Mid-infrared spectroscopy on a fiber tip for molecular monitoring

Abstract

Mid-infrared (mid-IR) fiber sensors offer highly specific and sensitive detection and analysis of various chemical species due to many molecular vibrations and fundamental absorption bands in this range. In this paper, we present a compact transflection-based mid-IR fiber probe, designed to be potentially suitable for in vivo sensing, with an optical pathlength that can be controlled during fabrication. The optical fiber probe was fabricated using a silver halide polycrystalline fiber positioned in front of a gold-coated short fiber to act as a mirror, and a connector to hold the two parts facing each other at a predetermined distance. The optical fiber probe was tested with a quantum cascade laser (QCL) in the mid-IR region. To investigate the impact of optical pathlength, the reflected signal was recorded at various distances between the cleaved fiber and the mirror in air and water. The influence of gold layer thickness was also examined to optimize the optical fiber probe’s performance. To validate its sensing capability, the probe was employed to detect glucose solutions at physiological concentrations, achieving a detection limit of 8.91 mmol/L. The results highlight the potential of the proposed optical fiber probe sensor for molecular detection and analysis, offering a promising platform for in situ monitoring for chemical and biomedical applications.

1. INTRODUCTION

The mid-infrared (mid-IR) spectral region, spanning wavelengths from 3 to 25 μm, holds significant potential for chemical sensing due to the presence of strong and characteristic absorption bands of molecular vibrations¹. These absorption features enable the detection and quantification of various analytes with high specificity and sensitivity², making mid-IR sensing technologies crucial in fields such as biomedical diagnostics³, environmental monitoring⁴, and industrial process control⁴. Among these, fiber-optic sensors have gained significant attention for their versatility and ability to perform in situ and provide real-time measurements. In biomedical diagnostics, they are used to detect metabolites such as glucose⁵, cholesterol⁶, and urea⁷ with high specificity. In environmental monitoring, these sensors enable the detection of pollutants⁸, and volatile organic compounds (VOCs)⁹. Furthermore, in process monitoring, mid-IR sensors are employed for real-time monitoring of chemical reactions¹⁰, ensuring product quality.

Evanescent wave sensing (EWS) is one of the most commonly reported types of fiber-optic sensors, where light propagates through the core of a fiber and interacts with the surrounding medium through the evanescent field¹¹. The sensing area is located along a fiber where one fiber end is used for light delivery and the other for delivering light to a detector. This approach has been widely employed in mid-IR sensing for detecting gases⁷, liquids¹², and biomolecules¹³. However, the size of such fibers is too large for tissue insertion and in vivo sensing. Moreover, EWS fibers are not sensitive enough for measurement of low analyte concentrations, unless the sensing length is relatively long, thus limiting the possibility of miniaturization. Researchers have explored various strategies to enhance the performance of EWS sensors. One approach to enhance the sensitivity is the deposition of nanoporous particles on the fiber’s surface¹⁴ to increases the number of molecules in the close proximity to the evanescent field through capillary condensation of volatile organic compounds. Other approaches, such as using U-shaped fiber profile¹⁵, tapered fibers¹⁶, and side-polished fibers¹⁷, have been proposed to increase the penetration depth of the evanescent field. System comprising of ZnS prism with silver halide fibers and integrated with an external-cavity quantum cascade laser was set up for in vitro glucose measurement at physiological concentrations^18,19. However, such EWS sensors face limitations in miniaturization, restricting their applicability to in vivo applications where minimizing the size of the sensor is critical. Fiber probes based on light reflection, such as total attenuated reflectance (ATR) probes²⁰, can be made small, however their sensitivity is low due to limited interaction with the sample.

In transmission mode, excellent sensitivity is achieved through relatively small optical pathlengths of 20 to 50 μm for compounds in aqueous media. An example of a transmission mid-IR fiber sensor coupled with a microfluidic platform²¹, where the system consists of two optical fibers facing each other. One of the fibers receives light from a light source and emits it into a predefined optical pathlength, i.e., the sensing region. The sensing region is designed with a predefined optical pathlength, where the light propagates through the analyte, enabling interaction and absorption by specific chemical components. The other fiber, positioned at the other end of the sensing region, collects the transmitted light and delivers it to a detector for analysis. With the integration of quantum cascade laser, the performance can be further enhanced due to their high spectral emission power¹⁸. Several glucose biosensors based on transmission mid-IR spectroscopy have been reported, which allows detection of physiological concentration of glucose^22–24. The system can be used for continuous measurements, however, due to the fact that two fibers need to be implanted into a tissue which can cause significant damage, it is not suitable for in vivo applications. Introducing two separate fibers facing each other into a biological tissue can cause significant damage. Furthermore, any motion caused by physiological activities, such as blood flow, respiration and muscle contraction can disrupt the optical alignment and compromise the stability of the measurement.

