Ladder Operators for the Harmonic Oscillator

A vibrating diatomic molecule stores energy in quanta, and a piano string does not. The difference is discreteness. The fastest way to see where it comes from is algebra, not calculus. Solving the harmonic oscillator Schrödinger Equation by series methods eventually produces the Hermite Functions and Oscillator Wavefunctions. These form a tower of polynomials dressed in Gaussians. But there is a second route, due essentially to Dirac. It never writes down a differential equation at all. Instead it factors the Hamiltonian into a product of two new operators. These are traditionally called a^\hat a and a^†\hat a^\dagger. Their commutation relation alone forces the energy spectrum to be evenly spaced by ℏω\hbar\omega. This matters far beyond a computational shortcut. The ladder-operator method is the prototype for how quantum field theory describes photons, phonons, and every other harmonic excitation. A "particle" is just one quantum of raising operator acting on a vacuum. It is also the cleanest illustration of a key idea. An abstract commutation relation, [x^,p^]=iℏ[\hat x,\hat p]=i\hbar, can by itself generate an entire spectrum. This happens without ever solving for a wavefunction in position space. Vibrational spectroscopy, laser cooling, cavity quantum electrodynamics, and the second-quantized treatment of molecular vibrations all lean on this algebra. This includes force-field and ab initio anharmonic calculations. The big picture is that the one-dimensional harmonic oscillator Hamiltonian,

H^=p^22m+12mω2x^2,\hat H = \frac{\hat p^2}{2m} + \frac{1}{2}m\omega^2\hat x^2,

can be rewritten as H^=ℏω(a^†a^+12)\hat H = \hbar\omega(\hat a^\dagger\hat a + \tfrac12). Here a^\hat a and a^†\hat a^\dagger lower and raise the number of vibrational quanta by exactly one. The eigenstates ∣n⟩|n\rangle built this way are identical to the Hermite-function eigenstates. Those are obtained by direct integration of the differential equation. The algebra is not a different theory. It is a different, more powerful bookkeeping system for the same physics. This article builds the operators from scratch and derives the commutator. It proves the spectrum is En=(n+12)ℏωE_n = (n+\tfrac12)\hbar\omega by pure operator algebra. It constructs the normalized eigenstates by repeated raising from the ground state. Finally, it connects the number-operator picture back to position-space wavefunctions. It also connects this picture to matrix elements needed for selection rules in vibrational spectroscopy.

Key takeaways
  • Define a^=mω2ℏ(x^+ip^mω)\hat a=\sqrt{\tfrac{m\omega}{2\hbar}}\big(\hat x+\tfrac{i\hat p}{m\omega}\big) and a^†\hat a^\dagger its adjoint; they obey [a^,a^†]=1[\hat a,\hat a^\dagger]=1.
  • The Hamiltonian becomes H^=ℏω(N^+12)\hat H=\hbar\omega(\hat N+\tfrac12) with number operator N^=a^†a^\hat N=\hat a^\dagger\hat a.
  • N^\hat N has non-negative integer eigenvalues only, forced by requiring norms ≥0\ge 0 at every rung of the ladder.
  • a^∣n⟩=n ∣n−1⟩\hat a|n\rangle=\sqrt{n}\,|n-1\rangle and a^†∣n⟩=n+1 ∣n+1⟩\hat a^\dagger|n\rangle=\sqrt{n+1}\,|n+1\rangle generate every eigenstate from ∣0⟩|0\rangle.
  • Matrix elements of x^\hat x and p^\hat p in the ∣n⟩|n\rangle basis vanish unless Δn=±1\Delta n=\pm1, which is the algebraic root of the harmonic-oscillator selection rule in IR spectroscopy.
  • Key idea: factoring H^\hat H into ℏω(a^†a^+12)\hbar\omega(\hat a^\dagger\hat a+\tfrac12) turns solving a differential equation into counting quanta with a commutator.
The core insight

The single commutation relation [a^,a^†]=1[\hat a,\hat a^\dagger]=1, together with positivity of the norm, forces the harmonic oscillator spectrum to be an infinite, evenly spaced ladder En=(n+12)ℏωE_n=(n+\tfrac12)\hbar\omega — no series solution, no Hermite polynomial, no boundary condition at infinity is needed to get the energies.

Where this lives today

Vibrational spectroscopy still uses the Δv=±1\Delta v=\pm1 harmonic selection rule derived above as the zeroth-order guide for interpreting infrared and Raman spectra of every diatomic and polyatomic molecule, with anharmonic corrections computed as perturbative mixing between number states via cubic and quartic terms in x^\hat x. Molecular dynamics force fields and normal-mode analyses represent a polyatomic molecule's small-amplitude vibrations as a set of independent harmonic oscillators (normal modes), each quantized with its own ladder operators, and thermodynamic partition functions for vibrations are built directly from the evenly spaced En=(n+12)ℏωE_n=(n+\tfrac12)\hbar\omega ladder. In condensed-matter physics, lattice vibrations are quantized as phonons using precisely this creation/annihilation algebra applied mode-by-mode to the crystal's normal coordinates; phonon population statistics, thermal conductivity, and Raman/infrared selection rules in solids all trace back to [a^k,a^k†]=1[\hat a_k,\hat a_k^\dagger]=1 for each phonon mode kk. Quantum optics and cavity quantum electrodynamics use the identical algebra for photon number states ∣n⟩|n\rangle of a cavity mode, with coherent states ∣α⟩|\alpha\rangle (eigenstates of a^\hat a) describing laser light and squeezed states built by non-trivial combinations of a^,a^†\hat a,\hat a^\dagger used in gravitational-wave detectors such as LIGO to beat the standard quantum limit. In trapped-ion quantum computing, a single ion's motional state in a harmonic trap is manipulated and cooled to its ground state (n=0n=0) using exactly the raising/lowering language developed here, and sideband cooling exploits the Δn=±1\Delta n=\pm1 coupling to laser fields to reach the vibrational ground state needed for high-fidelity quantum logic gates.

