Adaptively Optimized Gas Analysis Model with Deep Learning for Near-Infrared Methane Sensors.

Noise significantly limits the accuracy and stability of retrieving gas concentration with the traditional direct absorption spectroscopy (DAS). Here, we developed an adaptively optimized gas analysis model (AOGAM) composed of a neural sequence filter (NSF) and a neural concentration retriever (NCR) based on deep learning algorithms for extraction of methane absorption information from the noisy transmission spectra and obtaining the corresponding concentrations from the denoised spectra. The model was trained on two data sets, including a computationally generated one and the experimental one. We have applied this model for retrieving methane concentration from its transmission spectra in the near-infrared (NIR) region. The NSF was implemented through an encoder-decoder structure enhanced by the attention mechanism, improving robustness under noisy conditions. Further, the NCR was employed based on a combination of a principal component analysis (PCA) layer, which focuses the algorithm on the most significant spectral components, and a fully connected layer for solving the nonlinear inversion problem of the determination of methane concentration from the denoised spectra without manual computation. Evaluation results show that the proposed NSF outperforms widely used digital filters as well as the state-of-the-art filtering algorithms, improving the signal-to-noise ratio by 7.3 dB, and the concentrations determined with the NCR are more accurate than those determined with the traditional DAS method. With the AOGAM enhancement, the optimized methane sensor features precision and stability in real-time measurements and achieves the minimum detectable column density of 1.40 ppm·m (1σ). The promising results of the present study demonstrate that the combination of deep learning and absorption spectroscopy provides a more effective, accurate, and stable solution for a gas monitoring system.

Near infrared frequency comb vernier spectrometer for broadband trace gas detection.

We present a femtosecond frequency comb vernier spectrometer in the near infrared with a femtosecond Er doped fiber laser, a scanning high-finesse cavity and an InGaAs camera. By utilizing the properties of a frequency comb and a scanning high-finesse cavity such a spectrometer provides broad spectral bandwidth, high spectral resolution, and high detection sensitivity on a short time scale. We achieved an absorption sensitivity of ~8 × 10(-8) cm(-1)Hz(-1/2), corresponding to a detection limit of ~70 ppbv for acetylene, with a resolution of ~1.1 GHz in single images taken in 0.5 seconds and covering a frequency range of ~5 THz. Such measurements have broad applications for sensing greenhouse gases in this fingerprint near infrared region with a simple apparatus.

Femtosecond electron-lattice thermalization dynamics in a gold film probed by pulsed surface plasmon resonance.

The dynamics of electronic excitations and their relaxation in a gold film is studied on the femtosecond time scale with a pump-probe technique. For the pump beam we use pulses with wavelengths centered at 800 nm, 400 nm or both. The surface plasmon resonance (SPR) in Kretschmann's configuration is used as a sensitive and fast-response probe of the dynamics of the dielectric properties of the gold film. The quantity that is monitored is the intensity of the reflected light at an incidence angle close to the SPR. With changes of the dielectric properties induced by the pump beam and during subsequent relaxation, the amount of the reflected light of the probe beam, sent with a variable delay, also changes, thus providing information on the temporal characteristics of the thermalization process. Special features of SPR probing with short pulses are also accounted for in this work. The thermalization of the electronic subsystem and energy transfer to the lattice are discussed in connection with the two-temperature relaxation model that takes into account temperature dependences of the electronic heat capacity and the electron-phonon coupling.

High-power mid-infrared frequency comb source based on a femtosecond Er:fiber oscillator.

We report on a high-power mid-infrared (MIR) frequency comb source based on a femtosecond (fs) Er:fiber oscillator with a stabilized repetition rate of 250 MHz. The MIR frequency comb is produced through difference frequency generation in a periodically poled MgO-doped lithium niobate crystal. The output power is about 120 mW, with a pulse duration of about 80 fs and spectrum coverage from 2.9 to 3.6 µm, and the single comb mode power is larger than 0.3 µW over the range of 700 nm. The coherence properties of the produced high-power broadband MIR frequency comb are maintained, which was verified by heterodyne measurements. As the first application, the spectrum of a ~200 ppm methane-air mixture in a short 20 cm glass cell at ambient atmospheric pressure and temperature was measured.

Two-photon fluorescence of Coumarin 30 excited by optimally shaped pulses.

We optimized the two-photon fluorescence (TPF) of a Coumarin 30 dye by using a feedback-controlled femtosecond pulse shaping technique. For optimization we implemented an evolutionary algorithm with a liquid crystal phase-only pulse shaper in a folded 4f setup. The optimization procedure applied to the second harmonic generation, and TPF noticeably improved the output signals and demonstrated good convergence. In addition, signal ratios involving TPF and second harmonic generation (SHG) were successfully optimized. The correlation between TPF and SHG was studied, and it was found to decrease when the pulse shape was close to the optimum. These experimental results are of interest for potential applications of coherent control to complex molecular systems as well as in biomedical imaging.

Visible and UV coherent Raman spectroscopy of dipicolinic acid.

We use time-resolved coherent Raman spectroscopy to obtain molecule-specific signals from dipicolinic acid (DPA), which is a marker molecule for bacterial spores. We use femtosecond laser pulses in both visible and UV spectral regions and compare experimental results with theoretical predictions. By exciting vibrational coherence on more than one mode simultaneously, we observe a quantum beat signal that can be used to extract the parameters of molecular motion in DPA. The signal is enhanced when an UV probe pulse is used, because its frequency is near-resonant to the first excited electronic state of the molecule. The capability for unambiguous identification of DPA molecules will lead to a technique for real-time detection of spores.