Other approaches for designing a sensitive and miniature fiber probe involve refractive index measurement²⁵ and localized surface plasmon resonance (LSPR)^26–28. However, both refractive index sensing and LSPR are inherently non-selective and require (bio)chemical functionalization of the fiber surface to achieve selectivity, adding complexity to the sensor fabrication. Although these functionalized surfaces enhance selectivity to a particular analyte, surface functionalization degrades over time and limits the long-term stability and reliability of the sensor. On the other hand, mid-IR spectroscopy offers a direct measurement of the analyte of interest without a need for (bio)chemical functionalization. Moreover, it can be used to monitor more than one analyte/compound simultaneously^10,29.

In addition to mid-IR sensors, Raman fiber probes and multimodal platforms have been explored for in vivo and ex vivo chemical sensing^30,31. Raman-based probes offer label-free detection with relatively low water interference, but often require strong laser excitation and are prone to background fluorescence in biological media³². Multimodal probes combining Raman, fluorescence, and reflectance modalities are emerging but tend to be complex in design³³.

An optical fiber probe in transflection mode can address the challenges of in vivo measurements by combining the delivery and collection of light within a single fiber. In this configuration, light interacts with the analyte in a predefined sensing area and is reflected back to the same fiber, eliminating the need for two fibers or the use of a bent or extended fiber. This design minimizes tissue damage for in vivo applications. However, there is a very limited selection of commercially available mid-IR transflection probes, and they are typically large in size³⁴ (on the order of centimeters) and thus not suitable for in vivo sensing and/or for applications that require small probe dimensions. In this study, we present a novel mid-IR transflection optical fiber probe sensor that is small and sensitive. To the best of our knowledge, this is the smallest reported transflection-based optical fiber probe operating in the mid-IR region that is small enough (overall diameter of 1.59 mm) to be potentially suitable for in vivo sensing. In our probe configuration, the light travels through the sample/solution twice - once during transmission and once after reflection from a gold-coated fiber acting as a mirror. The probe enables direct interaction with the sample over a defined optical pathlength, which can be precisely controlled during the fabrication process. The schematics of the mid-IR transflection optical fiber probe and how it compares to existing probes is provided in Figure 1. To demonstrate the performance of the proposed probe, aqueous glucose solutions at physiological concentration levels were measured by monitoring the C-O stretching vibration at 1,035 cm⁻¹ (wavelength of ~9.7 μm), demonstrating its potential for mid-IR chemical analysis. In addition to presenting the novel probe design and its sensing application, we also addressed other challenges, such as identifying the source of signal interference and developing a regression model for calibration — aspects that have not been previously reported or published.

Figure 1.

Schematic of the different types of optical fiber probes for sensing.

2. MATERIALS AND METHODS

Materials

Polycrystalline mid-infrared fiber, which is made of AgCl-AgBr (PIR 500, Art Photonics GmbH), was used for the optical fiber probe fabrication. The silver halide fiber with a diameter of 500 μm has an IR transparent wavelength range 3–17 µm, numerical aperture (NA) of 0.3, and core refractive index of 2.1. D-glucose (≥99.5%, G7021 from Sigma Aldrich) was used for preparation of the standard solutions in deionized water having concentrations between 0 and 200 mmol/L.

Measurement of gold coating thickness

In our design, the end face of an optical fiber was coated with gold to serve as a reflective mirror, which is described in detail in the later section. Different coating duration time was applied to the fiber end to investigate the optimal thickness of the coating layer. The thickness of the gold layer was determined by placing a glass slide next to the fiber during the sputtering process. After sputtering, a scratch was made on the gold-coated surface of the glass slide using a razor blade, and the depth of the resulting valley was measured using an optical profilometer (Zeta-20, Zeta Instruments).