:::quiz[What is the value of [a^,a^†][\hat a,\hat a^\dagger] for the harmonic oscillator ladder operators as defined in this article?]

  • iℏi\hbar
  • ℏω\hbar\omega
  • 11 ✓
  • 00

Ask the tutor: How do the ladder operators change when treating a two-dimensional isotropic harmonic oscillator, and how does that foreshadow angular momentum quantization?

The physical picture

Think of the harmonic oscillator Hamiltonian as an unfactored quadratic form, H^∝x^2+p^2\hat H \propto \hat x^2+\hat p^2 in suitable units. If x^\hat x and p^\hat p were ordinary numbers this would factor as a difference of squares, (x^+ip^)(x^−ip^)(\hat x+i\hat p)(\hat x-i\hat p), using i2=−1i^2=-1. Because x^\hat x and p^\hat p do not commute, the factorization is not exact — but it is exact up to an extra additive constant that comes entirely from the commutator [x^,p^]=iℏ[\hat x,\hat p]=i\hbar. This leftover constant is the origin of the famous zero-point energy 12ℏω\tfrac12\hbar\omega. Physically, a^†\hat a^\dagger adds one quantum of vibrational energy ℏω\hbar\omega to the system, and a^\hat a removes one. Repeated application of a^\hat a on any state must eventually terminate — you cannot remove infinitely many quanta from a system whose energy is bounded below — and this termination condition is what pins the ground state energy to exactly 12ℏω\tfrac12\hbar\omega rather than zero. Everything above the ground state is built by climbing the ladder with a^†\hat a^\dagger, picking up a n+1\sqrt{n+1} normalization factor at each rung, exactly mirroring how angular momentum raising and lowering operators build the ∣j,m⟩|j,m\rangle ladder used later for the Particle on a Sphere and Spherical Harmonics problem. The position-space wavefunctions ψn(x)\psi_n(x) built from Hermite polynomials are not abandoned by this approach; they are recovered as the unique solution of a^ ψ0(x)=0\hat a\,\psi_0(x)=0, a first-order differential equation, followed by repeated differentiation-and-multiplication via a^†\hat a^\dagger. The ladder operators are thus a bridge: they give the spectrum algebraically and, when needed, regenerate the explicit real-space functions with far less labor than solving the second-order Hermite equation directly.

Figure
The ladder of harmonic-oscillator energy levels, evenly spaced by ħω
position, x ½mω²x² n=0, E₀=½ħω n=1, E₁=³⁄₂ħω n=2, E₂=⁵⁄₂ħω n=3, E₃=⁷⁄₂ħω ↠↠Each rung is exactly ħω above the last; spacing is independent of n.
Figure
Action of â and ↠as bra–ket flow diagrams on the number states
|0⟩ |1⟩ |2⟩ |3⟩ â†, ×1 â†, ×√2 â†, ×√3 â, ×1 â, ×√2 â, ×√3 â|0⟩=0

The core ideas

The starting Hamiltonian for a particle of mass mm on a spring of force constant k=mω2k=m\omega^2 is

H^=p^22m+12mω2x^2.\hat H = \frac{\hat p^2}{2m} + \frac{1}{2}m\omega^2\hat x^2.

where: x^,p^\hat x,\hat p are the position and momentum operators satisfying [x^,p^]=iℏ[\hat x,\hat p]=i\hbar, and ω=k/m\omega=\sqrt{k/m} is the classical angular frequency. Reading: this is exactly the same Hamiltonian solved by Hermite functions; nothing physical has changed, only the algebra used to diagonalize it. Define the dimensionless-in-spirit combinations

a^=mω2ℏ(x^+imωp^),a^†=mω2ℏ(x^−imωp^).\hat a = \sqrt{\frac{m\omega}{2\hbar}}\left(\hat x + \frac{i}{m\omega}\hat p\right), \qquad \hat a^\dagger = \sqrt{\frac{m\omega}{2\hbar}}\left(\hat x - \frac{i}{m\omega}\hat p\right).

where: a^\hat a is the annihilation (lowering) operator, a^†\hat a^\dagger its Hermitian-conjugate creation (raising) operator, and the prefactor mω/2ℏ\sqrt{m\omega/2\hbar} is chosen precisely so the resulting commutator is unity. Reading: a^\hat a is a complex combination of position and momentum, not itself Hermitian, so it is not an observable — but a^†a^\hat a^\dagger\hat a will be. The fundamental commutation relation, derived below, is

