🔬Condensed Matter Physics Unit 2 Review

The Fermi surface is the boundary in momentum space (or equivalently, reciprocal space) that separates occupied from unoccupied electron states at absolute zero. It governs how electrons respond to external perturbations, making it central to understanding conductivity, optical response, and magnetism in solids.

Fermi surface shapes range from nearly perfect spheres in simple metals to wildly complex, multi-sheeted structures in real crystals. These geometric details directly control measurable material properties.

Definition and significance

At $T = 0$ , electrons fill up states from the lowest energy upward until all $N$ electrons are accommodated. The surface in $\mathbf{k}$ -space where the energy equals the Fermi energy, $E(\mathbf{k}) = E_F$ , is the Fermi surface.

Why does this matter so much? Only electrons near the Fermi surface participate in low-energy processes like electrical conduction, heat transport, and magnetic response. Electrons deep inside the Fermi sea are "frozen out" by the Pauli exclusion principle. So the shape and topology of this surface control nearly all the interesting electronic behavior of a material.

The Fermi surface also determines how a material responds to external electric and magnetic fields, which is why mapping it experimentally is such a priority.

Fermi energy vs Fermi level

These two terms are related but not identical:

Fermi energy ( $E_F$ ): the energy of the highest occupied state at $T = 0$ . It's a ground-state property and stays fixed for a given electron density.
Fermi level ( $\mu$ ): the chemical potential of the electron system at finite temperature. At $T = 0$ , $\mu = E_F$ exactly. As temperature increases, $\mu$ shifts slightly because the Fermi-Dirac distribution smears out.

For metals at typical temperatures, the shift is tiny (on the order of $k_B T / E_F$ , which is $\sim 10^{-4}$ at room temperature). But in semiconductors with small carrier densities, the distinction between $E_F$ and $\mu(T)$ becomes significant and affects carrier statistics.

Brillouin zone relationship

The Fermi surface is conventionally drawn within the first Brillouin zone, the Wigner-Seitz cell of the reciprocal lattice. This is natural because the crystal's periodicity means $E(\mathbf{k})$ is periodic in reciprocal space, and the first Brillouin zone captures one full period.

Several important connections:

The Fermi surface inherits the point-group symmetry of the crystal lattice
At Brillouin zone boundaries, Bragg reflection opens energy gaps, which distort the Fermi surface away from the free-electron sphere
When the Fermi surface approaches or touches a zone boundary, new phenomena appear (neck orbits in noble metals, for example)
Techniques like ARPES directly map the Fermi surface within the Brillouin zone, making this relationship experimentally concrete

Fermi surface characteristics

Shape and topology

For a free electron gas, the Fermi surface is a perfect sphere with radius $k_F = (3\pi^2 n)^{1/3}$ , where $n$ is the electron density. Real materials deviate from this because the periodic lattice potential distorts the energy bands.

In practice, Fermi surfaces can be:

Nearly spherical with small distortions (alkali metals like Na, K)
Multi-sheeted, with several distinct surfaces from different bands crossing $E_F$
Open, extending across the entire Brillouin zone in certain directions (as in Cu, where the Fermi surface has "necks" touching the zone boundary along $\langle 111 \rangle$ )
Disconnected, with isolated pockets of electrons or holes

The topology of the Fermi surface (whether it's closed, open, multiply connected) has direct consequences for magnetoresistance, quantum oscillations, and transport anisotropy.

Electron vs hole surfaces

Whether a Fermi surface sheet is called an "electron surface" or a "hole surface" depends on which states it encloses:

Electron surfaces enclose occupied states. They tend to be convex and sit near band minima. Carriers on these sheets behave like negatively charged particles.
Hole surfaces enclose empty states. They tend to be concave and sit near band maxima. Carriers behave as if they have positive charge.

When both electron and hole pockets coexist at the Fermi level, the material is a semimetal (bismuth is a classic example). This coexistence produces compensated transport, where electron and hole contributions partially cancel in Hall effect measurements, and leads to large, non-saturating magnetoresistance.

Fermi surface nesting

Nesting occurs when large, parallel segments of the Fermi surface can be connected by a single wavevector $\mathbf{Q}$ . Mathematically, this means a translation $\mathbf{k} \to \mathbf{k} + \mathbf{Q}$ maps one portion of the Fermi surface onto another.

Why nesting matters:

It produces a peak in the electronic susceptibility $\chi(\mathbf{Q})$ at the nesting vector
This enhanced susceptibility can drive instabilities like charge density waves (CDWs) or spin density waves (SDWs)
Strong nesting also enhances electron-phonon coupling at $\mathbf{Q}$ , potentially softening phonon modes and triggering lattice distortions (Kohn anomalies)
In iron-based superconductors, nesting between electron and hole pockets is thought to mediate the pairing interaction

Nesting can be tuned by doping or applying pressure, which shifts the Fermi level and changes the Fermi surface geometry.

