Room-temperature continuous-wave lasing from monolayer molybdenum ditelluride integrated with a silicon nanobeam cavity

Abstract

Monolayer transition-metal dichalcogenides (TMDs) have the potential to become efficient optical-gain materials for low-energy-consumption nanolasers with the smallest gain media because of strong excitonic emission. However, until now TMD-based lasing has been realized only at low temperatures. Here we demonstrate for the first time a room-temperature laser operation in the infrared region from a monolayer of molybdenum ditelluride on a silicon photonic-crystal cavity. The observation is enabled by the unique combination of a TMD monolayer with an emission wavelength transparent to silicon, and a high-Q cavity of the silicon nanobeam. The laser is pumped by a continuous-wave excitation, with a threshold density of 6.6 W cm. Its linewidth is as narrow as 0.202 nm with a corresponding Q of 5,603, the largest value reported for a TMD laser. This demonstration establishes TMDs as practical materials for integrated TMD-silicon nanolasers suitable for silicon-based nanophotonic applications in silicon-transparent wavelengths.

1305 nm Few‐Layer MoTe2‐on‐Silicon Laser‐Like Emission

“…This may lead to deviations between the values shown here and those published, e.g. for, but it gives a better comparison and provides more insight.…”

“…Once the pump power is increased above 1.75 kW/cm 2 , the peak narrows down to below 0.495 nm. When increasing the pump power above threshold, linewidth re‐broadening is observed, which we explain with an increase in carrier‐density and corresponding loss with pump power . Notice that the linewidth change is not as obvious as expected.…”

“…At the wavelength of 1130 nm, the intrinsic absorption loss in silicon is approximately 6.5 dB/cm. Although this loss is sufficiently low to enable the demonstration of high quality (Q‐factor) photonic crystal (PhC) cavities, the practical telecommunication wavelength range starts at 1260 nm, where the loss in silicon drops to well below 0.01 dB/cm. Regarding the MoTe 2 gain material, we observe a similarly favourable trend with moving to longer wavelength, as the absorption coefficient of MoTe 2 drops from 2.4 × 10 5 cm −1 at 1130 nm to 6 × 10 3 cm −1 at 1300 nm .…”

“…Excitingly, several TMD‐based lasers have now been demonstrated, including a continuous‐wave (CW) WSe 2 monolayer laser emitting around 740 nm at 80 K, based on a gallium phosphide photonic crystal (PhC) cavity, a microdisk laser including a tungsten disulfide (WS 2 ) monolayer that is sandwiched by a Si 3 N 4 and hydrogen silsesquioxane (HSQ), operating at 612 nm and at 10 K, and a room temperature four layer MoS 2 laser emitting in the wavelength range of 600 to 800 nm, based on a vertically coupled microdisk and microsphere cavity . Whereas these initial demonstrations were focussed on the visible regime, a TMD‐based laser operating in the silicon transparency window at 1132 nm has been demonstrated very recently, using a silicon nanobeam cavity and a monolayer of MoTe 2 as the gain material. Our work now demonstrates that the operating wavelength can be pushed further into the infrared and into the 1260 nm to 1360 nm communications window (“O band”), which signifies a technological step‐change for this type of laser.…”

“…At 1300 nm, the photoluminescence (PL) efficiency, as well as the gain, of MoTe 2 monolayers is very similar to that of few‐layer sheets . Therefore, using few‐layer MoTe 2 as the gain medium for 1300 nm emission can benefit from multiple aspects: high‐quality, few‐layer MoTe 2 can now be grown by chemical vapor deposition (CVD) in a relatively large size (>2 cm 2 ), while monolayers obtained either by CVD growth or by mechanical exfoliation are currently limited to sizes of 100s of μm 2 ; (2) similarly, the fabrication tolerances increase when stringent layer control is not required, (3) the gain peak red‐shifts with increasing layer thickness and (4) the mode overlap Γ is significantly larger when employing few‐layer material . We believe that the latter point is pivotal to achieving the low‐threshold operation we report here.…”