IN Brief:
- QPT has launched qDesign to optimise its qAttach thermal interface layer.
- The service combines geometry generation, thermal-mechanical simulation, and design iteration.
- High-frequency GaN drives and dense AI power supplies are increasing pressure on power module heat removal.
QPT has launched qDesign, an AI-driven generative design service for optimising its qAttach thermal interface layer in power modules.
The Cambridge-based power electronics company is offering the service to semiconductor manufacturers, power module OEMs, and system integrators. Customers provide module specifications, including die geometry, substrate stack, heatsink interface, and thermal and stress targets, while QPT’s workflow generates and simulates a tuned qAttach geometry.
The service combines geometry generation, thermal-mechanical simulation, and design iteration. Candidate qAttach structures are evaluated through cloud-based finite-element analysis for thermal and stress performance, with results feeding into further design iterations until the geometry matches the required thermal and reliability conditions.
Each revision is captured as a traceable deliverable, including parametric inputs, simulation results, engineering drawings, and reports for review. QPT is offering qDesign as a partnered engineering service rather than a catalogue part because the performance of qAttach depends on the specific die, substrate, heatsink, bond line, and manufacturing constraints of each module.
qAttach was first introduced as a die-attach process intended to remove waste heat from power semiconductor dies more effectively than conventional sintered die-attach approaches. The company says the technology can deliver up to 15 times better thermal performance than existing alternatives and support 1MHz-plus hard-switching in GaN-based systems.
Thermal interface design is becoming a limiting factor in high-density power electronics. A semiconductor die may support faster switching and higher voltage operation, but module reliability is still governed by heat flow, thermal cycling, mechanical stress, and delamination risk. As wide-bandgap devices shrink the power stage, packaging becomes a larger share of the engineering problem.
QPT’s 1MHz GaN motor-drive demonstrations showed how high switching frequency can change drive control, diagnostics, and system behaviour. qDesign addresses the layer beneath that system: removing heat quickly enough for the semiconductor and package to maintain the intended performance.
AI-assisted design also fits the complexity of power module packaging. The available geometry space is too large for manual iteration alone when engineers are balancing ultra-thin bond layers, stress relief, thermal resistance, manufacturability, and reliability. Automating the generate-simulate-iterate loop allows more candidates to be explored before a design reaches physical prototyping.
The critical test will be correlation between model and hardware. Simulation-led optimisation depends on reliable material data, process control, and measurement feedback from real modules. If that loop holds, qDesign could reduce packaging development time while making the thermal interface a deliberate design variable rather than a fixed constraint.