Experimental techniques

de Haas-van Alphen effect

When a metal is placed in a strong magnetic field, electrons on the Fermi surface follow cyclotron orbits. As the field changes, Landau levels pass through the Fermi energy one by one, causing oscillations in the magnetization (and magnetic susceptibility).

The key result is the Onsager relation:

$F = \frac{\hbar}{2\pi e} A_{\text{ext}}$

where $F$ is the oscillation frequency (in units of inverse field, $1/B$ ) and $A_{\text{ext}}$ is the extremal cross-sectional area of the Fermi surface perpendicular to the applied field.

By rotating the sample and measuring how $F$ changes with field direction, you can reconstruct the full 3D Fermi surface geometry. The temperature dependence of the oscillation amplitude gives the effective mass through the Lifshitz-Kosevich formula, and the field dependence gives the Dingle temperature (related to scattering rate).

Requirements: low temperatures (to sharpen the Fermi-Dirac distribution) and high magnetic fields (so that $\omega_c \tau \gg 1$ , where $\omega_c$ is the cyclotron frequency).

Angle-resolved photoemission spectroscopy

ARPES uses the photoelectric effect to directly measure the band structure $E(\mathbf{k})$ . A photon ejects an electron from the sample, and by measuring the kinetic energy and emission angle of the photoelectron, you recover both the binding energy and the in-plane crystal momentum.

The Fermi surface appears as the locus of points where the spectral intensity crosses $E_F$
Modern ARPES achieves energy resolution below 1 meV and momentum resolution of $\sim 0.01 \, \text{Å}^{-1}$
Spin-resolved ARPES adds spin detection, which is essential for studying topological surface states with spin-momentum locking
The main limitation is surface sensitivity (probing depth of only a few angstroms), so it primarily measures surface electronic structure

ARPES has been indispensable for mapping the Fermi surface of cuprate superconductors, topological insulators, and other quantum materials.

Positron annihilation

A complementary bulk probe. When a positron enters a solid, it thermalizes and eventually annihilates with an electron, emitting two gamma rays. The slight deviation from perfect anti-collinearity of these gamma rays reflects the momentum of the annihilating electron.

2D-ACAR (two-dimensional angular correlation of annihilation radiation) maps the electron momentum density projected onto a plane
Discontinuities in this momentum density correspond to the Fermi surface
Unlike ARPES, positron annihilation probes the bulk, making it useful for studying buried interfaces and materials where surface effects are problematic
It's less widely used than dHvA or ARPES but provides unique information, especially in disordered or defective systems

Theoretical models

Nearly free electron model

This model starts from free electrons (plane waves) and treats the periodic lattice potential as a weak perturbation. It works well when the electron-ion interaction is screened, as in alkali metals (Na, K) and noble metals (Cu, Ag, Au).

The free-electron Fermi surface is a sphere of radius $k_F$ . When you turn on the lattice potential:

Energy gaps open at Brillouin zone boundaries (where Bragg diffraction is strong)
Near these boundaries, the Fermi surface distorts, bulging toward or away from the zone face
If the sphere is large enough to reach the zone boundary, pieces of the Fermi surface get "cut" and appear in higher zones

The resulting Fermi surface can be understood by starting with the free-electron sphere and folding it back into the first Brillouin zone using the Harrison construction. This gives a surprisingly good first approximation for many real metals.

Tight-binding approximation

The opposite starting point: electrons are assumed to be localized on atomic sites, with small overlap integrals allowing hopping between neighbors. The bandwidth is set by the hopping parameter $t$ .

For a simple cubic lattice with nearest-neighbor hopping, the dispersion is:

$E(\mathbf{k}) = \epsilon_0 - 2t(\cos k_x a + \cos k_y a + \cos k_z a)$

This model is well-suited for:

Transition metals, where narrow $d$ -bands dominate the Fermi surface
Strongly correlated systems, where on-site interactions (Hubbard $U$ ) compete with hopping
Materials with multiple relevant orbitals, since tight-binding naturally accommodates multi-orbital physics and spin-orbit coupling

The Fermi surfaces from tight-binding can be highly non-spherical, with features like saddle points and van Hove singularities that strongly affect the density of states.

Density functional theory

DFT is the workhorse first-principles method for computing Fermi surfaces in real materials. Instead of solving the full many-body Schrödinger equation, it maps the problem onto an effective single-particle system where the total energy is a functional of the electron density $n(\mathbf{r})$ .

