Ireta
Ireta
Angular Momentum
Angular momentum: is as important in classical mechanics as in quantum mechanics. It is
particularly useful for studying the dynamics of systems that move under the influence of
spherically symmetric, or central, potentials, 𝑉(𝑟⃗) = 𝑉(𝑟), for the orbital angular momenta of
these systems are conserved.
One of the cornerstones of Bohr’s model of the hydrogen atom (where the electron moves in
the proton’s Coulomb potential, a central potential) is based on the quantization of angular
momentum.
The total angular momentum,(𝐽⃗), combines both the spin and orbital angular momentum of a
⃗⃗ + 𝑆⃗.
particle (or a system), namely 𝐽⃗ = 𝐿
where 𝐿⃗⃗ is orbital angular momentum, 𝑆⃗ is spin angular momentum or just spin.
Additionally, angular momentum plays a critical role in the description of molecular rotations,
the motion of electrons in atoms, and the motion of nucleons in nuclei.
I. Orbital angular momentum (L)
A. General Formalism of angular momentum
In classical physics the angular momentum of a particle with momentum 𝑝⃗ and position 𝑟⃗⃗⃗ is
defined by
𝐿⃗⃗ = 𝑟⃗⃗ × 𝑝⃗ = −𝑝⃗ × 𝑟⃗⃗
𝐿⃗⃗ = (𝑦𝑝𝑧 − 𝑧𝑝𝑦 )𝑖̂ + (𝑧𝑝𝑥 − 𝑥𝑝𝑧 )𝑗̂ + (𝑥𝑝𝑦 − 𝑦𝑝𝑥 )𝑘̂
The orbital angular momentum operator 𝐿 ⃗̂⃗ can be obtained at once by replacing 𝑟⃗ and 𝑝⃗ by the
corresponding operators in the position representation, 𝑅⃗⃗̂ and 𝑃⃗̂⃗ = −𝑖ℏ∇
⃗⃗= ℏ ∇
⃗⃗:
𝑖
⃗̂⃗ = 𝑅⃗⃗̂ × 𝑃⃗⃗̂ = −𝑖ℏ(𝑅⃗⃗̂ × ⃗∇⃗)
𝐿
⃗̂⃗ are
The Cartesian components of 𝐿
𝜕 𝜕
𝐿̂𝑥 = 𝑌̂ 𝑃
̂𝑧 − 𝑍̂ 𝑃
̂𝑦 = −𝑖ℏ (𝑌̂ ̂
−𝑍 ),
𝜕𝑧 𝜕𝑦
𝜕 𝜕
𝐿̂𝑦 = 𝑍̂ ̂ ̂̂
𝑃𝑥 − 𝑋 ̂
𝑃𝑧 = −𝑖ℏ (𝑍 ̂
−𝑋 ),
𝜕𝑥 𝜕𝑧
𝜕 𝜕
𝐿̂𝑧 = 𝑋
̂𝑃
̂𝑦 − 𝑌̂ 𝑃
̂𝑥 = −𝑖ℏ (𝑋
̂ ̂
−𝑌 ),
𝜕𝑦 𝜕𝑥
Clearly, angular momentum does not exist in a one-dimensional space. We should mention that
⃗̂⃗,
the components 𝐿̂𝑥 , 𝐿̂𝑦 , 𝐿̂𝑧 , and the square of 𝐿
⃗̂⃗2 = 𝐿̂2𝑥 + 𝐿̂2𝑦 + 𝐿̂2𝑧
𝐿
⃗̂⃗2 = 𝑟 2 𝑝2 − (𝑟 ∙ 𝑝)2 + 𝑖ℏ𝑟 ∙ 𝑝
𝐿
are all Hermitian.
𝑅⃗⃗̂ → (𝑋
̂, 𝑌̂ , 𝑍̂ ), 𝑃⃗̂⃗ → (𝑃
̂𝑥 , 𝑃
̂𝑦 , 𝑃 ⃗̂⃗ → (𝐿̂ 𝑥 , 𝐿̂ 𝑦 , 𝐿̂ 𝑧 )
̂𝑧 ), 𝐿
Commutators
[𝐴, 𝐵] = 𝐴𝐵 − 𝐵𝐴
=[𝑌̂ 𝑃
̂𝑧 , 𝑍̂ 𝑃
̂𝑥 ] − [𝑌̂ 𝑃
̂𝑧 , 𝑋
̂𝑃
̂𝑧 ] − [𝑍̂ 𝑃
̂𝑦 , 𝑍̂ 𝑃
̂𝑥 ] + [𝑍̂ 𝑃
̂𝑦 , 𝑋
̂𝑃
̂𝑧 ]
=𝑌̂[𝑃̂𝑧 , 𝑍̂]𝑃̂𝑥 + 𝑋̂[𝑍̂, 𝑃̂𝑧 ]𝑃̂𝑦 = 𝑖ℏ(𝑋̂𝑃̂𝑦 − 𝑌̂𝑃̂𝑥 )
̂
= 𝑖ℏ𝐿 𝑧
Similar calculation yields the other two commutation relations; but it is much simpler to infer
them from equation above by means of a cyclic permutation of the 𝑥𝑦𝑧 components, 𝑥 → 𝑦 →
𝑧 → 𝑥:
[𝐿̂𝑥 , 𝐿̂𝑦 ] = 𝐿̂𝑥 𝐿̂𝑦 − 𝐿̂𝑦 𝐿̂𝑥 = 𝑖ℏ𝐿̂𝑧 ,
[𝐿̂𝑦 , 𝐿̂𝑧 ] = 𝐿̂𝑦 𝐿̂𝑧 − 𝐿̂𝑧 𝐿̂𝑦 = 𝑖ℏ𝐿̂𝑥 ,
[𝐿̂𝑧 , 𝐿̂𝑥 ] = 𝐿̂𝑧 𝐿̂𝑥 − 𝐿̂𝑥 𝐿̂𝑧 = 𝑖ℏ𝐿̂𝑦 ,
The three equations are equivalent to the vectorial commutation relation:
𝐿̂ × 𝐿̂ = 𝑖ℏ𝐿̂ ,
Note that this can only be true for operators; since, for regular vectors, 𝐿̂ × 𝐿̂ = 𝟎.