[a^,a^†]=a^a^†−a^†a^=1.[\hat a,\hat a^\dagger] = \hat a\hat a^\dagger - \hat a^\dagger\hat a = 1.

where: the right side is the identity operator (times ℏ0\hbar^0, i.e. a pure number in the chosen units). Reading: this single algebraic fact, inherited entirely from [x^,p^]=iℏ[\hat x,\hat p]=i\hbar, is the engine that produces the entire energy ladder. The Hamiltonian rewritten in terms of the number operator N^≡a^†a^\hat N \equiv \hat a^\dagger\hat a is

H^=ℏω(N^+12).\hat H = \hbar\omega\left(\hat N + \frac{1}{2}\right).

where: N^\hat N is Hermitian (N^†=(a^†a^)†=a^†a^=N^\hat N^\dagger = (\hat a^\dagger\hat a)^\dagger = \hat a^\dagger\hat a = \hat N) so its eigenvalues are real, and in fact turn out to be the non-negative integers n=0,1,2,…n=0,1,2,\dots. Reading: once you know N^∣n⟩=n∣n⟩\hat N|n\rangle = n|n\rangle, the energies fall out immediately as En=(n+12)ℏωE_n=(n+\tfrac12)\hbar\omega. The ladder action on normalized eigenstates ∣n⟩|n\rangle is

a^∣n⟩=n ∣n−1⟩,a^†∣n⟩=n+1 ∣n+1⟩.\hat a|n\rangle = \sqrt{n}\,|n-1\rangle, \qquad \hat a^\dagger|n\rangle = \sqrt{n+1}\,|n+1\rangle.

where: the square-root prefactors are exactly what keep ∣n⟩|n\rangle normalized to unity at every step; they are not free choices but forced by the algebra, as shown in the derivation below. Reading: this is the operational heart of the method — every matrix element, every selection rule, and every explicit wavefunction follows from these two lines. Position and momentum recovered from the ladder operators are

x^=ℏ2mω(a^+a^†),p^=iℏmω2(a^†−a^).\hat x = \sqrt{\frac{\hbar}{2m\omega}}\left(\hat a+\hat a^\dagger\right), \qquad \hat p = i\sqrt{\frac{\hbar m\omega}{2}}\left(\hat a^\dagger - \hat a\right).

where: these are just the inverted definitions of a^,a^†\hat a,\hat a^\dagger. Reading: because a^,a^†\hat a,\hat a^\dagger only connect states differing by one quantum, x^\hat x and p^\hat p can only have nonzero matrix elements between ∣n⟩|n\rangle and ∣n±1⟩|n\pm1\rangle — the algebraic origin of the Δv=±1\Delta v=\pm1 vibrational selection rule (in the harmonic approximation).

:::derivation[The commutator [a^,a^†]=1[\hat a,\hat a^\dagger]=1 from [x^,p^]=iℏ[\hat x,\hat p]=i\hbar] Goal. Show that the specific prefactor mω/2ℏ\sqrt{m\omega/2\hbar} chosen in the definitions of a^,a^†\hat a,\hat a^\dagger makes their commutator exactly the number 11, using only [x^,p^]=iℏ[\hat x,\hat p]=i\hbar.

Step 1 — Write the product a^a^†\hat a\hat a^\dagger. Substitute the definitions:

a^a^†=mω2ℏ(x^+imωp^)(x^−imωp^).\hat a\hat a^\dagger = \frac{m\omega}{2\hbar}\left(\hat x+\frac{i}{m\omega}\hat p\right)\left(\hat x-\frac{i}{m\omega}\hat p\right).

Expand the product term by term, being careful not to commute x^\hat x and p^\hat p past each other:

a^a^†=mω2ℏ[x^2−imωx^p^+imωp^x^+1m2ω2p^2].\hat a\hat a^\dagger = \frac{m\omega}{2\hbar}\left[\hat x^2 - \frac{i}{m\omega}\hat x\hat p + \frac{i}{m\omega}\hat p\hat x + \frac{1}{m^2\omega^2}\hat p^2\right].

Step 2 — Group the cross terms. The two middle terms combine as

−imωx^p^+imωp^x^=−imω(x^p^−p^x^)=−imω[x^,p^].-\frac{i}{m\omega}\hat x\hat p+\frac{i}{m\omega}\hat p\hat x = -\frac{i}{m\omega}(\hat x\hat p-\hat p\hat x) = -\frac{i}{m\omega}[\hat x,\hat p].

Insert [x^,p^]=iℏ[\hat x,\hat p]=i\hbar:

−imω(iℏ)=ℏmω.-\frac{i}{m\omega}(i\hbar) = \frac{\hbar}{m\omega}.

So

a^a^†=mω2ℏ[x^2+1m2ω2p^2+ℏmω]=mω2ℏx^2+12mℏωp^2+12.\hat a\hat a^\dagger = \frac{m\omega}{2\hbar}\left[\hat x^2+\frac{1}{m^2\omega^2}\hat p^2+\frac{\hbar}{m\omega}\right] = \frac{m\omega}{2\hbar}\hat x^2+\frac{1}{2m\hbar\omega}\hat p^2+\frac{1}{2}.

