How Quantum Phase Estimation Extracts Eigenvalues

Many quantum problems can be expressed in terms of a unitary operator whose eigenvalues have the form exp(2πiφ), where φ is a phase between zero and one. Quantum phase estimation, or QPE, is an algorithm for estimating φ when the processor can implement controlled applications of that operator. For Hamiltonian simulation, the measured phase can be related to an energy eigenvalue through the chosen evolution time.

In the standard circuit, an auxiliary register is placed into a superposition and used to control successive powers of the unitary. These powers cause each auxiliary qubit to accumulate phase information at a different binary scale. An inverse quantum Fourier transform then converts the relative phases into measurement probabilities concentrated around a binary approximation of φ.

Precision has a substantial resource cost. Estimating one more binary digit generally requires reaching a controlled power twice as large as before, or equivalently sustaining coherent evolution for about twice as long. Gate errors, decoherence and imperfect implementations of the unitary can therefore become more damaging as the desired precision increases.

QPE also requires attention to the input state. If the system register is an exact eigenstate, the algorithm returns the corresponding eigenphase with the accuracy allowed by the circuit. If it is a superposition of eigenstates, measurement selects among their phases with probabilities determined by the input state's squared overlaps with those eigenstates.

Several variants alter the implementation without removing the underlying constraints. Iterative phase estimation can reuse a single auxiliary qubit with measurements and classical feed-forward, reducing qubit requirements, while other formulations use statistical inference across many shorter circuits. QPE is a building block in algorithms including order finding and quantum energy estimation, but practical use still depends on preparing relevant states and implementing controlled evolution accurately.