Case Study 2: Graphene — A Quantum Mechanics Playground

Overview

Graphene — a single layer of carbon atoms arranged in a honeycomb lattice — is the thinnest material ever isolated. It is one atom thick, roughly $3.4\,\text{\AA}$ from top to bottom. Yet within this atomic monolayer, an extraordinary range of quantum mechanical phenomena can be observed: massless Dirac fermions, a Berry phase of $\pi$, an anomalous quantum Hall effect, Klein tunneling, and the seeds of topological physics.

This case study examines graphene as a laboratory for quantum mechanics — a system where textbook quantum phenomena become real, measurable, and sometimes surprising.

Part 1: From Pencil Lead to Nobel Prize

The Isolation of Graphene

Graphite — the stuff in pencil lead — is a stack of weakly bonded graphene layers. Physicists had known since the 1940s (P.R. Wallace's 1947 paper) that a single graphene layer would have remarkable electronic properties. But for decades, the prevailing view was that a 2D crystal could not be thermodynamically stable — it would crumple, roll up, or decompose.

In 2004, Andre Geim and Konstantin Novoselov at the University of Manchester used a method of disarming simplicity: they pressed a piece of adhesive tape onto graphite, peeled it off, and repeated until only a few layers remained. They transferred these flakes to a silicon wafer with a carefully chosen oxide thickness (300 nm of SiO$_2$) that made even a single monolayer visible under an optical microscope through thin-film interference effects.

The key measurements came quickly:

• The flakes were indeed single-layer: AFM (atomic force microscopy) confirmed a thickness of $\sim 3.4\,\text{\AA}$.
• The charge carriers exhibited ambipolar transport: by sweeping a gate voltage, the carriers could be continuously tuned from electrons to holes.
• The mobility was extraordinarily high: $\mu > 10{,}000\,\text{cm}^2/(\text{V}\cdot\text{s})$ at room temperature, far exceeding silicon.
• The quantum Hall effect revealed the half-integer quantization predicted for Dirac fermions.

Geim and Novoselov received the 2010 Nobel Prize in Physics.

🔵 Historical Note: Geim had previously won the 2000 Ig Nobel Prize for levitating a frog with magnets. He is the only person to have received both an Ig Nobel and a Nobel Prize. The graphene discovery is a reminder that important physics can emerge from playful curiosity and simple experimental techniques.

Part 2: The Tight-Binding Band Structure — Dirac Fermions

Why a Honeycomb Lattice Gives Dirac Cones

As derived in Section 26.6, the honeycomb lattice has two atoms per unit cell. The tight-binding Hamiltonian for the $p_z$ orbitals is a $2 \times 2$ matrix:

$$H(\mathbf{k}) = \begin{pmatrix} 0 & -tf(\mathbf{k}) \\ -tf^*(\mathbf{k}) & 0 \end{pmatrix}$$

The vanishing of the diagonal elements reflects the sublattice symmetry — A and B atoms are identical (both carbon), so their on-site energies are equal. Setting them to zero fixes the energy reference.

The off-diagonal element $f(\mathbf{k}) = \sum_j e^{i\mathbf{k}\cdot\boldsymbol{\delta}_j}$ vanishes at exactly two points in the Brillouin zone — the K and K' points. At these points, $E_+ = E_- = 0$: the two bands touch. Near these points, the dispersion is linear:

$$E_\pm(\mathbf{q}) = \pm \hbar v_F |\mathbf{q}|$$

This is a Dirac cone — the 2D analogue of the energy-momentum relation for a massless relativistic particle.

The Effective Dirac Equation

The low-energy Hamiltonian near the K point can be written as:

$$H_K = \hbar v_F \begin{pmatrix} 0 & q_x - iq_y \\ q_x + iq_y & 0 \end{pmatrix} = \hbar v_F \,\boldsymbol{\sigma} \cdot \mathbf{q}$$

where $\boldsymbol{\sigma} = (\sigma_x, \sigma_y)$ are Pauli matrices acting on the sublattice (A/B) degree of freedom, not real spin.

This is the 2D massless Dirac equation with the speed of light $c$ replaced by $v_F \approx 10^6\,\text{m/s} \approx c/300$. The sublattice index plays the role of spin — it is called pseudospin. The pseudospin direction is locked to the momentum direction (helicity).

💡 Key Insight: Graphene electrons are not actually relativistic — they travel at $v_F \approx c/300$, well below the speed of light. But their effective low-energy Hamiltonian has the mathematical structure of the Dirac equation. This means that relativistic quantum phenomena (Klein tunneling, Zitterbewegung, pair creation analogues) can be observed in a benchtop condensed matter experiment rather than at a particle accelerator.

