Derivation of L, X and P Operators from Commutation Relations
This document provides a bottom-up derivation starting with the commutation
relations of the operators for position X i X_i X i , momentum P i P_i P i , and angular
momentum L i L_i L i . We compute their form, demonstrate their infinite-dimensional
nature, and prove their uniqueness in the Hilbert space L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 ) .
Commutation Relations ¶ The four fundamental commutation relations are:
[ L i , L j ] = i ℏ ϵ i j k L k [L_i, L_j] = i \hbar \epsilon_{ijk} L_k [ L i , L j ] = i ℏ ϵ ijk L k
[ L i , X j ] = i ℏ ϵ i j k X k [L_i, X_j] = i \hbar \epsilon_{ijk} X_k [ L i , X j ] = i ℏ ϵ ijk X k
[ L i , P j ] = i ℏ ϵ i j k P k [L_i, P_j] = i \hbar \epsilon_{ijk} P_k [ L i , P j ] = i ℏ ϵ ijk P k
[ X i , P j ] = i ℏ δ i j [X_i, P_j] = i \hbar \delta_{ij} [ X i , P j ] = i ℏ δ ij
All generators are Hermitian. The relation 1. describes the angular momentum
Lie algebra, relations 2. and 3. indicate that X i X_i X i and P i P_i P i are vector
operators with repsect to L i L_i L i . The relation 4. is the canonical commutation
relation for position and momentum.
Here, ϵ i j k \epsilon_{ijk} ϵ ijk is the Levi-Civita symbol, δ i j \delta_{ij} δ ij is the Kronecker delta, and ℏ \hbar ℏ is the reduced Planck constant.
Infinite-Dimensional Nature of X i X_i X i and P i P_i P i ¶ The operators X i X_i X i and P i P_i P i must be infinite-dimensional due to the commutation relation [ X i , P j ] = i ℏ δ i j [X_i, P_j] = i \hbar \delta_{ij} [ X i , P j ] = i ℏ δ ij . In a finite-dimensional Hilbert space of dimension n n n , the trace of a commutator [ A , B ] [A, B] [ A , B ] is zero:
Tr ( [ A , B ] ) = Tr ( A B − B A ) = Tr ( A B ) − Tr ( B A ) = Tr ( A B ) − Tr ( A B ) = 0 . \text{Tr}([A, B])
=\text{Tr}(AB-BA)
=\text{Tr}(AB)-\text{Tr}(BA)
=\text{Tr}(AB)-\text{Tr}(AB)
=0\,. Tr ([ A , B ]) = Tr ( A B − B A ) = Tr ( A B ) − Tr ( B A ) = Tr ( A B ) − Tr ( A B ) = 0 . Thus:
Left side: Tr ( [ X i , P j ] ) = 0 \text{Tr}([X_i, P_j]) = 0 Tr ([ X i , P j ]) = 0 .
Right side: Tr ( i ℏ δ i j I ) = i ℏ δ i j n \text{Tr}(i \hbar \delta_{ij} I) = i \hbar \delta_{ij} n Tr ( i ℏ δ ij I ) = i ℏ δ ij n , which is non-zero for i = j i = j i = j and n ≠ 0 n \neq 0 n = 0 .
This contradiction implies that n n n cannot be finite; hence, the Hilbert space must be infinite-dimensional, such as L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 ) .
Both X i X_i X i and P i P_i P i must be infinite-dimensional, and consequently also L i L_i L i ,
otherwise the commutators like [ L i , X j ] [L_i, X_j] [ L i , X j ] would not make sense. All the
generators must be from the same space. When infinite-dimensional space is
needed, the L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 ) is usually used, so we will use it as well.
The Stone-von Neumann theorem says that X i X_i X i and P i P_i P i are uniquely determined
as below, except a unitary transformation.
In the position representation on L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 ) :
Position Operator : X i ψ ( x ) = x i ψ ( x ) X_i \psi(x) = x_i \psi(x) X i ψ ( x ) = x i ψ ( x ) , where x i x_i x i is the i i i -th coordinate.
Momentum Operator : P i = − i ℏ ∂ ∂ x i P_i = -i \hbar \frac{\partial}{\partial x_i} P i = − i ℏ ∂ x i ∂ .