Step 3 — Compute a^†a^\hat a^\dagger\hat a the same way. Swapping the order of the factors flips the sign of the cross term:

a^†a^=mω2ℏ[x^2+imωx^p^−imωp^x^+1m2ω2p^2]=mω2ℏx^2+12mℏωp^2−12.\hat a^\dagger\hat a = \frac{m\omega}{2\hbar}\left[\hat x^2+\frac{i}{m\omega}\hat x\hat p-\frac{i}{m\omega}\hat p\hat x+\frac{1}{m^2\omega^2}\hat p^2\right] = \frac{m\omega}{2\hbar}\hat x^2+\frac{1}{2m\hbar\omega}\hat p^2-\frac{1}{2}.

Notice the quadratic terms in x^2,p^2\hat x^2,\hat p^2 are identical in both expressions — only the sign of the constant flips.

Step 4 — Subtract. All the x^2\hat x^2 and p^2\hat p^2 pieces cancel exactly:

a^a^†−a^†a^=12−(−12)=1.\hat a\hat a^\dagger - \hat a^\dagger\hat a = \frac{1}{2}-\left(-\frac{1}{2}\right) = 1.

This confirms [a^,a^†]=1[\hat a,\hat a^\dagger]=1 as a pure number, independent of m,ω,ℏm,\omega,\hbar — those constants were absorbed entirely into the choice of prefactor in Step 1. Equivalently, adding the two boxed expressions from Steps 2–3 gives a^a^†+a^†a^=mωℏx^2+1mℏωp^2\hat a\hat a^\dagger+\hat a^\dagger\hat a = \frac{m\omega}{\hbar}\hat x^2+\frac{1}{m\hbar\omega}\hat p^2, and comparing to H^=p^22m+12mω2x^2\hat H=\frac{\hat p^2}{2m}+\frac12 m\omega^2\hat x^2 shows

a^†a^=H^ℏω−12⟹H^=ℏω(a^†a^+12),\hat a^\dagger\hat a = \frac{\hat H}{\hbar\omega}-\frac12 \quad\Longrightarrow\quad \hat H=\hbar\omega\left(\hat a^\dagger\hat a+\frac12\right),

exactly the factorization promised.

Derivation AThe spectrum En=(n+12)ℏωE_n=(n+\tfrac12)\hbar\omega from the commutator alone

Goal. Using only [a^,a^†]=1[\hat a,\hat a^\dagger]=1 and H^=ℏω(N^+12)\hat H=\hbar\omega(\hat N+\tfrac12), prove that N^=a^†a^\hat N=\hat a^\dagger\hat a has eigenvalues exactly 0,1,2,3,…0,1,2,3,\dots — an infinite ladder, evenly spaced, bounded below — without ever writing ψ(x)\psi(x).

Step 1 — Commutators of N^\hat N with the ladder operators. Using [A^B^,C^]=A^[B^,C^]+[A^,C^]B^[\hat A\hat B,\hat C]=\hat A[\hat B,\hat C]+[\hat A,\hat C]\hat B,

[N^,a^]=[a^†a^,a^]=a^†[a^,a^]+[a^†,a^]a^=0+(−1)a^=−a^,[\hat N,\hat a] = [\hat a^\dagger\hat a,\hat a] = \hat a^\dagger[\hat a,\hat a]+[\hat a^\dagger,\hat a]\hat a = 0+(-1)\hat a = -\hat a,
[N^,a^†]=a^†[a^,a^†]+[a^†,a^†]a^=a^†(1)+0=a^†.[\hat N,\hat a^\dagger] = \hat a^\dagger[\hat a,\hat a^\dagger]+[\hat a^\dagger,\hat a^\dagger]\hat a = \hat a^\dagger(1)+0 = \hat a^\dagger.

These say a^\hat a lowers the eigenvalue of N^\hat N by 1, and a^†\hat a^\dagger raises it by 1 — proven precisely in Step 2.

Step 2 — Show a^∣n⟩\hat a|n\rangle is an eigenstate of N^\hat N with eigenvalue n−1n-1. Let N^∣n⟩=n∣n⟩\hat N|n\rangle = n|n\rangle for some (as yet unknown, but real and non-negative — see Step 3) number nn. Then

N^(a^∣n⟩)=(a^N^+[N^,a^])∣n⟩=(a^N^−a^)∣n⟩=a^(n∣n⟩)−a^∣n⟩=(n−1)a^∣n⟩.\hat N(\hat a|n\rangle) = (\hat a\hat N+[\hat N,\hat a])|n\rangle = (\hat a\hat N-\hat a)|n\rangle = \hat a(n|n\rangle)-\hat a|n\rangle = (n-1)\hat a|n\rangle.

So a^∣n⟩\hat a|n\rangle is either the zero vector or an eigenvector of N^\hat N with eigenvalue n−1n-1. Identical algebra with [N^,a^†]=a^†[\hat N,\hat a^\dagger]=\hat a^\dagger shows a^†∣n⟩\hat a^\dagger|n\rangle has eigenvalue n+1n+1.

Step 3 — nn cannot be negative. The eigenvalue of N^\hat N equals a squared norm:

n=⟨n∣N^∣n⟩=⟨n∣a^†a^∣n⟩=∥a^∣n⟩∥2≥0,n = \langle n|\hat N|n\rangle = \langle n|\hat a^\dagger\hat a|n\rangle = \big\|\hat a|n\rangle\big\|^2 \ge 0,

since the norm of any vector (here a^∣n⟩\hat a|n\rangle) is non-negative. So every eigenvalue of N^\hat N satisfies n≥0n\ge 0.

