Ladder Operators as Lie Algebra Tools

A ladder operator is an operator whose commutator with a chosen reference operator shifts that reference operator’s eigenvalue by a fixed amount. Ladder operators turn spectral problems into algebra: once a starting state and normalization are known, neighboring states often follow without solving a differential equation again.

This page gives the general algebraic pattern. The angular-momentum-specific derivation belongs to Ladder Operators, while the oscillator application belongs to Ladder-Operator Solution: First Encounter.

The Basic Criterion

Let $Q$ be an operator with eigenvectors $\lvert q\rangle$:

$$Q\lvert q\rangle = q\lvert q\rangle.$$

Suppose another operator $L$ satisfies

$$[Q,L] = \lambda L,$$

where $\lambda$ is a scalar. Then, whenever $L\lvert q\rangle$ is nonzero,

$$Q(L\lvert q\rangle) = (q+\lambda)(L\lvert q\rangle).$$

Thus $L$ maps a $Q$-eigenvector with eigenvalue $q$ to a new $Q$-eigenvector with eigenvalue $q+\lambda$. If $\lambda\gt0$, $L$ is a raising operator for the $Q$ label. If $\lambda\lt0$, it is a lowering operator.

The proof is a one-line commutator calculation:

$$\begin{aligned} Q L\lvert q\rangle &= (LQ+[Q,L])\lvert q\rangle\\ &= Lq\lvert q\rangle+\lambda L\lvert q\rangle\\ &= (q+\lambda)L\lvert q\rangle. \end{aligned}$$

The statement is conditional. The vector $L\lvert q\rangle$ may vanish, and in infinite-dimensional examples it may lie outside the domain of some unbounded operator unless domains are controlled.

Adjoint Ladders

If $Q$ is self-adjoint and

$$[Q,L]=\lambda L$$

with real $\lambda$, then taking adjoints gives

$$[Q,L^\dagger] = -\lambda L^\dagger.$$

So the adjoint ladder moves in the opposite direction. This is why creation and annihilation operators, or raising and lowering operators, naturally come in adjoint pairs when the inner product is part of the structure.

The norm of the shifted state is determined by

$$\lVert L\lvert q\rangle\rVert^2 = \langle q\rvert L^\dagger L\lvert q\rangle.$$

This positivity is often what fixes coefficients and stops a ladder from continuing forever.

Lie Algebra Viewpoint

In a Lie-algebra representation, ladder operators usually appear after choosing a commuting reference operator, often called a Cartan generator in the semisimple case. Operators that have simple commutators with that reference operator shift its eigenvalue labels.

The schematic pattern is

$$[H,E_\alpha] = \alpha E_\alpha.$$

Then $E_\alpha$ raises the $H$ eigenvalue by $\alpha$, while $E_{-\alpha}$ lowers it. In the angular momentum algebra, the reference operator is $J_z$ and the ladder operators are $J_+$ and $J_-$.

This root-and-weight language becomes essential in larger Lie algebras. For basic quantum mechanics, the main point is already visible in $SU(2)$ and the harmonic oscillator: commutators can determine how a whole family of states is connected.

Angular Momentum Example

For angular momentum, define

$$J_\pm=J_x\pm iJ_y.$$

The defining ladder commutators are

$$[J_z,J_\pm]=\pm\hbar J_\pm, \qquad [J^2,J_\pm]=0.$$

Therefore $J_+$ and $J_-$ change the $m$ label while keeping the same $j$ label:

$$J_\pm: \lvert j,m\rangle \longmapsto \text{constant}\times \lvert j,m\pm1\rangle.$$

The constants are fixed by the Casimir operator $J^2$ and positivity of norms:

$$J_\pm\lvert j,m\rangle = \hbar \sqrt{j(j+1)-m(m\pm1)} \, \lvert j,m\pm1\rangle.$$

The sequence terminates because a finite-dimensional unitary $SU(2)$ representation has both a highest and a lowest weight:

$$J_+\lvert j,j\rangle=0, \qquad J_-\lvert j,-j\rangle=0.$$

This is the algebraic mechanism behind the allowed labels

$$m=-j,-j+1,\ldots,j.$$

For the full angular-momentum normalization and physical interpretation, use Angular Momentum Algebra and the canonical Ladder Operators page.