⃗̂⃗2 does commute with each of the three components: 𝐿̂𝑥 , 𝐿̂𝑦 𝑜𝑟 𝐿̂𝑧
It is easy to show that 𝐿
For example (using [𝐿̂2𝑥 , 𝐿̂𝑥 ] = 0):
⃗̂⃗2 , 𝐿̂𝑥 ] = [𝐿̂2𝑦 + 𝐿̂2𝑧 , 𝐿̂𝑥 ] = [𝐿̂2𝑦 , 𝐿̂𝑥 ] + [𝐿̂2𝑧 , 𝐿̂𝑥 ]
[𝐿
= 𝐿̂𝑦 [𝐿̂𝑦 , 𝐿̂𝑥 ] + [𝐿̂𝑦 , 𝐿̂𝑥 ]𝐿̂𝑦 + 𝐿̂𝑧 [𝐿̂𝑧 , 𝐿̂𝑥 ] + [𝐿̂𝑧 , 𝐿̂𝑥 ]𝐿̂𝑧
= −𝑖ℏ(𝐿̂𝑦 𝐿̂𝑧 + 𝐿̂𝑧 𝐿̂𝑦 ) + 𝒊ℏ(𝐿̂𝑧 𝐿̂𝑦 + 𝐿̂𝑦 𝐿̂𝑧 )=0
Similarly,
⃗̂⃗2 , 𝐿̂𝑦 ] = [𝐿
[𝐿 ⃗̂⃗2 , 𝐿̂𝑧 ] = 0
which can be summarized as
⃗⃗2 , 𝐿
[𝐿 ⃗⃗] = 0
Physically, this means that one can find simultaneous eigenfunctions of 𝐿 ⃗⃗2 and one of the
components of 𝐿 ⃗⃗, implying that both the magnitude of the angular momentum and one of its
components can be precisely determined. Once these are known, they fully specify the angular
momentum.
Example:
Solution:
(a) The only nonzero commutator which involves 𝑋, ̂ and the various components of , 𝐿̂𝑥
̂ ̂
, 𝐿𝑦 , 𝐿𝑧 is [𝑋 ̂
̂, 𝑃𝑥 ] = 𝑖ℏ. Having stated this result, we can easily evaluate the needed
commutators.
First, since 𝐿̂𝑥 = 𝑌̂ 𝑃 ̂𝑦 involves no 𝑃̂ , the operator 𝑋̂ commutes separately with 𝑌̂, 𝑃
̂𝑧 − 𝑍̂ 𝑃 ̂𝑧 ,
𝑍̂, and 𝑃̂𝑦 ; hence
̂, 𝐿̂𝑥 ] = [𝑋,
[𝑋 ̂ 𝑌̂ 𝑃
̂𝑧 − 𝑍̂ 𝑃
̂𝑦 ] = 0
The evaluation of the other two commutators is straightforward:
̂, 𝐿̂𝑦 ] = [𝑋,
[𝑋 ̂ 𝑍̂ 𝑃
̂𝑥 − 𝑋
̂𝑃 ̂𝑧 ] = [𝑋,
̂ 𝑍̂ 𝑃
̂𝑥 ] = 𝑍̂ [𝑋,
̂𝑃 ̂𝑥 ] = 𝑖ℏ𝑍̂ ,
̂, 𝐿̂𝑧 ] = [𝑋,
[𝑋 ̂𝑋 ̂𝑃̂𝑦 − 𝑌̂ 𝑃
̂𝑥 ] = −[𝑋,̂ 𝑌̂ 𝑃
̂𝑥 ] = −𝑌̂ [𝑋,
̂𝑃 ̂𝑥 ] = −𝑖ℏ𝑌̂ ,
(b) The only commutator between 𝑃̂𝑥 and the components of 𝐿̂𝑥 , 𝐿̂𝑦 , 𝐿̂𝑧 that survives is again
̂𝑥 , 𝑋
[𝑃 ̂] = −𝑖ℏ. We may thus infer
̂𝑥 , 𝐿̂𝑥 ] = [𝑃̂𝑥 , 𝑌̂ 𝑃
[𝑃 ̂𝑧 − 𝑍̂ 𝑃
̂𝑦 ] = 0,
̂𝑥 , 𝐿̂𝑦 ] = [𝑃̂𝑥 , 𝑍̂ 𝑃
[𝑃 ̂𝑥 − 𝑋̂𝑧 ] = −[𝑃̂𝑥 , 𝑋
̂𝑃 ̂𝑧 ] = −[𝑃̂𝑥 , 𝑋
̂𝑃 ̂ ]𝑃
̂𝑧 = 𝑖ℏ𝑃
̂𝑧 ,
