Complex Fugacity and the Zeros of the Partition Function

Phase transitions, such as water boiling into steam or a metal becoming a magnet, are among the most dramatic collective phenomena in nature. Yet, their mathematical description poses a profound challenge. In the formal framework of statistical mechanics, the thermodynamic properties of any finite-sized system are perfectly smooth, showing no sign of the sharp, singular behavior we associate with a phase transition. This creates a significant knowledge gap: if finite systems can't have transitions, how do they emerge in the macroscopic world we observe?

This article delves into the ingenious solution proposed by C. N. Yang and T. D. Lee, which involves a daring journey into the complex plane. By treating physical parameters like fugacity not as real numbers but as complex variables, a hidden mathematical structure is revealed. The reader will learn how this "unphysical" approach provides a powerful and unified picture of phase transitions. The first chapter, "Principles and Mechanisms," will introduce the core concept of complex fugacity, the partition function, and its crucial zeros. The second chapter, "Applications and Interdisciplinary Connections," will then demonstrate the remarkable utility of this idea across diverse fields, from classical magnetism and quantum physics to the abstract world of pure mathematics.

Imagine you are a god-like administrator of a vast cosmic city. Your city is made of countless building sites, and you have a population of particles eager to move in. You can't control every particle individually, but you can set the global rules. You set the temperature, which dictates how energetically your citizens behave. And you set a "welcome bonus," a sort of chemical encouragement, for any particle that decides to occupy a site. In physics, we call this bonus the ​​fugacity​​, denoted by the variable $z$. A low fugacity means particles are hesitant to join; a high fugacity means they rush in.

As the administrator, you're interested in the city's overall state. How many particles are there? What's the pressure? To figure this out, you perform a grand census. You consider every single possible configuration of particles on your sites: one particle here, two particles there, a cluster over yonder, even the completely empty city. For each configuration with $N$ particles, you weigh its contribution by a factor of $z^N$. Summing up all these weighted possibilities gives you a master function, the grand partition function, $\Xi$.

$\Xi(z) = \sum_{N=0}^{\infty} Z_N z^N$

Here, $Z_N$ is the "sub-census" for exactly $N$ particles, accounting for their interactions and the temperature. This grand function $\Xi(z)$ contains, in principle, everything you could ever want to know about your city's equilibrium state. All the thermodynamic properties, like pressure and density, can be extracted from it, typically from its logarithm, $\ln \Xi$.

Now, for any real-world city, this welcome bonus, the fugacity $z$, is a real, positive number. And if you look at the formula, you'll notice something curious. The term $Z_N$ is always positive (it's essentially a sum of probabilities), and so is $z^N$. This means $\Xi(z)$ is a sum of positive numbers. It can get very large, but it can never be zero. Why is this important? Because the interesting, dramatic events in a system—the boiling of a liquid, the magnetization of a metal—are associated with mathematical "singularities," or non-analytic points, in the thermodynamic functions. And the most common source of a singularity in a function like $\ln \Xi(z)$ is a point where its argument, $\Xi(z)$, becomes zero.

For any finite-sized city, since $\Xi(z)$ is never zero for any physical $z > 0$, its logarithm is perfectly smooth and well-behaved. There are no sharp transitions, no sudden boiling points, just gentle changes. This is a profound statement: ​​phase transitions, in their mathematical sharpness, do not happen in finite systems​​.

This is where the genius of C. N. Yang and T. D. Lee came in. They asked a question that seems, at first, like a physicist's idle fantasy: what if the fugacity $z$ were not a real number, but a ​​complex number​​? This is a purely mathematical trick; there's no lab where you can set a complex welcome bonus. But by stepping off the line of real numbers into the vast plane of complex numbers, they revealed a hidden landscape that governs the real-world behavior.

