IQE and Atomera agree RF GaN development

IQE and US semiconductor materials company Atomera have signed a joint development agreement to investigate radio-frequency gallium nitride structures grown on engineered silicon wafers. The programme combines Atomera’s Mears Silicon Technology with IQE’s compound semiconductor epitaxy capabilities, aiming to improve RF linearity and reduce losses while creating a potential route to volume manufacturing. The agreement starts development work rather than confirming a qualified production process.

The companies plan to produce and test GaN epitaxial structures on silicon wafers incorporating Atomera’s proprietary material. They expect to optimise the multilayer structure through several development cycles, using measurements to guide subsequent changes. IQE is identified as the intended manufacturer of MST-enabled wafers for foundry and integrated device manufacturer customers, with the partners also seeking samples suitable for engagement with fabless semiconductor businesses.

Radio-frequency GaN devices are used where substantial power must be handled at high operating frequencies. Their electrical characteristics influence the efficiency and fidelity of amplifiers and other RF circuits, while the underlying wafer can contribute losses that undermine the performance of the active device. High-performance RF GaN has commonly been produced on silicon carbide substrates, which have favourable properties for demanding applications but can carry higher manufacturing costs.

Silicon provides a different substrate option, with the prospect of larger wafer formats and access to established semiconductor manufacturing processes. Growing gallium nitride on silicon nevertheless introduces material and electrical difficulties. Mismatches between the materials must be controlled during epitaxy, and unwanted conductive paths can allow RF energy to leak through the substrate. Parasitic channel effects are particularly troublesome as frequency rises, reducing the benefit of otherwise capable transistor structures.

Atomera’s Mears Silicon Technology modifies the silicon material near its surface using an engineered epitaxial structure incorporating oxygen. The company proposes that this approach can suppress unwanted substrate conduction and provide improved RF performance beneath the GaN layers. The degree of improvement depends on the resulting material stack, its interfaces and the manufacturing conditions, which the new partners will now examine together.

Earlier work by Atomera provides the technical basis for the agreement. In June 2026 the company reported more than a tenfold reduction in parasitic channel charge in evaluated material, together with encouraging RF linearity testing. Atomera also cited measurements performed by Incize. Those results concern the company’s earlier research and cannot be treated as performance measurements from IQE’s forthcoming jointly developed wafers or as independent product qualification.

The distinction between substrate measurements and finished device performance is material. An epitaxial wafer may have promising characteristics, but the RF behaviour of a commercial amplifier also depends on transistor geometry, fabrication, interconnections and packaging. Thermal management and the conditions under which the RF circuit operates can affect both efficiency and linearity. Foundries will need their own measurements before adopting a material process into production.

IQE specialises in growing controlled compound semiconductor layers for downstream device manufacturers. Layer thickness, composition, crystal quality and interface conditions influence the characteristics of the finished wafer. The joint programme will use that manufacturing expertise to develop the combination of MST-enabled silicon and GaN structures most suitable for the intended RF applications, rather than assuming that earlier experimental results transfer directly into an industrial process.

Atomera says the initial development will focus on RF GaN-on-silicon and seek the best achievable balance of linearity and electrical losses. The companies have not published target thresholds, an agreed completion date, a manufacturing capacity or committed customer orders. Their stated intention is to make samples available for foundry, integrated device manufacturer and fabless customer assessment after further material development.

The joint agreement also allows the scope to expand to additional GaN applications, including power electronics. Those uses place different demands on voltage blocking, switching behaviour and thermal performance, and would require their own material and device evaluations. The possibility of extending the research should not be read as a launch of a qualified power semiconductor product.

IQE’s role connects the earlier substrate research with an established compound semiconductor wafer producer. The next evidence will come from jointly developed wafers and their electrical characterisation, followed by customer testing where appropriate. Until those stages are completed, the reported loss and linearity advantages remain objectives supported by Atomera’s preliminary results, not commercially established performance claims for the new collaboration.