Skip to content

Coupled and Uncoupled Bases

Composite angular momentum systems have two natural bases. The uncoupled basis labels each subsystem separately:

∣j1,m1⟩∣j2,m2⟩.|j_1,m_1\rangle|j_2,m_2\rangle.

The coupled basis labels total angular momentum:

∣j1,j2;J,M⟩.|j_1,j_2;J,M\rangle.

Both bases span the same tensor-product Hilbert space for fixed j1j_1 and j2j_2. The difference is which commuting observables are diagonal.

Let the first subsystem carry angular momentum J1\mathbf J_1 and the second carry J2\mathbf J_2. The composite Hilbert space is

H=Hj1⊗Hj2,\mathcal H = \mathcal H_{j_1}\otimes\mathcal H_{j_2},

with dimension

dim⁡H=(2j1+1)(2j2+1).\dim\mathcal H = (2j_1+1)(2j_2+1).

Operators for the two subsystems act on different tensor factors. The total angular momentum is

J=J1+J2,\mathbf J=\mathbf J_1+\mathbf J_2,

with

Jz=J1z+J2z.J_z=J_{1z}+J_{2z}.

The total components again satisfy the angular momentum algebra:

[Ji,Jj]=iℏ∑kϵijkJk.[J_i,J_j] = i\hbar\sum_k\epsilon_{ijk}J_k.

The purpose of angular momentum addition is to understand how the tensor product decomposes into irreducible total-JJ multiplets.

The uncoupled basis diagonalizes

J12,J1z,J22,J2z.J_1^2,\quad J_{1z},\quad J_2^2,\quad J_{2z}.

Its basis states obey

J12∣j1,m1⟩∣j2,m2⟩=ℏ2j1(j1+1)∣j1,m1⟩∣j2,m2⟩,J_1^2|j_1,m_1\rangle|j_2,m_2\rangle = \hbar^2j_1(j_1+1) |j_1,m_1\rangle|j_2,m_2\rangle,

and

J1z∣j1,m1⟩∣j2,m2⟩=ℏm1∣j1,m1⟩∣j2,m2⟩,J_{1z}|j_1,m_1\rangle|j_2,m_2\rangle = \hbar m_1 |j_1,m_1\rangle|j_2,m_2\rangle,

with analogous equations for J22J_2^2 and J2zJ_{2z}.

This basis is natural when the Hamiltonian treats the two angular momenta separately, for example in a field that couples independently to J1zJ_{1z} and J2zJ_{2z}.

The coupled basis diagonalizes

J12,J22,J2,Jz.J_1^2,\quad J_2^2,\quad J^2,\quad J_z.

Its basis states obey

J2∣j1,j2;J,M⟩=ℏ2J(J+1)∣j1,j2;J,M⟩,J^2|j_1,j_2;J,M\rangle = \hbar^2J(J+1)|j_1,j_2;J,M\rangle,

and

Jz∣j1,j2;J,M⟩=ℏM∣j1,j2;J,M⟩.J_z|j_1,j_2;J,M\rangle = \hbar M|j_1,j_2;J,M\rangle.

The labels j1j_1 and j2j_2 are usually kept in the ket because the same total JJ value can arise from different constituent angular momenta in larger problems.

For fixed j1j_1 and j2j_2, the allowed total angular momenta are

J=∣j1−j2∣,∣j1−j2∣+1,…,j1+j2.J = |j_1-j_2|, |j_1-j_2|+1, \ldots, j_1+j_2.

For each JJ,

M=−J,−J+1,…,J.M=-J,-J+1,\ldots,J.

The dimension check is

(2j1+1)(2j2+1)=∑J=∣j1−j2∣j1+j2(2J+1).(2j_1+1)(2j_2+1) = \sum_{J=|j_1-j_2|}^{j_1+j_2}(2J+1).

