Spherical Harmonics

Spherical harmonics are the orthonormal angular functions on the unit sphere. They are the spherical analogue of Fourier modes: a square-integrable function of direction can be expanded in them, and rotationally invariant kernels can be decomposed with them.

This page is the mathematical reference for normalization, completeness, and phase convention. The physical angular-momentum interpretation is developed in Spherical Harmonics as Angular-Momentum States.

Sphere Coordinates and Measure

Write a direction on the unit sphere as

$$\begin{gathered} \Omega=(\theta,\phi),\\ 0\le\theta\le\pi, \qquad 0\le\phi<2\pi. \end{gathered}$$

The sphere measure is

$$d\Omega = \sin\theta\,d\theta\,d\phi.$$

The inner product for angular functions is

$$\langle f,g\rangle_{S^2} = \int_{S^2} f(\Omega)^*g(\Omega)\,d\Omega.$$

The factor $\sin\theta$ is part of the geometry. Omitting it changes the problem.

Convention and Definition

This page uses the Condon–Shortley phase convention. With the associated Legendre convention used on Associated Legendre Functions, for $m\ge0$,

$$Y_\ell^m(\theta,\phi) = N_{\ell m} P_\ell^m(\cos\theta) e^{im\phi},$$

where

$$N_{\ell m} = (-1)^m \sqrt{ \frac{2\ell+1}{4\pi} \frac{(\ell-m)!}{(\ell+m)!} }.$$

The allowed nonnegative labels are

$$\ell=0,1,2,\ldots, \qquad m=0,1,\ldots,\ell.$$

For negative $m$, define

$$Y_\ell^{-m} = (-1)^m \left(Y_\ell^m\right)^*, \qquad m>0.$$

The $m=0$ case reduces to

$$Y_\ell^0(\theta,\phi) = \sqrt{\frac{2\ell+1}{4\pi}}\, P_\ell(\cos\theta).$$

The sign convention matters. Some references absorb the Condon–Shortley phase into $P_\ell^m$ instead of writing it explicitly in $Y_\ell^m$.

Low-Order Harmonics

The first few functions fix signs and normalization more reliably than a verbal convention:

$$\begin{aligned} Y_0^0 &= \frac{1}{\sqrt{4\pi}},\\ Y_1^0 &= \sqrt{\frac{3}{4\pi}}\cos\theta,\\ Y_1^{\mathord\pm1} &= \mp \sqrt{\frac{3}{8\pi}} \sin\theta\,e^{\mathord\pm i\phi},\\ Y_2^0 &= \sqrt{\frac{5}{16\pi}} \left( 3\cos^2\theta-1 \right). \end{aligned}$$

These formulas use the convention stated above. A table that gives $Y_1^1$ with the opposite sign is using a different associated-Legendre or spherical-harmonic phase convention.

Plots often use radius proportional to $\lvert Y_\ell^m\rvert$ and color to show sign or phase. Such a surface is a visualization convention, not the three-dimensional probability density of a particle. A physical orbital also contains a radial factor.

Eigenfunctions of the Sphere Laplacian

The angular Laplacian on the unit sphere is

$$\Delta_{S^2} = \frac{1}{\sin\theta} \frac{\partial}{\partial\theta} \left( \sin\theta \frac{\partial}{\partial\theta} \right) + \frac{1}{\sin^2\theta} \frac{\partial^2}{\partial\phi^2}.$$

Spherical harmonics satisfy

$$\Delta_{S^2}Y_\ell^m = -\ell(\ell+1)Y_\ell^m.$$

In quantum mechanics,

$$L^2 = -\hbar^2\Delta_{S^2}, \qquad L_z = -i\hbar\frac{\partial}{\partial\phi},$$

so the same functions also obey

$$\begin{aligned} L^2Y_\ell^m &= \hbar^2\ell(\ell+1)Y_\ell^m, \\ L_zY_\ell^m &= \hbar mY_\ell^m. \end{aligned}$$

The mathematical statement is the Laplacian eigenvalue problem on the sphere. The physical statement is that these functions carry orbital angular momentum labels.