Fourier-transform infrared spectroscopy (FTIR)

To compare the spectra measured using the optical fiber probe and a standard instrument, a Fourier-transform infrared (FTIR) spectrometer (Nicolet iS50, Thermo Fisher Scientific) equipped with a transmission-mode liquid cell (Pike Technologies) was utilized. The optical transmission cell (Oyster, Specac) has an optical pathlength of 50 µm and requires a minimum sample volume of 10 µL. Each spectrum was acquired for 300 scans, yielding a total acquisition time of 187 seconds, with a spectral resolution of 0.482 cm⁻¹.

Experimental setup

Figure 2 shows the schematics of the experimental setup for optical fiber probe measurements. A mid-IR quantum cascade laser (MIRcat, DRS Daylight Solutions, USA) was used as the light source, providing a spectral range of 800–1,800 cm⁻¹ (5,556–12,500 nm). The collimated laser beam was directed into a beam splitter (BSW705, Thorlabs) using three mirrors. After the beam splitter, an off-axis parabolic mirror (35–643, Edmund Optics) focused the beam onto the optical fiber probe. An optical fiber clamp (SM1F1–250, Thorlabs) was used to secure the optical fiber probe and was fixed on a 3-axis stage to adjust the position for signal optimization. The infrared light propagates through the optical fiber probe and interacts with the sample solution in the sensing area, which is between the optical fiber and the mirror in the probe sensor. The reflected signal is sent back through the fiber and directed to the detector (PVI-4TE-10.6, VIGO Photonics). The zoomed-in figure shows the optical fiber probe inserted into a 250 μL centrifuge tube, containing the analyte solution, with the sensing area fully submerged in the liquid when performing measurement for glucose.

Figure 2.

Experiment setup for the optical fiber probe.

A cleaved fiber was used instead of the optical fiber probe when quantifying the reflection signal. A mirror, mounted on a translational stage, was positioned directly behind the fiber’s end face, allowing the reflection signal to be measured at varying distances from the mirror.

3. RESULTS AND DISCUSSIONS

Estimation of the optical pathlength for optical fiber probe

One of the key advantages of the proposed design is the ability to fabricate the optical fiber probe sensor with a specific optical pathlength. The optical pathlength directly impacts the sensor’s performance based on Beer-Lambert’s law. Longer pathlengths increase absorbance, therefore enhancing the sensor’s sensitivity at low analyte concentrations. But if too long, light emerging from fiber diverges as it propagates, consequently, less light is reflected back to the detector. Determining the optimal optical pathlength is particularly important for aqueous samples due to the strong absorption of mid-IR light by water. Optical pathlengths, over ~100 μm, can result in complete absorption of the light by the sample, leaving no detectable signal reflected back to the detector³⁵. To estimate the optimal optical pathlength, a cleaved optical fiber and a mirror was used. Infrared light was introduced at one end of the fiber, and the reflected signal from the opposite end was measured at varying distances from the mirror.

Figure 3(a) shows the reflected signal from the optical fiber probe sensor, showing the summed signal in the 1,000–1,180 cm⁻¹ region at various distances from a mirror in air. The reflected signal was normalized by subtracting the minimum signal (recorded when the mirror was far away) and dividing by the difference between the maximum signal (recorded when the mirror was almost attached to the fiber) and the minimum signal. The reflected signal decreased almost linearly as the distance from the mirror increased from 0 to 1,000 μm. At a distance of 560 μm, the signal dropped to 50% of its maximum value and further reduced to 10% at a distance of 1,023 μm.

Figure 3.

Normalized mid-IR reflected signal as a function of distance between a silver halide fiber and a mirror in (a) air, and (b) water. Error bars represent the standard deviation (N=3).

When the optical fiber probe sensor was immersed in water, Figure 3(b), the reflected signal decreased sharply compared to when it was in air. The signal declined significantly when the distance between the fiber end-face and the mirror exceeded 40 μm (corresponding to a total optical pathlength of 80 μm) and dropped to half its maximum value at a distance of 53 μm. This is explained by the strong IR light absorption by the water.

This data serves as a guideline for designing pathlengths for both gaseous (in air) and aqueous samples. This is especially important for measuring low concentrations of analytes in aqueous solutions.

