Abstract. The empty lattice approximation sets the periodic potential exactly to zero while retaining the full translational symmetry of the crystal. The result is remarkable: the free-electron parabola $E = \hbar^2 k^2/2m$ must be folded back into the first Brillouin zone under all reciprocal lattice translations $\mathbf{K}$, producing a rich spectrum of band crossings and degeneracies. These accidental degeneracies are then resolved by the point-group symmetry of each high-symmetry wavevector. We work through the simple cubic lattice in detail, decomposing the degenerate states at the $\Gamma$, $X$, and $R$ points into irreducible representations of $O_h$ and $D_{4h}$ using the character-projection method.

§1 Free Electrons in a Periodic Box

The simplest conceivable model of an electron in a crystal is one where the ionic potential is ignored entirely. Despite its apparent triviality, the empty lattice model is the essential zeroth-order starting point for nearly-free-electron (NFE) band theory: once we understand the symmetry-enforced structure of the empty-lattice bands, we can treat the periodic potential as a perturbation that opens gaps at the zone boundaries.

The energy eigenvalues for a Bloch wavevector $\mathbf{k}$ in the first Brillouin zone, labelled by a reciprocal lattice vector $\mathbf{K}$, are simply the kinetic energies of plane waves shifted by $\mathbf{K}$:

\begin{equation} E_{\mathbf{K}}(\mathbf{k}) = \frac{\hbar^2}{2m}\,|\mathbf{k} + \mathbf{K}|^2. \label{eq:el_1} \end{equation}

Each choice of $\mathbf{K}$ defines one "branch" of the band structure; different choices of $\mathbf{K}$ can yield the same energy at a given $\mathbf{k}$, creating the degeneracies we will systematically decompose. The band index $n$ in standard notation corresponds to ordering these branches by energy.

Why bother? Real metals like aluminium and sodium have band structures that look remarkably close to the empty-lattice bands, especially at energies far from zone boundaries. The empty lattice is not merely an academic exercise — it is the correct leading-order description of simple metals.

§2 The Simple Cubic Lattice and Its Reciprocal

We specialise to the simple cubic (SC) crystal structure with lattice constant $a$. Its primitive direct lattice vectors are:

\begin{equation} \mathbf{a}_1 = a\hat{x}, \quad \mathbf{a}_2 = a\hat{y}, \quad \mathbf{a}_3 = a\hat{z}. \label{eq:el_2} \end{equation}

The reciprocal lattice is itself simple cubic, with primitive vectors:

\begin{equation} \mathbf{b}_1 = \frac{2\pi}{a}\hat{k}_x, \quad \mathbf{b}_2 = \frac{2\pi}{a}\hat{k}_y, \quad \mathbf{b}_3 = \frac{2\pi}{a}\hat{k}_z, \label{eq:el_3} \end{equation}

satisfying $\mathbf{a}_i \cdot \mathbf{b}_j = 2\pi\delta_{ij}$. A general direct-lattice point and a general reciprocal lattice vector are, respectively:

\begin{equation} \mathbf{R} = l_1\mathbf{a}_1 + l_2\mathbf{a}_2 + l_3\mathbf{a}_3, \qquad \mathbf{K} = n_1\mathbf{b}_1 + n_2\mathbf{b}_2 + n_3\mathbf{b}_3, \label{eq:el_45} \end{equation}

where $l_i, n_i \in \mathbb{Z}$. A general Bloch wavevector within the first Brillouin zone is written as:

\begin{equation} \mathbf{k} = \mu_x\mathbf{b}_1 + \mu_y\mathbf{b}_2 + \mu_z\mathbf{b}_3, \label{eq:el_6} \end{equation}

where the reduced coordinates $\mu_i \in [-\tfrac{1}{2}, \tfrac{1}{2})$ are not required to be integers — they parameterise the continuous interior of the Brillouin zone. Substituting into Eq. \eqref{eq:el_1} and using $|\mathbf{k}+\mathbf{K}|^2 = (2\pi/a)^2[(n_1+\mu_x)^2+(n_2+\mu_y)^2+(n_3+\mu_z)^2]$, we arrive at:

\begin{equation} E(\mu_x,\mu_y,\mu_z;\,n_1,n_2,n_3) = \frac{\hbar^2}{2m}\!\left(\frac{2\pi}{a}\right)^{\!2} \Bigl[(n_1+\mu_x)^2+(n_2+\mu_y)^2+(n_3+\mu_z)^2\Bigr]. \label{eq:el_7} \end{equation}