Origin of the Labels

Separating an angular function as

$$Y(\theta,\phi) = \Theta(\theta)\Phi(\phi)$$

first gives azimuthal modes

$$\Phi(\phi)=e^{im\phi}.$$

Single-valuedness under $\phi\mapsto\phi+2\pi$ requires $m\in\mathbb Z$. The polar equation becomes the associated Legendre equation after $x=\cos\theta$. Regularity at both poles selects

$$\ell=0,1,2,\ldots, \qquad \lvert m\rvert\le\ell.$$

For each $\ell$, there are $2\ell+1$ allowed values of $m$. The eigenvalue $\ell(\ell+1)$ is therefore $(2\ell+1)$-fold degenerate for the scalar Laplacian on the sphere.

Ladder Operators

In the angular representation,

$$L_\pm = \hbar e^{\mathord\pm i\phi} \left( \mathord\pm\frac{\partial}{\partial\theta} +i\cot\theta\frac{\partial}{\partial\phi} \right).$$

Their action is

$$\begin{aligned} L_\pm Y_\ell^m &= \hbar \sqrt{ \ell(\ell+1)-m(m\mathord\pm1) }\\ &\quad\times Y_\ell^{m\mathord\pm1}. \end{aligned}$$

The coefficient vanishes at $m=\pm\ell$, so a fixed-$\ell$ multiplet is finite. The positive square-root convention is tied to the Condon–Shortley phase. The algebraic derivation belongs in Angular Momentum Algebra.

Orthonormality

The normalized spherical harmonics obey

$$\int_{S^2} Y_\ell^m(\Omega)^* Y_{\ell'}^{m'}(\Omega)\,d\Omega = \delta_{\ell\ell'}\delta_{mm'}.$$

The orthogonality in $\phi$ comes from Fourier modes $e^{im\phi}$. The polar orthogonality comes from associated Legendre functions with the correct weight; the general weighted-polynomial viewpoint is summarized in Orthogonal Polynomials.

Completeness on the Sphere

For suitable square-integrable angular functions,

$$f(\Omega) = \sum_{\ell=0}^{\infty} \sum_{m=-\ell}^{\ell} c_{\ell m}Y_\ell^m(\Omega),$$

where

$$c_{\ell m} = \int_{S^2} Y_\ell^m(\Omega)^* f(\Omega)\,d\Omega.$$

The completeness relation can be written distributionally as

$$\sum_{\ell=0}^{\infty} \sum_{m=-\ell}^{\ell} Y_\ell^m(\Omega) Y_\ell^m(\Omega')^* = \delta(\Omega,\Omega').$$

This is the angular analogue of Fourier completeness. It should be interpreted under an integral, not as an ordinary pointwise equality.

The corresponding Parseval identity is

$$\int_{S^2} \lvert f(\Omega)\rvert^2\,d\Omega = \sum_{\ell=0}^{\infty} \sum_{m=-\ell}^{\ell} \lvert c_{\ell m}\rvert^2.$$

The $\ell=0$ coefficient extracts the spherical average:

$$\begin{aligned} c_{00} &= \frac{1}{\sqrt{4\pi}} \int_{S^2}f(\Omega)\,d\Omega,\\ \frac{1}{4\pi} \int_{S^2}f(\Omega)\,d\Omega &= \frac{c_{00}}{\sqrt{4\pi}}. \end{aligned}$$

In coordinates, the delta distribution on the sphere can be written

$$\delta(\Omega,\Omega') = \frac{ \delta(\theta-\theta') \delta_{2\pi}(\phi-\phi') }{ \sin\theta },$$

where $\delta_{2\pi}$ is the periodic delta in the azimuthal angle. The $1/\sin\theta$ factor compensates the integration measure. At a pole, an invariant or chart-based description avoids coordinate ambiguity.

Addition Theorem

For fixed $\ell$, summing over $m$ gives a rotationally invariant kernel:

$$\sum_{m=-\ell}^{\ell} Y_\ell^m(\Omega)^* Y_\ell^m(\Omega') = \frac{2\ell+1}{4\pi} P_\ell(\cos\gamma).$$

Here $\gamma$ is the angle between the two directions,

$$\begin{aligned} \cos\gamma &= \cos\theta\cos\theta' \\ &\quad+ \sin\theta\sin\theta'\cos(\phi-\phi'). \end{aligned}$$

The addition theorem explains why Legendre Polynomials appear in multipole expansions, central Green functions, and partial-wave scattering after angular variables are summed or averaged.

