystem, ein verfahren zur bestimmung einer korrekturfunktion zur abbildungskorrektur eines durch ein spektralanalysesystem aufgenommenen spektrums und computerprogramm
DE102019134456A1 (de) * 2019-12-16 2021-06-17 Carl Zeiss Spectroscopy Gmbh Spektrometer und Verfahren zu dessen Prüfung
DE102019219945A1 (de) * 2019-12-18 2021-06-24 Robert Bosch Gmbh Verfahren zum Betreiben einer Spektrometereinrichtung, Spektrometereinrichtung und Ladevorrichtung für eine Spektrometereinrichtung

Also Published As

Publication number Publication date
WO2023091709A3 (fr) 2023-07-27

Similar Documents

Publication Publication Date Title
JP7162066B2 (ja) 広帯域波長可変レーザー及びそのレーザーシステムの波長判定
CN106687781B (zh) 用于确定分光仪的光谱标度的方法及装置
Werner et al. Accurate determination of the spectral responsivity of silicon trap detectors between 238 nm and 1015 nm using a laser-based cryogenic radiometer
US20100246631A1 (en) Blackbody fitting for temperature determination
US9012833B2 (en) Terahertz wave measuring apparatus and measurement method
US10545049B2 (en) Method for stabilizing a spectrometer using single spectral notch
Tillman et al. Mid-infrared absorption spectroscopy across a 14.4 THz spectral range using a broadband femtosecond optical parametric oscillator
US20230152216A1 (en) On-line compensation of instrumental response drift in miniaturized spectrometers
Shim et al. TiO x/Ti/TiO x Tri-Layer Film-Based Waveguide Bolometric Detector for On-Chip Si Photonic Sensor
WO2023091709A2 (fr) Compensation en ligne de la dérive de réponse instrumentale dans des spectromètres miniaturisés
JP6439030B2 (ja) 受光装置及び受発光装置
US6830942B1 (en) Method for processing silicon workpieces using hybrid optical thermometer system
US10247608B2 (en) Spectroscopic module control method
US20220268635A1 (en) Method and device for monitoring radiation
JP6664367B2 (ja) 検出器、ならびに、検出器の補正方法、校正方法、検出装置
EP0961310A2 (fr) Appareil pour le traitement d'éléments de silicium avec un contrÔle amélioré de la température
Faria et al. A high-performance test-bed dedicated for responsivity measurements of infrared photodetectors in a wide band of low temperatures
Yang Design and fabrication of a tunable Fabry-Perot interferometer/photodiode micro-spectral image sensor
Hasan et al. Mid-Infrared Computational Spectrometry with Wavelength-Skewed Microbolometer Arrays
US11635330B2 (en) Microcavity-enhanced optical bolometer
Schrama et al. Monochromator-based cryogenic radiometry between 1 µm and 20 µm
Mudgal et al. Large optoelectronic chromatic dispersion in PN-type silicon photodiodes and
Rauhala Implementation of a Controller for a Piezo-Actuated Fabry-Perot Interferometer
Eppeldauer et al. Development and calibration of pyroelectric radiometer standards at NIST
Zurita et al. Photodiodes as optical radiation measurement standards

Legal Events

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

Ref document number: 22843456

Country of ref document: EP

Kind code of ref document: A2