Fig. 1. Experimental setup of MIR upconversion imaging based on the external-cavity pumping. The pump source from a YDFL operates at the single-longitudinal mode at 1064 nm. The QCL serves as the MIR signal at a tunable wavelength around 4.3μm. Then, the CW MIR beam is expanded with a pair of lenses before being steered into an optical cavity to implement the sum-frequency generation. The cavity is comprised of a flat mirror M1, a concave mirror M3, and a CPLN crystal. One crystal facet is coated with high reflection at 1.064μm, thus serving as the end mirror. This cavity is stabilized with a digital locking unit based on a programmed field programmable gate array (FPGA). Under the locked state, the pump power can be significantly enhanced. The MIR signal after transmitting the object is then steered into a 4f imaging system, where the crystal is placed at the Fourier plane. The upconverted image is recorded by an EMCCD after passing through a series of spectral filters. YDFL and YDFA, Yb-doped fiber laser and Yb-doped fiber amplifier; EOM, electro-optical modulator; L, lens; M, mirror; Col., collimator; ISO, isolator; PZT, piezoelectric actuator; PD, photodiode; HV, high-voltage amplifier; FG, spectral filter group.

Fig. 2. Optical cavity design and characterization. (a) Diagram of the semi-monolithic optical cavity. The cavity consists of a mirror M1, a concave mirror M3, and a crystal end M2. The end of the crystal is coated with high reflection at 1.064μm, which is used as a flat mirror for the cavity. (b) Beam radius within the nonlinear crystal as a function of the distances of D1 and D2. The red point denotes the beam waist used in the experiment. Note that the white area represents the unstable region for an optical cavity. (c) Evolution of the beam waist along the optical cavity, when M1 is defined as the starting position. The shaded red area denotes the nonlinear crystal. (d) The intracavity power varies with the injected pump power. The thermal locking technique is used to stabilize the optical cavity at high-power operation.

Fig. 3. Thermal effect on the crystal-embedded optical cavity. (a), (b) Experimental measurements of transmission peaks during the cavity-length scanning via the PZT in the case of various pump powers (a) and sweep rates (b). (c), (d) Corresponding numerical simulations at different pump powers (c) and sweep rates (d). (e) Process diagram of the optical locking based on the thermal effect. The locking operation initiates at 0.05 s and stabilizes after ∼3s. The signal vibration is due to the external disturbance for the purpose of verifying the robustness. Note that the behavior of the PDH locking is presented for the sake of direct comparison, which shows a pronounced thermal effect on the error signal.

Fig. 4. Performance characterization of the MIR upconversion imaging under coherent and incoherent illuminations. (a) Coherent MIR upconversion images for the USAF resolution target. (b) Representative cross-section traces are given for the line pairs of the first element in the zeroth group under coherent illumination. (c) Incoherent MIR upconversion images. (d) Corresponding cross sections, which show an enhanced contrast. Note that all the images are acquired by an EMCCD at an exposure time of 80 ms.

Fig. 5. Real-time MIR spectral imaging for gas flow monitoring. (a) Transmission spectrum of CO2 at standard atmospheric pressure, which is obtained from the HITRAN database. The optical spectrum of the used QCL light source centers at 4308 nm, coinciding with the gas absorption peak. (b) Real-time monitoring of the CO2 concentration within the imaging FOV. The CO2 gas is delivered from a prefilled reservoir bag and injected into the imaging area through a nozzle, as illustrated in the inset. By repeatedly compressing the gas bag, the gas concentration varies in time as expected. (c) Spectral images during the period of 2.28 to 2.44 s. The MIR illumination power is ∼10μW/mm2. The exposure time of the camera is set to 40 ms, corresponding to a frame rate of 25 fps. See Video 1 for the recorded dynamics (Video 1, MP4, 896 KB [URL: https://doi.org/10.1117/1.APN.4.5.056003.s1]).