Harmonic Oscillator Example

For the harmonic oscillator, the number operator is

$$N=a^\dagger a,$$

with

$$[a,a^\dagger]=1.$$

The key commutators are

$$[N,a]=-a, \qquad [N,a^\dagger]=a^\dagger.$$

Thus $a$ lowers the number eigenvalue and $a^\dagger$ raises it:

$$N\lvert n\rangle = n\lvert n\rangle \quad\Longrightarrow\quad N(a^\dagger\lvert n\rangle) = (n+1)(a^\dagger\lvert n\rangle).$$

The norm coefficients follow from positivity:

$$\lVert a\lvert n\rangle\rVert^2 = \langle n\rvert a^\dagger a\lvert n\rangle = n,$$

and

$$\lVert a^\dagger\lvert n\rangle\rVert^2 = \langle n\rvert aa^\dagger\lvert n\rangle = n+1.$$

For normalized number states,

$$a\lvert n\rangle = \sqrt n\,\lvert n-1\rangle, \qquad a^\dagger\lvert n\rangle = \sqrt{n+1}\,\lvert n+1\rangle.$$

The oscillator ladder is not finite in both directions. It stops below because $N=a^\dagger a$ is positive:

$$\langle\psi\rvert N\lvert\psi\rangle = \lVert a\lvert\psi\rangle\rVert^2 \ge 0.$$

There is a lowest state $\lvert0\rangle$ with

$$a\lvert0\rangle=0,$$

but there is no top state in the ideal harmonic oscillator. This produces the semi-infinite spectrum

$$n=0,1,2,\ldots.$$

The Hamiltonian is

$$H = \hbar\omega\left(N+\frac12\right),$$

so the energy levels are

$$E_n = \hbar\omega\left(n+\frac12\right).$$

Finite, Semi-Infinite, and Formal Ladders

Ladder methods work because algebra and positivity constrain a spectrum. The shape of the ladder depends on the representation.

Angular momentum has finite ladders in each irreducible multiplet. There is a top and a bottom because finite-dimensional unitary $SU(2)$ representations have highest and lowest weights.

The harmonic oscillator has a semi-infinite ladder. Positivity of $N$ gives a bottom state, while repeated application of $a^\dagger$ produces states with arbitrarily large $n$.

Momentum translations give a useful caution. Formally,

$$[P,e^{ikX}] = \hbar k\,e^{ikX},$$

so $e^{ikX}$ shifts momentum eigenvalue labels. But exact momentum eigenvectors are generalized, not normalizable Hilbert-space vectors. The algebraic shift is still useful, but the domain and normalization interpretation is different from finite multiplets or oscillator number states.

Worked Example: Why Norms Fix Coefficients

Suppose $N\lvert n\rangle=n\lvert n\rangle$, $\langle n\rvert n\rangle=1$, and $[a,a^\dagger]=1$. Since $N=a^\dagger a$,

$$\lVert a\lvert n\rangle\rVert^2 = \langle n\rvert N\lvert n\rangle = n.$$

Therefore

$$a\lvert n\rangle = c_n\lvert n-1\rangle$$

with

$$\lvert c_n\rvert^2=n.$$

Choosing the standard phase convention gives

$$c_n=\sqrt n.$$

The same logic fixes angular momentum coefficients after replacing $N$ by the relevant products $J_-J_+$ and $J_+J_-$.