To verify:
[ X i , P j ] ψ ( x ) = X i P j ψ ( x ) − P j X i ψ ( x ) = x i ( − i ℏ ∂ ψ ∂ x j ) − ( − i ℏ ∂ ∂ x j ( x i ψ ) ) [X_i, P_j] \psi(x) = X_i P_j \psi(x) - P_j X_i \psi(x) = x_i \left(-i \hbar \frac{\partial \psi}{\partial x_j}\right) - \left(-i \hbar \frac{\partial}{\partial x_j} (x_i \psi)\right) [ X i , P j ] ψ ( x ) = X i P j ψ ( x ) − P j X i ψ ( x ) = x i ( − i ℏ ∂ x j ∂ ψ ) − ( − i ℏ ∂ x j ∂ ( x i ψ ) ) Using the product rule:
= − i ℏ x i ∂ ψ ∂ x j + i ℏ ( ∂ x i ∂ x j ψ + x i ∂ ψ ∂ x j ) = − i ℏ x i ∂ ψ ∂ x j + i ℏ δ i j ψ + i ℏ x i ∂ ψ ∂ x j = i ℏ δ i j ψ = -i \hbar x_i \frac{\partial \psi}{\partial x_j} + i \hbar \left( \frac{\partial x_i}{\partial x_j} \psi + x_i \frac{\partial \psi}{\partial x_j} \right) = -i \hbar x_i \frac{\partial \psi}{\partial x_j} + i \hbar \delta_{ij} \psi + i \hbar x_i \frac{\partial \psi}{\partial x_j} = i \hbar \delta_{ij} \psi = − i ℏ x i ∂ x j ∂ ψ + i ℏ ( ∂ x j ∂ x i ψ + x i ∂ x j ∂ ψ ) = − i ℏ x i ∂ x j ∂ ψ + i ℏ δ ij ψ + i ℏ x i ∂ x j ∂ ψ = i ℏ δ ij ψ Thus, [ X i , P j ] = i ℏ δ i j [X_i, P_j] = i \hbar \delta_{ij} [ X i , P j ] = i ℏ δ ij , confirming the forms.
Unitary Rotation Operator U U U ¶ In a previous chapter we derived for any vector operator (we will use the
vector operator X i X_i X i ):
U X i U − 1 = R j i X j . U X_i U^{-1} = R^j{}_i X_j\,. U X i U − 1 = R j i X j . Now we can write, using the known properties of X i X_i X i derived above:
X i ∣ x ⟩ = x i ∣ x ⟩ , X_i |x\rangle = x_i |x\rangle\,, X i ∣ x ⟩ = x i ∣ x ⟩ , U X i ∣ x ⟩ = U x i ∣ x ⟩ , U X_i |x\rangle = U x_i |x\rangle\,, U X i ∣ x ⟩ = U x i ∣ x ⟩ , U X i ∣ x ⟩ = x i U ∣ x ⟩ , U X_i |x\rangle = x_i U |x\rangle\,, U X i ∣ x ⟩ = x i U ∣ x ⟩ , R j i X j U ∣ x ⟩ = x i U ∣ x ⟩ , R^j{}_i X_j U |x\rangle = x_i U |x\rangle\,, R j i X j U ∣ x ⟩ = x i U ∣ x ⟩ , X j U ∣ x ⟩ = ( R − 1 ) j i x i U ∣ x ⟩ . X_j U |x\rangle = (R^{-1})_j{}^i x_i U |x\rangle\,. X j U ∣ x ⟩ = ( R − 1 ) j i x i U ∣ x ⟩ . Comparing with X i ∣ x ′ ⟩ = x i ′ ∣ x ′ ⟩ X_i |x'\rangle = x_i' |x'\rangle X i ∣ x ′ ⟩ = x i ′ ∣ x ′ ⟩ we get
U ∣ x ⟩ = ∣ x ′ ⟩ U |x\rangle=|x'\rangle U ∣ x ⟩ = ∣ x ′ ⟩ and x i ′ = ( R − 1 ) j i x i x_i' = (R^{-1})_j{}^i x_i x i ′ = ( R − 1 ) j i x i , so:
U ∣ x ⟩ = ∣ x ′ ⟩ = ∣ R − 1 x ⟩ . U |x\rangle=|x'\rangle = | R^{-1} x\rangle\,. U ∣ x ⟩ = ∣ x ′ ⟩ = ∣ R − 1 x ⟩ . Then we can compute:
⟨ x ∣ ψ ′ ⟩ = ψ ′ ( x ) = U ψ ( x ) = ⟨ x ∣ U ∣ ψ ′ ⟩ = ⟨ R − 1 x ∣ ψ ′ ⟩ = ψ ( R − 1 x ) . \langle x | \psi' \rangle