Near the K' Point

At the other Dirac point K', the effective Hamiltonian is:

$$H_{K'} = \hbar v_F \begin{pmatrix} 0 & -q_x - iq_y \\ -q_x + iq_y & 0 \end{pmatrix} = -\hbar v_F \,\boldsymbol{\sigma}^* \cdot \mathbf{q}$$

The K and K' valleys have opposite helicity — they are related by time reversal. This valley degeneracy is an additional quantum number (valley pseudospin) that has been proposed as the basis for "valleytronics" — information processing using valley polarization rather than charge or spin.

Part 3: The Berry Phase and Its Observable Consequences

Berry Phase of $\pi$

The eigenstates of the Dirac Hamiltonian near K are:

$$|\mathbf{q}, +\rangle = \frac{1}{\sqrt{2}} \begin{pmatrix} 1 \\ e^{i\theta_\mathbf{q}} \end{pmatrix}, \qquad |\mathbf{q}, -\rangle = \frac{1}{\sqrt{2}} \begin{pmatrix} 1 \\ -e^{i\theta_\mathbf{q}} \end{pmatrix}$$

where $\theta_\mathbf{q} = \arctan(q_y/q_x)$ is the angle of $\mathbf{q}$ in the Brillouin zone.

As $\mathbf{q}$ traces a closed loop around the K point, $\theta_\mathbf{q}$ advances by $2\pi$. The Berry phase (Chapter 32) accumulated by the upper band state is:

$$\gamma = \oint \mathcal{A} \cdot d\mathbf{q} = \pi$$

where $\mathcal{A} = i\langle \mathbf{q}, +| \nabla_\mathbf{q} | \mathbf{q}, +\rangle$ is the Berry connection.

This Berry phase of $\pi$ is directly measurable. It has three major consequences:

1. Anomalous Quantum Hall Effect

In the integer quantum Hall effect (Section 26.8), the Hall conductivity is quantized as $\sigma_{xy} = ie^2/h$ with $i = 0, \pm 1, \pm 2, \ldots$ for a normal 2D electron gas. In graphene, the Berry phase shifts the quantization to half-integer values:

$$\sigma_{xy} = \pm \frac{4e^2}{h}\left(n + \frac{1}{2}\right), \quad n = 0, 1, 2, \ldots$$

The factor of 4 comes from the valley (2) and spin (2) degeneracies. The crucial $+1/2$ offset is a direct consequence of the Berry phase $\pi$.

This was one of the first experimental confirmations of the Dirac fermion nature of graphene's carriers. Kim et al. and Novoselov et al. independently observed this anomalous quantization in 2005.

2. Suppressed Backscattering (Klein Tunneling)

The Berry phase of $\pi$ means that the wavefunction acquires a minus sign when the momentum is reversed ($\mathbf{q} \to -\mathbf{q}$). This leads to destructive interference for backscattering — an electron approaching a potential barrier head-on is perfectly transmitted, regardless of the barrier height.

This is the solid-state analogue of Klein tunneling in relativistic quantum mechanics: a Dirac fermion can tunnel through an arbitrary potential barrier with unit probability. It was directly observed in graphene p-n junctions.

🧪 Experiment: In 2006, Katsnelson, Novoselov, and Geim predicted that graphene p-n junctions would exhibit Klein tunneling. In 2009, A.F. Young and P. Kim at Columbia University created a gate-tunable p-n junction in graphene and measured the transmission probability as a function of angle. The results confirmed near-perfect transmission at normal incidence, consistent with Klein tunneling.

3. Absence of Anderson Localization

In ordinary 2D systems, any amount of disorder localizes all states (Anderson localization). Graphene is different: the Berry phase protects extended states at the Dirac point from localization. The interference pathways that normally lead to localization acquire an extra phase of $\pi$ from the pseudospin, converting constructive interference (localization) into destructive interference (delocalization). This is called weak anti-localization.

Part 4: Breaking the Dirac Cone — Gaps and Masses

Opening a Gap

The Dirac cones in graphene are gapless — the valence and conduction bands touch at the K and K' points. This makes graphene a zero-gap semiconductor (or semimetal), which is wonderful for fundamental physics but problematic for digital electronics (a transistor needs an off-state, which requires a gap).

Several methods have been used to open a gap in graphene, each corresponding to breaking a specific symmetry:

Sublattice symmetry breaking: If the A and B sublattices are inequivalent (different atoms or different environments), the diagonal elements of the Hamiltonian become $\pm \Delta$:

$$H = \begin{pmatrix} \Delta & -tf \\ -tf^* & -\Delta \end{pmatrix}$$

The eigenvalues become $E_\pm = \pm\sqrt{\Delta^2 + t^2|f|^2}$, and a gap of $2\Delta$ opens at the K points. This happens naturally in hexagonal boron nitride (hBN), where the two sublattice atoms are boron and nitrogen.

Quantum confinement (nanoribbons): Cutting graphene into narrow strips (width $W$) introduces quantization of the transverse momentum, opening a gap $E_g \propto \hbar v_F/W \sim 1/W$. A ribbon of width 10 nm has a gap of $\sim 0.1$ eV.