𝑃𝑥 , 𝐿̂𝑧 ] = [𝑃̂𝑥 , 𝑋
[̂ ̂̂ 𝑃𝑥 ] = [𝑃̂𝑥 , 𝑋
̂̂
𝑃𝑦 − 𝑌 𝑃𝑦 ] = [𝑃̂𝑥 , 𝑋
̂̂ ̂ ]̂
𝑃𝑦 = −𝑖ℏ̂
𝑃𝑦 ,
(c) Using the commutators derived in (a) and (b), we infer
̂𝐿
[𝑋, ⃗̂⃗2 ] = [𝑋,
̂ 𝐿̂2𝑥 ] + [𝑋,
̂ 𝐿̂2𝑦 ] + [𝑋,
̂ 𝐿̂2𝑧 ]
= 0 + 𝐿̂𝑦 [𝑋,
̂ 𝐿̂𝑦 ] + [𝑋,
̂ 𝐿̂𝑦 ]𝐿̂𝑦 + 𝐿̂𝑧 [𝑋,
̂ 𝐿̂𝑧 ] + [𝑋,
̂ 𝐿̂𝑧 ]𝐿̂𝑧
= 0 + 𝐿̂𝑦 [𝑃̂𝑥 , 𝐿̂𝑦 ] + [𝑃̂𝑥 , 𝐿̂𝑦 ]𝐿̂𝑦 + 𝐿̂𝑧 [𝑃̂𝑥 , 𝐿̂𝑧 ] + [𝑃̂𝑥 , 𝐿̂𝑧 ]𝐿̂𝑧
= 𝑖ℏ(𝐿̂𝑦 𝑃̂𝑧 + 𝑃̂𝑧 𝐿̂𝑦 − 𝐿̂𝑧 𝑃̂𝑦 − 𝑃̂𝑦 𝐿̂𝑧 )
⃗⃗2 and one of the components of 𝐿
In order to obtain the eigenvalues of 𝐿 ⃗⃗ (typically, 𝐿̂𝑧 ), it is
convenient to express the angular momentum operators in spherical polar coordinates: 𝑟, 𝜃, 𝜙,
rather than the Cartesian coordinates 𝑥, 𝑦, 𝑧. The spherical coordinates are related to the
Cartesian ones via
𝑥 = 𝑟𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜙;
𝑦 = 𝑟𝑠𝑖𝑛𝜃𝑠𝑖𝑛𝜙;
𝑧 = 𝑟𝑐𝑜𝑠𝜃;
After some algebra, one gets:
𝜕 𝜕
𝐿̂𝑥 = 𝑖ℏ (𝑠𝑖𝑛𝜙 + 𝑐𝑜𝑡𝜃𝑐𝑜𝑠𝜙 )
𝜕𝜃 𝜕𝜙
̂𝐿𝑦 = 𝑖ℏ (−𝑐𝑜𝑠𝜙 𝜕 + 𝑐𝑜𝑡𝜃𝑠𝑖𝑛𝜙 𝜕 )
𝜕𝜃 𝜕𝜙
𝜕
𝐿̂𝑧 = −𝑖ℏ 𝜕𝜙
1 𝜕 1 𝜕 1 𝜕2
⃗⃗2 = −ℏ2 (
𝐿 ( )+ )
𝑠𝑖𝑛𝜃 𝜕𝜃 𝑠𝑖𝑛𝜃 𝜕𝜃 𝑠𝑖𝑛2 𝜃 𝜕∅2
We thus find that the operators 𝐿̂𝑥 , 𝐿̂𝑦 , 𝐿̂𝑧 and 𝐿 ⃗⃗2 depend on 𝜃 and 𝜙 only, that is they are
independent on the radial coordinate 𝑟. All these operators therefore commute with any
function of 𝑟,
[𝐿̂𝑥 , 𝑓(𝑟)] = [𝐿̂𝑦 , 𝑓(𝑟)] = [𝐿̂𝑧 , 𝑓(𝑟)] = [𝐿 ⃗⃗2 , 𝑓(𝑟)] = 0
Also, obviously, if a wavefunction depends only on 𝑟 (but not on 𝜃, 𝜙) it can be simultaneously
an eigenfunction of 𝐿̂𝑥 , 𝐿̂𝑦 , 𝐿̂𝑧 and 𝐿
⃗⃗2 . In all cases, the corresponding eigenvalue will be 0. (This
is the only exception to the rule that that eigenvalues of one component (e.g., 𝐿̂𝑥 ) cannot be
simultaneously eigenfunctions of the two other components of 𝐿̂).