The Kohn-Sham equations are solved self-consistently to obtain band structures
Exchange and correlation effects are approximated (LDA, GGA, hybrid functionals, etc.)
DFT handles complex crystal structures and multi-element compounds without adjustable parameters
Predicted Fermi surfaces agree well with experiment for most conventional metals and many correlated systems

DFT does struggle with strongly correlated materials (where $U/t$ is large), and extensions like DFT+U or dynamical mean-field theory (DMFT) are needed in those cases.

Fermi surface in materials

Metals vs semiconductors

In a metal, at least one band is partially filled, so the Fermi level cuts through it and a Fermi surface exists. The Fermi surface can be simple (single sheet in alkali metals) or complex (multiple sheets in transition metals like Fe or Mo).

In a semiconductor at $T = 0$ , the valence band is completely full and the conduction band is completely empty. There is no Fermi surface in the usual sense. However, when you dope a semiconductor, small Fermi pockets appear near the band edges. These pockets are typically ellipsoidal, reflecting the effective mass anisotropy (as in Si, where the conduction band minima are ellipsoids along $\langle 100 \rangle$ ).

Semimetals (Bi, WTe $_2$ ) sit in between: they have small, compensated electron and hole pockets, leading to unusual magnetotransport like extremely large magnetoresistance.

Superconductors and Fermi surface

Superconductivity is fundamentally a Fermi surface phenomenon. In BCS theory, electrons with opposite momenta near the Fermi surface form Cooper pairs via an attractive interaction (typically phonon-mediated).

The superconducting gap $\Delta(\mathbf{k})$ opens on the Fermi surface, and its symmetry reflects the pairing mechanism:

s-wave (conventional superconductors): isotropic gap around the entire Fermi surface
d-wave (cuprates): gap has nodes along specific directions, vanishing along the zone diagonals

Fermi surface nesting plays a role in several unconventional superconductors. In iron pnictides, the nesting vector connecting electron and hole pockets enhances spin fluctuations that may mediate pairing. In cuprates, the pseudogap partially destroys the Fermi surface above $T_c$ , leaving disconnected "Fermi arcs" that remain one of the field's biggest puzzles.

Topological materials

Topological materials have Fermi surface features protected by symmetry and band topology:

Dirac semimetals (Cd $_3$ As $_2$ , Na $_3$ Bi): the Fermi "surface" shrinks to discrete points where bands cross linearly in all directions, producing massless Dirac fermions
Weyl semimetals (TaAs): similar point-like crossings, but with broken inversion or time-reversal symmetry, leading to pairs of Weyl nodes connected by surface Fermi arcs
Topological insulators (Bi $_2$ Se $_3$ ): insulating bulk, but metallic surface states with a single Dirac cone and spin-momentum locking at the Fermi level
Nodal line semimetals: the band crossing forms a continuous line in $\mathbf{k}$ -space rather than discrete points

These topological Fermi surface features produce measurable signatures like the anomalous Hall effect, chiral anomaly-driven negative magnetoresistance, and quantized surface transport.

Quantum oscillations

Quantum oscillations are periodic variations in physical properties as a function of $1/B$ (inverse magnetic field). They arise because the quantized cyclotron orbits (Landau levels) sequentially pass through the Fermi energy as $B$ changes.

Shubnikov-de Haas effect

This is the resistivity counterpart of the de Haas-van Alphen effect. As Landau levels cross $E_F$ , the density of states at the Fermi level oscillates, causing oscillations in $\rho_{xx}$ .

The oscillation frequency again satisfies the Onsager relation ( $F = \hbar A_{\text{ext}} / 2\pi e$ ), so it gives the same Fermi surface cross-sectional areas as dHvA. The SdH effect is especially useful for:

2D electron systems (quantum wells, heterostructures), where only a single extremal orbit exists
Extracting the Berry phase: a non-trivial Berry phase of $\pi$ shifts the oscillation pattern by a half-period, serving as a diagnostic for topological bands
Measuring quantum scattering times, which can differ from transport scattering times

Quantum Hall effect

In a 2D electron gas under strong perpendicular magnetic field, the energy spectrum collapses into discrete Landau levels separated by $\hbar \omega_c$ . When the Fermi level sits between Landau levels, the Hall conductance is quantized:

$\sigma_{xy} = \nu \frac{e^2}{h}$

where $\nu$ is an integer (integer quantum Hall effect) or a rational fraction (fractional quantum Hall effect).

The integer QHE is a single-particle topological effect, with each filled Landau level contributing one quantum of conductance. The fractional QHE requires electron-electron interactions and gives rise to exotic quasiparticles (composite fermions, anyons with fractional charge and statistics).

Berry phase and topology

When an electron completes a closed orbit on the Fermi surface in a magnetic field, it can accumulate a geometric (Berry) phase in addition to the usual dynamical phase.

