Skip to content

Harmonic Oscillator Spectrum

For the one-dimensional Hamiltonian

H=P22m+12mω2X2,m>0,ω>0,H = \frac{P^2}{2m} +\frac12 m\omega^2X^2, \qquad m>0, \quad \omega>0,

the exact energy spectrum is

En=ℏω(n+12),n=0,1,2,…E_n = \hbar\omega \left( n+\frac12 \right), \qquad n=0,1,2,\ldots

Adjacent levels have constant spacing

En+1−En=ℏω,E_{n+1}-E_n = \hbar\omega,

and the ground-state energy is

E0=12ℏω.E_0 = \frac12\hbar\omega.
QuantityFormula
Oscillator lengthℓ=ℏ/(mω)\ell=\sqrt{\hbar/(m\omega)}
Momentum scalepℓ=ℏ/ℓ=mℏωp_\ell=\hbar/\ell=\sqrt{m\hbar\omega}
Energy scaleℏω\hbar\omega
SpectrumEn=ℏω(n+1/2)E_n=\hbar\omega(n+1/2)
Level spacingΔE=ℏω\Delta E=\hbar\omega
Ground-state widthΔx0=ℓ/2\Delta x_0=\ell/\sqrt2
Ground-state momentum widthΔp0=ℏ/(2 ℓ)\Delta p_0=\hbar/(\sqrt2\,\ell)
Parity(−1)n(-1)^n
Number of real nodesnn
Classical turning pointsxt(n)=±ℓ2n+1x_{\mathrm t}^{(n)}=\pm\ell\sqrt{2n+1}

The length convention matters. Some authors call Δx0\Delta x_0 the oscillator length; this card uses

ℓ=ℏmω.\ell = \sqrt{\frac{\hbar}{m\omega}}.

The quadratic oscillator has a purely discrete, nondegenerate one-dimensional spectrum. Its levels are equally spaced and begin at a nonzero zero-point energy.

The formula is exact only for an exactly quadratic Hamiltonian on the real line. Near a smooth stable minimum of a more general potential, it gives the leading approximation, with anharmonic terms producing corrections.

The angular frequency is ω\omega. If ordinary frequency is ν\nu, then

ω=2πν,\omega=2\pi\nu,

and the spacing can equivalently be written

ℏω=hν.\hbar\omega = h\nu.

The parameters mm, ω\omega, and ℏ\hbar define

ℓ=ℏmω,\ell = \sqrt{\frac{\hbar}{m\omega}}, pℓ=ℏℓ=mℏω,p_\ell = \frac{\hbar}{\ell} = \sqrt{m\hbar\omega}, Eℓ=ℏω.E_\ell = \hbar\omega.

With

ξ=xℓ,π=ℓpℏ,\xi = \frac{x}{\ell}, \qquad \pi = \frac{\ell p}{\hbar},

the canonical commutator is

[ξ,π]=i[\xi,\pi]=i

and the Hamiltonian becomes

Hℏω=12(π2+ξ2).\frac{H}{\hbar\omega} = \frac12 \left( \pi^2+\xi^2 \right).

All one-dimensional oscillators therefore share the same dimensionless eigenvalue problem. Changing mm and ω\omega rescales position, momentum, energy, and time.

Define

a=12(Xℓ+iℓPℏ),a = \frac{1}{\sqrt2} \left( \frac{X}{\ell} +\frac{i\ell P}{\hbar} \right), a†=12(Xℓ−iℓPℏ).a^\dagger = \frac{1}{\sqrt2} \left( \frac{X}{\ell} -\frac{i\ell P}{\hbar} \right).

Then

[a,a†]=I,[a,a^\dagger]=I, N=a†a,N = a^\dagger a,

and

H=ℏω(N+12I).H = \hbar\omega \left( N+\frac12I \right).

Number states satisfy

N∣n⟩=n∣n⟩,N\lvert n\rangle = n\lvert n\rangle, a∣n⟩=n∣n−1⟩,a\lvert n\rangle = \sqrt n \lvert n-1\rangle, a†∣n⟩=n+1∣n+1⟩.a^\dagger\lvert n\rangle = \sqrt{n+1} \lvert n+1\rangle.

The normalized construction from the vacuum is

∣n⟩=(a†)nn!∣0⟩,a∣0⟩=0.\lvert n\rangle = \frac{ (a^\dagger)^n }{ \sqrt{n!} } \lvert0\rangle, \qquad a\lvert0\rangle=0.