I. Eigenvalues of 𝑳𝒛
Since, in spherical coordinates 𝐿𝑧 depends only on ∅, we can denote its eigenvalue by
𝑚ℏ and the corresponding eigenfunctions by Φ𝑚 (𝜙). We thus have:
𝑳𝒛 Φ𝑚 (𝜙) = 𝑚ℏΦ𝑚 (𝜙)
Namely
𝜕
−i Φ (𝜙) = mΦ𝑚 (𝜙)
𝜕∅ 𝑚
The solutions to this equation are
1
Φ𝑚 (𝜙) = 𝒆𝒊𝒎∅
√2𝜋
This is satisfied for any value of m; however, physically we require the wave function to be
single valued (alternatively: continuous), namely Φ𝑚 (2𝜋) = Φ𝑚 (0), from which we find
𝒆𝒊2𝜋𝒎∅ = 𝟎
This equation is satisfied for 𝑚 = 0, ±1, ±2, ±3, . . .. The eigenvalues of the operator 𝑳𝒛 are
thus 𝑚ℏ, with m being integer (positive or negative) or zero. The number 𝑚 is called the
magnetic quantum number, due to the role it plays in the motion of charged particles in
magnetic fields.
This means, that when measuring the z-component of an orbital angular momentum,
one can only obtain 0,± ℏ,±2ℏ, .... Since the choice of the z direction was arbitrary, we see that
the component of the orbital angular momentum about any axis is quantized.
10 1 0 01 −𝑖 0 1 0 0
𝑙𝑥 = (1 0 1) ℏ, 𝑙𝑦 = (𝑖 0 −𝑖 ) ℏ, 𝑙𝑧 = (0 0 0 ) ℏ,
√2 0 1 0 √2 0 𝑖 0 0 0 −1
There are other ways to express these matrices in subspace 𝐶 3 of an infinite dimensional
Hilbert space. Relations above are dominantly the most popular. Since the three operators do
not commute, we arbitrarily have selected a basis for one of them, and then expressed the
other two in that basis. Notice 𝑙𝑧 is diagonal. That means the basis selected is natural for 𝑙𝑧 .
The terminology usually used is the operators in equations above are in the 𝑙𝑧 basis.
10 1 0
(1 0 1 ) ℏ
√2 0 1 0
𝑙𝑥
1 0 −𝑖 0
𝑙 = (𝑙𝑦 ) = ( 𝑖 0 −𝑖 ) ℏ
𝑙𝑧 √2 0 𝑖 0
1 0 0
(0 0 0 ) ℏ
( 0 0 −1 )
Again, this operator will normally be denoted just 𝑙. The 𝑙 operator is a different sort of object
than the component operators. It is a different object in a different space. Yet, we would like a
way to address angular momentum with a 3 X 3 matrix which is in the same subspace asthe
components. We can do this if we use 𝑙 2 . This operator is
1 0 0
𝑙 2 = 2ℏ2 𝐼 = 2ℏ2 (0 1 0)
0 0 1
Where I is the identity matrix.
It is convenient to define the raising and lowering operators (note the similarity to the
Harmonic oscillator!):
𝐿± = 𝐿̂𝑥 ± 𝑖𝐿̂𝑦
Which satisfy the commutation relations;
[𝐿+ , 𝐿− ] = 2ℏ𝑙𝑧 , [𝐿𝑍 , 𝐿± ] = ±ℏ𝐿± , [𝐿± , 𝑙2 ] = 0,
These relations are relatively easy to prove using the commutation relations we’ve already
mentioned:
[𝐿̂𝑥 , 𝐿̂𝑦 ] = 𝑖ℏ𝐿̂𝑧 , [𝐿̂𝑦 , 𝐿̂𝑧 ] = 𝑖ℏ𝐿̂𝑥 , [𝐿̂𝑧 , 𝐿̂𝑥 ] = 𝑖ℏ𝐿̂𝑦 , [𝐿̂2 , 𝐿̂𝑧 ] = 0
For example:
[𝐿̂𝑧 , 𝐿± ] = [𝐿̂𝑧 , 𝐿̂𝑥 ] ± 𝑖[𝐿̂𝑧 , 𝐿̂𝑦 ]
= 𝑖ℏ𝐿̂𝑦 ± 𝑖(−𝑖ℏ𝐿̂𝑥 ) = ±ℏ(𝐿̂𝑥 ± 𝑖𝐿̂𝑦 ) = ±ℏ𝐿±
The raising and lowering operators have a peculiar effect on the eigenvalue of 𝑙𝑧 :
𝛽 𝛽 𝛽 𝛽
𝐿̂𝑧 (𝐿± 𝑌𝛼 ) = ([𝐿̂𝑧 , 𝐿± ] + 𝐿± 𝐿̂𝑧 )𝑌𝛼 = ( ±ℏ𝐿± + 𝐿± 𝛽)𝑌𝛼 = (𝛽 ±ℏ )(𝐿± 𝑌𝛼 )
