Complex Analysis in Physics: The Residue Theorem

The Magic of the Complex Plane

Many integrals over the real axis are elegantly computed by closing a contour in the complex plane. The idea: the closed integral along contour C of an analytic function equals 0 (Cauchy's theorem). But if there are singular points (poles) inside C, the integral equals the sum of residues multiplied by $2\pi i$. By choosing a clever contour, we “collect” the desired residue and obtain the real integral.

In physics, this method appears everywhere: computation of propagators in quantum field theory, dispersion relations in optics, Green's functions in statistical mechanics.

Cauchy's Theorem and Residues

Cauchy Integral: For analytic $f$ in a simply connected domain $D$ and contour $C \subset D$:

[ \oint_C f(z),dz = 0 \quad \text{(Cauchy's theorem)} ]

But if $f$ has a pole at $z_0$ inside $C$:

[ \oint_C f(z),dz = 2\pi i, \operatorname{Res}_{z=z_0} f ]

Residue — the coefficient at $(z - z_0)^{-1}$ in the Laurent expansion:

[ f(z) = \dots + \frac{a_{-2}}{(z-z_0)^2} + \frac{a_{-1}}{(z-z_0)} + a_0 + a_1(z-z_0) + \dots ]

$\operatorname{Res} = a_{-1}$.

For a simple pole: $\operatorname{Res}{z=z_0} f(z) = \lim{z\to z_0} (z-z_0)f(z)$. For a fraction $f = g/h$ with $h(z_0)=0$, $g(z_0)\neq 0$: $\operatorname{Res} = g(z_0)/h'(z_0)$.

Residue theorem: $\oint_C f(z),dz = 2\pi i \sum_{\text{poles inside } C} \operatorname{Res} f$

Calculation of Real Integrals

Integrals $\int_{-\infty}^{+\infty} f(x),dx$: We close the contour with a semicircle in the upper (or lower) half-plane.

Jordan's lemma: for $e^{iaz} f(z)$ with $a > 0$ and $|f(z)| \to 0$ along the semicircle — the integral over the arc $\to 0$.

Result: $\int_{-\infty}^{+\infty} f(x),dx = 2\pi i \times \sum_{\operatorname{Im}(z_0)>0} \operatorname{Res} f(z_0)$

Numerical Example 1: $I = \int_{-\infty}^{+\infty} \frac{dx}{x^2 + 1}$

Poles $f(z) = 1/(z^2+1) = 1/((z+i)(z−i))$ at $z = \pm i$. In the upper half-plane: $z = i$.

$\operatorname{Res}{z=i} = \lim{z\to i} \frac{z-i}{z^2+1} = \frac{1}{2i}$

$I = 2\pi i \times \frac{1}{2i} = \pi \ \checkmark$ (Check: $\int dx/(x^2+1) = \arctan(x)|_{-\infty}^{+\infty} = \pi$)

Numerical Example 2: $I = \int_{-\infty}^{+\infty} \frac{\cos(ax)}{x^2 + b^2} dx$, $a, b > 0$

We write as $\operatorname{Re}[\int e^{iax}/(x^2+b^2),dx]$. The pole in the upper half-plane: $z = ib$.

$\operatorname{Res}_{z=ib} \frac{e^{iaz}}{z^2+b^2} = \frac{e^{-ab}}{2ib}$

$I = \operatorname{Re}[2\pi i \times \frac{e^{-ab}}{2ib}] = \operatorname{Re}[\frac{\pi e^{-ab}}{b}] = \frac{\pi e^{-ab}}{b}$

Kramers–Kronig Dispersion Relations

Physical justification: the response function $\chi(\omega)$ of a system is analytic in the upper half-plane — this follows from causality (response after excitation, not before). Applying Cauchy's theorem to $\chi(\omega)/(\omega' − \omega)$:

[ \operatorname{Re}\chi(\omega) = \frac{2}{\pi}, \textrm{P.V.} \int_0^\infty \frac{\omega' \operatorname{Im}\chi(\omega')}{\omega'^2 - \omega^2} d\omega' ]

[ \operatorname{Im}\chi(\omega) = -\frac{2\omega}{\pi}, \textrm{P.V.} \int_0^\infty \frac{\operatorname{Re}\chi(\omega')}{\omega'^2 - \omega^2} d\omega' ]

P.V. — principal value according to Cauchy (integral in the sense of limits when bypassing the pole).

