# What are functions of a complex variable used for in physics?

Whenever someone asks "Why are complex numbers important?" the answer, at least in the context of physics, usually includes things like quantum mechanics, oscillators and AC circuits. This is all very fine, but I've never seen anyone talk about functions of a complex variable. Complex functions of real variables are used often enough, but I do not yet see (with one minor exception; see below) why my university would decide to dedicate half a semester to the theory of holomorphic functions if there are no physics applications.

Don't get me wrong; I don't regret learning about complex functions. I think it is one of the most beautiful subjects within math, but my question still stands. Are there any applications of functions $f: \mathbb{C} \to \mathbb{C}$ within physics?

About the exception: If a function $f$ is holomorphic, then its components $u,v$ are automatically harmonic. This is a quick way to find solutions to Laplace's equation $\nabla^2 u = 0$, but surely this minor trick doesn't justify having to learn about the whole theory.

• You mean aside from the bajillion problems you can solve by analytic extension and the method of residuals? – dmckee --- ex-moderator kitten Feb 24 '14 at 1:36
• @dmckee: I actually don't know about the bajillion problems you mention. I know about the method of residues, but is it really that important? It's obviously pretty neat, but I didn't imagine it would be so useful. Maybe you should mention all of this in an answer. – Javier Feb 24 '14 at 1:38
• Just found this very closely related question: physics.stackexchange.com/q/89072 – user10851 Feb 24 '14 at 2:35
• @ChrisWhite also this physics.stackexchange.com/q/45981 – joshphysics Feb 24 '14 at 2:56
• @nervxxx: I'm not saying it isn't applicable, I'm just saying that I haven't seen any applications. That's why I'm asking this question instead of going to the math department to complain about being taught useless stuff. – Javier Feb 24 '14 at 2:58

Complex analysis is more than just a tool that can be used for computing difficult integrals. For example:

1. In quantum field theory, one of the most popular regularization schemes relies on the theory of complex functions. In particular, it relies on the concept of analytic continuation of functions $f:D\to\mathbb C$ for some $D\subseteq \mathbb C$. In particular, one often encounters integrals that are divergent in four dimensions, so one first computes the same integral in arbitrary positive integer dimension $d$, and then one regards this expression as a function of a complex variable $d$ and analytically continues it to a meromorphic function on the complex plane. This allows one to then parameterize the divergence that appears in the integral in the parameter $\epsilon = d-4$, conduct renormalization, and then take $\epsilon\to 0$. This is a case in which complex analysis isn't being used just as a trick, it's being used to render the expressions in perturbative quantum field theory well-defined. The procedure outlined above is called dimensional regularization.

2. In the context of two-dimensional conformal field theory (CFT), the fields of the theory can be regarded as functions on the complex plane. If you open any CFT text, it'll basically just seem like an advanced complex analysis textbook.

3. You mention Laplace's equation (which comes up in lots of physical contexts like electrostatics, heat flow, fluid flow, etc.), but here's another facet you didn't mention. Laplace's equation is invariant under conformal transformations which are themselves mappings $\mathbb C\to \mathbb C$, an the theory of such functions can be used to solve Laplace's equation boundary value problems.

4. In quantum field theory and string theory, zeta regularization can be used to give meaning to divergent sums by using the Riemann zeta function which is a function of a complex variable. For example, some derivations of the Casimir effect use zeta regularization.

This is all very fine, but I've never seen anyone talk about functions of a complex variable.

The Laplace transform is a widely used integral transform in mathematics with many applications in physics and engineering. It is a linear operator of a function f(t) with a real argument t (t ≥ 0) that transforms f(t) to a function F(s) with complex argument s, given by the integral

$$F(s) = \mathcal L\{f(t)\}= \int_0^\infty f(t)e^{-st}dt, \, s = \sigma +i\omega$$

Since, by the above

$$\mathcal L\{\frac{d}{dt}f(t)\} = sF(s) - f(0^-)$$

and

$$\mathcal L\{\int_0^tf(\tau)d\tau\} = \frac{1}{s}F(s)$$

the Laplace transform of a differential or integral equation is an algebraic equation in the complex variable $s$.

An elementary application is to the Tautochrone problem.