Thus, 𝐿+ (𝐿− ) raises (lowers) the eigenvalue of 𝐿̂𝑧 by ℏ, hence the names. Since the raising and
lowering operators commute with 𝐿 ̂2 they do not change the value of 𝛼 and so we can write
𝛽 𝛽±ℏ
𝐿± 𝑌𝛼 ∝ 𝑌𝛼
̂
and so the eigenvalues of 𝐿𝑧 are evenly spacedǃ
What are the limits on this ladder of eigenvalues? Recall that for the harmonic oscillator, we
found that there was a minimum eigenvalue and the eigenstates could be created by successive
applications of the raising operator to the lowest state. There is also a minimum eigenvalue in
this case. To see this, note that
〈𝐿̂2𝑥 + 𝐿̂2𝑦 〉 = 〈𝐿̂2𝑥 〉 + 〈𝐿̂2𝑦 〉 ≥ 0
This result simply reflects the fact that if you take any observable operator and square it, you
must get back a positive number. To get a negative value for the average value of 𝐿̂2𝑥 or 𝐿̂2𝑦
would imply an imaginary eigenvalue of 𝐿̂𝑥 or 𝐿̂𝑦 , which is impossible since these operators
are Hermitian. Besides, what would an imaginary angular momentum mean? We now apply
the above equation for the specific wavefunction 𝑌𝛼𝛽 :
𝛽∗ 2 2 𝛽 𝛽∗ 2 2 𝛽 𝛽∗ 𝛽
̂𝑥 + 𝐿
∫ 𝑌𝛼 (𝐿 ̂𝑦 ) 𝑌𝛼 = ∫ 𝑌𝛼 (𝐿
̂ −𝐿
̂𝑧 ) 𝑌𝛼 = ∫ 𝑌𝛼 (𝛼 − 𝛽 2 )𝑌𝛼 = 𝛼 − 𝛽 2
Hence 𝛽2 ≤α and therefore − 𝛼 ≤ 𝛽 ≤ 𝛼 . Which means that there are both maximum and
minimum values that 𝛽 can take on for a given 𝛼. If we denote these values by 𝛽𝑚𝑎𝑥 and 𝛽𝑚𝑖𝑛 ,
respectively, then it is clear that
𝛽 𝛽
𝐿+ 𝑌𝛼 𝑚𝑎𝑥 = 0, 𝐿− 𝑌𝛼 𝑚𝑖𝑛 = 0,
We can then use this knowledge and some algebra tricks trick to determine the relationship
between 𝛼 and 𝛽𝑚𝑎𝑥 (or 𝛽𝑚𝑖𝑛 ). First note that:
𝛽 𝛽
⇒ 𝐿+ 𝐿− 𝑌𝛼 𝑚𝑎𝑥 = 0 𝐿− 𝐿+ 𝑌𝛼 𝑚𝑖𝑛 = 0
We can expand this explicitly in terms of 𝐿̂𝑥 and 𝐿̂𝑦 :
𝛽
⇒ (𝐿̂2𝑥 + 𝐿̂2𝑦 − 𝑖(𝐿̂𝑦 𝐿̂𝑥 − 𝐿̂𝑥 𝐿̂𝑦 )) 𝑌𝛼 𝑚𝑎𝑥 = 0 (𝐿̂2𝑥 + 𝐿̂2𝑦 + 𝑖(𝐿̂𝑦 𝐿̂𝑥 − 𝐿̂𝑥 𝐿̂𝑦 )) 𝑌𝛽𝛼 𝑚𝑖𝑛 = 0
However, this is not the most convenient form for the operators, because we don’t know what
𝐿̂𝑥 or 𝐿̂𝑦 gives when acting on 𝑌𝛼𝛽 . However, we can rewrite the same expression in terms of
𝑙 2 and 𝐿̂𝑧 :
𝐿̂2𝑥 + 𝐿̂2𝑦 ± 𝑖 (𝐿
⏟ ̂ 𝑦 𝐿̂𝑥 − 𝐿̂𝑥 𝐿̂𝑦 )
⏟
𝐿̂2 −𝐿̂2𝑧 −𝑖ℏ𝐿̂𝑧
So then we have;
𝛽 𝛽
⇒ (𝐿̂2 − 𝐿̂2𝑧 − ℏ𝐿̂𝑧 )𝑌𝛼 𝑚𝑎𝑥 = 0 (𝐿̂2 − 𝐿̂2𝑧 + ℏ𝐿̂𝑧 )𝑌𝛼 𝑚𝑖𝑛 = 0
2 𝛽 2 𝛽
⇒ (𝛼 − 𝛽𝑚𝑎𝑥 − ℏ𝛽𝑚𝑎𝑥 )𝑌𝛼 𝑚𝑎𝑥 = 0 (𝛼 − 𝛽𝑚𝑖𝑛 + ℏ𝛽𝑚𝑖𝑛 )𝑌𝛼 𝑚𝑖𝑛 = 0
⇒ 𝛼 = 𝛽𝑚𝑎𝑥 (𝛽𝑚𝑎𝑥 + ℏ) = 𝛽𝑚𝑖𝑛 (𝛽𝑚𝑖𝑛 − ℏ)
𝛽𝑚𝑎𝑥 = −𝛽𝑚𝑖𝑛 = ℏ𝑙
where in the last line we have simply defined a new variable, 𝑙, that is dimensionless (notice
that ℏ has the units of angular momentum). So combining these minimum and maximum values
we have that −ℏ𝑙 ≤ 𝛽 ≤ ℏ𝑙 . Further, since we can get from the lowest to the highest
eigenvalue in increments of ℏ by successive applications of the raising operator, it is clear that
the difference between the highest and lowest values [ ℏ𝑗 − (−ℏ𝑗) = 2ℏ𝑙 ] must be an
integer multiple of ℏ . Thus, 𝑙 itself must either be an integer or a half-integer.