=\psi'(x)
= U\psi(x)
= \langle x | U | \psi' \rangle
= \langle R^{-1} x | \psi' \rangle
= \psi(R^{-1} x)\,. ⟨ x ∣ ψ ′ ⟩ = ψ ′ ( x ) = U ψ ( x ) = ⟨ x ∣ U ∣ ψ ′ ⟩ = ⟨ R − 1 x ∣ ψ ′ ⟩ = ψ ( R − 1 x ) . So we derived:
U ψ ( x ) = ψ ( R − 1 x ) . U\psi(x) = \psi(R^{-1} x)\,. U ψ ( x ) = ψ ( R − 1 x ) . For a rotation by angle θ \theta θ around axis n \mathbf{n} n , the unitary operator U U U is:
U = exp ( − i ℏ θ n k L k ) U = \exp\left(-\frac{i}{\hbar} \theta n^k L_k\right) U = exp ( − ℏ i θ n k L k ) Acting on a wave function:
U ψ ( x ) = ψ ( R − 1 x ) U \psi(x) = \psi(R^{-1} x) U ψ ( x ) = ψ ( R − 1 x ) where R R R is the rotation matrix, and R − 1 R^{-1} R − 1 is its inverse.
Derivation of L i L_i L i ¶ Consider an infinitesimal rotation by δ θ \delta \theta δ θ around n \mathbf{n} n . The rotation matrix approximates as R i j ≈ δ i j + δ θ ϵ i j k n k R_{ij} \approx \delta_{ij} + \delta \theta \epsilon_{ijk} n_k R ij ≈ δ ij + δ θ ϵ ijk n k , and:
R − 1 x i ≈ x i − δ θ ( n × x ) i , ( n × x ) i = ϵ i j k n j x k R^{-1} x_i \approx x_i - \delta \theta (\mathbf{n} \times \mathbf{x})_i, \quad (\mathbf{n} \times \mathbf{x})_i = \epsilon_{ijk} n_j x_k R − 1 x i ≈ x i − δ θ ( n × x ) i , ( n × x ) i = ϵ ijk n j x k Expand ψ ( R − 1 x ) \psi(R^{-1} x) ψ ( R − 1 x ) :
ψ ( R − 1 x ) ≈ ψ ( x ) − δ θ ϵ i j k n j x k ∂ ψ ∂ x i \psi(R^{-1} x) \approx \psi(x) - \delta \theta \epsilon_{ijk} n_j x_k \frac{\partial \psi}{\partial x_i} ψ ( R − 1 x ) ≈ ψ ( x ) − δ θ ϵ ijk n j x k ∂ x i ∂ ψ For U U U :
U ≈ 1 − i ℏ δ θ n m L m , U ψ ( x ) ≈ ψ ( x ) − i ℏ δ θ n m L m ψ ( x ) U \approx 1 - \frac{i}{\hbar} \delta \theta n_m L_m, \quad U \psi(x) \approx \psi(x) - \frac{i}{\hbar} \delta \theta n_m L_m \psi(x) U ≈ 1 − ℏ i δ θ n m L m , U ψ ( x ) ≈ ψ ( x ) − ℏ i δ θ n m L m ψ ( x ) Equate:
ψ ( x ) − i ℏ δ θ n m L m ψ = ψ ( x ) − δ θ ϵ i j k n j x k ∂ i ψ \psi(x) - \frac{i}{\hbar} \delta \theta n_m L_m \psi = \psi(x) - \delta \theta \epsilon_{ijk} n_j x_k \partial_i \psi ψ ( x ) − ℏ i δ θ n m L m ψ = ψ ( x ) − δ θ ϵ ijk n j x k ∂ i ψ Subtract ψ ( x ) \psi(x) ψ ( x ) and simplify:
i ℏ n m L m ψ = ϵ i j k n j x k ∂ i ψ \frac{i}{\hbar} n_m L_m \psi = \epsilon_{ijk} n_j x_k \partial_i \psi ℏ i n m L m ψ = ϵ ijk n j x k ∂ i ψ Since this holds for all n \mathbf{n} n , equate coefficients:
i ℏ L i ψ = ϵ i j k x k ∂ j ψ \frac{i}{\hbar} L_i \psi = \epsilon_{ijk} x_k \partial_j \psi ℏ i L i ψ = ϵ ijk x k ∂ j ψ Substitute P j = − i ℏ ∂ j P_j = -i \hbar \partial_j P j = − i ℏ ∂ j , so ∂ j ψ = 1 − i ℏ P j ψ = i ℏ P j ψ \partial_j \psi = \frac{1}{-i \hbar} P_j \psi = \frac{i}{\hbar} P_j \psi ∂ j ψ = − i ℏ 1 P j ψ = ℏ i P j ψ :
i ℏ L i ψ = ϵ i j k x k ( i ℏ P j ψ ) = i ℏ ϵ i j k X k P j ψ \frac{i}{\hbar} L_i \psi = \epsilon_{ijk} x_k \left( \frac{i}{\hbar} P_j \psi \right) = \frac{i}{\hbar} \epsilon_{ijk} X_k P_j \psi ℏ i L i ψ = ϵ ijk x k ( ℏ i P j ψ ) = ℏ i ϵ ijk X k P j ψ Multiply by ℏ i \frac{\hbar}{i} i ℏ :
L i ψ = ϵ i j k X j P k ψ L_i \psi = \epsilon_{ijk} X_j P_k \psi L i ψ = ϵ ijk X j P k ψ Thus, L i = ϵ i j k X j P k L_i = \epsilon_{ijk} X_j P_k L i = ϵ ijk X j P k .
Uniqueness of L i L_i L i ¶ Suppose L i ′ L_i' L i ′ also satisfies the commutation relations. Define D i = L i − L i ′ D_i = L_i - L_i' D i = L i − L i ′ :
[ D i , X j ] = [ L i , X j ] − [ L i ′ , X j ] = 0 [D_i, X_j] = [L_i, X_j] - [L_i', X_j] = 0 [ D i , X j ] = [ L i , X j ] − [ L i ′ , X j ] = 0
[ D i , P j ] = [ L i , P j ] − [ L i ′ , P j ] = 0 [D_i, P_j] = [L_i, P_j] - [L_i', P_j] = 0 [ D i , P j ] = [ L i , P j ] − [ L i ′ , P j ] = 0
D i D_i D i commutes with all X j X_j X j and P j P_j P j . In the irreducible representation on L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 ) , Schur’s lemma implies D i = c i I D_i = c_i I D i = c i I . Check the algebra:
[ L i , L j ] = [ L i ′ + D i , L j ′ + D j ] = [ L i ′ , L j ′ ] = i ℏ ϵ i j k L k ′ , [ L i , L j ] = i ℏ ϵ i j k ( L k ′ + c k I ) [L_i, L_j] = [L_i' + D_i, L_j' + D_j] = [L_i', L_j'] = i \hbar \epsilon_{ijk} L_k', \quad [L_i, L_j] = i \hbar \epsilon_{ijk} (L_k' + c_k I) [ L i , L j ] = [ L i ′ + D i , L j ′ + D j ] = [ L i ′ , L j ′ ] = i ℏ ϵ ijk L k ′ , [ L i , L j ] = i ℏ ϵ ijk ( L k ′ + c k I ) Thus, c k = 0 c_k = 0 c k = 0 , so D i = 0 D_i = 0 D i = 0 , and L i = L i ′ L_i = L_i' L i = L i ′ .
Summary ¶ The four commutation relations uniquely determine the generators:
X i = x i X_i = x_i X i = x i (position operator)
P i = − i ℏ ∂ ∂ x i P_i = -i \hbar \frac{\partial}{\partial x_i} P i = − i ℏ ∂ x i ∂ (momentum operator)
L i = ϵ i j k X j P k L_i = \epsilon_{ijk} X_j P_k L i = ϵ ijk X j P k (angular momentum operator)
Acting in the infinite-dimensional Hilbert space L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 )
No other solution is possible. One can then extend this Hilbert space for
example to also include spin, but in the smallest space L 2 ( R 3 ) L^2(\mathbb{R}^3) L 2 ( R 3 ) the
above solution is unique.