The coupled and uncoupled bases are related by Clebsch–Gordan coefficients:

∣j1,j2;J,M⟩=∑m1,m2⟨j1,m1;j2,m2∣J,M⟩∣j1,m1⟩∣j2,m2⟩.|j_1,j_2;J,M\rangle = \sum_{m_1,m_2} \langle j_1,m_1;j_2,m_2|J,M\rangle |j_1,m_1\rangle|j_2,m_2\rangle.

The inverse expansion is

∣j1,m1⟩∣j2,m2⟩=∑J,M⟨J,M∣j1,m1;j2,m2⟩∣j1,j2;J,M⟩.|j_1,m_1\rangle|j_2,m_2\rangle = \sum_{J,M} \langle J,M|j_1,m_1;j_2,m_2\rangle |j_1,j_2;J,M\rangle.

The transformation is unitary. It is a change of basis, not a physical approximation.

Because

Jz=J1z+J2z,J_z=J_{1z}+J_{2z},

the coefficient vanishes unless

M=m1+m2.M=m_1+m_2.

The detailed coefficient conventions, signs, and table-reading rules are in Clebsch–Gordan Coefficients.

For three or more angular momenta, different orders of pairwise coupling are related by Recoupling and Wigner Symbols. Which order is physically useful is discussed in Angular Momentum Coupling Schemes.

The preferred basis depends on the observables or Hamiltonian.

The uncoupled basis is natural for questions such as:

  • What are the two individual zz components?
  • What happens when each spin couples to a different local field?
  • What is the tensor-product state before an interaction is turned on?

The coupled basis is natural for questions such as:

  • What is the total angular momentum?
  • Which irreducible rotational multiplet is the state in?
  • How does an interaction depending on J1⋅J2\mathbf J_1\cdot\mathbf J_2 split the space?

The key identity is

J1⋅J2=12(J2−J12−J22).\mathbf J_1\cdot\mathbf J_2 = \frac12 \left( J^2-J_1^2-J_2^2 \right).

Therefore an interaction proportional to J1⋅J2\mathbf J_1\cdot\mathbf J_2 is diagonal in the coupled basis, because J2J^2, J12J_1^2, and J22J_2^2 are all diagonal there.

For two spin-1/21/2 systems,

12⊗12=1⊕0.\frac12\otimes\frac12 = 1\oplus0.

The uncoupled basis is

∣↑↑⟩,∣↑↓⟩,∣↓↑⟩,∣↓↓⟩.|\uparrow\uparrow\rangle,\quad |\uparrow\downarrow\rangle,\quad |\downarrow\uparrow\rangle,\quad |\downarrow\downarrow\rangle.

The coupled basis is a triplet plus a singlet:

∣1,1⟩=∣↑↑⟩,∣1,0⟩=12(∣↑↓⟩+∣↓↑⟩),∣1,−1⟩=∣↓↓⟩,∣0,0⟩=12(∣↑↓⟩−∣↓↑⟩).\begin{aligned} |1,1\rangle &= |\uparrow\uparrow\rangle, \\ |1,0\rangle &= \frac{1}{\sqrt2} \left( |\uparrow\downarrow\rangle + |\downarrow\uparrow\rangle \right), \\ |1,-1\rangle &= |\downarrow\downarrow\rangle, \\ |0,0\rangle &= \frac{1}{\sqrt2} \left( |\uparrow\downarrow\rangle - |\downarrow\uparrow\rangle \right). \end{aligned}

The full derivation starts in Two Spin-1/2 Particles, and the rotational scalar/vector structure is developed in Singlet and Triplet States.

For one electron in an orbital state with angular momentum ℓ\ell and spin s=1/2s=1/2, the uncoupled basis is

∣ℓ,mℓ⟩∣s,ms⟩.|\ell,m_\ell\rangle|s,m_s\rangle.

The coupled basis is

∣ℓ,s;j,m⟩,|\ell,s;j,m\rangle,

where

j=ℓ+12orj=ℓ−12j=\ell+\frac12 \quad \text{or} \quad j=\ell-\frac12

when ℓ>0\ell>0. For ℓ=0\ell=0, only j=1/2j=1/2 occurs.