For the purpose of classifying degeneracies by symmetry, it is convenient to work with the dimensionless reduced energy

\begin{equation} \tilde{E} \equiv \frac{2ma^2}{\hbar^2(2\pi)^2}\,E = (n_1+\mu_x)^2+(n_2+\mu_y)^2+(n_3+\mu_z)^2, \label{eq:reduced_E} \end{equation}

which we use throughout the symmetry analysis below. The overall energy scale $\hbar^2(2\pi)^2/2ma^2$ sets the width of the lowest band.

First Brillouin zone of the simple cubic lattice Isometric cube showing Γ, X, M, R high-symmetry points with dashed symmetry paths and a legend. Γ (0,0,0) X (½,0,0) X M (½,½,0) R (½,½,½) k_x k_z k_y Point Site sym. Γ O_h X D_4h M D_4h R O_h
Fig. 1: First Brillouin zone of the simple cubic lattice (truncated-octahedron for SC is actually a cube). High-symmetry points $\Gamma$, $X$, $M$, $R$ are marked with their point groups. Dashed lines indicate high-symmetry paths.

§3 Zone Folding and Degeneracy Structure

Zone folding is the process of taking the infinite free-electron parabola and mapping every state back into the first Brillouin zone by subtracting an appropriate reciprocal lattice vector $\mathbf{K}$. Wherever two or more branches — labelled by different $\mathbf{K}$ vectors — have the same energy at the same $\mathbf{k}$, we have a band crossing or degeneracy.

At a general point $\mathbf{k}$ in the zone, these crossings are accidental and will typically be lifted by any real potential. However, at high-symmetry points, crossings are enforced by the little group of $\mathbf{k}$ — the subgroup of the full point group $G$ that leaves $\mathbf{k}$ invariant. States belonging to different irreducible representations (irreps) of this little group cannot hybridise, and their crossing is a symmetry-protected degeneracy.

Key principle. The dimensionality of the irreps of the little group sets an upper bound on degeneracy. For the $O_h$ point group ($m\bar{3}m$), the maximum irrep dimension is 3 (the $T$ representations). Any band touching of order $> 3$ at a point with $O_h$ symmetry must split when a periodic potential is turned on.

§4 Symmetry Analysis: The Character-Projection Method

To decompose an $N$-fold degenerate level into irreps, we treat the $N$ degenerate plane-wave states $\{|\mathbf{k}+\mathbf{K}_i\rangle\}$ as a basis for a reducible representation $\Gamma_{\rm red}$ of the little group $G_{\mathbf{k}}$. The character of each symmetry operation $g \in G_{\mathbf{k}}$ in this representation is simply:

\begin{equation} \chi(g) = \#\{\mathbf{K}_i : g(\mathbf{k}+\mathbf{K}_i) = \mathbf{k}+\mathbf{K}_i\} = \text{number of states fixed by } g. \label{eq:character} \end{equation}

States that are merely permuted by $g$ contribute 0 to the character (their diagonal matrix element vanishes). The multiplicity $n_\alpha$ of each irrep $\Gamma^{(\alpha)}$ in the decomposition is then given by the Great Orthogonality Theorem:

\begin{equation} n_\alpha = \frac{1}{|G_{\mathbf{k}}|}\sum_{g \in G_{\mathbf{k}}} \chi(g)\,\chi^{(\alpha)}(g)^*, \label{eq:GOT} \end{equation}

where $\chi^{(\alpha)}(g)$ is the character of $g$ in the irrep $\Gamma^{(\alpha)}$, read directly from the character table, and $|G_{\mathbf{k}}|$ is the order of the little group.

