An instrument for material-process codesign.

The Intelligent Codesign Cloud Lab for Non-Equilibrium and Low-Dimensional Materials (ICoN-PCL) is a Purdue-led NSF Programmable Cloud Lab test bed. It integrates programmable laboratories spanning synthesis through fabrication with a data and AI platform for designing, optimizing, and executing research workflows.

What we are building: a set of programmable laboratories, from synthesis through fabrication, integrated with a data and AI platform for designing, optimizing, and running research workflows.

The bet: that opening these resources to far more people makes material–fabrication codesign possible for the first time, and shortens the path from a new material to a validated device or alloy. Testing a newly synthesized MXene as a strengthening dispersion in a metallic alloy takes months or years today. It should take days.

Area 1: 2D materials for electronic devices

TSMC, Intel, Samsung and imec all point to 2D materials as a leading path for future logic devices. The reason is a thickness problem: silicon's carrier transport degrades sharply once the channel body thins below a few nanometers, while TMDs stay semiconducting and keep useful transport down to a single monolayer under 1 nm thick. That makes them the leading candidate channel material at the 5 nm node and below. Van der Waals 2D–2D contacts help on the injection side, reducing the Fermi-level pinning and metal-induced gap states that dominate conventional metal–semiconductor junctions. Hole injection is the harder half: low-resistance p-type contacts to TMDs are not solved, and that is one of the interface problems this test bed exists to attack.

MXenes are candidates for ultra-scaled, highly conductive source and drain materials, carrying current densities on the order of 10⁸ A/cm², which is well above what electromigration permits in copper interconnects. TMDs are also under study for advanced interconnects and non-volatile memories.

The central obstacle is interface control. Because these materials are two-dimensional, interfaces dominate performance and ultimately determine device properties, interfacial defects are not well controlled, and the design space across materials, stacks and processing runs to millions of configurations. Direct bottom-up growth of MXenes and their wafer-level deposition must both be solved before 2D device manufacturing is realistic.

Searching a space this large is not a throughput problem alone; it is a planning problem. Predictive twins and agent-driven design are what make the search tractable, and we will publish experiments-per-staff-hour as the test bed comes online.

Area 2: advanced manufacturing of alloys with nanoparticle dispersions

Advanced structural alloys today rely on precipitating thermodynamically favourable second phases or mixing in block ceramics. Oxide dispersion strengthening is the reference case: NASA's GRX-810 disperses ~20 nm yttria particles through a NiCoCr-based alloy and reports roughly twice the tensile strength and over a thousandfold longer creep life than comparable superalloys. The same approach using 2D dispersions has not been achieved, because processing conditions destroy them.

The approach here is to wrap metal powder particles in MXene before melting, so few-atom-thick metal carbide layers survive into the finished part and sit at the grain boundaries. What remains unclear is the mechanism governing strengthening, stability, and degradation, how much control manufacturing can exert over dispersion and phase evolution, and which of a vast space of low-dimensional additives is worth using at all.

ICoN-PCL is assembling an autonomous workflow spanning additive manufacturing, sample preparation, heat treatment, room- and high-temperature mechanical testing, aqueous corrosion testing, and surface integrity measurement, so base alloy, low-dimensional phase and microstructure can be co-optimized rather than chosen one at a time.