The operators aa and a†a^\dagger are dimensionless under this convention.

Using the physicists’ Hermite polynomials HnH_n,

ψn(x)=12nn!(1πℓ2)1/4Hn(xℓ)exp⁡(−x22ℓ2).\psi_n(x) = \frac{1}{\sqrt{2^n n!}} \left( \frac{1}{\pi\ell^2} \right)^{1/4} H_n\left( \frac{x}{\ell} \right) \exp\left( -\frac{x^2}{2\ell^2} \right).

They obey

∫−∞∞ψm∗(x)ψn(x) dx=δmn,\int_{-\infty}^{\infty} \psi_m^*(x)\psi_n(x)\,dx = \delta_{mn}, ψn(−x)=(−1)nψn(x).\psi_n(-x) = (-1)^n\psi_n(x).

The state ψn\psi_n has exactly nn simple nodes on the real line. Its tails extend beyond the classical turning points; those turning points are not boundary conditions.

The full stationary solution is

Ψn(x,t)=ψn(x)e−iEnt/ℏ.\Psi_n(x,t) = \psi_n(x) e^{-iE_nt/\hbar}.

Its probability density is time independent even though the state vector carries a phase.

The normalized ground-state wavefunction is

ψ0(x)=1π1/4ℓexp⁡(−x22ℓ2).\psi_0(x) = \frac{1}{ \pi^{1/4}\sqrt{\ell} } \exp\left( -\frac{x^2}{2\ell^2} \right).

Its uncertainties are

Δx0=ℓ2,\Delta x_0 = \frac{\ell}{\sqrt2}, Δp0=ℏ2 ℓ,\Delta p_0 = \frac{\hbar}{\sqrt2\,\ell},

so

Δx0Δp0=ℏ2.\Delta x_0 \Delta p_0 = \frac{\hbar}{2}.

The ground state saturates the position-momentum uncertainty bound. Excited number states do not.

For the number state ∣n⟩\lvert n\rangle,

⟨X⟩n=0,⟨P⟩n=0,\langle X\rangle_n = 0, \qquad \langle P\rangle_n = 0, ⟨X2⟩n=(n+12)ℓ2,\langle X^2\rangle_n = \left( n+\frac12 \right) \ell^2, ⟨P2⟩n=(n+12)mℏω.\langle P^2\rangle_n = \left( n+\frac12 \right) m\hbar\omega.

Therefore

ΔXnΔPn=ℏ(n+12).\Delta X_n \Delta P_n = \hbar \left( n+\frac12 \right).

The virial theorem gives

⟨P22m⟩n=⟨12mω2X2⟩n=En2.\left\langle \frac{P^2}{2m} \right\rangle_n = \left\langle \frac12m\omega^2X^2 \right\rangle_n = \frac{E_n}{2}.

These equations are strong checks on analytic and numerical eigenstates.

The inverse ladder relations are

X=ℓ2(a+a†),X = \frac{\ell}{\sqrt2} \left( a+a^\dagger \right), P=iℏ2 ℓ(a†−a).P = \frac{i\hbar}{\sqrt2\,\ell} \left( a^\dagger-a \right).

Hence

⟨m∣X∣n⟩=ℓ2[n+1 δm,n+1+n δm,n−1],\langle m\rvert X\lvert n\rangle = \frac{\ell}{\sqrt2} \left[ \sqrt{n+1}\, \delta_{m,n+1} + \sqrt n\, \delta_{m,n-1} \right], ⟨m∣P∣n⟩=iℏ2 ℓ[n+1 δm,n+1−n δm,n−1].\langle m\rvert P\lvert n\rangle = \frac{i\hbar}{\sqrt2\,\ell} \left[ \sqrt{n+1}\, \delta_{m,n+1} - \sqrt n\, \delta_{m,n-1} \right].

A perturbation proportional to XX therefore has the ideal-oscillator selection rule

Δn=±1.\Delta n=\pm1.

Other couplings have different rules. For example, X2X^2 connects Δn=0,±2\Delta n=0,\pm2.

Setting the potential energy equal to EnE_n gives

12mω2(xt(n))2=ℏω(n+12),\frac12 m\omega^2 \left( x_{\mathrm t}^{(n)} \right)^2 = \hbar\omega \left( n+\frac12 \right),

so

xt(n)=±ℓ2n+1.x_{\mathrm t}^{(n)} = \pm \ell \sqrt{2n+1}.