For conventional parabolic bands, the Berry phase is $0$
For Dirac-like linear dispersions (as in graphene or topological surface states), the Berry phase is $\pi$
This $\pi$ phase shift is directly observable as an offset in the Landau level fan diagram from quantum oscillation data

Measuring the Berry phase through quantum oscillations has become a standard method for identifying topological character in new materials. It connects the geometry of the Fermi surface to the global topological properties of the band structure.

Applications and implications

Transport properties

The Fermi surface dictates electrical transport because only electrons near $E_F$ carry current. Specifically:

The Fermi velocity $\mathbf{v}_F = \nabla_\mathbf{k} E / \hbar$ at each point on the Fermi surface determines how fast carriers move
Anisotropic Fermi surfaces produce anisotropic conductivity (layered materials like graphite conduct much better in-plane than out-of-plane)
Open Fermi surface orbits in a magnetic field lead to non-saturating magnetoresistance along certain crystallographic directions
The density of states at $E_F$ controls the electronic specific heat coefficient $\gamma$ and the Pauli susceptibility
Thermoelectric performance depends on the energy derivative of the density of states near $E_F$ , which is shaped by Fermi surface features like van Hove singularities

Optical properties

Optical transitions involving electrons near the Fermi surface determine a material's interaction with light:

Intraband (Drude) transitions: free-carrier absorption at low frequencies, governed by the plasma frequency $\omega_p$ , which depends on the Fermi surface geometry
Interband transitions: absorption onset occurs when photon energy matches the gap between occupied and unoccupied bands near $E_F$ . In noble metals, the threshold for $d$ -to- $s$ transitions gives copper and gold their characteristic colors
Plasmon resonances in metallic nanostructures depend on carrier density and effective mass, both Fermi surface properties

Magnetic susceptibility

The magnetic response of conduction electrons has two competing contributions, both tied to the Fermi surface:

Pauli paramagnetism: proportional to the density of states at $E_F$ , $\chi_{\text{Pauli}} = \mu_B^2 g(E_F)$
Landau diamagnetism: arises from the orbital motion of electrons, typically $-1/3$ of the Pauli term for free electrons, but can be enhanced or reduced depending on band curvature

In materials with strong spin-orbit coupling, the Fermi surface spin texture (how the spin direction varies around the surface) affects magnetic anisotropy and is relevant for spintronics applications.

Advanced concepts

Fermi liquid theory

Landau's Fermi liquid theory explains why the independent-electron picture works so well for metals, even though electrons interact strongly via Coulomb repulsion.

The central idea: the low-energy excitations of the interacting system are quasiparticles that have a one-to-one correspondence with the non-interacting electron states. These quasiparticles:

Have the same charge and spin as bare electrons
Carry a renormalized effective mass $m^*$ that can be much larger than the band mass (heavy fermion systems like CeCoIn $_5$ have $m^*/m_e \sim 100$ )
Have a finite lifetime $\tau \propto 1/(E - E_F)^2$ that diverges at the Fermi surface, meaning quasiparticles are well-defined only near $E_F$
Interact through Landau parameters ( $F_0^s, F_0^a,$ etc.) that renormalize thermodynamic quantities

Fermi liquid theory predicts $C_V \propto T$ , $\rho \propto T^2$ , and a well-defined Fermi surface that satisfies Luttinger's theorem.

Non-Fermi liquids

In some systems, the quasiparticle picture breaks down entirely. The hallmarks of non-Fermi liquid behavior include:

Resistivity that scales as $T^\alpha$ with $\alpha \neq 2$ (the "strange metal" phase in cuprates shows $\rho \propto T$ over a wide range)
Broad, incoherent spectral functions in ARPES without sharp quasiparticle peaks
Anomalous power laws in thermodynamic quantities

Non-Fermi liquid behavior arises near quantum critical points, in systems with strong coupling to critical fluctuations, and in certain low-dimensional systems. Understanding these states is one of the major open problems in condensed matter physics.

Luttinger's theorem

Luttinger's theorem states that the volume enclosed by the Fermi surface is determined solely by the total electron density:

$V_{FS} = (2\pi)^3 \frac{n}{2}$

(the factor of 2 accounts for spin degeneracy). This holds even in the presence of arbitrarily strong electron-electron interactions, as long as the system remains a Fermi liquid.

This is a powerful constraint: interactions can reshape the Fermi surface and renormalize the effective mass, but they cannot change the enclosed volume. Apparent violations of Luttinger's theorem (as seen in underdoped cuprates and certain heavy fermion compounds) signal that something fundamentally different is happening, such as fractionalization of the electron or the emergence of a non-Fermi liquid ground state.