Skip to content

Coherent States

A coherent state is a harmonic-oscillator state that remains a minimum-uncertainty Gaussian while its center follows the classical oscillator motion. Algebraically, it is an eigenstate of the annihilation operator:

a^∣α⟩=α∣α⟩,α∈C.\hat a\lvert\alpha\rangle=\alpha\lvert\alpha\rangle, \qquad \alpha\in\mathbb C.

Coherent states are the canonical oscillator states closest to classical motion. They are not energy eigenstates; they are special superpositions of number states.

For electromagnetic modes, photon statistics, optical coherence, realistic laser fields, and photodetection, continue to Coherent Light.

Expand a coherent state in number states:

∣α⟩=∑n=0∞cn∣n⟩.\lvert\alpha\rangle =\sum_{n=0}^{\infty}c_n\lvert n\rangle.

The eigenvalue equation a^∣α⟩=α∣α⟩\hat a\lvert\alpha\rangle=\alpha\lvert\alpha\rangle and the ladder action a^∣n⟩=n ∣n−1⟩\hat a\lvert n\rangle=\sqrt n\,\lvert n-1\rangle imply

cn=αnn!c0.c_n=\frac{\alpha^n}{\sqrt{n!}}c_0.

Normalization fixes

c0=e−∣α∣2/2,c_0=e^{-\lvert\alpha\rvert^2/2},

so

∣α⟩=e−∣α∣2/2∑n=0∞αnn!∣n⟩.\lvert\alpha\rangle =e^{-\lvert\alpha\rvert^2/2} \sum_{n=0}^{\infty} \frac{\alpha^n}{\sqrt{n!}}\lvert n\rangle.

This formula shows that a coherent state contains many oscillator number states unless α=0\alpha=0.

The probability to measure nn oscillator quanta is

P(n)=∣⟨n∣α⟩∣2=e−∣α∣2∣α∣2nn!.P(n) =\lvert\langle n\vert\alpha\rangle\rvert^2 =e^{-\lvert\alpha\rvert^2} \frac{\lvert\alpha\rvert^{2n}}{n!}.

This is a Poisson distribution with mean

⟨N^⟩=∣α∣2,\langle \hat N\rangle=\lvert\alpha\rvert^2,

and variance

(ΔN)2=∣α∣2.(\Delta N)^2=\lvert\alpha\rvert^2.

The relative number fluctuation is

ΔN⟨N^⟩=1∣α∣.\frac{\Delta N}{\langle \hat N\rangle} =\frac{1}{\lvert\alpha\rvert}.

For large ∣α∣\lvert\alpha\rvert, the relative fluctuation is small, one reason coherent states can look classical.

Coherent states can also be constructed by displacing the oscillator ground state. Define the displacement operator

D^(α)=exp⁡(αa^†−α∗a^).\hat D(\alpha) =\exp\left( \alpha\hat a^\dagger-\alpha^*\hat a \right).

Then

∣α⟩=D^(α)∣0⟩.\lvert\alpha\rangle=\hat D(\alpha)\lvert0\rangle.

This makes the geometry clear: a coherent state is the ground-state Gaussian shifted in phase space, without changing its minimum-uncertainty width.

Using

x^=ℓ2(a^+a^†),p^=ℏiℓ2(a^−a^†),\hat x =\frac{\ell}{\sqrt2} \left(\hat a+\hat a^\dagger\right), \qquad \hat p =\frac{\hbar}{i\ell\sqrt2} \left(\hat a-\hat a^\dagger\right),

with ℓ=ℏ/(mω)\ell=\sqrt{\hbar/(m\omega)}, the coherent-state means are

⟨x⟩α=2 ℓ Re⁡α,⟨p⟩α=2 ℏℓIm⁡α.\langle x\rangle_\alpha =\sqrt2\,\ell\,\operatorname{Re}\alpha, \qquad \langle p\rangle_\alpha =\frac{\sqrt2\,\hbar}{\ell}\operatorname{Im}\alpha.

Thus the complex number α\alpha labels a point in the oscillator phase plane.

The uncertainties are the same as the ground-state uncertainties:

Δx=ℓ2,Δp=ℏ2 ℓ.\Delta x=\frac{\ell}{\sqrt2}, \qquad \Delta p=\frac{\hbar}{\sqrt2\,\ell}.

Therefore

Δx Δp=ℏ2.\Delta x\,\Delta p=\frac{\hbar}{2}.

Coherent states are minimum-uncertainty states, but they are not the only minimum-uncertainty states.

Under the harmonic-oscillator Hamiltonian,

H^=ℏω(N^+12),\hat H=\hbar\omega\left(\hat N+\frac12\right),

a coherent state remains coherent. Apart from an overall phase,

α(t)=α(0)e−iωt.\alpha(t)=\alpha(0)e^{-i\omega t}.