Practical rule. To apply Eq. \eqref{eq:character} at a high-symmetry point: list all $N$ degenerate $\mathbf{K}$ vectors, apply each symmetry operation to every vector, count how many are left fixed. That count is $\chi(g)$. Then apply Eq. \eqref{eq:GOT} with the character table of the little group.

§5 The $\Gamma$-Point $(0,0,0)$ — Little Group $O_h$

At the zone centre $\boldsymbol{\mu} = (0,0,0)$, the little group is the full cubic point group $O_h$ ($m\bar{3}m$) of order 48. The reduced energy simplifies to:

\begin{equation} \tilde{E}_\Gamma = n_1^2 + n_2^2 + n_3^2. \label{eq:EGamma} \end{equation}

The lowest shells of reciprocal lattice points, ordered by $\tilde{E}_\Gamma$, are:

Fig 2: Interactive visualization of degenerate K-points for different $E_\Gamma$.
$\tilde{E}$ Degeneracy Irrep Decomposition
$0$ 1 $A_{1g}$
$1$ 6 $A_{1g}+E_g+T_{1u}$
$2$ 12 $A_{1g}+E_g+T_{2g}+T_{1u}+T_{2u}$
$3$ 8 $A_{1g}+A_{2u}+T_{1u}+T_{2g}$
Counting check. The sum of dimensions of all irreps must equal the degeneracy. For $\tilde{E}_\Gamma=2$ we get $1+2+3+3+3=12$ ✓. For $\tilde{E}_\Gamma=3$: $1+1+3+3=8$ ✓. This is a necessary consistency condition.

5.1 Significance of the $A_{1g}$ appearance at every level

Notice that $A_{1g}$ (the totally symmetric, one-dimensional irrep) appears in every shell. This is because the $\mathbf{K}=\mathbf{0}$ state — which always exists at the zone centre from the $\Gamma$ branch itself — always transforms as $A_{1g}$. The other irreps in each shell arise from the symmetrised combinations of the remaining degenerate states.

§6 The X-Point $(\tfrac{1}{2},0,0)$ — Little Group $D_{4h}$

Moving to the face-centre of the Brillouin zone along $\hat{k}_x$, the little group reduces from $O_h$ to $D_{4h}$ (order 16), because the $C_4$ axis along $\hat{k}_x$ is preserved but the three-fold $C_3$ axes are broken. The reduced energy becomes:

\begin{equation} \tilde{E}_X = \left(n_1+\tfrac{1}{2}\right)^2 + n_2^2 + n_3^2. \label{eq:EX} \end{equation}

The lowest degenerate shells are:

Fig 3: Interactive visualization of degenerate K-points for different $E_X$.
$\tilde{E}$ Degeneracy Irrep Decomposition
$0$ 1 $A_{1g}$
$\tfrac{1}{4}$ 2 $A_{1g}+A_{2u}$
$\tfrac{5}{4}$ 8 $A_{1g}+A_{2u}+B_{1g}+B_{2u}+E_g+E_u$
$\tfrac{9}{4}$ 10 $2A_{1g}+2A_{2u}+B_{1u}+B_{2g}+E_g+E_u$
$\tfrac{13}{4}$ 8 $A_{1g}+A_{2u}+B_{1g}+B_{2u}+E_g+E_u$

§7 The R-Point $(\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2})$ — Little Group $O_h$

The corner of the simple cubic Brillouin zone at $\boldsymbol{\mu} = (\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2})$ regains the full $O_h$ symmetry. Physically, this is because the R-point is equally distant from all six face centres and all eight neighbouring corners — the full octahedral symmetry is restored. The reduced energy is:

\begin{equation} \tilde{E}_R = \left(n_1+\tfrac{1}{2}\right)^2 + \left(n_2+\tfrac{1}{2}\right)^2 + \left(n_3+\tfrac{1}{2}\right)^2. \label{eq:ER} \end{equation}

