摘要： Laser spectrometers based on Quantum Cascade Lasers (QCL) provide good results for gas sensing in terms of sensibility and selectivity thanks to the characteristics of these sources. Distributed feedback (DFB) configuration provides narrow linewidth that enhances selectivity and their emission in the mid-IR enables to reach the fundamental absorption bands of molecules. The main limitation of these sources is their narrow tuning range (~ 10 cm-1) that prevents from monitoring complex species with broad absorption spectra or realizing multi-gas sensing. To obtain a broader tuning range, one solution consists of QCL DFB array with an arrayed waveguide grating to perform multi-species spectroscopy [1]. A more common technique is to implement the laser in an external cavity system. In GSMA, a commercial Extended-Cavity Quantum Cascade Laser emitting at 10.5 μm has been used to demonstrate photoacoustic gas sensing of heavy molecules such as butane [2] and a lab-made EC-QCL emitting at 7.5 μm was developed for detection of acetone and POCl3 in gas phase [3]. Moreover, the use of an external cavity allows the possibility to perform Intra-Cavity Laser Absorption Spectroscopy (ICLAS). In this method the sample is placed within the laser resonator. Absorption lines of the sample imprint signatures on the spectrum because they influence it during many round trips. The cavity output light is detected to perform spectroscopy. The used QCL developed by mirSense is operated in cw operation at room temperature and emits from 7.4 to 7.8 μm. A technique using the QCL compliance voltage [4] called EVIS (EC-QCL Voltage Intracavity Sensing) is used to retrieve the spectrum of the gas inside the cavity. This way no detector outside the cavity is needed. With gas inside the cavity, losses rise implies a variation of intracavity intensity thus a variation of laser voltage. The large wavenumber sweeping is performed through the grating rotation and the variation of QCL current supply permits to obtain a small wavelength variation. In order to suppress the QCL mode hops influence, a current ramp is applied so that the small wavelength total variation is equal to 1 QCL mod hop. Then QCL voltage is recorded during the current ramp, for all grating positions, and a numerical treatment is performed to improve the signal. Fig. 1 presents a result of this measurement (on the left) compared with the absorption coefficient calculated from the HITRAN database [www.hitran.org] (on the right). This measurement corresponds to a cavity filled with 0.3% of CH4 in air (containing water vapor) at atmospheric pressure. This preliminary result shows a good agreement between the voltage difference and the calculated absorption coefficient. The resolution is estimated to approx. 0.1 cm-1.