Bilayer graphene with electric field: In Bernal-stacked bilayer graphene, applying a perpendicular electric field breaks the interlayer symmetry and opens a tunable gap up to $\sim 0.3$ eV. This is the most promising route for graphene electronics.

The Mass Analogy

In the Dirac equation, a mass term opens a gap: $E = \pm\sqrt{m^2c^4 + p^2c^2}$. The graphene analogue is exact: the sublattice asymmetry $\Delta$ plays the role of a rest mass:

$$E_\pm = \pm\sqrt{\Delta^2 + (\hbar v_F q)^2}$$

Massless Dirac fermions ($\Delta = 0$): linear dispersion, no gap. Massive Dirac fermions ($\Delta > 0$): hyperbolic dispersion, gap of $2\Delta$.

This analogy is not merely pedagogical — it connects condensed matter physics to high-energy physics and has inspired deep theoretical cross-pollination.

Part 5: Graphene as a Platform for Exotic Quantum States

Twisted Bilayer Graphene and Flat Bands

In 2018, Pablo Jarillo-Herrero's group at MIT discovered that when two layers of graphene are stacked with a small twist angle of $\theta \approx 1.1°$ (the "magic angle"), the resulting moire superlattice has nearly flat bands. Flat bands mean:

• Huge density of states at the Fermi energy
• Electrons effectively localized (heavy effective mass)
• Electron-electron interactions dominate kinetic energy

The result: unconventional superconductivity at $T_c \approx 1.7$ K, appearing in a system with only two carbon atoms per microscopic unit cell. This "magic-angle twisted bilayer graphene" (MATBG) is now one of the most active research areas in condensed matter physics, because:

1. The superconductivity appears near a correlated insulating state, reminiscent of high-temperature cuprate superconductors.
2. The twist angle provides a continuously tunable parameter — no chemistry required.
3. The system is simple enough (two layers of carbon) that complete theoretical description should be achievable.

Fractional Quantum Hall Effect in Graphene

Graphene's high mobility and the four-fold valley/spin degeneracy produce a rich FQHE phenomenology. States at filling fractions $\nu = 1/3, 2/3, 4/3$, and more exotic fractions have been observed, providing a platform to study:

• Fractional charge excitations (anyons)
• Non-abelian statistics (at $\nu = 5/2$ in bilayer graphene)
• Topological quantum computing proposals

Graphene Quantum Dots

Confining graphene to nanometer-scale regions creates quantum dots with discrete energy levels. The relativistic dispersion means that the level spacing scales as $v_F/L$ rather than $\hbar^2/(mL^2)$, giving larger level spacings for the same size — advantageous for quantum information applications.

Discussion Questions

1. Graphene's carriers obey the same equation as massless relativistic particles, but they travel at $v_F \approx c/300$. In what sense is graphene "relativistic"? Is calling its carriers "Dirac fermions" a metaphor or a precise statement?

2. The Berry phase of $\pi$ in graphene leads to suppressed backscattering and enhanced conductivity. Yet in practice, graphene's conductivity is limited by disorder (charged impurities, ripples, phonons). How can the Berry phase protect against backscattering while disorder still limits conductivity? (Hint: the Berry phase forbids exact backscattering, but what about scattering at other angles?)

3. Magic-angle twisted bilayer graphene exhibits superconductivity at 1.7 K — incredibly low by any practical standard. Yet this discovery was hailed as a breakthrough. Why? What makes this superconductor scientifically more interesting than, say, aluminum (which superconducts at a comparable temperature)?

4. Graphene was isolated in 2004 and the Nobel Prize was awarded in 2010 — an unusually short interval. Yet graphene devices have not replaced silicon transistors. Why not? Is the scientific importance of graphene separate from its technological impact, and is that acceptable?

5. The connection between the graphene Dirac equation and the relativistic Dirac equation has enabled condensed matter physicists to observe analogues of relativistic phenomena (Klein tunneling, Zitterbewegung) that are inaccessible at particle accelerators. What are the limits of this analogy? What phenomena of the true Dirac equation have no analogue in graphene, and vice versa?

Connections

• Section 26.6: The tight-binding derivation of graphene's band structure is the starting point for this case study.
• Chapter 13 (Spin): The pseudospin in graphene has the same algebra as real spin-1/2, using Pauli matrices on the sublattice degree of freedom.
• Chapter 29 (Dirac Equation): The effective low-energy Hamiltonian of graphene is the 2D massless Dirac equation — the condensed matter realization of a relativistic quantum theory.
• Chapter 32 (Berry Phase): The Berry phase of $\pi$ in graphene is a concrete example of geometric phase with observable physical consequences.
• Chapter 36 (Topological Phases): Graphene is the prototype for topological band theory; the Chern number, valley Chern number, and $\mathbb{Z}_2$ invariant all have clear expressions in the graphene tight-binding model.