The mean position and momentum therefore obey

⟨x⟩(t)=2 ℓ Re⁡(α(0)e−iωt),\langle x\rangle(t) =\sqrt2\,\ell\, \operatorname{Re}\left(\alpha(0)e^{-i\omega t}\right),

and

⟨p⟩(t)=2 ℏℓ Im⁡(α(0)e−iωt).\langle p\rangle(t) =\frac{\sqrt2\,\hbar}{\ell}\, \operatorname{Im}\left(\alpha(0)e^{-i\omega t}\right).

These are the classical oscillator equations for the packet center. The probability density does not spread because the harmonic potential refocuses the packet exactly.

Coherent-State Dynamics develops the exact state-vector phase, driven evolution, survival probability, and the criterion separating coherence-preserving motion from squeezing and nonlinear shearing.

Coherent states are normalized but not mutually orthogonal:

⟨β∣α⟩=exp⁡[−12∣β∣2−12∣α∣2+β∗α].\langle\beta\vert\alpha\rangle =\exp\left[ -\frac12\lvert\beta\rvert^2 -\frac12\lvert\alpha\rvert^2 +\beta^*\alpha \right].

The overlap magnitude is

∣⟨β∣α⟩∣2=e−∣α−β∣2.\lvert\langle\beta\vert\alpha\rangle\rvert^2 =e^{-\lvert\alpha-\beta\rvert^2}.

Widely separated coherent states are nearly orthogonal, but distinct coherent states are never exactly orthogonal at finite separation in phase space.

The family is nevertheless complete in the overcomplete sense:

∫Cd2απ∣α⟩⟨α∣=I.\int_{\mathbb C}\frac{d^2\alpha}{\pi} \lvert\alpha\rangle\langle\alpha\rvert=I.

This resolution of the identity is not an orthonormal-basis expansion. Its continuous coefficients are redundant, which is both useful and a source of care when coherent-state path integrals are constructed.

The displacement operator is unitary and translates the ladder operators:

D(α)†=D(−α),D(α)†aD(α)=a+α.D(\alpha)^\dagger=D(-\alpha), \qquad D(\alpha)^\dagger aD(\alpha)=a+\alpha.

Because the commutator in the exponent is a scalar, the Baker–Campbell–Hausdorff formula gives

D(α)=e−∣α∣2/2eαa†e−α∗a.D(\alpha) =e^{-\lvert\alpha\rvert^2/2} e^{\alpha a^\dagger}e^{-\alpha^*a}.

Two translations compose with a phase,

D(α)D(β)=exp⁡ ⁣[αβ∗−α∗β2]D(α+β),D(\alpha)D(\beta) =\exp\!\left[ \frac{\alpha\beta^*-\alpha^*\beta}{2} \right]D(\alpha+\beta),

so their noncommutativity records the oriented phase-space area enclosed by the translations.

Coordinate Wavefunction and Energy Moments

Section titled “Coordinate Wavefunction and Energy Moments”

With qα=⟨x⟩q_\alpha=\langle x\rangle and pα=⟨p⟩p_\alpha=\langle p\rangle, the normalized coordinate wavefunction may be written

ψα(x)=(1πℓ2)1/4exp⁡ ⁣[−(x−qα)22ℓ2+ipαxℏ−iqαpα2ℏ].\psi_\alpha(x) =\left(\frac{1}{\pi\ell^2}\right)^{1/4} \exp\!\left[-\frac{(x-q_\alpha)^2}{2\ell^2} +\frac{ip_\alpha x}{\hbar} -\frac{iq_\alpha p_\alpha}{2\hbar}\right].

The last term is an xx-independent phase fixed by the displacement convention. It cancels from one-state position probabilities but matters in interference and overlap calculations.

For H=ℏω(N+1/2)H=\hbar\omega(N+1/2),

⟨H⟩=ℏω(∣α∣2+12),(ΔH)2=(ℏω)2∣α∣2.\langle H\rangle =\hbar\omega\left(\lvert\alpha\rvert^2+\frac12\right), \qquad (\Delta H)^2=(\hbar\omega)^2\lvert\alpha\rvert^2.

Thus a nonvacuum coherent state is never an energy eigenstate.

For a normally ordered function,

⟨α∣:f(a†,a):∣α⟩=f(α∗,α).\langle\alpha\rvert{:}f(a^\dagger,a){:}\lvert\alpha\rangle =f(\alpha^*,\alpha).

Normal ordering is essential: ⟨aa†⟩=∣α∣2+1\langle aa^\dagger\rangle =\lvert\alpha\rvert^2+1, not ∣α∣2\lvert\alpha\rvert^2. The extra term is the commutator contribution.

