Phase Kickback & Phase Estimation
How a control qubit learns the eigenvalue of an operation.
A controlled operation acts on its target, yet when the target is an eigenstate it is the control that changes. That reversal, phase kickback, followed by one layer of interference, lets a register of qubits read off an eigenvalue digit by digit. It is the engine of Shor's algorithm and of energy estimation on future fault-tolerant machines.
What you’ll learn
- Derive phase kickback: a controlled- on with .
- Explain Deutsch's algorithm as kickback plus interference.
- Derive the phase-estimation register state and show that the inverse QFT returns exactly when it has a -bit expansion.
- Compute the outcome distribution for general and the number of counting qubits needed for a given precision and success probability.
- Relate the cost of phase estimation to (Heisenberg) scaling.
The control learns the eigenvalue
Let be an eigenstate of a unitary , with . The eigenvalue is a global phase for the target alone, so measuring the target could never reveal it. Put a control qubit in superposition and watch where that phase goes.
Controlled-: apply to the target only on the branch of the control.
Linearity.
The eigenvalue factors out of the target and becomes a relative phase of the control. The target is untouched and stays unentangled.
A final Hadamard turns the phase into a bias: this is the Mach–Zehnder formula with as the arm phase.
Deutsch's algorithm is kickback
A bit oracle acts on the target as . Since , the state is an eigenvector with eigenvalue :
With the control in , the oracle writes (up to normalization) onto the control, and maps that to . One query determines , which classically needs two. Every hidden-string algorithm (Deutsch–Jozsa, Bernstein–Vazirani, Simon) starts with this move.
In Deutsch's algorithm the target register is both before and after the oracle. How, then, does the circuit learn anything about ?
- It doesn't; the answer comes from measuring the target
- The oracle's eigenvalue is kicked back into the control's relative phase, and the final converts that phase into a bit
- The oracle secretly entangles control and target
- The control qubit is copied into the target and read later
Show the answer
B. The oracle's eigenvalue is kicked back into the control's relative phase, and the final converts that phase into a bit
Because is an eigenvector of every , the target factors out unchanged, and all information about ends up in the control's relative phase , which makes measurable.
Phase estimation
One kickback gives one blurry bit. To read to binary digits, use counting qubits, each in , and let counting qubit control . It picks up : doubling the power shifts the binary expansion of one place to the left.
Expand the product with . The amplitudes are a discrete plane wave with frequency .
The quantum Fourier transform produces exactly such plane waves.
If has a -bit expansion, the inverse QFT returns its digits with certainty: all paths to wrong outcomes cancel.
The cost is applications of for precision . Precision for uses of is the Heisenberg scaling; repeating a single kickback times and averaging gives only .
and counting qubits, so outcomes are integers with estimate . Which is the most likely outcome?
Show the answer
2
, so the nearest grid point is (estimate ), with probability about ; the next is with about . Check it in the simulator.
How many counting qubits?
Apply and sum the geometric series.
For the nearest : use in the denominator, then the fact that on .
The same bound applied to both neighbors, minimized at .
Counting qubits sufficient to get to bits with probability at least (Nielsen & Chuang §5.2.1). The tail of falls off like , so a few extra qubits buy a lot of confidence.
Using the formula above, how many counting qubits suffice to estimate to bits with success probability at least ?
Show the answer
C.
gives , rounded up to , so .
Uses, and going deeper
| Application | Unitary | Eigenphase reveals |
|---|---|---|
| Shor's order finding | for random ; continued fractions give the order | |
| Energy estimation | for an eigenvalue of the Hamiltonian | |
| Quantum counting | Grover iterate | the rotation angle, hence the number of marked items |
| Hidden-string problems | phase oracle | a global property of via one round of kickback |
Kitaev's original formulation (1995) measured the bits one at a time with a single control qubit, feeding earlier results forward as classical phase corrections. Griffiths and Niu (1996) showed that the inverse QFT followed by measurement can always be done this way, with one-qubit gates and classical control. This "iterative" phase estimation needs only one ancilla, which matters on small or early fault-tolerant machines.
Key takeaways
- : an eigenvalue becomes a relative phase on the control, and the target is untouched.
- Deutsch, Deutsch–Jozsa and Bernstein–Vazirani are kickback onto followed by Hadamard interference.
- Phase estimation writes ; the inverse QFT returns exactly if it has bits, otherwise the nearest -bit value with probability .
- bits with probability need counting qubits, and precision scales as : Heisenberg scaling.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
. What does do to ?
- , a global phase that changes nothing
- , an entangled state
- : the control acquires a relative phase
- , with the control unchanged
Show the answer
C. : the control acquires a relative phase
Only the branch receives , so the phase is relative between the control's branches. The target factors out unchanged; the state is a product state.
A controlled- acts on , then is applied to the control. What is the probability that the control reads ?
Show the answer
B.
kicks back the phase , and then gives .
Textbook phase estimation with counting qubits uses how many applications of in total?
Show the answer
D.
Counting qubit controls , and . Precision for about uses is Heisenberg scaling.
exactly and . What does phase estimation output?
- with certainty
- with probability at least
- a uniformly random 3-bit string
- with certainty
Show the answer
A. with certainty
is a 3-bit fraction, so the register state is exactly and the inverse QFT returns deterministically.
Go deeper
- Eigenvalue Kickback & Controlled Unitaries · The mechanism inside Deutsch, Shor and phase estimation.
- Quantum Metrology & the Heisenberg Limit · Entanglement-enhanced sensing and its limits.
Further reading
- M. A. Nielsen & I. L. Chuang, Quantum Computation and Quantum Information, §5.2 (Cambridge, 10th anniv. ed., 2010).
- R. Cleve, A. Ekert, C. Macchiavello & M. Mosca, "Quantum algorithms revisited", Proc. R. Soc. Lond. A 454, 339 (1998).
- A. Yu. Kitaev, "Quantum measurements and the Abelian stabilizer problem", arXiv:quant-ph/9511026 (1995).
- R. B. Griffiths & C.-S. Niu, "Semiclassical Fourier transform for quantum computation", Phys. Rev. Lett. 76, 3228 (1996).