Qubit ZooThe merged-element transmon (mergemon or MET) is a superconducting qubit in which the Josephson junction’s intrinsic self-capacitance replaces the large external shunt capacitor of a conventional transmon. This merges the two defining circuit elements into one, shrinking the qubit footprint by ~100× and confining >90% of the electromagnetic energy to the tunnel barrier region — away from the lossy interfaces that limit coherence in standard transmons.
Figure
Description
Bottom-up origin
The mergemon concept traces to the proposal by Shim et al. (2014) to build superconducting devices inside a group-IV semiconductor (Si or Ge) using precision atomic-scale doping. By creating heavily hole-doped regions within a single crystal, the doped regions become superconducting below ~1 K, and the undoped semiconductor between them forms a Josephson tunnel barrier. This “bottom-up” approach makes the entire qubit — junction and capacitor — from a single semiconductor crystal, naturally merging the elements.
Modern implementations
The Pappas group (NIST/CU Boulder) demonstrated the first mergemon prototypes using:
- Al/AlOₓ/Al overlap junctions (2020–2021): Micrometer-scale junctions with long oxidation, achieving = 10–90 μs. Junction loss is not the dominant limiter.
- FinMET: Silicon fin-based merged transmons compatible with CMOS foundry processing. Uses mature semiconductor fin-etch processes.
- Crystalline barriers: All-crystalline mergemons using WSe₂ or epitaxial semiconductor barriers, aiming to eliminate amorphous dielectric loss entirely.
Design advantage
In a standard transmon, the shunt capacitor (typically ~100 fF, ~300 μm across) dominates the device area. Surface dielectric loss at the capacitor’s metal-substrate-vacuum interfaces is the primary coherence limiter. The mergemon eliminates this capacitor: the junction itself provides the required capacitance (20–100 fF for transmon operation), and the electric field is concentrated inside the junction barrier rather than at exposed surfaces.
Hamiltonian
Same as the transmon:
but with set by the junction self-capacitance rather than an external shunt: . The merged element enforces , concentrating field energy in the barrier.
Performance Metrics
| Metric | Value | Notes | Fidelity reference |
|---|---|---|---|
| 10–90 μs | Al/AlOₓ overlap mergemon | madon-2021-mergemon | |
| (annealed) | >100 μs | With ABAA annealing | madon-2021-mergemon |
| Footprint reduction | ~100× | vs. standard transmon | zhao-2020-mergemon |
| Energy confinement | >90% in barrier | Reduced surface participation | zhao-2020-mergemon |
| Junction capacitance | 20–100 fF | Self-capacitance of large-area junction | — |
| CMOS compatibility | Yes | FinMET variant | — |
Scaling Considerations
- Density: ~100× smaller footprint enables dramatically denser qubit arrays.
- Loss mechanism shift: Moves dominant loss from surface dielectrics to junction barrier quality — a more controllable parameter.
- Fabrication: Compatible with semiconductor foundry processes (fin etch, trilayer deposition). Path to industrial-scale manufacturing.
- Materials frontier: Crystalline barriers (epitaxial semiconductors, 2D materials) may push coherence beyond amorphous AlOₓ limits.
- Super-semi integration: Natural platform for merging spin qubit and transmon functionality on a single chip.