T1: Energy Relaxation Time
T1 is the timescale over which an excited qubit loses its energy and decays to the ground state, capping how long computation can run.
Amplitude decay
T1, the energy relaxation or longitudinal relaxation time, measures how long a qubit stays excited before decaying from |1> to |0>. Prepare the qubit in |1>, wait a variable time, and measure; the excited-state probability falls off as exp(-t/T1). It is the qubit analog of an atom's spontaneous-emission lifetime.
What causes it
- Coupling to lossy dielectrics and two-level-system defects in oxides and interfaces
- Radiative loss into control and readout lines (the Purcell effect)
- Quasiparticles in superconductors
- For atoms and spins, spontaneous emission and spin-lattice relaxation
Why it matters
Every gate and every idle moment consumes a fraction of T1. The number of operations a qubit can perform before decaying scales with the ratio of T1 to the gate time. Superconducting transmons now reach T1 of tens to a few hundred microseconds; trapped ions and nuclear spins reach seconds or longer. Longer T1 means deeper circuits and lower error per gate.
Improving it
Progress has come from cleaner materials, better surface treatments, improved geometries that reduce participation of lossy interfaces, Purcell filters that block radiative decay through readout lines, and shielding against stray infrared and magnetic fields. T1 also fluctuates in time as individual defects switch, so it is reported as a distribution, not a single number.
T1 sets a hard ceiling: no amount of clever control recovers energy already lost to the environment, which is why relaxation is the first quantity a hardware team fights to extend.