Step 4 — The ladder must terminate exactly at zero. Suppose some eigenvalue nn were not a non-negative integer, e.g. n=2.7n=2.7. Applying a^\hat a repeatedly generates eigenvalues 1.7, 0.7, −0.3,…1.7,\ 0.7,\ -0.3,\dots By Step 3 none of these can be negative — yet the sequence, if it never hits exactly zero, must eventually cross below zero after finitely many steps (since each step subtracts exactly 1). This is a contradiction. The only way to avoid it is for the descending chain to land exactly on n=0n=0 and stop there, which requires a^∣0⟩=0\hat a|0\rangle = 0 (the null vector, not an eigenvector — the chain simply terminates). Hence every allowed nn must be a non-negative integer: n=0,1,2,3,…n=0,1,2,3,\dots

Step 5 — Read off the energies. Since H^=ℏω(N^+12)\hat H=\hbar\omega(\hat N+\tfrac12) and N^∣n⟩=n∣n⟩\hat N|n\rangle=n|n\rangle with n=0,1,2,…n=0,1,2,\dots,

H^∣n⟩=ℏω(n+12)∣n⟩⟹En=(n+12)ℏω.\hat H|n\rangle = \hbar\omega\left(n+\frac12\right)|n\rangle \quad\Longrightarrow\quad E_n = \left(n+\frac12\right)\hbar\omega.

The zero-point energy E0=12ℏωE_0=\tfrac12\hbar\omega is not an assumption; it is forced by the requirement a^∣0⟩=0\hat a|0\rangle=0 together with the +12+\tfrac12 that appeared automatically in the factorization of H^\hat H.

Step 6 — Normalization constants on the ladder. Write a^∣n⟩=cn∣n−1⟩\hat a|n\rangle = c_n|n-1\rangle. Taking the norm-squared,

∣cn∣2=⟨n∣a^†a^∣n⟩=⟨n∣N^∣n⟩=n⟹cn=n|c_n|^2 = \langle n|\hat a^\dagger\hat a|n\rangle = \langle n|\hat N|n\rangle = n \quad\Longrightarrow\quad c_n=\sqrt{n}

(choosing the phase real and positive by convention). An identical computation with a^†∣n⟩=dn∣n+1⟩\hat a^\dagger|n\rangle=d_n|n+1\rangle gives ∣dn∣2=⟨n∣a^a^†∣n⟩=⟨n∣(N^+1)∣n⟩=n+1|d_n|^2=\langle n|\hat a\hat a^\dagger|n\rangle=\langle n|(\hat N+1)|n\rangle=n+1, so dn=n+1d_n=\sqrt{n+1}. This reproduces exactly the boxed ladder relations stated in the core ideas.

Worked examples

Example 1 — Foundational: building ∣2⟩|2\rangle from the vacuum and checking its energy. Starting from ∣0⟩|0\rangle, defined by a^∣0⟩=0\hat a|0\rangle=0 and ⟨0∣0⟩=1\langle 0|0\rangle=1, construct ∣2⟩|2\rangle and verify H^∣2⟩=52ℏω∣2⟩\hat H|2\rangle = \tfrac52\hbar\omega|2\rangle. Apply a^†\hat a^\dagger once:

a^†∣0⟩=0+1 ∣1⟩=∣1⟩.\hat a^\dagger|0\rangle = \sqrt{0+1}\,|1\rangle = |1\rangle.

Apply a^†\hat a^\dagger a second time, using the ladder rule with n=1n=1:

a^†∣1⟩=1+1 ∣2⟩=2 ∣2⟩.\hat a^\dagger|1\rangle = \sqrt{1+1}\,|2\rangle = \sqrt{2}\,|2\rangle.

So (a^†)2∣0⟩=2 ∣2⟩(\hat a^\dagger)^2|0\rangle = \sqrt{2}\,|2\rangle, i.e.

∣2⟩=12(a^†)2∣0⟩.|2\rangle = \frac{1}{\sqrt{2}}(\hat a^\dagger)^2|0\rangle.

Now check the energy directly using H^=ℏω(a^†a^+12)\hat H=\hbar\omega(\hat a^\dagger\hat a+\tfrac12). First lower: a^∣2⟩=2 ∣1⟩\hat a|2\rangle=\sqrt{2}\,|1\rangle. Then raise again: a^†(2∣1⟩)=2⋅2 ∣2⟩=2∣2⟩\hat a^\dagger(\sqrt2|1\rangle) = \sqrt2\cdot\sqrt{2}\,|2\rangle = 2|2\rangle. So a^†a^∣2⟩=N^∣2⟩=2∣2⟩\hat a^\dagger\hat a|2\rangle = \hat N|2\rangle = 2|2\rangle, confirming n=2n=2 directly by explicit operator action rather than by assumption. Therefore

H^∣2⟩=ℏω(2+12)∣2⟩=52ℏω ∣2⟩,\hat H|2\rangle = \hbar\omega\left(2+\frac12\right)|2\rangle = \frac{5}{2}\hbar\omega\,|2\rangle,

matching En=(n+12)ℏωE_n=(n+\tfrac12)\hbar\omega at n=2n=2 exactly, with no reference to Hermite polynomials.