Putting all these facts together, we conclude (Define 𝑚 ≡ 𝛽 / ℏ ):
1 3
𝐿̂2 𝑌𝑚 2 𝑚
𝑙 = ℏ 𝑙(𝑙 + 1)𝑌𝑙 𝑙 = 0, , 1, , 2, ….
2 2
And
̂𝑧 𝑌𝑙𝑚
𝐿 = 𝑚ℏ𝑌𝑙𝑚 𝑚 = −𝑙, −𝑙 + 1, … , 𝑙 − 1, 𝑙
where we have replaced 𝛼 with 𝑙 and 𝛽 with 𝑚 so that 𝑌𝛼𝛽 becomes 𝑌𝑙𝑚 . Also, in the first
equation, we have noted that 0 ≤ 〈𝐿̂2 〉 = ℏ2 𝑙 (𝑙 + 1) implies 𝑙 ≥ 0 . These are the
fundamental eigenvalue equations for all forms of angular momentum.
At first, you might think this means we made a mistake in our derivation above and that 𝑙
should only be an integer and not a half integer. However, there is no error. The difference
arises because our derivation above is valid for any kind of angular momentum. Thus, while
certain values of 𝑙 may not appear for certain types of angular momentum, we will see later on
that they can appear for other types of angular momentum. Most notably, electrons have an
1
intrinsic spin angular momentum with 𝑙 = 2 . Thus, while individual systems may have
additional restrictions on the allowed values of 𝑙, angular momentum states always obey the
above eigenvalue relations.
Important remark:
The four angular momentum operators are related as
and they would if the factors were scalars, but they are operators which do not commute, so
this is not factoring. Just like the SHO, it is a good mnemonic, nevertheless.
[𝐿⃗̂⃗2 , 𝐿∓ ] = 0
[𝐿̂𝑧 , 𝐿∓ ] = ∓ℏ𝐿∓
⃗̂⃗2 , 𝐿+ ] = 0
Example: Show [𝐿
⃗̂⃗2 , 𝐿+ ] = [𝐿
[𝐿 ⃗̂⃗2 , 𝐿̂𝑥 + 𝑖𝐿̂𝑦 ] = [𝐿
⃗̂⃗2 , 𝐿̂𝑥 ] + 𝑖 [𝐿
⃗̂⃗2 , 𝐿̂𝑦 ] = 0 + 𝑖(0) = 0
Example: Given the spherical coordinate representations of 𝐿𝑥 and 𝐿𝑦 , show that, we can
write the ladder operator of orbital angular momentum as;
𝜕 𝜕
𝐿± = ±ℏ𝑒 ±𝑖𝜙 (𝜕𝜃 ± 𝑖𝑐𝑜𝑡𝜃 𝜕∅) Sherical coordinate,
𝐿± = 𝐿̂𝑥 + 𝑖𝐿̂𝑦 Cartesian coordinate,
𝜕 𝜕 𝜕 𝜕
𝐿̂𝑥 + 𝑖𝐿̂𝑦 = 𝑖ℏ (𝑠𝑖𝑛𝜙 + 𝑐𝑜𝑡𝜃𝑐𝑜𝑠𝜙 ) + 𝑖 [𝑖ℏ (−𝑐𝑜𝑠𝜙 + 𝑐𝑜𝑡𝜃𝑠𝑖𝑛𝜙 )]
𝜕𝜃 𝜕𝜙 𝜕𝜃 𝜕𝜙
𝜕 𝜕 𝜕 𝜕
= ℏ [𝑖𝑠𝑖𝑛𝜙 + 𝑖𝑐𝑜𝑠𝜙𝑐𝑜𝑡𝜃 + 𝑐𝑜𝑠𝜙 − 𝑠𝑖𝑛𝜙𝑐𝑜𝑡𝜃 ]
𝜕𝜃 𝜕𝜙 𝜕𝜃 𝜕𝜙
𝜕 𝜕
= ℏ [(𝑐𝑜𝑠𝜙 + 𝑖𝑠𝑖𝑛𝜙) 𝜕𝜃 + 𝑖(𝑐𝑜𝑠𝜙 − 𝑠𝑖𝑛𝜙)𝑐𝑜𝑡𝜃 𝜕𝜙]
𝜕 𝜕
= ℏ [(𝑐𝑜𝑠𝜙 + 𝑖𝑠𝑖𝑛𝜙) 𝜕𝜃 + 𝑖(𝑐𝑜𝑠𝜙 + 𝑖𝑠𝑖𝑛𝜙)𝑐𝑜𝑡𝜃 𝜕𝜙]
𝜕 𝜕
= ℏ [(𝑒 𝑖𝜙 ) 𝜕𝜃 + 𝑖(𝑒 𝑖𝜙 )𝑐𝑜𝑡𝜃 𝜕𝜙]
𝜕 𝜕
= ℏ𝑒 𝑖𝜙 [𝜕𝜃 + 𝑖𝑐𝑜𝑡𝜃 𝜕𝜙]
Figure: Geometrical representation of the angular momentum 𝐿⃗̂⃗ : the vector 𝐿⃗̂⃗ rotates along the surface
of a cone about its axis; the cone’s height is equal to 𝑚ℏ, the projection of 𝐿⃗̂⃗on the cone’s axis. The tip
of 𝐿⃗̂⃗ lies, within the 𝐿̂ 𝐿 plane, on a circle of radius ℏ√𝑙(𝑙 + 1).