Assembly of the optical fiber probe sensor

Figure 4(a) shows the schematics of the optical fiber probe design in transflection mode. The probe sensor consists of an optical fiber, a short piece of gold-coated silver halide fiber which is used as mirror, and a connector to hold and align the fiber and the gold-coated fiber facing each other. The surface of the shorter fiber was polished using aluminum oxide lapping film (50–20060, Allied High Tech Products, Inc.) prior to gold sputtering (6002, Ted Pella, Inc.). This polishing process ensured a smooth surface, minimizing signal loss due to light scattering. The connector is made of a 1 cm long hollow polyetheretherketone (PEEK) tube with an inner diameter of 0.51 mm and an outer diameter of 1.59 mm. A 3 mm wide groove in the connector was made using a hand drill, while a needle was inserted into the tubing during the drilling process to prevent the hole from collapsing. The mirror was inserted with its coated face oriented toward the cleaved optical fiber. The optical pathlength – two times of the distance between the fiber and the mirror – was adjusted under a microscope after inserting the fiber and mirror into the PEEK connector. Figure 4(b) shows the photo of the fabricated optical fiber probe. Figure 4(c) shows the close-up view of the sensing area and the optical pathlength.

Figure 4.

(a) Schematic of the transflection optical fiber probe where the optical interaction pathlength is twice the gap d. (b) Photograph of the transflection optical fiber probe. (c) Zoom-in of the sensing area and the optical pathlength.

Optimization of the gold thickness of the mirror coating

The transflection optical fiber probe sensor relies on light interacting with the analyte in the sensing region having a total optical pathlength of 2d and carrying the spectral information after reflecting off the mirror. Consequently, the quality of the mirror significantly affects the sensor’s performance. To optimize the reflected signal from the mirror, the optical fiber probe with various gold coating thicknesses was investigated.

Short fibers with flat end faces were first polished and then sputtered with gold as fiber mirror for durations of 0, 60, 90, 180, 300, and 600 seconds, corresponding to coating thicknesses of approximately 0, 35, 54, 122, 175, and 267 nm. Figure 5(a) presents the reflected spectra measured using the optical fiber probe assembled with fiber mirrors of different gold coating thicknesses. The distance between the two fibers, the cleaved fiber and the fiber mirror, was approximately 40 μm, with the gap filled with ambient air for all measurements. The yellow spectrum at the bottom was measured using an optical fiber probe with a mirror made from a cleaved fiber without a gold coating. As the gold coating thickness increases, the reflection signal improves significantly.

Figure 5.

(a) Reflected signal was measured for the optical fiber probes with fiber mirror coated with different gold thickness, with ambient air in the optical pathlength of approximately 80 μm. (b) Area under spectral curve measured with mirror with different coating thickness. Error bars represent the standard deviation (N=3).

The obtained spectra consist of multiple peaks originating from back reflections at the fiber-air interface before the light was reflected off the mirror. It is also due to multiple reflections within the small gap between the two fibers. To better understand the periodic modulations observed in the spectral data, a Fast Fourier Transform (FFT) was performed on the spectra in Figures 5(a). The dominant frequency was identified at 0.0218, which corresponds to an optical pathlength of approximately 109 µm. This value is close to the optical pathlength of the probe, approximately 80 μm (twice the air gap between the fiber tip and the mirror), which supports that the modulations result from multiple reflections within the small cavity formed between the fiber end face and the mirror.

To further evaluate the reflectance in the 1,000 – 1,180 cm⁻¹ region (a range critical for detecting absorption peaks of molecules such as glucose) the area under each spectral curve in Figure 5(a) was calculated. This analysis provides a quantitative measure of the mirror’s performance. Figure 5(b) shows the area under each spectral curve for different coating thicknesses. The reflected signal initially increased significantly with the thickness of the gold coating and then reached a plateau at a thickness of around 170 nm. A previous study showed a maximum reflectance of 90.5% using a gold coating thickness of 153 nm, and a 88.5% reflectance using gold coating thickness of 235 nm³⁶. These findings aligned with our measurements. To ensure optimal performance, in the subsequent sensors, the mirror of the fiber sensor was fabricated using fibers coated with gold to a thickness of approximately 267 nm.

