Imec presented a three-metal-level (3ML) niobium-titanium-nitride (NbTiN)-based Josephson junction circuit at the 2026 Applied Superconductivity Conference (ASC).
The 3ML NbTiN Josephson junction circuit debuted at ASC is notable for achieving circuit design densities of 3.8 million junctions per cm2.
The routing was also demonstrated with superconducting wires scaled down to 30nm linewidth, further strengthening Imec’s NbTiN-based superconducting technology research platform for next-generation HPC and AI applications, which is open to industry partners.
Overcoming the restrictions of Nb-based technology
Conventional niobium (Nb)-based superconducting technology has traditionally lacked the scalability to be applicable to other domains, but Imec’s NbTiN-based circuit uses 300mm CMOS-compatible processes to enable competitive scaling.
Using three metal levels and junction diameters as small as 150nm, as well as three layers of high-density and low-loss NbTiN routing (wires and vias), the platform can be applied to passive transmission lines, incorporating inductors, ground planes, and clock or power resonators.
NbTiN metal wires were scaled down to 30nm linewidth, which is about ten times smaller than what can be achieved with conventional Nb-based technology.
For both Josephson junction circuits and interconnects, the superconducting properties (i.e., critical current density (Jc) and critical current (Ic), respectively) can be controllably tuned to address a broad range of target applications.
“A heterogenous approach” that “opens opportunities beyond HPC and AI”
Richard Rouse, Director Superconducting Digital Program at imec, commented: “With our superconducting digital program, we are targeting foundries, hyperscalers and system companies to take our NbTiN superconducting technology to the next level. We have structured our program around three pillars: (1) process technology, (2) design and EDA enablement, and (3) system-level scaling centered around 2.5D and 3D integration addressing the co-integration of superconducting devices with other technologies. This heterogeneous approach opens opportunities beyond HPC and AI, including quantum control and readout, neuromorphic computing, and high-resolution single photon detection supporting e.g. space and biomedical applications.”
