Telecom-band multiwavelength vertical emitting quantum well nanowire laser arrays

Highly integrated optoelectronic and photonic systems underpin the development of next-generation advanced optical and quantum communication technologies, which require compact, multiwavelength laser sources at the telecom band. Here, we report on-substrate vertical emitting lasing from ordered InGaAs/InP multi-quantum well core–shell nanowire array epitaxially grown on InP substrate by selective area epitaxy. To reduce optical loss and tailor the cavity mode, a new nanowire facet engineering approach has been developed to achieve controlled quantum well nanowire dimensions with uniform morphology and high crystal quality. Owing to the strong quantum confinement effect of InGaAs quantum wells and the successful formation of a vertical Fabry–Pérot cavity between the top nanowire facet and bottom nanowire/SiO2 mask interface, stimulated emissions of the EH11a/b mode from single vertical nanowires from an on-substrate nanowire array have been demonstrated with a lasing threshold of ~28.2 μJ cm−2 per pulse and a high characteristic temperature of ~128 K. By fine-tuning the In composition of the quantum wells, room temperature, single-mode lasing is achieved in the vertical direction across a broad near-infrared spectral range, spanning from 940 nm to the telecommunication O and C bands. Our research indicates that through a carefully designed facet engineering strategy, highly ordered, uniform nanowire arrays with precise dimension control can be achieved to simultaneously deliver thousands of nanolasers with multiple wavelengths on the same substrate, paving a promising and scalable pathway towards future advanced optoelectronic and photonic systems.


Growth method
Before growth, ~30 nm of SiO2 was first deposited on InP (111)A substrates at 300 °C by plasma-enhanced chemical vapor deposition.Then hexagonal arrays of holes were patterned by electron beam lithography, followed by reactive ion etching.After trim etching with phosphoric acid and hydrogen peroxide, the substrate was loaded into the MOCVD system (Aixtron CCS 3×2) with a close-coupled shower head reactor for SAE growth.Trimethylindium, trimethylgallium, phosphine and arsine were used as precursors for the In, Ga, P, and As, respectively.To adjust the lasing wavelength, the flux ratio of In/(In+Ga) in the NW quantum wells was increased from 0.45 to 0.6.  1 , while {112 ̅ 0} facets have lower surface energy at low-temperature condition [1][2][3] .Since facets with lower surface energy are more stable, the {11 ̅ 00} to {112 ̅ 0} facet rotation happens for InP core-shell NW to maintain the lowest possible surface energy during the lower temperature growth step.at the wavelength of ~875 nm, whereas the ZB InP substrate exhibits a longer wavelength peak at ~920 nm, indicating that the crystal structure of InP shell is predominantly WZ-based.On the lower energy shoulder of InP core-shell nanowire PL spectra, several smaller and periodic peaks could be observed, which are attributed to the Fabry-Perot cavity effect, as more optical modes can be supported in the InP core-shell nanowire due to their larger diameter.The minority carrier lifetime of the WZ InP nanowire and InP core-shell nanowire can be extracted to be 0.74 and 1.10 ns, by fitting the TRPL decay curve with a single exponential model.Here, we note a high concentration of P atoms in the QWs, indicating quaternary InGaAsP formation instead of intended InGaAs growth.Some residual Ga and As atoms can also be observed in the InP barrier layers, suggesting atomic interdiffusion (In-Ga and As-P) at the QW/barrier interface.On the other hand, the Ga concentration is much lower than In concentration in the QW layers, even though a much higher molar fraction of Ga precursor (that of than In precursor) was used during the growth, which may be attributed to the shorter diffusion length of Ga atoms compared with that of the In atoms.

Mode simulations
The two-dimensional (2D) finite difference eigenmode (FDE) solver was used to calculate the transverse modes supported in the as-grown NW.The model with an InP NW standing on the SiO2 substrate with a 120 nm hole opening was used.A square shaped 2D FDE solver region was set at the center of the cross-section of NW (Size: 5 times of NW diameter; Boundary setting: Metal).reasonable reflectivity can still be obtained for the TE01, HE21a, HE21b, and EH11a and EH11b modes.

Fig
Fig. S1.(a, c, e) Schematics of the lateral (left) and vertical (right) cross-section of WZ based InP

Fig
Fig. S2.(a) Normalized PL spectra of WZ InP NW, WZ InP core-shell NW and ZB InP substrate at

Fig
Fig. S3 STEM-HAADF image taken along [112] zone axis from the vertical cross-section of

Fig.
Fig. S4 STEM-HAADF images of the lateral cross-section of four different InGaAs/InP QW NWs

Fig. S6 EDXArea 5 .
Fig. S6 EDX maps of the NW lasing at 1550 nm.The QW region shown in the Area #1 of HAADF

Fig. S10
Fig. S10 Bottom surface reflectivity versus the thickness of SiO2 for each transverse mode of the NW with

Fig
Fig. S11 Simulated electric field distributions of the different transverse modes.

Fig
Fig. S13 Temperature dependent PL intensities under weak pump fluence, showing a slight red shift of ~ 15

Fig. S14
Fig. S14 Telecom-band emission spectra at different pump fluences from different NWs.

Fig. S15 PL
Fig. S15 PL images from three other locations in the NW array after lasing threshold (a-c).(d) An attenuated image on (c).

Table S1 .
Summary of different growth conditions and facet information for WZ, mixed ZB/WZ, and WZ core-shell InP nanowires.

Table S3 .
Summary of the various properties of different InGaAs/InP MQW nanowire structures achieved under different growth strategies.