Charge Noise and Anharmonicity
Two competing quantities govern superconducting qubit design: anharmonicity, which lets you address one transition, and charge dispersion, which invites noise.
The trade-off
A superconducting qubit must be nonlinear enough to isolate its 0-1 transition (anharmonicity) yet insensitive enough to background electric fields (low charge dispersion). These pull in opposite directions, and each qubit family picks a different compromise along the axis set by the ratio E_J/E_C.
Charge dispersion
A bare charge qubit's frequency depends on stray offset charge on nearby islands, which fluctuates as trapped charges hop in the substrate. This charge noise directly dephases the qubit. Raising E_J/E_C suppresses the sensitivity exponentially: the transmon exploits exactly this, becoming essentially flat against offset charge.
Anharmonicity
Anharmonicity alpha is the frequency difference between the 0-1 and 1-2 transitions. Large negative alpha lets gates be short because a pulse can be spectrally narrow enough to skip the 1-2 transition. Small alpha forces longer or more carefully shaped pulses (DRAG) to avoid leakage into the 2 state. The transmon's alpha of a few hundred MHz is a workable middle ground.
Power-law versus exponential
- Charge dispersion decays exponentially with E_J/E_C
- Anharmonicity decays only as a power law with E_J/E_C
- So a large ratio kills charge noise while keeping usable anharmonicity
Fluxonium takes a different route, using a large superinductance to reach very high anharmonicity and long coherence at a lower frequency. NV centers, spins, and ions face different noise channels entirely, but the same principle recurs across modalities: a qubit must be nonlinear enough to control and quiet enough to trust, and the designer's job is to find where those demands meet.