Group ¶ To determine the group elements, we can define:
T ( a ) = e i a ⋅ P . T(a) = e^{i {\bf a}\cdot{\bf P}}\,. T ( a ) = e i a ⋅ P . That’s a translation operator: T ( a ) ψ ( x ) = ψ ( x + a ) T(a)\psi(x) = \psi(x+a) T ( a ) ψ ( x ) = ψ ( x + a ) .
We derived that P {\bf P} P is a translation generator.
Next we compute from BCH:
e i a ⋅ X P j e − i a ⋅ X = P j − a j e^{i {\bf a}\cdot{\bf X}}
P_j
e^{-i {\bf a}\cdot{\bf X}}
=P_j - a_j e i a ⋅ X P j e − i a ⋅ X = P j − a j This is a momentum translation element, generated by a momentum translation
generator X {\bf X} X .
Finally:
U = exp ( − i ℏ θ n k L k ) U = \exp\left(-\frac{i}{\hbar} \theta n^k L_k\right) U = exp ( − ℏ i θ n k L k ) is an operator of rotation, due to U ψ ( x ) = ψ ( R − 1 x ) U\psi(x) = \psi(R^{-1} x) U ψ ( x ) = ψ ( R − 1 x ) .
Explicit Example: Rotation Operator Around the Z-Axis ¶ For a rotation by angle θ \theta θ around the z-axis, with n = ( 0 , 0 , 1 ) \mathbf{n} = (0, 0, 1) n = ( 0 , 0 , 1 ) :
Angular Momentum Operator L z L_z L z ¶ L z = − i ℏ ( x ∂ ∂ y − y ∂ ∂ x ) L_z = -i \hbar \left( x \frac{\partial}{\partial y} - y \frac{\partial}{\partial x} \right) L z = − i ℏ ( x ∂ y ∂ − y ∂ x ∂ ) In cylindrical coordinates: L z = − i ℏ ∂ ∂ ϕ L_z = -i \hbar \frac{\partial}{\partial \phi} L z = − i ℏ ∂ ϕ ∂
Rotation Operator U U U ¶ U = exp ( − i ℏ θ L z ) U = \exp\left(-\frac{i}{\hbar} \theta L_z\right) U = exp ( − ℏ i θ L z ) Explicitly:
Cartesian: U = exp ( − θ ( x ∂ ∂ y − y ∂ ∂ x ) ) U = \exp\left(-\theta \left( x \frac{\partial}{\partial y} - y \frac{\partial}{\partial x} \right)\right) U = exp ( − θ ( x ∂ y ∂ − y ∂ x ∂ ) )
Cylindrical: U = exp ( − θ ∂ ∂ ϕ ) U = \exp\left(-\theta \frac{\partial}{\partial \phi}\right) U = exp ( − θ ∂ ϕ ∂ )
Action on Wave Function ¶ ( U ψ ) ( x , y , z ) = ψ ( x cos θ + y sin θ , − x sin θ + y cos θ , z ) (U \psi)(x, y, z) = \psi(x \cos \theta + y \sin \theta, -x \sin \theta + y \cos \theta, z) ( U ψ ) ( x , y , z ) = ψ ( x cos θ + y sin θ , − x sin θ + y cos θ , z ) ( U ψ ) ( ρ , ϕ , z ) = ψ ( ρ , ϕ − θ , z ) (U \psi)(\rho, \phi, z) = \psi(\rho, \phi - \theta, z) ( U ψ ) ( ρ , ϕ , z ) = ψ ( ρ , ϕ − θ , z ) Matches ψ ( R − 1 x ) \psi(R^{-1} \mathbf{x}) ψ ( R − 1 x ) , where:
R = ( cos θ − sin θ 0 sin θ cos θ 0 0 0 1 ) R = \begin{pmatrix} \cos \theta & -\sin \theta & 0 \\ \sin \theta & \cos \theta & 0 \\ 0 & 0 & 1 \end{pmatrix} R = ⎝ ⎛ cos θ sin θ 0 − sin θ cos θ 0 0 0 1 ⎠ ⎞