7.1 Lowest level: $\tilde{E}_R = \tfrac{3}{4}$, 8-fold

R — Ẽ = 3/4 (8-fold)

The eight degenerate states correspond to all combinations $n_i \in \{0, -1\}$:

$(0,0,0),\;(-1,0,0),\;(0,-1,0),\;(0,0,-1),\;(-1,-1,0),\;(-1,0,-1),\;(0,-1,-1),\;(-1,-1,-1).$

These are exactly the eight corners of a cube in reciprocal space centred at $-(\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2})$. The $O_h$ symmetry permutes them among themselves.

Applying the character projection: under $E$ all 8 are fixed ($\chi=8$); under $C_3$ (body diagonal rotation) cycles of length 3 appear among the 6 non-diagonal states, leaving only 2 fixed; under $C_2$ (face diagonal) none are fixed... Evaluating the full set and projecting onto $O_h$ irreps gives:

Fig 4: Interactive visualization of degenerate K-points for \(E = 3/4\).
\begin{equation} \Gamma_{R,\,3/4} = A_{1g} + A_{2u} + T_{1u} + T_{2g}. \label{eq:GammaR1} \end{equation}

Dimensions check: $1+1+3+3 = 8$ ✓.

7.2 Second shell: $\tilde{E}_R = \tfrac{11}{4}$, 24-fold

The next shell at $\tilde{E}_R = \tfrac{11}{4}$ arises from all permutations of $(n_1,n_2,n_3)$ containing one entry equal to $+1$ and two entries in $\{0,-1\}$, plus permutations with one entry equal to $-2$. These 24 states form a larger orbit of $O_h$. The irrep decomposition is:

Fig 5: Interactive visualization of degenerate K-points for \(E = 11/4\).
\begin{equation} \Gamma_{R,\,11/4} = A_{1g} + A_{2u} + E_g + E_u + T_{1g} + 2T_{2g} + 2T_{1u} + T_{2u}. \label{eq:GammaR2} \end{equation}

Dimensions check: $1+1+2+2+3+6+6+3 = 24$ ✓.

7.3 Third shell: $\tilde{E}_R = \tfrac{19}{4}$, 24-fold

Remarkably, the third shell at $\tilde{E}_R = \tfrac{19}{4}$ is also 24-fold degenerate and shares the identical irrep decomposition as the second shell:

Fig 6: Interactive visualization of degenerate K-points for \(E = 19/4\).
\begin{equation} \Gamma_{R,\,19/4} = A_{1g} + A_{2u} + E_g + E_u + T_{1g} + 2T_{2g} + 2T_{1u} + T_{2u}. \label{eq:GammaR3} \end{equation}

This is not a coincidence — both 24-fold shells arise from the same orbit structure under $O_h$ (a set of 24 points related by the 24 proper rotations of the cube), and consequently they carry identical symmetry content. The energy difference $\tfrac{19}{4} - \tfrac{11}{4} = 2$ simply reflects the different radii of the two shells in reciprocal space.

