Simplified Superconducting Circuits Boost Performance with Single-Step Fabrication

The bottleneck in superconducting quantum computing is often not the qubits themselves, but the messy interfaces required to connect them. New research from Prakiran Baidya at Friedrich-Alexander University Erlangen-Nuremberg and colleagues at Quint Computing GmbH replaces traditional ‘bandages’ with solely airbridges for all electrical interconnects. This shift in architecture removes the material interfaces that previously caused unwanted energy dissipation.

By eliminating bandages, the team has addressed a primary source of decoherence. Previous designs relied on both airbridges and bandages, often struggling to surpass 100 microseconds in relaxation times. The new process uses a single-step gray-scale electron-beam lithography technique to create connections that range from 0.5 to 4μm in width and 5 to 40μm in length. These high-yield, mechanically stable connections allow the resulting transmon qubits to exhibit relaxation times exceeding 250μs.

This advancement changes the math for quantum error correction. The relaxation time (T1) dictates how long a qubit retains its excited state before decaying; longer T1 values translate directly to more computational steps possible before errors accumulate. Because the new technique maintains high coherence while streamlining device fabrication, it provides a path toward building more powerful quantum computers.

The implications for scaling are clear: complexity in manufacturing usually leads to instability. By using electron-beam lithography to precisely control superconducting material deposition and forming airbridges without the need for additional material layers, the researchers have removed the interfaces that previously made longer coherence times impossible. As we move toward larger-scale processors, the ability to maintain phase coherence (T2) and relaxation time (T1) through simplified fabrication will determine which architectures survive the transition from laboratory experiments to functional machines.

Consider how much of your current hardware roadmap is dictated by the limitations of your manufacturing process rather than the logic of your algorithms.

Source