Upconversion time-stretch infrared spectroscopy

Broadband mid-infrared (MIR) spectroscopy is a valuable tool for identifying molecules and analyzing their structural changes. However, increasing the measurement speed is crucial for studying fast phenomena and analyzing large datasets. While techniques like dual-comb spectroscopy have improved speed, they are limited by signal-to-noise ratio (SNR). Time-stretch infrared spectroscopy (TSIR) offers higher SNR than FTIR, but previous implementations had limited spectral resolution and element count. This work introduces UC-TSIR, aiming to overcome these limitations by using nonlinear upconversion to shift the spectrum to the near-infrared (NIR) for improved time-stretching and detection. The use of NIR allows for low-loss time-stretching with optical fibers and low-noise detection with high-bandwidth photoreceivers, ultimately leading to significantly improved speed and resolution.

The introduction reviews existing methods for broadband MIR spectroscopy, including sensor arrays and advanced FTIR techniques like MIR dual-comb spectroscopy (MIR-DCS), rapid-scan FTIR, and phase-controlled FTIR. It highlights the limitations of these methods, especially the SNR limitations encountered by high-speed FTIR techniques. The superior SNR of frequency-swept spectroscopy (FSS) is discussed, but existing MIR-FSS implementations have had limitations in speed and the number of measurable spectral elements. The authors position their work within this context, emphasizing the potential of TSIR to overcome SNR limitations but acknowledging its previous limitations in spectral resolution and element count. The paper highlights the advantages of the upconversion technique compared to previous approaches.

The UC-TSIR spectrometer uses a femtosecond MIR optical parametric oscillator (OPO) as a broadband MIR light source. The MIR beam passes through the sample, and then is combined with a 1.064-µm continuous-wave (CW) laser beam. Difference frequency generation (DFG) in a periodically poled lithium niobate (PPLN) waveguide upconverts the MIR pulses to NIR pulses around 1.5 µm. The NIR pulses are coupled into a single-mode fiber, amplified, and temporally stretched using dispersion-compensating fibers (DCFs). The stretched pulses are detected with a high-bandwidth InGaAs photodetector and a high-speed oscilloscope. For longer DCFs, a pulse picker and a Raman amplifier are used to avoid pulse overlap and maintain signal intensity. The spectral resolution is determined by the pulse duration before stretching and the detector's impulse response. A gradient-descent (GD) algorithm is used to demodulate systematic spectral distortions arising from the near-field effect of dispersive propagation in the long fiber. The wavenumber axis is downconverted from the NIR region using the dispersion values for pulse stretching. The group-delay dispersion (GDD) and third-order dispersion (TOD) values are calibrated and estimated from the DCF's datasheet.

The UC-TSIR system demonstrated high-speed broadband MIR spectroscopy of methane gas. Operating at 80 MSpectra s⁻¹ with a 10-km DCF, the system achieved a spectral resolution of 0.10 cm⁻¹ and 200 spectral elements. With longer 30-km and 60-km DCFs and a reduced measurement rate of 10 MSpectra s⁻¹, resolutions of 0.034 cm⁻¹ (760 elements) and 0.017 cm⁻¹ (1000 elements) were achieved, respectively. The single-shot SNR was 10 with the 30-km DCF and 6 with the 60-km DCF. The GD algorithm successfully demodulated spectral distortions, recovering accurate absorption lines of methane. Averaging of 180 spectra further improved the signal quality, demonstrating the system's stability. The measured spectra matched well with theoretical calculations based on the HITRAN database, showing excellent agreement with the known absorption lines of methane. The accuracy of the peak positions of the absorption lines from the HITRAN data was within 0.007 cm⁻¹ (200 MHz).

The results demonstrate that UC-TSIR successfully achieves high-speed and high-resolution broadband MIR spectroscopy. The significantly increased number of spectral elements (over 1000) and high spectral resolution (0.017 cm⁻¹) compared to previous TSIR implementations are significant advancements. The ability to measure at rates above 10 MSpectra s⁻¹ opens up new possibilities for studying ultrafast dynamics of irreversible phenomena, statistical analysis of heterogeneous data, and high frame-rate hyperspectral imaging. The agreement between experimental and theoretical results validates the accuracy and reliability of the technique. The achievement of this high SNR and speed surpasses what’s achievable with other methods, offering a promising route for various applications.

This work presents a groundbreaking advancement in broadband MIR spectroscopy with the development of UC-TSIR. The technique achieves unprecedented speed and resolution, exceeding limitations of previous methods. The demonstrated high-resolution measurement of methane gas molecules showcases the system's potential for various scientific applications. Future research could focus on improving the SNR, further miniaturizing the system, and exploring applications in diverse fields such as chemical analysis and biomedical imaging.

The current SNR is limited by shot noise and amplified spontaneous emission (ASE) noise. Systematic spectral distortions due to the near-field propagation effect require the use of the GD algorithm for demodulation. The accuracy of the wavenumber axis relies on precise calibration of the dispersion values and estimations of higher-order dispersion terms.