7.4 Summary: High-Symmetry Point Analysis

Point Little Group $\tilde{E}$ Degeneracy Irrep Decomposition
$\Gamma$ $O_h$ $0$ 1 $A_{1g}$
$\Gamma$ $O_h$ $1$ 6 $A_{1g}+E_g+T_{1u}$
$\Gamma$ $O_h$ $2$ 12 $A_{1g}+E_g+T_{2g}+T_{1u}+T_{2u}$
$\Gamma$ $O_h$ $3$ 8 $A_{1g}+A_{2u}+T_{1u}+T_{2g}$
$X$ $D_{4h}$ $\tfrac{1}{4}$ 2 $A_{1g}+A_{2u}$
$X$ $D_{4h}$ $\tfrac{5}{4}$ 8 $A_{1g}+A_{2u}+B_{1g}+B_{2u}+E_g+E_u$
$X$ $D_{4h}$ $\tfrac{9}{4}$ 10 $2A_{1g}+2A_{2u}+B_{1u}+B_{2g}+E_g+E_u$
$X$ $D_{4h}$ $\tfrac{13}{4}$ 8 $A_{1g}+A_{2u}+B_{1g}+B_{2u}+E_g+E_u$
$R$ $O_h$ $\tfrac{3}{4}$ 8 $A_{1g}+A_{2u}+T_{1u}+T_{2g}$
$R$ $O_h$ $\tfrac{11}{4}$ 24 $A_{1g}+A_{2u}+E_g+E_u+T_{1g}+2T_{2g}+2T_{1u}+T_{2u}$
$R$ $O_h$ $\tfrac{19}{4}$ 24 $A_{1g}+A_{2u}+E_g+E_u+T_{1g}+2T_{2g}+2T_{1u}+T_{2u}$

§8 Physical Consequences and Beyond the Empty Lattice

The empty-lattice analysis establishes the symmetry skeleton of the full band structure. When a periodic ionic potential $V(\mathbf{r})$ is switched on:

Symmetry-protected crossings and topology. Some crossings enforced by nonsymmorphic symmetries (glide planes, screw axes) cannot be gapped even by an arbitrary potential. These protected band touchings are the crystalline analogues of topological semimetals. The simple cubic has only symmorphic symmetry, so all its empty-lattice crossings are generically gapped — but the FCC and diamond lattices, with nonsymmorphic elements, exhibit robust band touchings that persist in real materials like silicon.

8.1 Connection to Nearly-Free-Electron Theory

In the nearly-free-electron (NFE) limit, the energy correction to the empty-lattice bands from a weak periodic potential $V$ is calculated by degenerate perturbation theory within each degenerate subspace. The $2\times 2$ secular determinant for states connected by a single reciprocal lattice vector $\mathbf{K}$ gives:

\begin{equation} E_\pm = \bar{E} \pm \sqrt{\delta^2 + |V_{\mathbf{K}}|^2}, \label{eq:NFE} \end{equation}

where $\bar{E} = \tfrac{1}{2}(E_{\mathbf{K}}+E_{\mathbf{K}'})$ is the average energy of the two degenerate states and $\delta = \tfrac{1}{2}(E_{\mathbf{K}}-E_{\mathbf{K}'})$ is their energy difference (zero exactly at the zone boundary). The gap $\Delta E = 2|V_\mathbf{K}|$ opens precisely at the crossing point predicted by the empty-lattice analysis, affecting only states with the same $\mathbf{k}$ but different $\mathbf{K}$ labels.

This is why the empty-lattice band structure, combined with the symmetry decomposition into irreps, gives a complete map of where gaps open and how bands reconnect when a potential is added. The group-theory analysis tells us which states mix; the NFE perturbation theory tells us by how much.

Experimental relevance. Angle-resolved photoemission spectroscopy (ARPES) directly images the band structure of crystals. The empty-lattice bands are clearly visible as the skeleton around which real bands organise. For simple metals like Na, K, and Al, the nearly-free-electron picture works quantitatively well, with the Fourier components $V_{\mathbf{K}}$ extractable from the measured gaps.

References

  1. Tinkham, M. (2003). Group Theory and Quantum Mechanics. Dover Publications.
  2. Dresselhaus, M. S., Dresselhaus, G. & Jorio, A. (2008). Group Theory: Application to the Physics of Condensed Matter. Springer.
  3. Ashcroft, N. W. & Mermin, N. D. (1976). Solid State Physics. Holt, Rinehart & Winston. Ch. 9.
  4. Kittel, C. (2005). Introduction to Solid State Physics, 8th ed. Wiley. Ch. 7.
  5. Bradley, C. J. & Cracknell, A. P. (1972). The Mathematical Theory of Symmetry in Solids. Clarendon Press.