A coherent state is often the right first model for a classical-looking oscillator with quantum width. It has:

  • a localized Gaussian wave packet;
  • minimum position-momentum uncertainty;
  • a phase-space center following classical motion;
  • a Poisson distribution of oscillator quanta;
  • fixed shape under harmonic time evolution.

In quantum optics, coherent states model idealized laser-like field modes. In mechanical and molecular settings, they model displaced oscillator motion. Field-theory coherent states require additional many-mode structure and belong in the relevant field and quantum-optics treatments.

  • Treating a coherent state as an energy eigenstate.
  • Treating α\alpha as a position rather than a dimensionless phase-space label.
  • Forgetting the Poisson number distribution.
  • Assuming coherent states with different α\alpha are exactly orthogonal.
  • Confusing coherent states with squeezed states; coherent states have ground-state variances in both quadratures.
  • Saying coherent states are fully classical. They are quantum states with particularly classical-looking first moments.
  • D. J. Griffiths and D. F. Schroeter, Introduction to Quantum Mechanics, 3rd ed., Cambridge University Press, 2018.
  • R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994.
  • R. J. Glauber, ‘Coherent and Incoherent States of the Radiation Field,’ Physical Review 131, 2766-2788 (1963).
  • E. C. G. Sudarshan, ‘Equivalence of Semiclassical and Quantum Mechanical Descriptions of Statistical Light Beams,’ Physical Review Letters 10, 277-279 (1963).
  • J. R. Klauder and B.-S. Skagerstam, Coherent States: Applications in Physics and Mathematical Physics, World Scientific, 1985.
  • M. O. Scully and M. S. Zubairy, Quantum Optics, Cambridge University Press, 1997.
  1. Derive the Poisson number distribution from the number-basis expansion.
Solution

The expansion coefficient is

⟨n∣α⟩=e−∣α∣2/2αnn!.\langle n\vert\alpha\rangle =e^{-\lvert\alpha\rvert^2/2} \frac{\alpha^n}{\sqrt{n!}}.

Therefore

P(n)=∣⟨n∣α⟩∣2=e−∣α∣2∣α∣2nn!.P(n) =\lvert\langle n\vert\alpha\rangle\rvert^2 =e^{-\lvert\alpha\rvert^2} \frac{\lvert\alpha\rvert^{2n}}{n!}.

This is a Poisson distribution with mean ∣α∣2\lvert\alpha\rvert^2.

  1. Show that ⟨x⟩α=2 ℓ Re⁡α\langle x\rangle_\alpha=\sqrt2\,\ell\,\operatorname{Re}\alpha.
Solution

Use

x^=ℓ2(a^+a^†).\hat x =\frac{\ell}{\sqrt2} \left(\hat a+\hat a^\dagger\right).

Since a^∣α⟩=α∣α⟩\hat a\lvert\alpha\rangle=\alpha\lvert\alpha\rangle, one has

⟨a^⟩=α,⟨a^†⟩=α∗.\langle\hat a\rangle=\alpha, \qquad \langle\hat a^\dagger\rangle=\alpha^*.

Therefore

⟨x⟩α=ℓ2(α+α∗)=2 ℓ Re⁡α.\langle x\rangle_\alpha =\frac{\ell}{\sqrt2}(\alpha+\alpha^*) =\sqrt2\,\ell\,\operatorname{Re}\alpha.
  1. Explain why a coherent state with large ∣α∣\lvert\alpha\rvert can look classical even though it is not an energy eigenstate.
Solution

The state remains a localized Gaussian whose center follows the classical oscillator trajectory. Its mean number is ∣α∣2\lvert\alpha\rvert^2, while its number uncertainty is ΔN=∣α∣\Delta N=\lvert\alpha\rvert. The relative fluctuation is therefore

ΔN⟨N^⟩=1∣α∣.\frac{\Delta N}{\langle\hat N\rangle} =\frac{1}{\lvert\alpha\rvert}.

For large ∣α∣\lvert\alpha\rvert, relative fluctuations are small, and the first moments follow classical motion. The state is still a quantum superposition of number states.

  1. Explain why replacing a→αa\to\alpha and a†→α∗a^\dagger\to\alpha^* in aa†aa^\dagger gives the wrong expectation value, and compute the correct one.
Solution

The substitution rule applies to normally ordered products. Since

aa†=a†a+I,aa^\dagger=a^\dagger a+I,

one obtains

⟨aa†⟩=⟨a†a⟩+1=∣α∣2+1.\langle aa^\dagger\rangle =\langle a^\dagger a\rangle+1 =\lvert\alpha\rvert^2+1.

Direct substitution before normal ordering misses the commutator term.