Quasiparticles

Quasiparticles

A quasiparticle is a long-lived excitation of an interacting system that behaves approximately like a particle with renormalized properties. It is not a bare electron: the surrounding medium responds, and that dressing changes the excitation energy, spectral weight, velocity, and lifetime.

Energy shiftRe Σ
Lifetime−2 Im Σ
Coherent weightZ ≤ 1

From Dyson’s equation to a pole

For a single band, the retarded Green’s function can be written as

$$ G^R(\mathbf k,\omega)= \frac{1}{\omega+\mu-\epsilon_{\mathbf k}-\Sigma^R(\mathbf k,\omega)}. $$

The non-interacting pole lies at $\epsilon_{\mathbf k}-\mu$. Interactions replace it with the solution of the quasiparticle equation

$$ E_{\mathbf k}+\mu-\epsilon_{\mathbf k} -\operatorname{Re}\Sigma^R(\mathbf k,E_{\mathbf k})=0. $$

Because the self-energy depends on frequency, this equation is nonlinear and may have more than one solution. A recognizable quasiparticle requires one solution to carry appreciable spectral weight and have a width small compared with its characteristic energy scale.

Renormalization factor

Expand the real part of the self-energy near the quasiparticle energy. The coherent pole has residue

$$ Z_{\mathbf k}=\left[ 1-\left.\frac{\partial\operatorname{Re}\Sigma^R(\mathbf k,\omega)}{\partial\omega} \right|{\omega=E{\mathbf k}} \right]^{-1}. $$

$Z_{\mathbf k}$ measures the overlap between the bare-particle state and the dressed excitation. Interactions transfer the missing weight $1-Z_{\mathbf k}$ into incoherent backgrounds and satellite features. In a matrix problem, the derivative is evaluated in the relevant quasiparticle state rather than treated as a scalar.

Spectral function and lifetime

The quantity measured by charged-excitation spectroscopies is related to the spectral function

$$ A(\mathbf k,\omega)=-\frac{1}{\pi}\operatorname{Im}G^R(\mathbf k,\omega). $$

Near a well-isolated pole, the coherent contribution is approximately Lorentzian:

$$ A_{\mathrm{qp}}(\mathbf k,\omega)\approx \frac{Z_{\mathbf k}}{\pi} \frac{\Gamma_{\mathbf k}/2} {(\omega-E_{\mathbf k})^2+(\Gamma_{\mathbf k}/2)^2}, $$

with linewidth

$$ \Gamma_{\mathbf k}\approx -2Z_{\mathbf k}\operatorname{Im}\Sigma^R(\mathbf k,E_{\mathbf k}), \qquad \tau_{\mathbf k}\approx\frac{\hbar}{\Gamma_{\mathbf k}}. $$

Thus the two parts of the self-energy have complementary roles: $\operatorname{Re}\Sigma$ shifts the excitation, while $\operatorname{Im}\Sigma$ gives decay and broadening.

What a spectrum is telling you

FeatureInterpretation
Sharp peak with large $Z$Long-lived, particle-like excitation
Broad peakShort lifetime or strong scattering
Satellite peakCoupling to another excitation, such as a plasmon
Strong incoherent continuumSpectral weight not captured by a single quasiparticle

Photoemission probes electron removal; inverse photoemission probes electron addition. Neither experiment measures a Kohn–Sham eigenvalue directly. In Green’s-function methods, quasiparticle energies emerge from the interacting propagator and its self-energy.

When the picture breaks down

The quasiparticle approximation is controlled only when $\Gamma_{\mathbf k}$ is small and the self-energy varies smoothly near the pole. It becomes unreliable when peaks strongly overlap, $Z$ becomes very small, satellites compete with the main peak, or the Green’s function has no isolated pole. Strongly correlated metals, systems near quantum criticality, and fractionalized phases can require a description centered on the full spectral function rather than individual quasiparticles.

Connection to GW

The GW approximation constructs $\Sigma\approx iGW$ from a propagator and a dynamically screened interaction. Its most common use is to correct charged-excitation energies, but the same frequency-dependent self-energy also contains spectral weights, lifetimes, and satellites. A one-shot $G_0W_0$ calculation evaluates these corrections from a fixed reference; self-consistent GW updates the propagator and screening together.

Further reading