Physical consequences: The real part (dispersion = deviation of waves) is determined by the imaginary part (absorption) and vice versa. This is a fundamental restriction: a material not absorbing waves at any frequency also does not alter their speed. Dispersion relations are used in optics, acoustics, nuclear scattering, for checking causality of models.

Green's Functions via Residues

The Green's function for a damped oscillator: $G(\omega) = 1/[m(\omega_0^2 - \omega^2 - i\gamma\omega)]$.

Poles: $\omega_\pm = -i\gamma/2 \pm \sqrt{\omega_0^2 - \gamma^2/4}$. For $\gamma < 2\omega_0$: two poles with $\operatorname{Im}(\omega_\pm) < 0$ (in the lower half-plane). The causal $G(t)$ = inverse Fourier transform, closing the contour downward for $t > 0$:

[ G(t) = \theta(t) \times \frac{1}{m\omega_d} \sin(\omega_d t) e^{-\gamma t/2},\quad \omega_d = \sqrt{\omega_0^2 - \gamma^2/4} ]

Physically: this is a damped oscillator, rising from zero at $t = 0$.

Applications in Quantum Field Theory

Feynman propagator for free scalar field: $G(k) = i/(k^2 - m^2 + i\epsilon)$. The pole at $k^2 = m^2$ — “mass shell” (particles on the mass surface). The +$i\epsilon$ shift (Feynman's prescription) selects the causal (retarded) propagator. All Feynman diagrams are products of such propagators, integrated over momenta using the residue theorem.

Complex Analysis in Signal Processing and Quantum Mechanics

Complex analysis and dispersion relations are tools for signal processing and quantum theory. In system theory, the transfer function $H(\omega)$ (impedance, filter, amplifier) is analytically continued to the upper half-plane: causality of the system (response occurs after excitation) is equivalent to the fact that the poles of $H(\omega)$ lie in the lower half-plane. This is precisely the condition for filter stability. Kramers–Kronig dispersion relations connect the real and imaginary parts of dielectric permittivity $\varepsilon(\omega)$: knowing absorption at all frequencies ($\operatorname{Im} \varepsilon$) completely determines the dispersion ($\operatorname{Re} \varepsilon$). This is used in material spectroscopy: from the measured absorption, the complete refractive index is recovered. In nuclear physics, dispersion relations for the scattering amplitude connect elastic and inelastic scattering. In quantum chromodynamics, sum rules from dispersion relations allow determining proton structure functions from scattering data. The saddle-point method (method of steepest descent) is used for asymptotics of eigenvalues and spectral coefficients — in quantum mechanics for computation of WKB amplitudes for passage through a barrier in tunneling.

Kramers–Kronig dispersion relations are applied not only in electrodynamics, but also in acoustics (dispersion of sound waves in viscous media), optics of metamaterials (design of negative refractive index with minimal losses), and financial option theory (connection between implied volatility and fair price via integral relations in strike space). All these applications are united by causality and analyticity of the system's response.

Assignment: (a) Calculate $\int_0^\infty \frac{\sin(x)}{x} dx$ via a contour with a cut around zero. (b) For $G(\omega)$ of the oscillator with $\gamma = \omega_0/10$: find the poles in the complex plane. Construct the causal $G(t)$. At what $t$ is the maximum? (c) Check the dispersion relation for the simple model $\chi(\omega) = \omega_0/(\omega_0^2 - \omega^2 - i\epsilon \omega)$ — compute $\operatorname{Re} \chi$ and $\operatorname{Im} \chi$ and numerically verify the Kramers–Kronig formula.