Measurement of glucose at different concentrations

The measurement of glucose solutions at various concentrations demonstrates the sensing performance of the optical fiber probe. An optical fiber probe with a distance of 38 μm between two fibers (optical pathlength of 76 μm) was fabricated to avoid total absorption of signal due to water absorption. Figure 6(a) shows the absorbance spectra of glucose solutions at different concentrations, using water as background. Each spectrum represents an average of three measurements. The spectrum of a 100 mmol/L glucose aqueous solution, measured with a FTIR with a 50 μm optical pathlength, was also included for reference. The discrepancy between the glucose spectra obtained from the fiber probe and the FTIR can be attributed to the presence of interference patterns caused by multiple reflections within the sensing cavity. Furthermore, because the absorbance is calculated using signals that include back reflections, glucose-induced spectral features are partially masked according to Equation (1). Furthermore, when the magnitude of back reflection becomes larger, the net absorbance can be significantly reduced, decreasing the sensitivity. These two factors, interference and back reflection, distort the overall spectral shape and are absent in the FTIR reference spectrum.

Figure 6.

(a) Absorbance of glucose solutions at different concentrations between 0 and 200 mmol/L measured using the fiber probe with a pathlength of 76 μm, and a reference FTIR spectrum of 100 mmol/L glucose using 50 μm pathlength. Water was used as background to calculate the absorbance spectra. (b) Absorbance at 1,035 cm⁻¹ for different glucose concentrations. Error bars represent the standard deviation (N=3).

To verify the origin of the short-period modulation, the same FFT analysis described in the previous section was applied to the spectrum in Figure 6(a). The dominant frequency component was identified at 0.0216, which corresponds to an optical pathlength of approximately 81 μm. This value is close to the designed optical pathlength of 76 μm for the optical fiber probe, indicating that the short-period modulations observed in Figure 6(a) originate from multiple reflections within the small cavity between the fiber and the gold-coated fiber mirror.

The absorbance at 1,035 cm⁻¹, corresponding to C–O stretching, was used for concentration quantification. Figure 6(b) shows the peak heights at various concentrations. The absorbance at 1,035 cm⁻¹ at different concentrations was not linear as shown in Figure 6(b), primarily due to the presence of back reflection present in both water and glucose spectra. Specifically, the signal received by the detector is a combination of analyte-specific signal reflected from the gold-coated mirror and from back reflections at the fiber–liquid interface. In order to account for the influence of back reflections, the absorbance calculated from the spectra collected with the optical fiber probe can be expressed as:

$$Abs\left(\lambda ,c\right)={\log }_{10}\frac{S_w\left(\lambda \right)+R_w\left(\lambda \right)}{S_g\left(\lambda ,c\right)+R_g\left(\lambda ,c\right)}$$ (1)

where R_w and R_g represent the back reflection from the end surfaces when measuring water and glucose solutions, respectively, while S_w and S_g represent the obtained unprocessed spectra. All parameters are functions of wavenumber or wavelength λ, while S_g and R_g also vary with glucose concentration c.

At a specific wavenumber of 1,035 cm⁻¹ to analyze the glucose peak, S_g and R_g depend only on concentration.

Experimental observations show that the back reflection signal varies by less than 5% (at 1035 cm⁻¹) when measuring water and glucose at different concentrations as summarized in Table 1.

Table 1.

Back reflection signals at different concentrations of glucose in aqueous solution

Concentration (mmol/L)	Intensity at 1035 cm−1 (a. u.)
0	1.032 ± 0.037
50	1.041 ± 0.023
100	1.005 ± 0.041
200	1.047 ± 0.028

As a result, we can assume that R_w, R_e and R are constants, and equal to each other, in the range of 1000–1200 cm⁻¹, and thus:

$$R_w=R_g=R$$ (2)

Based on Beer’s Lambert’s law, ideally, the absorbance value at 1,035 cm⁻¹ is linearly proportional to glucose concentration c. Therefore, the glucose absorbance signal can be expressed as:

$$Abs{\left(c\right)}_{ideal}={\log }_{10}\frac{S_w}{S_g\left(c\right)}=C_1\cdot c$$ (3)

and therefore

$$S_g\left(c\right)=S_w\times {10}^{-C_1\cdot c}$$ (4)

where C₁ is a constant. By substituting experimental data and Equation (4) into Equation (1), the rearranged equation for absorbance becomes:

$$Abs\left(c\right)={\log }_{10}\frac{C_0}{S_w\times {10}^{-C_1\cdot c}+R}$$ (5)

where C₀ is the total signal for water measurement at 1,035 cm⁻¹ (including both water and back reflection signals).