Example 2 — Applied: vibrational excitation energy of gas-phase CO and the algebraic origin of its IR selection rule. Carbon monoxide has a harmonic vibrational wavenumber ν~=2170 cm−1\tilde\nu = 2170\ \text{cm}^{-1} (a genuine, oft-tabulated value; note it is not a round number, by design). Using E=hcν~E=hc\tilde\nu, the fundamental quantum is

ℏω=hcν~=(6.626×10−34 J s)(2.998×1010 cm s−1)(2170 cm−1).\hbar\omega = hc\tilde\nu = (6.626\times10^{-34}\ \text{J s})(2.998\times10^{10}\ \text{cm s}^{-1})(2170\ \text{cm}^{-1}).

Compute the product stepwise:

hc=6.626×10−34×2.998×1010=1.986×10−23 J cm.hc = 6.626\times10^{-34}\times 2.998\times10^{10} = 1.986\times10^{-23}\ \text{J cm}.
ℏω=1.986×10−23 J cm×2170 cm−1=4.310×10−20 J.\hbar\omega = 1.986\times10^{-23}\ \text{J cm}\times 2170\ \text{cm}^{-1} = 4.310\times10^{-20}\ \text{J}.

The energy of the v=0→v=1v=0\to v=1 transition is exactly this quantum, since

E1−E0=(1+12)ℏω−(0+12)ℏω=ℏω,E_1-E_0 = \left(1+\tfrac12\right)\hbar\omega-\left(0+\tfrac12\right)\hbar\omega = \hbar\omega,

regardless of which rung you start from — the ladder spacing is uniform, a direct consequence of N^\hat N having integer eigenvalues one apart. Why does only Δv=±1\Delta v=\pm1 appear (in the harmonic/double-harmonic approximation)? The transition dipole matrix element for absorption involves ⟨v′∣x^∣v⟩\langle v'|\hat x|v\rangle (since the dipole moment is expanded linearly in the bond-length displacement for the harmonic approximation). Using x^=ℏ/2mω (a^+a^†)\hat x=\sqrt{\hbar/2m\omega}\,(\hat a+\hat a^\dagger),

⟨v′∣x^∣v⟩=ℏ2mω(v δv′,v−1+v+1 δv′,v+1).\langle v'|\hat x|v\rangle = \sqrt{\frac{\hbar}{2m\omega}}\left(\sqrt{v}\,\delta_{v',v-1}+\sqrt{v+1}\,\delta_{v',v+1}\right).

Every term on the right is nonzero only when v′=v±1v'=v\pm1; there is no way to get v′=v±2v'=v\pm2 or higher from a single power of a^+a^†\hat a+\hat a^\dagger. This is the operator-algebra proof of the harmonic selection rule, sharper and shorter than doing the corresponding Hermite-polynomial integral by hand.

Example 3 — Competition-level: matrix elements of x^2\hat x^2 and the anharmonic correction structure. Compute ⟨3∣x^2∣1⟩\langle 3|\hat x^2|1\rangle exactly (a nonzero, non-diagonal, non-adjacent-looking matrix element that tests whether the ladder algebra is applied correctly), for a particle of mass mm and frequency ω\omega. Write x^2\hat x^2 in terms of ladder operators:

x^2=ℏ2mω(a^+a^†)2=ℏ2mω(a^2+a^a^†+a^†a^+(a^†)2).\hat x^2 = \frac{\hbar}{2m\omega}(\hat a+\hat a^\dagger)^2 = \frac{\hbar}{2m\omega}\left(\hat a^2+\hat a\hat a^\dagger+\hat a^\dagger\hat a+(\hat a^\dagger)^2\right).

Act term by term on ∣1⟩|1\rangle. Term a^2∣1⟩\hat a^2|1\rangle: first a^∣1⟩=1∣0⟩=∣0⟩\hat a|1\rangle=\sqrt1|0\rangle=|0\rangle, then a^∣0⟩=0\hat a|0\rangle=0. So a^2∣1⟩=0\hat a^2|1\rangle=0, contributing nothing to ⟨3∣a^2∣1⟩\langle3|\hat a^2|1\rangle. Term a^a^†∣1⟩\hat a\hat a^\dagger|1\rangle: first a^†∣1⟩=2∣2⟩\hat a^\dagger|1\rangle=\sqrt2|2\rangle, then a^(2∣2⟩)=2⋅2∣1⟩=2∣1⟩\hat a(\sqrt2|2\rangle)=\sqrt2\cdot\sqrt2|1\rangle=2|1\rangle. This produces ∣1⟩|1\rangle, orthogonal to ∣3⟩|3\rangle, so it contributes 00 to ⟨3∣⋯∣1⟩\langle3|\cdots|1\rangle. Term a^†a^∣1⟩\hat a^\dagger\hat a|1\rangle: a^∣1⟩=∣0⟩\hat a|1\rangle=|0\rangle, then a^†∣0⟩=∣1⟩\hat a^\dagger|0\rangle=|1\rangle. Again produces ∣1⟩|1\rangle, contributing 00 to the ⟨3∣\langle3| overlap. Term (a^†)2∣1⟩(\hat a^\dagger)^2|1\rangle: first a^†∣1⟩=2∣2⟩\hat a^\dagger|1\rangle=\sqrt2|2\rangle, then a^†(2∣2⟩)=2⋅3∣3⟩=6∣3⟩\hat a^\dagger(\sqrt2|2\rangle)=\sqrt2\cdot\sqrt3|3\rangle=\sqrt6|3\rangle. This is the only term with nonzero overlap onto ⟨3∣\langle3|, giving ⟨3∣(a^†)2∣1⟩=6\langle3|(\hat a^\dagger)^2|1\rangle=\sqrt6. Collecting only the surviving term,