Let's make this concrete. Imagine a tiny system, a "hard-core" gas on a $2 \times 2$ checkerboard where particles can't occupy adjacent squares. We can do the census by hand:

• ​​0 particles:​​ There is 1 way (the empty board). Contribution: $1 \cdot z^0 = 1$.
• ​​1 particle:​​ There are 4 sites, so 4 ways. Contribution: $4 \cdot z^1 = 4z$.
• ​​2 particles:​​ They can only occupy opposite corners. There are 2 such pairs. Contribution: $2 \cdot z^2 = 2z^2$.
• ​​3 or more particles:​​ Impossible due to the hard-core rule.

The grand partition function is just a simple polynomial: $\Xi(z) = 1 + 4z + 2z^2$. For real $z>0$, this is always positive. But in the complex plane, we can ask: where is it zero? We just solve the quadratic equation $2z^2 + 4z + 1 = 0$. The roots, the ​​Yang-Lee zeros​​, are $z = -1 \pm \frac{\sqrt{2}}{2}$. They are not on the positive real axis; they are on the negative real axis, safely away from the physical world.

In a tiny system, the zeros are just a few scattered points in the complex plane. But what happens in a truly vast, macroscopic system? The partition function becomes a polynomial of an enormous degree. The number of zeros becomes immense, and they are no longer scattered randomly. They begin to trace out sharp lines and curves in the complex plane.

A phase transition occurs at the precise moment when, in the limit of an infinitely large system, this line of zeros marches up and finally ​​touches the positive real axis​​ at some critical fugacity, $z_c$. At that point, the function $\ln \Xi(z)$ suddenly becomes singular, creating the sharp "break" in thermodynamic properties that we identify as a phase transition. The placid, smooth behavior of the finite system gives way to the dramatic, singular behavior of the infinite one.

This connection is not just qualitative. The mathematical theory of power series tells us that the series expansion for the pressure, known as the activity expansion, converges only up to the first singularity it encounters. The Yang-Lee zeros are precisely these singularities. The distance from the origin to the nearest zero in the complex plane dictates the radius of convergence of our series expansions, limiting the range where a single description (like "gas") is valid. For an ideal gas, which has no interactions and no phase transitions, the partition function turns out to be an exponential, $\Xi(z) = \exp(Az)$, which has no zeros anywhere in the complex plane. Correspondingly, its activity expansion has an infinite radius of convergence.

The theory reached a pinnacle of elegance with the ​​Lee-Yang circle theorem​​, which applies to a cornerstone model of magnetism called the ferromagnetic Ising model. In this model, the role of fugacity is played by a variable related to the external magnetic field. The theorem makes a startling claim: for any ferromagnetic system, regardless of its size or the temperature, all the Yang-Lee zeros lie perfectly on the unit circle in the complex fugacity plane, $|z|=1$.

This beautiful geometric constraint has profound physical consequences. Real magnetic fields correspond to the positive real axis ($z \ne 1$). Since the zeros are all on the unit circle, they never touch the real axis except possibly at the point $z=1$ (which corresponds to zero magnetic field). This means that a ferromagnet can never undergo a phase transition in the presence of an external magnetic field.

The magic happens only at zero field ($z=1$). The behavior of the zeros changes dramatically at the critical temperature, $T_c$. Above $T_c$, the distribution of zeros on the unit circle is dense enough to "pinch" the real axis at $z=1$ in the thermodynamic limit, preventing spontaneous magnetization. Below $T_c$, a ​​gap​​ opens up in the circle of zeros around $z=1$. This gap is the signature of the ordered phase; it stabilizes the spontaneous magnetization against small external fields. The phase transition at $T_c$ is precisely the point where this gap closes as the temperature is raised. For the simple one-dimensional Ising model, one can even calculate the edge of this gap exactly, and see how it closes as the temperature is raised. The location and density of these zeros carry all the information about the phase transition. The theorem is also incredibly specific: it relies on the interactions being ferromagnetic. If they are not, the zeros are no longer constrained to the circle. This specificity highlights the deep link between the physical interactions and the mathematical structure of the theory.

This idea of finding truth in the complex plane is far more general. We complexified the fugacity (related to chemical potential or magnetic field) to understand density-driven transitions. But what if we complexify the temperature?