A three-parameter regression model was used to fit the experimental data, as shown in Figure 6(b), to derive the calibration curve. This curve enables the calculation of unknown concentration of the glucose solution.

The limit of detection is around 8.91 mmol/L, determined as three times the standard deviation of the peak height of water over the slope around the origin, calculated using the fitting curve. This concentration is in the physiological glucose levels observed after eating or in the early stages of traumatic brain injury, typically around 10 mmol/L^37,38. For physiological glucose sensing application, the limit of detection needs to be further improved, preferably down to 0.5 mmol/L. Additionally, in complex biological fluids such as blood, the LOD would likely be further compromised due to overlapping infrared absorption from proteins, lipids, and other endogenous compounds. Future improvements to the probe will focus on enhancing detection performance through several approaches. First, optimization the optical design, which includes minimizing back reflections and optimizing optical alignment. Second, incorporation of a semi-permeable membrane which would exclude large interfering biomolecules, such as proteins, from the sensing region. Finally, applying signal processing techniques, such as spectral smoothing and multivariate calibration (e.g., partial least squares regression, PLSR) may further improve detection limit.

4. CONCLUSION

An optical fiber probe sensor was successfully fabricated to operate in transflection mode, utilizing a mid-IR silver halide fiber, a lab-made PEEK connector, and an additional gold-coated silver halide fiber to act as a reflective mirror. The overall diameter of the probe is 1.59 mm, which is, to the best of our knowledge, is the smallest transflection probe reported in literature. For optimal performance, a 267 nm gold coating was applied to the fiber end face to maximize signal reflection back to the detector. For aqueous samples, a distance of less than 50 μm between the fiber and the fiber mirror is recommended due to the divergence of light exiting the fiber and the strong absorption of water in the mid-infrared region. To validate the probe’s performance, glucose solutions with concentrations ranging from 3.13 to 200 mmol/L were measured using the probe with an optical pathlength of 76 μm (corresponding to a distance of 38 μm between the fiber end and the mirror fiber). A regression model was developed to address the non-linearity of the calibration curve, caused by back reflection at the fiber-air interface, yielding a detection limit of approximately 8.91 mmol/L. Due to the high selectivity of the mid-IR range, the probe can also be used to detect other analytes instead or in addition to glucose. Future work will aim to further reduce the overall size of the probe and enhance the probe’s performance by optimizing the optical design to reduce back reflections, integrating a semi-permeable membrane to block interfering biomolecules, and applying signal processing techniques to extend probe’s applicability for complex biological fluids.

Biographies

Tse-Ang Lee is a Ph.D. candidate in Mechanical Engineering at The University of Texas at Austin. Tse-Ang obtained his BSc and MSc in Mechanical Engineering at National Taiwan University. His research interests focus on microdialysis techniques and development of mid-infrared optical fiber sensors. He is a recipient of the Fred Murphy Jones and Homer Lindsey Bruce Endowed Fellowship from the Waggoner Center for Alcohol and Addiction Research at The University of Texas at Austin.

Zhenyang Xiao is a Ph.D. student in the Chandra Department of Electrical and Computer Engineering at The University of Texas at Austin. He earned his bachelor’s degree in Electrical Engineering from Tianjin University and his master’s degree in Optical Engineering from the Shanghai Institute of Technical Physics at the Chinese Academy of Sciences. His research interests focus on terahertz spectroscopy techniques and quantum cascade lasers.

Tanya Hutter is a faculty member at the Walker Department of Mechanical Engineering at The University of Texas at Austin and a member of the Advanced Materials Science and Engineering research area. Dr. Hutter has a B.Sc. in Chemical Engineering, M.Sc. in Materials Science and Engineering and Ph.D. in Physical Chemistry. Her research focuses on emerging molecular sensing technologies, nanomaterials, microfabrication and photonics with applications in environmental and industrial sensing, homeland security and medical diagnostics.

David Burghoff is faculty in the Department of Electrical and Computer Engineering at the University of Texas at Austin. He received his doctorate from MIT, his bachelor’s from UIUC, and has received early career awards from the AFOSR, NSF, ONR, and Moore Foundation. He currently leads PRISM, a MURI developing high-performance radiometry. His research focuses on nonlinear and quantum optoelectronics, primarily at long wavelengths, and also on using these technologies to develop novel sensing modalities.

CODE AND DATA AVAILABILITY

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.