⟨3∣x^2∣1⟩=ℏ2mω×6=6 ℏ2mω.\langle3|\hat x^2|1\rangle = \frac{\hbar}{2m\omega}\times\sqrt6 = \sqrt6\,\frac{\hbar}{2m\omega}.

This nonzero Δv=±2\Delta v=\pm2 matrix element of x^2\hat x^2 (as opposed to x^\hat x) is exactly why overtone and combination-band-inducing terms in real anharmonic potentials — which contain x^3,x^4,…\hat x^3,\hat x^4,\dots — connect states two, three, or four rungs apart, and why Raman scattering (which couples to x^2\hat x^2-type polarizability terms) shows different selection rules from IR absorption.

Limiting cases, scope & common traps

Regime / checkBehaviorWhy
n=0n=0 (ground state)a^∣0⟩=0\hat a|0\rangle=0 exactly; E0=12ℏω≠0E_0=\tfrac12\hbar\omega\ne0zero-point energy is forced by the +12+\tfrac12 from the commutator, not a boundary-condition artifact
Large nnLevel spacing stays exactly ℏω\hbar\omega, independent of nnN^\hat N has strictly evenly spaced integer eigenvalues; classical correspondence emerges as relative spacing ℏω/En→0\hbar\omega/E_n\to0
ℏ→0\hbar\to 0 (formal classical limit)Spacing ℏω→0\hbar\omega\to0; energy becomes effectively continuousclassical mechanics has no intrinsic quantum, matching the correspondence principle
Anharmonic real potentials (Morse, etc.)Ladder operators only diagonalize the harmonic part; anharmonic terms mix different ∣n⟩|n\ranglea^,a^†\hat a,\hat a^\dagger are defined relative to one specific ω\omega; a cubic term like x^3\hat x^3 connects Δn=±1,±3\Delta n=\pm1,\pm3
Two or three dimensions (isotropic oscillator)Factorizes into independent x,y,zx,y,z ladders, each with its own a^i,a^i†\hat a_i,\hat a_i^\daggerHamiltonian is separable; total n=nx+ny+nzn=n_x+n_y+n_z gives degenerate shells, foreshadowing angular-momentum shell structure
Coherent/quasi-classical states ∣α⟩|\alpha\rangleEigenstates of a^\hat a (not N^\hat N) with a^∣α⟩=α∣α⟩\hat a|\alpha\rangle=\alpha|\alpha\rangleminimum-uncertainty wavepackets that oscillate classically; not stationary states of H^\hat H

The chemist behind it

Paul Adrien Maurice Dirac introduced the operator method for the harmonic oscillator in the earliest years of matrix and operator quantum mechanics, treating x^\hat x and p^\hat p as abstract non-commuting objects rather than differential operators acting on wavefunctions. His approach — factor the Hamiltonian, extract the algebra, read off the spectrum from positivity of the norm — became the template not only for the oscillator but for angular momentum, spin, and eventually the creation/annihilation operator formalism of quantum electrodynamics, where photons themselves are literally treated as quanta of a harmonic-oscillator-like field mode.

Origins · 1 minPaul Adrien Maurice Dirac and how this was discovered
PA
Paul Adrien Maurice Dirac
1902–1984 · Cambridge, early quantum mechanics and quantum electrodynamics

Dirac unified Heisenberg's matrix mechanics and Schrödinger's wave mechanics into a single operator formalism, introduced bra–ket notation, predicted the positron from the relativistic electron equation that bears his name, and pioneered the second-quantized description of the electromagnetic field in which photon creation and annihilation operators obey exactly the algebra derived here for a single vibrational mode.

History
- 1925 — Heisenberg, Born, and Jordan formulate matrix mechanics; non-commuting observables enter quantum theory for the first time.
- 1926 — Schrödinger derives the wave equation and solves the harmonic oscillator using Hermite polynomials via series methods.
- 1927 — Dirac develops the transformation theory and operator calculus that make the algebraic factorization of H^\hat H natural and general.
- 1930 — Dirac's Principles of Quantum Mechanics presents the ladder-operator solution of the oscillator as a textbook method.
- 1932-onward — the same creation/annihilation algebra is generalized to quantized electromagnetic field modes, founding quantum electrodynamics and, later, phonon and photon counting throughout condensed-matter and molecular spectroscopy.
Are the ladder-operator eigenstates ∣n⟩|n\rangle literally the same functions as the Hermite-polynomial wavefunctions ψn(x)\psi_n(x)?