• This answer would benefit from an example physics problem for which Laplace transforms are useful. – DanielSank Feb 24 '14 at 5:31
• @DanielSank Choose any from almost all the integrable physic problems modelled with differential equations with given initial conditions: it's a quite wide spectrum! – DarioP Feb 24 '14 at 10:18
• @DarioP: I know that, but other people coming here may not. To make your answer better I'm suggesting you actually give an example. – DanielSank Feb 24 '14 at 16:19
• @DanielSank, I've provided additional context however, the physical applications of the Laplace transform are so numerous and well known that I'm not convinced any more detailed example is required. – Alfred Centauri Feb 25 '14 at 0:06
• @AlfredCentauri, your call, of course. I just figure that since OP is asking how functions of a complex variable are useful it would be appropriate to give a concrete example. Like I said, your call. Just trying to help. – DanielSank Feb 25 '14 at 0:08

Integration over complex contours is also crucial to one of the main methods of obtaining asymptotic approximations for a large class of special functions and theoretical results, the stationary phase approximation and the method of steepest descent.

Those methods are a huge aid in calculating integrals like $$\int_0^\pi \cos(x\cos(\theta))\text d\theta \quad \text{when }x\gg1,$$ which are hard to calculate, even numerically, because the integrand is highly oscillatory and you need to deal with a lot of cancellations. The idea is to express the integral as a contour integral, and then change the contour so that the integrand will look like a gaussian with a flat phase, which is easy to integrate. This sort of method is the backbone of most of the special function asymptotics, e.g. those for Bessel, Airy, Gamma or hypergeometric functions, among many others.

This sort of analysis is crucial in directly analysing many physical problems, and it is certainly my own daily bread and butter. To give some examples:

• It permits a very intuitive treatment of non-adiabatic transitions for a two-level system in the presence of a not-quite-adiabaticly varying perturbation, in the form of the Landau-Zener formula (doi).

• It is the driving force in any simple treatment of tunnelling and particularly tunnelling ionization. The main analytical and conceptual tool for dealing with High Harmonic Generation, for example, the strong-field approximation, is built using these tools.

• The method of stationary phase is behind the passage from quantum theory to classical mechanics in the form of the principle of stationary phase, as in e.g. this SE question.

• Similarly, it plays a large role in optics and particularly in describing caustics.

In fact, this part of mathematical physics is growing so rapidly that a part of the method is one of the newest additions to the Digital Library of Mathematical Functions (also known as the NIST Handbook of Mathematical Functions, the successor of Abramowitz and Stegun).

I'll answer from my own experience taking the mathematical methods course required of all physicists (and nearly everyone else too) at my undergrad.

Analytic continuation of functions' domains from $\mathbb{R}$ to $\mathbb{C}$ allows one to use what is in my opinion one of the most beautiful and simultaneously useful theorems in mathematics: Cauchy's Residue Theorem. With this method on hand, the class of integrals you can solve opens up enormously. In fact, ever since learning the subject, I've never come across an integral that e.g. Mathematica can solve but I can't. There might be a few, but they don't come up often in physics. Most integrands I've seen are analytic almost everywhere, so they lend themselves to such methods.

In a very simplified view of the world, complex analysis of the residue sort allows one to integrate by differentiating. Continuing with my own experience, the following term of that same course can be summarized as "how to solve ODE's by integrating," and the term after that was "how to turn PDE's into ODE's." Chain it all together and you see how simple these tools can make a physicist's life.

There are some other, less universally applicable uses of functions on the complex plain. You already mentioned the harmonicity/Laplace connection. Another example is conformal mapping: solve your differential equation on a domain where it is tractable, then map the results back to the domain you are given.

Note that if you've taken a pure math version of complex analysis, rather than an applied math version (I've taken both), you'll likely focus on other things that are not as universally powerful for application to physics as adeptness with residues is.

the functions of complex variables enjoy very rich analytic properties which their real counterparts do not;

to mention one such property- if a complex valued function is analytic at a given point in its domain of definition, it is analytic in the entire domain, this simple property gives functions of complex variable a very rich analytic structure and this property is certainly not enjoyed by the functions of real variables, wherein analyticity at one point has no connection to the analyticity at another;

an application of complex variables which looks very beautiful is the study of fractals, eg- mandelbrot set

• I know all those properties. This doesn't answer the question. – Javier Feb 26 '14 at 4:40