Minimal-gain-printed silicon nanolaser

Abstract

While there have been notable advancements in Si-based optical integration, achieving compact and efficient continuous-wave (CW) III-V semiconductor nanolasers on Si at room temperature remains a substantial challenge. This study presents an innovative approach: the on-demand minimal-gain-printed Si nanolaser. By using a carefully designed minimal III-V optical gain structure and a precise on-demand gain-printing technique, we achieve lasing operation with superior spectral stability under pulsed conditions and observe a strong signature of CW operation at room temperature. These achievements are attributed to addressing both fundamental and technological issues, including carrier diffusion, absorption loss, and inefficient thermal dissipation, through minimal-gain printing in the nanolaser. Moreover, our demonstration of the laser-on-waveguide structure emphasizes the integration benefits of this on-demand gain-printed Si nanolaser, highlighting its potential significance in the fields of Si photonics and photonic integrated circuits.

Fig. 1.. On-demand minimal-gain-printed Si nanolaser: Numerical characterization.

(A) Schematic of device exhibiting an on-demand μ-transferred III-V NB on 1D PhC structures defined on a single Si WG. (B) Top and side views of the obtained |E|² profiles of a resonant mode at 1536.8 nm. (C) Half-cut top views of a logarithmic plot of the steady-state charge carrier density (n) in III-V homogeneous (left) and III-V/Si heterogeneous models (right). (D) Half-cut side views of the steady-state temperature distribution (color plots) and vectorial thermal flux (red arrows) in III-V homogeneous (left) and III-V/Si heterogeneous models (right). In the heat transfer simulation, a plane of heat source with 1 mW radiation was introduced at the center of the NB. For all simulations, InGaAsP was considered as the III-V material. All scale bars denote 2 μm.

Fig. 2.. Device fabrication using the μ-transfer-enabled gain-printing technique.

(A) SEM images of the free-standing InGaAsP NB array (left) and single NB (right). (B) SEM images of the 1D Si PhC array (top) and single Si PhCs (bottom). Scale bars for the array and single devices denote 20 and 2 μm, respectively. (C) Schematic of the μ-transfer-printing setup. A cubical PDMS μ-tip with one side of w was used in the conventional transfer-print setup. High-precision XYZ stages with nanoscale step sizes were used. (D) Image of the fabricated PDMS μ-tip (left). Optical image of a single μ-tip (right). Scale bar denotes 20 μm. (E) On-demand addressable gain printing: (1) aligning, (2) breaking terminals, (3) pick-up, (4) align-registration, (5) peeling-off, and (6) pulling-out. (F) Optical image of completely fabricated on-demand minimal-gain-printed nanolasers. Scale bar denotes 10 μm. (G) Tilted SEM image of the white dotted box in (F). Scale bar represents 2 μm.

Fig. 3.. Pulsed lasing characteristics.

(A) SEM image of the single laser device and NIR camera images presenting SE, ASE, and lasing actions with increasing P_peak. The repetition rate and pulse width of a 976-nm pump laser were 1 MHz and 100 ns, respectively. Scale bars denote 2 μm. (B) Recorded spectra at below-, near-, and above-threshold levels revealing modal competition (Mode-A, Mode-B, and Mode-C) and the selection of the lasing mode (Mode-B). (C) Laser characteristic curve with P_th ~ 22.8 μW. Inset: Recorded lasing spectrum at near P_th. (D) Rate-equation analysis. L-L curve fittings for β = 1 (red), 0.5 (yellow), 0.13 (green), and 0.01 (blue). Experimental data (green-colored open circles) were best fitted with β = 0.13. (E) Recorded spectra (dots) and Lorentzian fits (lines) from far-below- (P_peak = 8.5 μW) to far-above-threshold (P_peak = 40.7 μW) levels. (F) Plots of the center wavelength (top) and linewidth (bottom) as a function of P_peak.

Fig. 4.. Emission characteristics under CW pumping.

(A) Normalized above-threshold lasing spectra at 10% (black) and 50% (blue) duty cycles and under CW (red) conditions at RT. Here, P_peak values for 10 and 50% conditions were 93 and 128 μW, respectively, and P_cw was 133 μW. The relative ratios of the peak intensity for black, blue, and red are 1.0, 2.5, and 0.6, respectively. (B) Emission characteristic curve (black) and linewidth plot (blue) under CW conditions. (C) Plots of the peak wavelength as a function of the normalized power P_cw/P_th, where P_th ~ 50 μW. Blue and red arrows represent types of peak shifts occurring while the phase transitions from SE to ASE and stimulated emission (StE). (D) Emission spectra of the TEC-cooled device under pulsed (blue) and CW (red) conditions. The chip was noncryogenic and Peltier-cooled at 289 K. For both conditions, P_peak and P_cw can be visualized inside the panel. (E) Emission characteristic curve (black) and linewidth plot (blue) for the TEC-cooled device. (F) Plots of the peak wavelength as a function of P_cw/P_th. The gray arrow represents a suppressed peak shift from ASE to StE transition.

Fig. 5.. Laser-on-WG application.

(A) Optical and SEM images of the on-demand WG-coupled laser device. Scale bar denotes 20 μm. Top and side views of the magnified SEM images depict the device in a white dotted box in the top panel and highlight structural details including modified air holes (left) and the seamless integration of NB with Si PhCs (right). Scale bars denote 1 μm. (B) Co plots of the calculated Q-factor (black) and transmission efficiency (P_trans/P_tot, blue) as a function of the number of air holes. Here, Mode-C is considered as the lasing mode. (C) Laser-to-WG coupling experimental results. NIR camera images captured the complete device (top) and either the WG end (blue dotted box, center) or the laser (red dotted box, bottom) screened devices. Scale bar denotes 5 μm. (D) Spectra recorded at the center of the laser (black) (middle panel in C) and at the end of the WG (red) (bottom panel in C). Inset: Spectra in logarithmic scale.