Atomic Layer Etching Sharpens MISHEMT Transistor Production

·wetetched.com

The Numbers That Matter

This story reports a measured change such as 0.5%, 30% and 2025. Figures like this show direction and scale, so it helps to keep them separate from the surrounding commentary.

  • Change / rate: 0.5% The ability of ALE to maintain critical dimension uniformity across full wafers—often better than 0.5% 3-sigma variation—is a compelling advantage for high-mix, low-volume runs typical of compound semiconductor manufacturing.
  • Change / rate: 30% Key performance indicators such as wafer-to-wafer repeatability and defect density are being tracked in qualification lots, and early data suggest that ALE can slash device-to-device threshold voltage spread by more than 30%…
  • Date / period: 2025 Near-Term Outlook: Milestones and Supply-Chain Readiness The coming months will see equipment manufacturers release dedicated ALE recipes for GaN MISHEMTs, with full process qualification expected by mid-2025.

The adoption of atomic layer etching marks a pivotal shift in the manufacture of metal-insulator-semiconductor high electron mobility transistors, delivering the atomic-scale precision essential for next-generation high-frequency and power electronics.

Why MISHEMTs are Critical for Next-Generation Electronics

Chemical Etching 6
Chemical Etching 6

Metal-Insulator-Semiconductor High Electron Mobility Transistors (MISHEMTs) are advanced field-effect transistors that incorporate a thin dielectric layer between the gate metal and the semiconductor. This insulator, often aluminium oxide or silicon nitride, suppresses gate leakage compared to conventional Schottky-gate HEMTs, enabling higher gate voltages and improved noise performance. Built on wide-bandgap materials such as gallium nitride (GaN) and increasingly on ultrawide-bandgap semiconductors like aluminium gallium nitride, MISHEMTs are key enablers for 5G and future 6G infrastructure, military radar, satellite transceivers, and compact power supplies. Their capability to operate at frequencies exceeding 100 GHz with high power-added efficiency makes them attractive for both commercial and defence applications.

However, fabricating reliable MISHEMTs requires exceptional control over the gate recess region. The depth and profile of the recess directly affect threshold voltage, transconductance, and current collapse behaviour. Microscopic variations can lead to broad parametric distributions across a wafer, reducing yield and performance in systems that depend on tightly matched transistors.

How Atomic Layer Etching Achieves Unmatched Precision

metal etching_
metal etching_

Atomic layer etching (ALE) is a subtractive manufacturing process that removes material through alternating, self-limiting chemical steps. In a typical plasma ALE cycle, a reactive gas is first introduced to form a modified surface layer without sputtering the bulk material; a subsequent step then selectively removes this modified layer using low-energy ion bombardment or a secondary gas. Because each cycle is self-limiting and removes roughly one monolayer, total etch depth can be controlled by the number of cycles, not by time. This eliminates the aspect-ratio-dependent effects and centre-to-edge non-uniformity that plague traditional reactive ion etching.

For MISHEMT gate recess, ALE offers the ability to stop precisely on an ultrathin etch-stop layer or at an exact barrier thickness, preserving the critical interface. The technique is compatible with a range of chemistries, including chlorine-based processes for GaN and fluorinated plasmas for dielectric layers, and it can be seamlessly integrated into existing cluster tools already used for dielectric deposition. Consequently, semiconductor fabs can adapt their process flows without major infrastructure changes, accelerating adoption. In contrast to conventional chemical etching machines, which may lack the necessary consistency for advanced compound semiconductors, ALE tools are engineered to handle the delicate heterostructures of GaN-based devices with repeatable atomic-layer fidelity.

Manufacturing Integration and Industry Momentum

Equipment makers now offer ALE chambers alongside traditional etch modules on single-wafer platforms, enabling hybrid process sequences. Several research groups have already demonstrated GaN MISHEMTs with record-low off-state leakage and significantly improved linearity using ALE-defined gate recesses. Meanwhile, foundries are conducting pilot runs on 150 mm and 200 mm wafers to validate throughput and defectivity for volume production. The ability of ALE to maintain critical dimension uniformity across full wafers—often better than 0.5% 3-sigma variation—is a compelling advantage for high-mix, low-volume runs typical of compound semiconductor manufacturing.

From a supply-chain perspective, the shift towards ALE is stimulating demand for high-purity process gases and advanced endpoint detection systems, as well as for metrology tools capable of in-situ monitoring at the atomic scale. Standards organisations such as SEMI are beginning to develop guidelines for ALE process qualification, a sign that the technology is moving from research into mainstream production.

Near-Term Outlook: Milestones and Supply-Chain Readiness

The coming months will see equipment manufacturers release dedicated ALE recipes for GaN MISHEMTs, with full process qualification expected by mid-2025. Key performance indicators such as wafer-to-wafer repeatability and defect density are being tracked in qualification lots, and early data suggest that ALE can slash device-to-device threshold voltage spread by more than 30% compared to inductively coupled plasma etching. As the technology matures, the industry is also eyeing its extension to gate-last integration schemes and to emerging materials like scandium aluminium nitride, which could push operating frequencies even higher. With several tier-one foundries already announcing plans to offer ALE-enabled processes by late 2025, the technology is on a clear trajectory to become the standard for precision compound semiconductor fabrication.

Atomic Layer Etching for MISHEMT Manufacturing at a Glance
Category Key Detail
Process Self-limiting cyclic etch with atomic-layer control
Target Device GaN MISHEMTs for RF and power applications
Typical Removal per Cycle <1 nm
Advantages Ultra-high uniformity, low damage, precise depth control
Wafer Sizes 100 mm to 200 mm, scaling demonstrated
Equipment Availability Modules integrated on cluster tools, stand-alone systems by 2025
Next Milestone High-volume manufacturing ramp starting late 2025

Why This Matters

As the semiconductor industry pushes towards higher frequencies and power densities, atomic layer etching provides the atomic-scale control needed for next-generation compound semiconductor devices, ensuring better uniformity and performance in RF and power applications.

FAQ

What is a MISHEMT?

A MISHEMT is a Metal-Insulator-Semiconductor High Electron Mobility Transistor that inserts a thin dielectric layer under the gate to reduce leakage current and improve voltage handling compared to a standard HEMT.

How does atomic layer etching improve MISHEMT fabrication?

ALE removes material one atomic layer at a time through self-limiting cycles, enabling precise gate recess depth control and excellent uniformity across the wafer, which directly translates into better device performance and higher yield.

Why is GaN commonly used for MISHEMTs?

GaN is a wide-bandgap semiconductor with high electron mobility and breakdown field, allowing MISHEMTs to operate at high frequencies and power levels essential for 5G, radar, and efficient power conversion.

When will atomic layer etching be widely adopted in production?

Equipment suppliers expect ALE modules to be fully qualified for GaN MISHEMT manufacturing by mid-2025, with volume production ramping up later that year as foundries complete pilot runs and qualification.

Source: "precision etching" – Google News