Dynamical Decoupling
Dynamical decoupling applies timed control pulses to average away slow environmental noise, extending a qubit's usable coherence.
Refocusing noise
Much of the noise that dephases a qubit is slow: the qubit frequency drifts over microseconds to milliseconds. Dynamical decoupling exploits this by periodically flipping the qubit so that phase accumulated in one interval is undone in the next. Noise slower than the pulse spacing averages toward zero, while the encoded quantum information is preserved.
From spin echo to sequences
The idea began as the Hahn spin echo in nuclear magnetic resonance: a single pi pulse halfway through a free evolution cancels static frequency offsets. Adding more pulses extends the effect to time-varying noise. The Carr-Purcell-Meiboom-Gill (CPMG) sequence uses a train of evenly spaced pi pulses; more elaborate schemes handle pulse imperfections and noise along multiple axes.
Common sequences
- Hahn echo: one refocusing pulse, cancels static dephasing
- CPMG: many equally spaced pulses, robust to slow noise
- XY-4, XY-8: alternate rotation axes to tolerate pulse errors
- Uhrig sequences: unevenly spaced pulses optimized for a noise spectrum
Limits and uses
Decoupling cannot fix energy relaxation (T1) or noise faster than the pulse rate, and each pulse carries its own small error, so there is an optimal density. Beyond extending coherence, the response of a qubit to different pulse spacings is used as noise spectroscopy, mapping the environment's noise spectrum. It is also a building block inside gate and memory operations on real hardware.
Dynamical decoupling is a passive, low-overhead complement to full quantum error correction, buying coherence time without extra qubits.