Yes. Solving a^ψ0(x)=0\hat a\psi_0(x)=0 as a first-order ODE in position representation gives the Gaussian ground state ψ0(x)∝e−mωx2/2ℏ\psi_0(x)\propto e^{-m\omega x^2/2\hbar}, identical to the n=0n=0 Hermite function. Repeatedly applying a^†\hat a^\dagger in position representation (a combination of multiplying by xx and differentiating) generates the same sequence of Hermite polynomials times a Gaussian that the series-solution method produces. The two approaches are completely equivalent descriptions of the same eigenstates; the operator method just avoids ever writing the second-order differential equation explicitly.

Why is the zero-point energy exactly 12ℏω\tfrac12\hbar\omega and not zero?

It comes directly from the algebra: H^=ℏω(a^†a^+12)\hat H=\hbar\omega(\hat a^\dagger\hat a+\tfrac12), and the 12\tfrac12 appears because a^a^†−a^†a^=1\hat a\hat a^\dagger-\hat a^\dagger\hat a=1 is nonzero — a direct consequence of the position–momentum commutator [x^,p^]=iℏ[\hat x,\hat p]=i\hbar. Physically, the uncertainty principle forbids a particle from having simultaneously exact position (at the potential minimum) and exact momentum (zero), so the true minimum-energy state must have some residual kinetic and potential energy, and that residual works out to precisely 12ℏω\tfrac12\hbar\omega.

Can a^\hat a and a^†\hat a^\dagger be measured directly in an experiment?

Not directly, since they are non-Hermitian and thus not observables. What is measured are Hermitian combinations: the position x^∝(a^+a^†)\hat x\propto(\hat a+\hat a^\dagger), the momentum p^∝i(a^†−a^)\hat p\propto i(\hat a^\dagger-\hat a), or the number operator N^=a^†a^\hat N=\hat a^\dagger\hat a, whose eigenvalue is the vibrational (or phonon, or photon) quantum number actually measured in spectroscopy or photon-counting experiments. The individual operators a^,a^†\hat a,\hat a^\dagger are best understood as bookkeeping devices that make the algebra transparent, analogous to how complex exponentials simplify Fourier analysis even though "complex amplitude" is not itself a direct observable.

How does this connect to angular momentum raising and lowering operators used for the rigid rotor and spherical harmonics?

The mathematical pattern is identical: define non-Hermitian combinations (J^±=J^x±iJ^y\hat J_\pm = \hat J_x\pm i\hat J_y for angular momentum, versus a^,a^†\hat a,\hat a^\dagger here) that shift a quantum number by a fixed step, derive their commutator with the relevant Casimir-like operator, and use positivity of the norm to force quantization. For angular momentum this produces the integer/half-integer jj ladder and, ultimately, the spherical harmonics that solve the Particle on a Sphere and Spherical Harmonics problem; the harmonic oscillator ladder is the simplest, cleanest example of the same general strategy.

Do real molecular vibrations ever behave like true harmonic oscillators, or is this all an approximation?

Real molecular potentials are anharmonic — bonds dissociate, so the potential flattens at large displacement rather than rising forever as 12kx2\tfrac12 kx^2. The harmonic ladder-operator treatment is the leading-order approximation, valid for small-amplitude vibrations near the potential minimum and for low vibrational quantum numbers. Anharmonic corrections are typically added perturbatively using the same a^,a^†\hat a,\hat a^\dagger basis, since matrix elements of x^3,x^4,…\hat x^3,\hat x^4,\dots in the harmonic basis are easy to compute exactly using the algebra demonstrated in the worked examples.

Why does raising or lowering pick up a n\sqrt{n} or n+1\sqrt{n+1} factor instead of just moving to the neighboring state unchanged?

The square-root factors are required to keep every ∣n⟩|n\rangle normalized to one. If a^†∣n⟩\hat a^\dagger|n\rangle simply equaled ∣n+1⟩|n+1\rangle with no prefactor, the norm of a^†∣n⟩\hat a^\dagger|n\rangle would have to equal the norm of ∣n⟩|n\rangle, namely 1 — but a direct computation of ⟨n∣a^a^†∣n⟩=⟨n∣(N^+1)∣n⟩=n+1\langle n|\hat a\hat a^\dagger|n\rangle=\langle n|(\hat N+1)|n\rangle=n+1 shows the actual norm-squared is n+1n+1, not 1. The n+1\sqrt{n+1} prefactor is exactly what reconciles these two numbers, and it is derived rather than assumed.

What would go wrong if someone used a^\hat a and a^†\hat a^\dagger defined without the specific prefactor mω/2ℏ\sqrt{m\omega/2\hbar}?

Any other constant multiplying x^+ip^/(mω)\hat x+i\hat p/(m\omega) would rescale the commutator to some value other than exactly 1, for example [a^,a^†]=ℏ[\hat a,\hat a^\dagger]=\hbar or 2ℏω2\hbar\omega depending on the chosen normalization. The final physics (the energy spectrum) would still come out correctly after adjusting the algebra consistently, but the clean statement H^=ℏω(N^+12)\hat H=\hbar\omega(\hat N+\tfrac12) with integer eigenvalues of N^\hat N specifically requires the prefactor derived here; that is why the convention is universal in textbooks and research literature alike.

Before this

Where this leads