Time-stretch infrared spectroscopy

Vibrational molecular spectroscopy is crucial in various fields. While pump-probe spectroscopy offers high temporal resolution, it's limited to reproducible phenomena. High-speed continuous measurement is needed for studying fast, non-repeating events like combustion or protein conformational changes, and for analyzing large spectral datasets in applications like raster scan imaging and flow cytometry. Mid-infrared (MIR) absorption spectroscopy, with its larger cross-section than coherent Raman scattering, holds great potential for high-speed continuous measurement. Fourier-transform spectroscopy (FTS) techniques, including dual-comb spectroscopy, have improved measurement rates but are fundamentally limited by the signal-to-noise ratio (SNR). Wavelength-swept spectroscopy offers an SNR advantage proportional to the number of spectral elements (N), particularly beneficial for broadband spectroscopy with large N. Time-stretch (TS) spectroscopy, a single-pulse technique, is ideal for high-speed frequency-swept spectroscopy but hasn't been demonstrated in the MIR region due to technological limitations. This work overcomes these limitations by developing time-stretch infrared (TS-IR) spectroscopy.

The authors review existing techniques in vibrational spectroscopy, highlighting the limitations of pump-probe techniques for non-repeatable phenomena and the advancements in broadband coherent Raman scattering spectroscopy, noting its speed limitations due to low SNR. They discuss the advantages of MIR absorption spectroscopy and the current state-of-the-art using FTS and dual-comb spectroscopy, emphasizing the SNR limitations of these methods. Wavelength-swept spectroscopy is presented as a potential solution due to its SNR advantage, but existing limitations in MIR wavelength-swept lasers are discussed. Time-stretch spectroscopy is introduced as an ideal technique, previously demonstrated in the near-infrared, but its lack of implementation in the MIR region is explained due to the unavailability of suitable components (high-repetition-rate MIR fs laser source, low-loss TS dispersive optics, and a fast MIR photodetector).

The researchers developed a TS-IR spectrometer using a synchronously pumped fs optical parametric oscillator (fs-OPO) as the MIR light source, a free-space angular-chirp-enhanced delay (FACED) system for pulse stretching, and a high-bandwidth quantum cascade detector (QCD) as a fast MIR photodetector. The fs-OPO generates MIR idler pulses tunable from 2.1 to 5.1 µm; in this experiment, the center wavelength was set at 4.6 µm. The FACED system, consisting of a diffraction grating, concave mirrors, and flat mirrors, stretches the pulses from ~100 fs to ~10 ns, enabling wavelength-time conversion. The stretched pulses were then focused onto the QCD, amplified, and digitized using a high-speed oscilloscope. The system was characterized by measuring the stretched pulse waveform under varying conditions of mirror distance and angle, confirming agreement with theoretical relations. The spectroscopic capability was demonstrated by measuring notch-filtered spectra and the absorption spectrum of liquid phenylacetylene. Further evaluations involved measuring a mixture of phenylacetylene and toluene, assessing spectral resolution using narrow notch filters, and verifying the linear concentration dependence of absorbance.

The developed TS-IR spectrometer achieved a spectral acquisition rate of 80 MSpectra s⁻¹, significantly faster than previous methods. Broadband absorption spectroscopy of phenylacetylene from 4.4 to 4.9 µm (2040–2270 cm⁻¹) was demonstrated with a resolution of 15 nm (7.7 cm⁻¹), a signal-to-noise ratio (SNR) of 85 without averaging, and a shot-to-shot fluctuation of only 1.3%. The system's performance was further validated by measuring multiple absorption lines from different molecular species (phenylacetylene and toluene), verifying its multi-species detection capability. Careful evaluation using narrow notch filters confirmed the spectral resolution to be 15 nm (7.7 cm⁻¹), matching the theoretical expectation. Finally, linear concentration dependence of absorbance was demonstrated, confirming the validity of the measurements.

The 80 MSpectra s⁻¹ acquisition rate of the TS-IR spectrometer represents a significant advancement in mid-infrared spectroscopy, exceeding previous records by 1-2 orders of magnitude. The high SNR and low shot-to-shot fluctuation demonstrate the system's robustness. While the current resolution is lower than some dual-comb systems, it is limited by the detector bandwidth and can be improved using faster detectors or reducing the laser repetition rate. The demonstrated capabilities open exciting possibilities in various fields. The high speed and continuous measurement capabilities address the need for studying non-repeatable events.

This research successfully demonstrated a time-stretch infrared (TS-IR) spectrometer operating at an unprecedented 80 MSpectra s⁻¹, exceeding previous state-of-the-art by one to two orders of magnitude. The system exhibited high signal-to-noise ratio and low shot-to-shot fluctuation, paving the way for various applications, particularly in studying fast, non-repeatable events and high-throughput measurements. Future improvements could involve using higher repetition-rate lasers, faster detectors, and broader spectral coverage.

The current spectral resolution (15 nm or 7.7 cm⁻¹) is a limitation compared to some dual-comb systems. This is primarily due to the detector bandwidth and can be improved with faster detectors or by reducing the laser repetition rate. The relatively low power throughput of the FACED system (9%) is another factor that could be addressed in future developments. Further optimization of the system's components and parameters can improve overall performance.