𝑧 𝑥𝑦
In classical terms, we can think of 𝐿⃗̂⃗ as representable graphically by a vector, whose endpoint
lies on a circle of radius ℏ√𝑙(𝑙 + 1), rotating along the surface of a cone of half-angle
𝑚
𝜃 = 𝑐𝑜𝑠 −1 ( )
√𝑙(𝑙 + 1)
such that its projection along the z-axis is always 𝑚ℏ. Notice that, as the values of the
quantum number 𝑚 are limited to 𝑚 = −𝑙, −𝑙 + 1, … , 𝑙 − 1, 𝑙 , the angle 𝜃 is quantized; the
only possible values of 𝜃 consist of a discrete set of 2𝑙 + 1 values:
−𝑙 −𝑙 + 1 𝑙−1 𝑙
𝜃 = 𝑐𝑜𝑠 −1 ( ) , 𝑐𝑜𝑠 −1 ( ) , … , 𝑐𝑜𝑠 −1 ( ) , 𝑐𝑜𝑠 −1 ( ).
√𝑙(𝑙 + 1) √𝑙(𝑙 + 1) √𝑙(𝑙 + 1) √𝑙(𝑙 + 1)
Since all orientations of 𝐿⃗̂⃗ on the surface of the cone are equally likely, the projection of 𝐿⃗̂⃗ on
both the x and y axes average out to zero:
〈𝐿̂𝑥 〉 = 〈𝐿̂𝑦 〉 = 0
where 〈𝐿̂𝑥 〉 stands for 〈𝑙, 𝑚|𝐿̂𝑥 |𝑙, 𝑚〉.
As an example, Figure below shows the graphical representation for the 𝑙 = 2 case. As
specified in the equation 𝜃, 𝜃 takes only a discrete set of values. In this case where 𝑙 = 2 the
angle 𝜃 takes only five values corresponding respectively to 𝑚 = −2, −1,0,1,2 ; they are given
by;
° ° ° ° °
𝜃 = −35.26 , −65.91 , 90 , 35.26 , 65.91
Figure: Graphical representation of the angular momentum 𝑙 = 2 for the state |𝑠, 𝑚⟩ with 𝑚 =
−2, −1,0,1,2. The radius of the circle is ℏ√2(2 + 1) = √6ℏ.
∑ |𝑠, 𝑚𝑠 ⟩ = 𝐼
𝑚𝑠 =−𝑠
where I is the unit matrix.
Each elementary particle has a fixed magnitude of the spin vector, given by the quantum
number 𝑠. However, the projection of the spin onto one axis, typically chosen to be the z-axis,
is needed in addition to the coordinates (or momenta) to fully specify the state of the particle.
A complete description of spin requires relativistic Quantum Mechanics.
1
Figure: Graphical representation of spin 2 the tip of 𝑆̂ lies on a circle of radius |𝑆̂| = √3ℏ⁄2 so that its
ℏ
projection on the z-axis takes only two values, ± , with 𝜃 = 54.73° .