This basis becomes natural when the Hamiltonian contains a spin–orbit term proportional to

L⋅S=12(J2−L2−S2).\mathbf L\cdot\mathbf S = \frac12 \left( J^2-L^2-S^2 \right).

The angular momentum algebra makes the structure transparent even before one studies the detailed atomic fine-structure Hamiltonian. The dedicated angular-momentum treatment is Spin–Orbit Coupling.

  • Thinking coupled states live in a different Hilbert space from uncoupled states. They are different bases of the same space.
  • Forgetting that M=m1+m2M=m_1+m_2 for every nonzero Clebsch–Gordan coefficient.
  • Treating JJ as if it were always j1+j2j_1+j_2; lower total values also occur.
  • Dropping the constituent labels j1,j2j_1,j_2 when they are needed to distinguish sectors.
  • Assuming the coupled basis is always better. It is better for rotationally invariant interactions, not for every Hamiltonian.
  • Confusing a product state with a state of definite total angular momentum.
  • 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.
  • D. A. Varshalovich, A. N. Moskalev, and V. K. Khersonskii, Quantum Theory of Angular Momentum, World Scientific, 1988.
  • R. N. Zare, Angular Momentum: Understanding Spatial Aspects in Chemistry and Physics, Wiley, 1988.
  1. For j1=1j_1=1 and j2=1j_2=1, list the allowed total JJ values and check the dimension count.
Solution

The allowed values are

J=∣1−1∣,…,1+1,J=|1-1|,\ldots,1+1,

so

J=0,1,2.J=0,1,2.

The uncoupled space has dimension

(2j1+1)(2j2+1)=3⋅3=9.(2j_1+1)(2j_2+1)=3\cdot3=9.

The coupled multiplet dimensions add to

(2⋅0+1)+(2⋅1+1)+(2⋅2+1)=1+3+5=9.(2\cdot0+1)+(2\cdot1+1)+(2\cdot2+1) = 1+3+5 = 9.
  1. Why must ⟨j1,m1;j2,m2∣J,M⟩\langle j_1,m_1;j_2,m_2|J,M\rangle vanish unless M=m1+m2M=m_1+m_2?
Solution

The product state is an eigenstate of

Jz=J1z+J2zJ_z=J_{1z}+J_{2z}

with eigenvalue

ℏ(m1+m2).\hbar(m_1+m_2).

The coupled state ∣j1,j2;J,M⟩|j_1,j_2;J,M\rangle is an eigenstate of JzJ_z with eigenvalue ℏM\hbar M. States with different JzJ_z eigenvalues are orthogonal, so the overlap can be nonzero only when

M=m1+m2.M=m_1+m_2.
  1. Show why J1⋅J2\mathbf J_1\cdot\mathbf J_2 is diagonal in the coupled basis.
Solution

Use

J2=(J1+J2)2=J12+J22+2J1⋅J2.J^2 = (\mathbf J_1+\mathbf J_2)^2 = J_1^2+J_2^2+2\mathbf J_1\cdot\mathbf J_2.

Therefore

J1⋅J2=12(J2−J12−J22).\mathbf J_1\cdot\mathbf J_2 = \frac12 \left( J^2-J_1^2-J_2^2 \right).

The coupled basis diagonalizes J2J^2, J12J_1^2, and J22J_2^2, so it also diagonalizes this scalar product.

  1. For ℓ=1\ell=1 and s=1/2s=1/2, list the allowed total jj values and their multiplet dimensions.
Solution

The allowed total values are

j=ℓ+s, ℓ−s,j=\ell+s,\ \ell-s,

so

j=32,12.j=\frac32,\frac12.

Their dimensions are

2j+1=4and2j+1=2.2j+1=4 \quad \text{and} \quad 2j+1=2.

The total dimension is 4+2=64+2=6, matching

(2ℓ+1)(2s+1)=3⋅2=6.(2\ell+1)(2s+1)=3\cdot2=6.