Quantum Zeno Effect

The quantum Zeno effect is the inhibition of transitions by sufficiently frequent suitable interventions. It tests the short-time structure of quantum dynamics and the disturbance produced by a measurement. The intervention is a physical coupling to a detector or environment. Quantum Zeno Dynamics develops repeated projections, continuous monitoring, and evolution inside constrained subspaces.

Survival under repeated tests

For a state $\lvert\psi\rangle$ with finite energy variance and a time-independent Hamiltonian, the short-time survival probability is

$$p(\tau)=|\langle\psi|e^{-iH\tau/\hbar}|\psi\rangle|^2 =1-\frac{(\Delta H)^2\tau^2}{\hbar^2}+o(\tau^2).$$

For ideal tests of the projector $\lvert\psi\rangle\langle\psi\rvert$ after each interval $t/N$, the probability that all $N$ tests return survival is

$$P_N(t)=[p(t/N)]^N \simeq \exp\!\left[-\frac{(\Delta H)^2t^2}{N\hbar^2}\right].$$

At fixed $t$, this tends to one as $N$ grows in the ideal mathematical limit. The quadratic initial law is essential: repeated application of a strictly exponential law would give $[e^{-\Gamma t/N}]^N=e^{-\Gamma t}$, with no inhibition.

A driven qubit as an operational example

Let $H=\hbar\Omega X/2$, prepare $\lvert0\rangle$, and measure $Z$ repeatedly. A single interval gives a survival probability $\cos^2(\Omega t/2N)$. Thus

$$P_N(t)=\cos^{2N}\!\left(\frac{\Omega t}{2N}\right).$$

For an unmeasured $\pi$ pulse, $t=\pi/\Omega$, survival is zero. Ten equally spaced survival tests instead give $\cos^{20}(\pi/20)\simeq0.781$. The intervention changes the statistics even if the resulting detector record is never read.

Final population includes return histories

Keep the same ideal interventions, but ignore their intermediate outcomes. After each $Z$ test the state is diagonal. If $z_k$ is its population difference after the $k$th test, one more rotation followed by an unread test gives

$$z_{k+1}=z_k\cos(\Omega t/N),\qquad z_0=1.$$

Consequently the probability of obtaining $0$ at the final test is

$$p_{0,\rm final}(t)=\frac{1+[\cos(\Omega t/N)]^N}{2}.$$

For the ten-test $\pi$ pulse this is about $0.803$, larger than the all-survival probability $0.781$. The extra probability comes from records in which the qubit flips and later returns. Both numbers describe the same intervention: $P_N$ selects the all-zero record, while $p_{0,\rm final}$ sums every record ending in zero. This is the selective versus nonselective distinction described by Facchi and Pascazio (2008, Section 13.1).

Ignoring the records does not undo the measurement interactions. The unmeasured $\pi$ pulse still has final population zero. An experiment must specify whether it reports an all-survival trajectory or a final population, especially with imperfect detection.

Physical measurement and interpretation

A measurement pulse entangles the system with a meter or environment; a strong coupling can separate dynamical sectors; a dissipative channel can suppress transitions. These mechanisms can implement Zeno constraints without inserting a human observer. In a multidimensional protected subspace, nontrivial evolution can continue: Zeno dynamics need not mean complete immobility.

Finite pulse duration, detector bandwidth, transitions during readout, and the bath spectrum bound the accessible regime. Some interventions accelerate decay, the anti-Zeno effect. Extrapolating an ideal $N\to\infty$ projection formula to arbitrary finite apparatuses is unjustified.

Agreement with Zeno predictions tests a specified dynamical measurement model. It does not select a unique interpretation of state update. An account using unitary system–apparatus dynamics can reproduce the same observable populations as a projection description over the applicable regime.

Exercise

For the driven qubit, find the leading all-survival failure probability at fixed $t$ and large $N$. How must $N$ scale to keep it below a small target $\delta$? As a finite-$N$ check, compare the all-survival probability and final $0$ population for a $\pi$ pulse interrupted by two equally spaced tests.

Solution

Using $\log\cos x=-x^2/2+O(x^4)$ gives $P_N\simeq e^{-\Omega^2t^2/(4N)}$. Therefore $1-P_N\simeq\Omega^2t^2/(4N)$, so $N\gtrsim\Omega^2t^2/(4\delta)$ within this ideal regime. This mathematical requirement must also be compatible with the finite time needed for each physical measurement.

For $N=2$ and $\Omega t=\pi$, each interval has flip probability $1/2$. The record $00$ has probability $1/4$, so $P_2=1/4$. The record $10$ also has probability $1/4$: the qubit flips and returns. Summing these two records gives $p_{0,\rm final}=1/2$. A final zero alone does not establish uninterrupted survival.

References

• Facchi, P., and Pascazio, S., “Quantum Zeno dynamics: mathematical and physical aspects,” Journal of Physics A: Mathematical and Theoretical 41, 493001 (2008), doi:10.1088/1751-8113/41/49/493001 — projection limits and dynamical implementations.