Coulomb Multipole Expansion

For $r\ne r'$, the addition theorem gives

$$\begin{aligned} \frac{1}{ \lvert\mathbf r-\mathbf r'\rvert } &= 4\pi \sum_{\ell=0}^{\infty} \frac{1}{2\ell+1} \frac{r_<^\ell}{r_>^{\ell+1}}\\ &\quad\times \sum_{m=-\ell}^{\ell} Y_\ell^m(\Omega) Y_\ell^m(\Omega')^*, \end{aligned}$$

where

$$r_<=\min(r,r'), \qquad r_>=\max(r,r').$$

The radial coefficient depends only on $\ell$, while the angular dependence is resolved into the complete multiplet. This separation is central to electrostatic multipoles, central-potential Green functions, and partial-wave methods.

Rotations

For fixed $\ell$, the span of

$$\left\lbrace Y_\ell^{-\ell}, \ldots, Y_\ell^\ell \right\rbrace$$

is invariant under rotations. A rotation mixes the $m$ components through a unitary $(2\ell+1)$-dimensional matrix. Those matrices are the irreducible rotation matrices described in Wigner D-Matrices.

The precise index order and complex conjugation depend on whether a source uses active rotations of functions or passive rotations of coordinates. The invariant statement is that rotations never mix different $\ell$ sectors for the sphere Laplacian.

Conjugation and Parity

With the convention on this page,

$$\left(Y_\ell^m\right)^* = (-1)^mY_\ell^{-m}.$$

Under spatial inversion, the coordinates transform as

$$\begin{gathered} \Omega\mapsto-\Omega,\\ (\theta,\phi) \mapsto (\pi-\theta,\phi+\pi), \end{gathered}$$

and

$$Y_\ell^m(-\Omega) = (-1)^\ell Y_\ell^m(\Omega).$$

Thus scalar central-potential angular wavefunctions with orbital label $\ell$ have parity $(-1)^\ell$.

Real Spherical Harmonics

Complex harmonics diagonalize $L_z$ and are usually the natural quantum basis. For real scalar fields and visualization, one can instead use real linear combinations. For $m>0$, one convention is

$$\begin{aligned} Y_{\ell m}^{(c)} &= \frac{ Y_\ell^{-m} +(-1)^mY_\ell^m }{\sqrt2},\\ Y_{\ell m}^{(s)} &= \frac{ Y_\ell^{-m} -(-1)^mY_\ell^m }{i\sqrt2}. \end{aligned}$$

Together with $Y_\ell^0$, these form a real orthonormal basis for the same fixed-$\ell$ subspace. Different communities attach different signs and names to the real combinations. A real harmonic with $m>0$ is not an eigenfunction of $L_z$, because it combines the $+m$ and $-m$ sectors.

Relation to Separation of Variables

For a central potential, the angular part of the stationary Schrödinger equation separates on the sphere. The angular functions are $Y_\ell^m$, while the radial dynamics is carried by $R(r)$ or the reduced radial function $u(r)=rR(r)$.

The full wavefunction often has the form

$$\psi(r,\theta,\phi) = R_{n\ell}(r)Y_\ell^m(\theta,\phi)$$

for a bound central-potential problem. The spherical harmonic is not hydrogen-specific; it is the angular basis dictated by rotation symmetry.

Numerical Practice

For angular quadrature, the substitution

$$x=\cos\theta$$

turns the measure into

$$d\Omega=d\phi\,dx.$$

This makes Gauss–Legendre nodes in $x$ and a uniform periodic grid in $\phi$ a natural product rule. For a function band-limited at $\ell_{\max}$, $N_\theta\ge\ell_{\max}+1$ Gauss–Legendre nodes and $N_\phi\ge2\ell_{\max}+1$ equally spaced azimuthal nodes are sufficient for many exact transform and orthogonality checks, provided the implementation uses consistent normalization and indexing.

Useful diagnostics are:

• reconstruct a known low-order harmonic from its coefficients;
• verify the orthonormality matrix numerically;
• check Parseval’s identity;
• rotate a fixed-$\ell$ coefficient vector and confirm norm preservation;
• increase both angular resolutions to expose aliasing.

Sampling uniformly in $\theta$ while omitting $\sin\theta$ overweights the poles and does not approximate the sphere measure.

