940-nm VCSELs grown by molecular beam epitaxy on Ge(001)

This paper reports the first demonstration of monolithically integrated 940-nm VCSELs grown by molecular beam epitaxy on Ge(001) substrates, achieving room-temperature continuous-wave lasing with threshold currents below 3 mA through the use of a GaAs-on-Ge virtual substrate and real-time in situ process monitoring.

The Gist

Imagine you are trying to build a tiny, super-efficient flashlight (a laser) that is so small it can fit inside a smartphone or a self-driving car's sensor. For decades, scientists have built these flashlights on a specific type of "ground" made of Gallium Arsenide (GaAs). It's like building a house on a perfectly flat, familiar foundation.

But there's a problem: The electronics inside our phones and computers are built on a different material called Germanium (Ge). To make these devices smaller and faster, scientists want to build the laser directly on top of the computer chip's foundation (the Germanium) instead of gluing a separate piece on top. It's like trying to build a skyscraper directly on a different type of bedrock.

This paper is the story of a team of scientists who successfully built one of these tiny lasers (a VCSEL) directly on a Germanium foundation using a very precise construction method called Molecular Beam Epitaxy (MBE).

Here is the breakdown of their journey, explained simply:

1. The Challenge: Building on "Rocky" Ground

Building a laser on Germanium is tricky. Germanium and the laser materials don't naturally fit together perfectly. It's like trying to stack bricks that are slightly different sizes; if you aren't careful, the tower will crack or fall over. Usually, scientists use a different method (MOVPE) to build these, but that method is like using a sledgehammer—it gets the job done, but it's hard to control the tiny details.

The team decided to use MBE, which is like using a robotic arm that places atoms one by one. This gives them incredible control, but they needed a way to "watch" the building process in real-time to make sure the tower didn't lean.

2. The Construction Site: Watching the "Bending"

To ensure the laser was built correctly, the scientists used two special "eyes" inside their vacuum chamber:

• The Curvature Monitor (The "Bending" Eye): As they built the layers, the wafer (the silicon-like disk) would bend slightly due to stress, just like a ruler bends when you press on it. The scientists watched this bending in real-time. They noticed something weird: when building on Germanium, the wafer bent in a different pattern than when building on the traditional material. It was like the ground was "squeezing" the building differently than expected. They realized the materials were relaxing and adjusting their stress as they cooled down.
• The Light Monitor (The "Mirror" Eye): They also shined a broad spectrum of light at the growing layers to see how the light bounced back. This told them if the "mirrors" inside the laser (which trap the light) were being built at the exact right thickness. It was like tuning a guitar string; they needed to know exactly when the pitch was perfect.

3. The Blueprint: The Laser Sandwich

The laser they built is essentially a sandwich:

• The Bottom Mirror: A stack of 35 pairs of layers that reflect light perfectly.
• The Filling (The Active Region): A tiny layer where the magic happens. This is where electricity turns into light. They put three very thin "pancakes" of a special material here to generate the laser beam.
• The Top Mirror: Another stack of mirrors, but fewer, to let some light escape as the laser beam.
• The Aperture: They created a tiny hole (about the width of a human hair) in the middle to force the light to go straight up, like a funnel.

4. The Result: A Working Laser

After the construction was done, they tested the device.

• The Test: They turned on the power at room temperature (no need for a freezer!).
• The Success: The laser turned on with very little electricity (less than 3 milliamps). That's like the power of a tiny LED.
• The Catch: When they pushed too much power, the laser got hot and stopped working efficiently (a "thermal rollover"). This is expected because Germanium doesn't pull heat away as well as the traditional material yet. But for a first attempt, it was a huge success.

Why This Matters

This paper is a "first" in the world of science. It is the first time anyone has successfully built this specific type of laser on Germanium using this precise, atom-by-atom method.

The Big Picture:
Think of this as the first time someone successfully built a high-speed train track directly onto a city's subway foundation. Before, the train and the subway had to be separate systems. Now, they can be integrated.

This breakthrough means we can eventually put these lasers directly onto computer chips. This will lead to:

• Faster internet: Data moving at the speed of light inside your computer.
• Better sensors: Self-driving cars that can "see" better and cheaper.
• Smaller devices: Phones and gadgets that are more powerful but use less battery.

In short, the scientists proved that you can build these complex lasers on the "wrong" foundation (Germanium) and still get them to work, opening the door for a new generation of super-integrated technology.

Technical Summary

Here is a detailed technical summary of the paper "940-nm VCSELs grown by molecular beam epitaxy on Ge(001)".

1. Problem Statement

Vertical-cavity surface-emitting lasers (VCSELs) operating near 940 nm are critical for consumer electronics, automotive sensing, and free-space optical communications. Historically, these devices are fabricated on Gallium Arsenide (GaAs) substrates using Metal-Organic Vapour Phase Epitaxy (MOVPE) or Molecular Beam Epitaxy (MBE).

