Molecule Detection by Resonantly Enhanced Third Harmonic Generation of Mid-Infrared Laser Pulses

In this work we presented thorough experimental studies on resonantly enhanced third harmonic generation (THG) driven by mid-infrared, spectrally narrowband, nanosecond laser pulses tuned in the vicinity of rovibrational molecular resonances with hydrogen chloride (HCl) molecules as a test gas. In contrast to other approaches for nonlinear spectroscopy, the technique requires only a single driving laser beam, which facilitates applications. As a major advantage of the THG approach among rovibrational states in the electronic ground state of the molecules, we drive the frequency conversion process simultaneously close to single-, two-, and three-photon resonances, with typical detunings in the range of 10 - 100 cm⁻¹ only. Tuning exactly to a single-photon resonance is, due to absorption of the fundamental radiation, no good choice. Tuning to two- and three-photon resonances permits large resonance enhancements of the THG yield by up to four orders of magnitude with regard to the background signal. The main technological challenge of this project was the setup of a laser system, which supplies continuously tunable mid-infrared laser pulses with pulse energies up to 2 mJ. Based upon a commercial optical parametric oscillator (OPO), we set up a three stage optical parametric amplifier chain. This allows stable operation and significantly reduces maintenance, when compared with typical pulsed mid-infrared laser sources, which usually include dye lasers. From transmission spectra in HCl we estimate a small spectral bandwidth ≤ 100 MHz (FWHM), i.e., only a factor of 2 above the Fourier transform limit. We reach peak intensities of up to 450 MW/cm² at a pulse energy of 1.9 mJ and a pulse duration of 8 ns. We compare the experimental results on the THG spectrum in HCl to a numerical simulation, which agrees well with the data. We systematically study and compare THG via two- and three-photon resonances in HCl. Two-photon resonances yield enhancement factors of more than three orders of magnitude with regard to offresonant THG. We find, that saturation of the two-photon transition and resonance enhancements occur for laser intensities beyond 100 MW/cm², while we do not yet observe such saturation at our available laser intensities for the higher-order three-photon transition. Measurements of the THG signal vs. sample pressure confirm the quadratic dependence expected for a nonlinear optical process. At large pressures beyond a few 10 mbar collisional broadening leads to saturation of the signal vs. sample pressure. A detailed analysis of the spectral line shape shows an asymmetric, Fano-type line profile at lower gas pressures, which is due to interference between the resonant signal from the test gas and nonresonant background. The asymmetry vanishes at higher pressures due to collisional broadening, giving rise to a symmetric line profile. From the pressure dependence, we determine a detection limit for the THG approach in the HCl sample in the range of ≤ 1 mbar, corresponding to roughly 3×10¹³ molecules in the confocal volume of the driving laser field. This paves the way towards applications of the approach under realistic conditions of trace gas detection. As an extension of our investigations in resonantly enhanced THG in HCl, we conducted first tests of resonantly enhanced sum frequency mixing (SFM) in HCl. Here, we probe the vibrational two-photon resonance with near-infrared nanosecond radiation. With this approach we both benefit from resonance enhancement and from a lower detuning to the first excited electronic state of HCl. Also, the visible SFM signal is detectable with photomultiplier tubes. We conducted measurements with collimated probe radiation at 1064 nm wavelength. We identified the same two-photon resonances as with THG, albeit with large background signals, approximately one order of magnitude larger, than for THG. Due to this background we did not improve the detection limit for HCl molecules. The intensity dependence of the SFM signal shows the expected dependencies for SFM. The spectral linewidth of the two-photon resonance appears to be narrower, when measured with SFM. An analysis of the background signal indicates, that it is off-resonant SFM signal, which emerges from the CaF₂ windows of the gas cell, i.e., a technical artefact of our experimental setup. Nevertheless, we conducted first of its kind comparisons of vibrational two-photon resonances with different two-photon resonant frequency conversion processes. We found in particular, that THG has the potential for realistic applications of molecule detection