This article dives into a breakthrough: integrating high-performance III–V photonic crystal lasers right on silicon substrates. Researchers pulled this off using selective lateral heteroepitaxy.
They grew InP membranes, coplanar with the silicon waveguide layer, inside a single MOCVD reactor. That’s pretty wild—it means a wafer-scale route to on-chip lasers that are tightly confined and couple efficiently to silicon photonics.
Monolithic integration of III–V PhC lasers on SOI
Researchers from Sun Yat-sen University, led by Yu Han and Siyuan Yu, showed how to monolithically integrate III–V photonic crystal lasers right onto silicon-on-insulator (SOI) substrates. Their approach revolves around selective lateral heteroepitaxy, growing large-area InP membranes that sit flush with the silicon waveguide layer in just one growth step.
This structure sandwiches the III–V material between silicon oxide layers, locking in strong optical confinement—no need for suspended membranes here. By combining III–V gain materials with silicon photonics, they’ve made it possible to generate light efficiently on a single chip.
Since the III–V membrane and the Si layer are on the same plane and share thickness, the design supports efficient optical coupling to silicon waveguides. The team embedded vertical InGaAs/InP quantum wells right at the optical field maximum of the cavity.
This precise alignment puts the laser’s gain region directly in the cavity mode, which boosts efficiency and cuts down on losses. Also, they avoided etching air holes through the active layer—a move that keeps surface non-radiative recombination low and reduces waste in pumped areas.
Key technical advances
The team used a single-growth-step method, skipping standard fabrication steps like substrate undercut, membrane transfer, suspended structures, or tricky bonding and regrowth. That’s a big deal for mechanical stability and could make factory-scale manufacturing a lot more realistic.
- Monolithic integration on SOI using selective lateral heteroepitaxy—no need to transfer active regions or mess with suspended membranes.
- Coplanar InP membranes and Si waveguide layer with matching thickness for efficient optical coupling to silicon circuits.
- Placing vertical InGaAs/InP quantum wells at the cavity’s optical field maximum to squeeze out the most gain.
- Skipping air holes in the active region helps cut down surface recombination and keeps pumped-area losses minimal, so pumping is more efficient.
- This method supports low-threshold, single-mode lasing in the telecom band—plus, it looks promising for electrically pumped setups integrated with Si photonics.
- It’s compatible with wafer-scale production and flexible enough for different microcavity designs, including edge- and surface-emitting devices.
Implications for silicon photonics and on-chip interconnects
This development marks a real step forward for cost-effective, scalable integration of laser sources right on silicon. It’s something folks in on-chip optical interconnects have been chasing for ages.
The wafer-scale compatibility and easier fabrication could knock down manufacturing barriers. That means denser photonic circuits, and you don’t have to fuss so much with alignment or bonding headaches.
By letting us put electrically pumped III–V lasers together with Si waveguides, this approach could make it a lot easier to roll out laser-powered transceivers for data communications and computing interconnects. That’s a big deal if you’re trying to scale up.
- You could get high-volume production by using the SOI fabrication lines we already have—no need to spend big on new processes.
- It works with a bunch of different laser geometries, so designers get more flexibility for chip-scale photonics.
- There’s better thermal and mechanical stability compared to some suspended-membrane tricks, which should help reliability in commercial gear.
- This opens doors to fully integrated optoelectronic systems—imagine light generation, modulation, and detection all on one silicon platform.
Here is the source article for this story: Monolithic III-V membrane photonic crystal lasers on soi using selective lateral heteroepitaxy