Common Mistakes

• Forgetting the measure $d\Omega=\sin\theta\,d\theta\,d\phi$.
• Mixing phase conventions for $P_\ell^m$ and $Y_\ell^m$.
• Confusing the magnetic quantum number $m$ with particle mass.
• Treating $P_\ell(\cos\theta)$ as the whole spherical harmonic when $m\ne0$.
• Interpreting orbital plots as literal particle trajectories rather than angular probability-density visualizations.
• Using completeness as a pointwise identity instead of a distributional identity under an integral.
• Forgetting that spin states are not functions on the ordinary sphere; spin uses representation theory, not scalar spherical harmonics.
• Treating a real tesseral harmonic as an $L_z$ eigenfunction.
• Comparing rotated coefficients without checking active versus passive conventions.
• Sampling uniformly in $\theta$ without the geometric weight.

Cross-Links

• Associated Legendre Functions
• Spherical Coordinates
• Angular and Radial Separation
• Legendre Polynomials
• Orthogonal Polynomials
• Hypergeometric Functions
• Separation of Variables
• Angular Momentum Algebra
• Wigner D-Matrices
• Spherical Harmonics as Angular-Momentum States
• Symmetry Notation Conventions
• Hydrogen Atom

References

• NIST Digital Library of Mathematical Functions, Chapter 14, Legendre and Related Functions.
• 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.
• G. B. Arfken, H. J. Weber, and F. E. Harris, Mathematical Methods for Physicists, 7th ed., Academic Press, 2013.
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.

Exercises

1. Expand $f(\theta,\phi)=\cos\theta$ in spherical harmonics.

Solution

From the low-order formula,

$$Y_1^0(\theta,\phi) = \sqrt{\frac{3}{4\pi}}\cos\theta.$$

Therefore,

$$\cos\theta = \sqrt{\frac{4\pi}{3}} Y_1^0(\theta,\phi).$$

All other coefficients vanish. Direct projection gives

$$\begin{aligned} c_{10} &= \int_{S^2} Y_1^0(\Omega)^* \cos\theta\,d\Omega\\ &= \sqrt{\frac{4\pi}{3}}. \end{aligned}$$

2. Verify explicitly that

   $$Y_1^1 = -\sqrt{\frac{3}{8\pi}} \sin\theta\,e^{i\phi}$$

   is normalized.

Solution

Compute

$$\begin{aligned} \int_{S^2} \left\lvert Y_1^1 \right\rvert^2 d\Omega &= \frac{3}{8\pi} \int_0^{2\pi}d\phi \int_0^\pi \sin^3\theta\,d\theta\\ &= \frac{3}{8\pi} (2\pi) \left(\frac{4}{3}\right)\\ &= 1. \end{aligned}$$

The phase $e^{i\phi}$ has unit modulus, and the extra $\sin\theta$ comes from the sphere measure.

3. Use the addition theorem with $\Omega'=\Omega$ to evaluate $\sum_{m=-\ell}^{\ell}\lvert Y_\ell^m(\Omega)\rvert^2$.

Solution

When $\Omega'=\Omega$, the angle between the directions is $\gamma=0$, so

$$P_\ell(\cos\gamma) = P_\ell(1) =1.$$

The addition theorem gives

$$\sum_{m=-\ell}^{\ell} \left\lvert Y_\ell^m(\Omega) \right\rvert^2 = \frac{2\ell+1}{4\pi}.$$

The result is independent of direction, as rotational invariance requires. It also shows that the total density of a complete fixed-$\ell$ multiplet is isotropic.

4. Use the differential form of $L_+$ to verify

   $$L_+Y_1^0 = \sqrt2\,\hbar Y_1^1.$$

Solution

Because $Y_1^0$ has no $\phi$ dependence,

$$\begin{aligned} L_+Y_1^0 &= \hbar e^{i\phi} \frac{\partial}{\partial\theta} \left( \sqrt{\frac{3}{4\pi}}\cos\theta \right)\\ &= -\hbar \sqrt{\frac{3}{4\pi}} \sin\theta\,e^{i\phi}. \end{aligned}$$

Using

$$Y_1^1 = -\sqrt{\frac{3}{8\pi}} \sin\theta\,e^{i\phi},$$

the right side becomes

$$\sqrt2\,\hbar Y_1^1.$$

This agrees with the algebraic coefficient

$$\hbar \sqrt{ \ell(\ell+1)-m(m+1) } = \sqrt2\,\hbar$$

for $\ell=1$ and $m=0$.