2
1
Let us now study the matrix representation of the spin 𝑠 = 2. Using above we can represent
the operators 𝑆̂ 2 and 𝑆̂𝑧 within the {|𝑠, 𝑚𝑠 ⟩}basis by the following matrices:
1 1 1 1 1 1 1 1
⟨2 , 2 |𝑆̂ 2 | 2 , 2⟩ ⟨2 , 2 |𝑆̂ 2 | 2 , − 2⟩ 3ℏ2 1 0
𝑆̂ 2 = ( )= ( )
1 1 1 1 1 1 1 1 4 0 1 (A)
⟨2 , − 2 |𝑆̂ 2 | 2 , 2⟩ ⟨2 , − 2 |𝑆̂ 2 | 2 , − 2⟩
0 ℏ 1
𝑆̂𝑧 =
( ) (B)
2 0 −1
The matrices 𝑆̂+ and 𝑆̂− of can be inferred from (A)
0 1 0 0
𝑆̂+ = ℏ ( ), 𝑆̂− = ℏ ( ), (C)
0 0 1 0
ℏ ℏ
and since 𝑆̂𝑥 = 2 (𝑆̂+ + 𝑆̂− ) and 𝑆̂𝑦 = 2 (𝑆̂+ − 𝑆̂− ), we have
ℏ 0 1 ℏ 0 −𝑖
𝑆̂𝑥 = 2 ( ), 𝑆̂𝑦 = 2 ( ), (D)
1 0 𝑖 0
The joint eigenvectors of 𝑆̂ 2 and 𝑆̂𝑧 are expressed in terms of two-element column matrices,
known as spinors:
1 1 1 1 1 0
|2 , 2⟩ = ( ), |2 , − 2⟩ = ( ), (E)
0 1
It is easy to verify that these eigenvectors form a basis that is complete,
1
2
1 1 1 1 1 0
∑ | , 𝑚𝑠 ⟩ ⟨ , 𝑚𝑠 | = ( ) (0 1) + ( ) ( 1 0) = ( ), (G)
2 2 0 0 0 1
1
𝑚𝑠 =
2
And orthonormal
1 1 1 1 1
⟨2 , 2 | 2 , 2⟩ = ( ) (0 1) = 1 (H)
0
1 1 1 1 0
⟨ 2 , − 2 | 2 , − 2 ⟩ = ( 0 1) ( ) = 1 (I)
1
1 1 1 1 1 1 1 1
⟨2 , 2 | 2 , − 2⟩ = ⟨2 , − 2 | 2 , 2⟩ = 0 (J)
Let us now find the eigenvectors of 𝑆̂𝑥 and 𝑆̂𝑦 . First, note that the basis vectors |𝑠, 𝑚𝑠 ⟩ are
eigenvectors of neither 𝑆̂𝑥 nor 𝑆̂𝑦 ; their eigenvectors can, however, be expressed in terms of
|𝑠, 𝑚𝑠 ⟩ as follows:
1 1 1 1 1
|𝜓𝑥 ⟩± = [| , ⟩ ± | , − ⟩] = 0 (K)
√2 2 2 2 2
1 1 1 1 1
|𝜓𝑦 ⟩± = [| , ⟩ ± 𝑖 | , − ⟩] = 0 (L)
√2 2 2 2 2
̂ ̂
The eigenvalue equations for 𝑆𝑥 and 𝑆𝑦 are thus given by
ℏ ℏ
𝑆̂𝑥 |𝜓𝑦 ⟩± = 2 |𝜓𝑥 ⟩± , 𝑆̂𝑦 |𝜓𝑦 ⟩± = 2 |𝜓𝑦 ⟩± , (M)
D. Spin operators
By analogy with orbital angular momentum operators, spin operators 𝑆̂𝑥 , 𝑆̂𝑦 , and 𝑆̂𝑧 can be
defined with analogous commutation relations. More commonly, the operators
𝜎̂𝑥 =2𝑆̂𝑥 ⁄ℏ, 𝜎̂𝑦 =2𝑆̂𝑦 ⁄ℏ, 𝜎̂𝑧 =2𝑆̂𝑧 ⁄ℏ,
are used, with commutation relations
[𝜎̂𝑥 , 𝜎̂𝑦 ] = 2𝑖𝜎̂𝑧
[𝜎̂𝑦 , 𝜎̂𝑧 ] = 2𝑖𝜎̂𝑥
[𝜎̂𝑧 , 𝜎̂𝑥 ] = 2𝑖𝜎̂𝑦
These operators can be written as the Pauli spin matrices
0 1 0 −𝑖 1 0
𝜎̂𝑥 = [ ], 𝜎̂𝑦 = [ ], 𝜎̂𝑧 = [ ],
1 0 𝑖 0 0 −1
The vector spin operator
0 1 0 −𝑖 1 0
𝜎̂ = 𝑖𝜎̂𝑥 + 𝑗𝜎̂𝑦 +𝑘𝜎̂𝑧 = 𝑖 [ ]+𝑗[ ]+𝑘[ ]
1 0 𝑖 0 0 −1
Spinor
A spinor is a vector in the direct product space of the spatial (or space and time) and spin basis
functions, and corresponds to a vector with a possibly different spatial (or space and time)
function for each spin direction, i.e.,
𝜓↑ (𝑟, 𝑡) 1 0
|Ψ⟩ = [ ] = 𝜓↑ (𝑟, 𝑡) [ ] + 𝜓↓ (𝑟, 𝑡) [ ]
(𝑟,
𝜓↓ 𝑡) 0 1
A spinor can represent any possible state of a single electron, including spin.