Common Mistakes

• Thinking every pair of operators called raising and lowering operators must have the oscillator commutator $[a,a^\dagger]=1$.
• Forgetting that a ladder action may vanish at a highest or lowest state.
• Assuming the ladder coefficient is always one; normalization usually gives square-root factors.
• Treating a formal shift of generalized eigenvectors as an ordinary Hilbert-space action without domain care.
• Confusing the reference operator being shifted, such as $J_z$ or $N$, with the Hamiltonian itself.
• Applying angular-momentum ladder formulas to the oscillator or oscillator formulas to angular momentum without checking the algebra.

Cross-Links

• Commutators and Anticommutators
• Lie Algebras
• Representations
• Angular Momentum Algebra
• SU(2)
• SU(2) versus SO(3)
• Angular Momentum Ladder Operators
• Harmonic Oscillator Ladder Solution
• Quantum Harmonic Oscillator
• Canonical Commutation Relations

References

• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.
• R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994.
• A. R. Edmonds, Angular Momentum in Quantum Mechanics, Princeton University Press, 1957.
• B. C. Hall, Lie Groups, Lie Algebras, and Representations, 2nd ed., Springer, 2015.
• B. C. Hall, Quantum Theory for Mathematicians, Springer, 2013.
• H. Georgi, Lie Algebras in Particle Physics, 2nd ed., Westview Press, 1999.

Exercises

1. Prove the basic ladder criterion: if $[Q,L]=\lambda L$ and $Q\lvert q\rangle=q\lvert q\rangle$, then $L\lvert q\rangle$ has $Q$ eigenvalue $q+\lambda$ when it is nonzero.

Solution

Use $QL=LQ+[Q,L]$:

$$\begin{aligned} Q(L\lvert q\rangle) &= (LQ+\lambda L)\lvert q\rangle\\ &= qL\lvert q\rangle+\lambda L\lvert q\rangle\\ &= (q+\lambda)L\lvert q\rangle. \end{aligned}$$

If $L\lvert q\rangle=0$, there is no new eigenvector.

2. Show that $[N,a]=-a$ follows from $N=a^\dagger a$ and $[a,a^\dagger]=1$.

Solution

Use the product rule for commutators:

$$[N,a] = [a^\dagger a,a] = a^\dagger[a,a]+[a^\dagger,a]a.$$

The first term vanishes, and $[a^\dagger,a]=-[a,a^\dagger]=-1$. Therefore

$$[N,a] = -a.$$

3. Use norm positivity to derive the oscillator coefficient $a^\dagger\lvert n\rangle=\sqrt{n+1}\lvert n+1\rangle$ up to phase.

Solution

Since $aa^\dagger=a^\dagger a+1=N+1$,

$$\lVert a^\dagger\lvert n\rangle\rVert^2 = \langle n\rvert aa^\dagger\lvert n\rangle = \langle n\rvert (N+1)\lvert n\rangle = n+1.$$

If

$$a^\dagger\lvert n\rangle = c_n\lvert n+1\rangle,$$

then $\lvert c_n\rvert^2=n+1$. The standard phase convention chooses $c_n=\sqrt{n+1}$.

4. Explain why angular-momentum ladders terminate above and below, while the oscillator ladder terminates only below.

Solution

For fixed $j$, angular momentum has a finite-dimensional unitary representation with $m=-j,-j+1,\ldots,j$. The ladder coefficients vanish at $m=j$ and $m=-j$, so both ends terminate.

For the oscillator, $N=a^\dagger a$ is positive, so there must be a lowest allowed eigenvalue. The lowering operator annihilates the ground state. But the ideal oscillator has no upper bound on $N$, so repeated application of $a^\dagger$ continues indefinitely.

5. If $[Q,L]=0$, is $L$ a ladder operator for $Q$?

Solution

It does not shift the $Q$ eigenvalue. If $Q\lvert q\rangle=q\lvert q\rangle$, then

$$Q(L\lvert q\rangle) = LQ\lvert q\rangle = qL\lvert q\rangle.$$

So $L$ preserves the $Q$ eigenvalue rather than raising or lowering it. It may still act nontrivially inside a degenerate eigenspace.