• Integration Challenge: As photonic-electronic integration advances, there is a strong drive to transfer III-V VCSEL technology onto Group IV substrates (specifically Germanium, Ge) to enable monolithic integration with CMOS electronics.
• Material Constraints: While Ge offers large wafer sizes and lattice matching with AlGaAs, growing high-quality AlGaAs VCSELs on Ge presents significant challenges regarding epitaxial strain management and thermal mismatch.
• Methodological Gap: Previous demonstrations of VCSELs on Ge have been exclusively grown via MOVPE. There is a lack of reports on MBE growth of these structures, which is crucial because MBE offers superior control over layer thickness, abrupt interfaces, and digital alloying. Furthermore, real-time monitoring of strain evolution and optical cavity alignment during MBE growth on Ge has remained largely unexplored.

2. Methodology

The authors successfully fabricated 940-nm VCSELs using a hybrid growth approach combined with advanced in situ diagnostics.

• Substrate Preparation:
  • Used a 4-inch N-doped Ge(001) wafer with a 6° miscut.
  • Hybrid Growth: A 100 nm GaAs buffer layer was first grown via MOVPE to ensure high-quality nucleation and minimize defects at the Ge interface.
  • MBE Growth: The remainder of the VCSEL structure was grown using a Riber 412 MBE system.
• VCSEL Structure Design:
  • Bottom DBR: 35 pairs of n-doped Al$_{0.9}$Ga$_{0.1}$As/GaAs.
  • Active Region: A half-wavelength cavity containing three 7.5 nm undoped In$_{0.1}$Ga$_{0.9}$As quantum wells separated by GaAs barriers.
  • Top DBR: 17 pairs of p-doped AlGaAs/GaAs.
  • Aperture: A 30 nm Al$_{0.98}$Ga$_{0.02}$As layer for lateral oxidation to define the optical aperture.
  • Digital Alloying: Used for high-aluminum content layers and graded transition layers to smooth compositional changes.
• In Situ Monitoring (Real-Time Control):
  • Strain Monitoring: Used the EZ-CURVE system with Magnification-Inferred Curvature (MIC) imaging to track wafer bowing and strain evolution with micrometer-level sensitivity (up to 100 Hz).
  • Optical Monitoring: Used the EZ-REF system for broadband reflectometry (320–1700 nm) to monitor Fabry–Pérot oscillations and Distributed Bragg Reflector (DBR) stopband formation in real time.
• Device Fabrication:
  • Mesa etching (ICP-RIE) down to the bottom DBR.
  • Selective lateral oxidation at 430°C to form oxide apertures (approx. 11–16 µm).
  • Metal contact deposition and annealing.

3. Key Contributions

• First MBE Demonstration: This work presents the first reported monolithic integration of 940-nm VCSELs grown on Ge substrates using Molecular Beam Epitaxy.
• Advanced Process Control: The study demonstrates the effective use of combined in situ curvature and reflectometry to manage the complex strain dynamics and optical thickness during MBE growth on Ge, a capability previously unexplored for this material system.
• Strain Analysis: The authors provide a detailed analysis of the strain evolution differences between GaAs and Ge substrates, identifying non-linear curvature variations and potential strain relaxation mechanisms specific to the Ge substrate.

4. Results

• Structural and Optical Quality:
  • Reflectivity: Post-growth FTIR measurements confirmed a high-reflectivity stopband centered at 942.7 nm with a full width at 90% reflectivity of 93.6 nm, matching the design target.
  • Surface Morphology: AFM scans showed continuous surfaces without cracks or step bunching, though the roughness ($R_q \approx 8.5-9.7$ nm) was higher than typical GaAs-grown VCSELs. This did not significantly degrade optical performance.
  • In Situ Data: Real-time reflectometry confirmed uniform layer growth and accurate stopband formation. Curvature data revealed progressive compressive stress and distinct non-linear strain evolution on Ge compared to the linear behavior on GaAs.
• Lasing Performance:
  • Devices with mesa diameters of 35–40 µm (oxide apertures ~11–16 µm) achieved room-temperature continuous-wave (CW) lasing.
  • Threshold Current: Below 3 mA (specifically ~2.8 mA for a 35 µm device).
  • Output Power: Reached ~0.7 mW at 5.2 mA drive current.
  • Thermal Limitations: A pronounced thermal rollover was observed at higher currents, attributed to self-heating and limited heat extraction through the Ge substrate, highlighting a need for improved thermal management.

5. Significance

This research represents a major breakthrough in the field of heterogeneous integration:

1. Viability of MBE on Ge: It proves that MBE is a viable and precise method for growing complex III-V/Ge VCSEL structures, offering advantages in interface control and lower growth temperatures compared to MOVPE.
2. Scalable Integration: By demonstrating low-threshold lasing on Ge, the work paves the way for monolithic integration of high-performance light sources directly onto CMOS-compatible Ge substrates, essential for future high-speed optical interconnects and sensing arrays.
3. Process Diagnostics: The successful application of real-time strain and optical monitoring provides a critical roadmap for controlling the growth of strain-sensitive devices on lattice-mismatched substrates, ensuring high structural quality and optical performance.