Fig. 1. (a) Schematic of an MIMWG QCL frequency comb. (b) False-colored scanning electron microscopy (SEM) image of the front facet of a device (upper) and simulated mode profile of TM− at frequency of 2170  cm−1 (lower); PR, passive region. (c) GVD of the different order transverse modes in the passive WG (12 μm wide), GVD of the TM00,AR mode in the AR, and GVD of the TM− supermode from the MIMWG via coupling the two WGs. Inset: mode profile for TM10 in the passive WG. (d) Modal indices of TM00, TM10, and TM20 in the passive WG and TM00,AR in the AR, and model indices of the two supermodes TM+/− in the MIMWG. neff of the two TM+/− modes exhibits an anti-crossing effect at crossing point near frequency of ∼2200  cm−1. The shaded area indicates the spectral range of the anti-crossing effect. (e) Calculated overlap factor and GVD of TM− at 2170  cm−1 as a function of the width of passive WG. (f) Cross-section thermal distributions of the MIMWG QCLs with the passive region beneath (up) and on the top (down) of active region. The simulated structures are both arranged into epi-side down mounting on diamond submount which is soldered to copper heatsink with the temperature controlled by a TEC.

Fig. 2. (a) Measured GVD of an MIMWG QCL at different currents below threshold and that of a device without the passive WG (purple) and the subthreshold spectrum at 0.95Ith (red). (b) Power-current-voltage (P-I-V) characterization in CW operation and the corresponding (c) wall plug efficiency (WPE) at 15°C, 20°C, and 25°C, respectively. (d) Measured two-dimensional far-field profile at 0.8 A. (e) Measured and simulated far-field profiles along the vertical direction.

Fig. 3. (a) Dual-comb multiheterodyne characterization at T=20°C. LO FC, local oscillator FC; sample FC, reference FC; BS, beamsplitter; MCT, mercury–cadmium–telluride detector. (b) RF intermode beatnotes frequency as a function of bias current. (c) Lasing spectra measured at different bias currents of 0.78, 0.85, 0.88, and 0.95 A at 20°C and corresponding intermode beating spectra with linewidths of 473, 1000, 769, and 1300 Hz at beatnote frequencies of 9.056, 9.052, 9.049, and 9.045 GHz, respectively. RBWs of 380, 600, 1000, and 1000 Hz are used in the measurement, respectively. (d) Optical spectrum spanning ∼15  cm−1 measured for the reference device based on the same active design without dispersion engineering and corresponding intermode beatnote measured with RBW of 3 kHz at 20°C at biasing currents of 0.95 A; the linewidth is 13.09 kHz.

Fig. 4. (a) RF beatnotes of the LO comb and sample comb measured with a spectrum analyzer; both linewidths of the combs are less than 1 kHz. (b) Spectra of LO comb (blue) and sample comb (red) obtained with an FTIR spectrometer (0.2  cm−1 resolution, 20°C, ILO=800  mA, ISignal=804  mA). (c) Multiheterodyne spectra (acquisition time 2 ms) of the two combs with testing condition corresponding to (b). Inset is the zoomed-in multiheterodyne spectrum with a measured repetition frequency of 20 MHz. (d) The FWHM of a typical dual-comb tooth is 0.63 MHz.