The interval between these points is classically allowed. Quantum eigenfunctions remain nonzero outside it and decay in the classically forbidden region.

For

H=P22m+12mω2(X−x0)2+C,H = \frac{P^2}{2m} +\frac12 m\omega^2 (X-x_0)^2 +C,

the spectrum is

En=C+ℏω(n+12).E_n = C +\hbar\omega \left( n+\frac12 \right).

The displacement x0x_0 shifts the eigenfunctions but does not change level spacings. The constant CC shifts every energy equally.

If a Hamiltonian is instead written with a linear term,

H=P22m+12mω2X2−FX,H = \frac{P^2}{2m} +\frac12m\omega^2X^2 -FX,

completing the square gives

x0=Fmω2,x_0 = \frac{F}{m\omega^2},

and

C=−F22mω2.C = -\frac{F^2}{2m\omega^2}.

For independent modes,

H=∑j=1d[Pj22mj+12mjωj2Xj2],H = \sum_{j=1}^{d} \left[ \frac{P_j^2}{2m_j} +\frac12 m_j\omega_j^2X_j^2 \right],

the spectrum is

En=∑j=1dℏωj(nj+12),E_{\mathbf n} = \sum_{j=1}^{d} \hbar\omega_j \left( n_j+\frac12 \right),

with

nj=0,1,2,…n_j=0,1,2,\ldots

For the isotropic case ωj=ω\omega_j=\omega, define

N=∑j=1dnj.N = \sum_{j=1}^{d}n_j.

Then

EN=ℏω(N+d2),E_N = \hbar\omega \left( N+\frac d2 \right),

and, ignoring additional internal degrees of freedom, the degeneracy is

gN=(N+d−1d−1).g_N = \binom{N+d-1}{d-1}.

In three dimensions,

gN=(N+1)(N+2)2.g_N = \frac{(N+1)(N+2)}{2}.

Anisotropic frequencies generally lift this shell degeneracy, although rationally related frequencies can produce accidental degeneracies.

Let a smooth potential have a stable nondegenerate minimum at x0x_0:

V′(x0)=0,V′′(x0)>0.V'(x_0)=0, \qquad V''(x_0)>0.

Then

V(x)=V(x0)+12V′′(x0)(x−x0)2+⋯V(x) = V(x_0) +\frac12 V''(x_0) (x-x_0)^2 +\cdots

with

ω=V′′(x0)m.\omega = \sqrt{ \frac{V''(x_0)}{m} }.

The leading energy estimate is

En≈V(x0)+ℏω(n+12).E_n \approx V(x_0) +\hbar\omega \left( n+\frac12 \right).

This approximation is controlled only when cubic and higher terms are small over the region occupied by the state. Higher states typically probe farther from the minimum and become less harmonic.

Within an isolated oscillator, adding a constant to HH changes all energies by the same amount and leaves eigenstates and transition frequencies unchanged. One may therefore choose a shifted zero and write excitation energies as

En=nℏω.\mathcal E_n = n\hbar\omega.

This notation does not mean the original Hamiltonian lacks zero-point energy. It means E0E_0 has been subtracted. State which Hamiltonian or energy zero is being used, especially when comparing different systems, boundary conditions, or gravitational settings.

SymbolMeaningUnits or range
mmOscillator massmass, m>0m>0
ω\omegaAngular frequencyinverse time, ω>0\omega>0
ν\nuOrdinary frequencyinverse time
ℓ\ellOscillator lengthlength
pℓp_\ellNatural momentum scalemomentum
nnOne-dimensional excitation number0,1,2,…0,1,2,\ldots
NNTotal excitation number in several dimensionsnonnegative integer
aa, a†a^\daggerLowering and raising operatorsdimensionless
N=a†aN=a^\dagger aNumber operatordimensionless
HnH_nPhysicists’ Hermite polynomialdimensionless function
EnE_nEnergy eigenvalueenergy

The symbol NN is overloaded: it denotes the number operator in an operator equation and a numerical total excitation label in an eigenvalue formula. Context and hats may be used to separate them.

The energy scale has dimensions

[ℏω]=energy.[\hbar\omega] = \text{energy}.

The oscillator length and momentum scale satisfy

[ℓ]=L,[pℓ]=MLT−1,ℓpℓ=ℏ.[\ell]=L, \qquad [p_\ell]=MLT^{-1}, \qquad \ell p_\ell=\hbar.

In one-dimensional coordinate normalization,

[ψn(x)]=L−1/2.[\psi_n(x)] = L^{-1/2}.

The Hermite-polynomial argument x/ℓx/\ell and ladder operators are dimensionless.

  • The one-dimensional formula uses the full line L2(R)L^2(\mathbb R).
  • The potential is exactly quadratic and confining.
  • The mass and angular frequency are positive.
  • Position and momentum obey [X,P]=iℏI[X,P]=i\hbar I.
  • The Hamiltonian has the standard nonrelativistic kinetic term.
  • Eigenfunctions are square-integrable and decay at both infinities.
  • The displayed Hermite normalization uses the physicists’ polynomials and the stated ℓ\ell convention.
  • Multidimensional degeneracies assume independent isotropic modes and no extra internal degrees of freedom.

The one-dimensional spectrum is exact for the displayed quadratic Hamiltonian. It is not exact for:

  • anharmonic potentials;
  • a harmonic well cut off by hard boundaries;
  • an inverted oscillator with ω2<0\omega^2<0;
  • a time-dependent frequency without a separate dynamical treatment;
  • coupled coordinates before normal-mode diagonalization;
  • nonlinear or driven systems unless transformed to an equivalent quadratic form.

A local quadratic expansion is an approximation whose error depends on the higher derivatives of the potential and the spatial support of the state.

  • The quantum number begins at n=0n=0.
  • Every energy has units of ℏω\hbar\omega.
  • Adjacent one-dimensional levels differ by exactly ℏω\hbar\omega.
  • The one-dimensional spectrum is nondegenerate.
  • The nnth eigenfunction has parity (−1)n(-1)^n and nn nodes.
  • The ground-state width is ℓ/2\ell/\sqrt2, not ℓ\ell.
  • Kinetic and potential expectations each equal En/2E_n/2.
  • Numerical eigenstates should be orthonormal and negligible at artificial box boundaries.
  • Grid or basis results should converge toward the exact energies as resolution increases.
  • A displaced quadratic potential changes the center but not the spacing.

If the mass changes to m′=4mm'=4m while ω\omega is fixed, then

ℓ′=ℏ4mω=ℓ2.\ell' = \sqrt{ \frac{\hbar}{4m\omega} } = \frac{\ell}{2}.

The wavefunctions become narrower in position, while

pℓ′=4mℏω=2pℓ.p_\ell' = \sqrt{ 4m\hbar\omega } = 2p_\ell.

The energies do not change:

En′=ℏω(n+12)=En.E_n' = \hbar\omega \left( n+\frac12 \right) = E_n.

This separates the spatial scale, which depends on mωm\omega, from the level spacing, which depends only on ω\omega.

Quantum Harmonic Oscillator owns the physical overview, scales, spectrum, eigenstates, moments, and approximation role.

Differential-Equation Solution derives the spectrum from square-integrability and Hermite-series termination. Ladder-Operator Solution derives it from factorization, positivity, and the lowest-weight state.

  • Writing En=nℏωE_n=n\hbar\omega without stating that the ground energy was subtracted.
  • Starting the quantum number at n=1n=1.
  • Confusing angular frequency ω\omega with ordinary frequency ν\nu.
  • Confusing ℓ\ell with the ground-state standard deviation.
  • Treating classical turning points as quantum walls.
  • Assuming equally spaced levels for a generic confining potential.
  • Calling a†a^\dagger a particle-creation operator in every context; here it raises oscillator excitation.
  • Forgetting the additive constant when completing the square.
  • Applying one-dimensional nondegeneracy to an isotropic multidimensional oscillator.
  • Using the probabilists’ Hermite-polynomial convention with the physicists’ normalization.
  • Applying the local harmonic approximation without estimating anharmonic corrections.
  • D. J. Griffiths and D. F. Schroeter, Introduction to Quantum Mechanics, 3rd ed., Cambridge University Press, 2018, ch. 2.
  • R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994, ch. 7.
  • J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020, sec. 2.3.
  • C. Cohen-Tannoudji, B. Diu, and F. Laloë, Quantum Mechanics, vol. I, Wiley, 1977, complement V A.