If we consider the canonical partition function $Z_N(\beta)$, where $\beta = 1/(k_B T)$ is the inverse temperature, and allow $\beta$ to be a complex variable, we find a new set of zeros known as ​​Fisher zeros​​. These zeros live in the complex temperature plane. Just like their Yang-Lee counterparts, they stay off the real axis for any finite system. A temperature-driven phase transition, like boiling water by heating it, occurs only when these Fisher zeros pinch the real temperature axis in the thermodynamic limit.

So we have a grand, unified picture. Yang-Lee zeros, in the complex field plane, govern field-driven transitions. Fisher zeros, in the complex temperature plane, govern temperature-driven transitions. Phase transitions are the physical manifestations of the mathematical singularities caused by these zeros.

You might still think this is all abstract mathematics. But these complex zeros have real, measurable consequences, even in the finite systems we study in labs and computer simulations. The distance of the nearest zero to the real axis determines the "sharpness" of a transition. If the zero is far away, the transition will look smooth and rounded in a finite system. If it's very close, the transition will appear sharp.

This distance also governs the lifetime of ​​metastable states​​, like supercooled water or a supersaturated vapor. These states are not the true thermodynamic equilibrium, but they can persist for a long time. They can be described by analytically continuing a thermodynamic function (like the logarithm of the partition function) from a stable region into an unstable one. This mathematical continuation is ultimately blocked by the first complex zero you run into. The closer the zero is to the real axis, the more tenuous the metastable state, and the shorter its lifetime.

Even more remarkably, the way these zeros approach the real axis as the system size $L$ grows follows universal laws, a theory known as ​​finite-size scaling​​. For a sharp, first-order transition (like boiling), the imaginary part of the closest zero scales as $L^{-d}$, where $d$ is the spatial dimension of the system. For a more subtle, continuous phase transition, it scales as $L^{-1/\nu}$, where $\nu$ is a famous "critical exponent" that describes the divergence of the correlation length at the transition.

Thus, by studying the locations of these ghostly zeros in finite systems, we can extract the universal exponents that characterize the phase transition in the infinite limit. The abstract journey into the complex plane comes full circle, providing concrete predictions and a profound understanding of the collective behaviors that shape our world. The census of our imaginary city, when viewed through the lens of complex numbers, reveals the hidden mathematical laws that govern its very social fabric.

In our journey so far, we have been bold. We took a perfectly sensible physical quantity—the fugacity, which is tied to the chemical potential—and allowed it to wander off the real number line into the vast, open landscape of the complex plane. You might rightly ask: what was the point of this seemingly "unphysical" excursion? The answer, it turns out, is that it gives us a completely new and profoundly insightful way to think about one of the most dramatic phenomena in nature: the phase transition. By looking for the "non-physical" complex values of fugacity where the partition function vanishes, we gain an almost magical ability to see the seeds of a phase transition long before it happens. Let's now explore the astonishingly diverse fields where this strange and beautiful idea bears fruit.

Let's start with the familiar worlds of classical statistical mechanics—the models of simple fluids and magnets that are the bedrock of the subject. Imagine a gas of particles that cannot overlap, like marbles on a checkerboard. This "hard-core lattice gas" is a beautifully simple model for the basic fact that atoms take up space. If we write down its partition function, we find it is a simple polynomial in the fugacity, $z$. While for any real, physical fugacity the partition function is strictly positive (after all, it's a sum of positive Boltzmann weights), we can ask: for which complex values of $z$ would it be zero? Even for a tiny chain of just three sites, we find these zeros exist as specific points in the complex plane. These are our first examples of Lee-Yang zeros. If we add a bit more realism, like an attractive force between particles that are not immediate neighbors, the partition function becomes a bit more complex, and its zeros move to new, interesting locations in the complex plane.

This picture becomes truly spectacular when we consider the ferromagnetic Ising model, the physicist's favorite model of a magnet. Here, the celebrated Lee-Yang circle theorem tells us something remarkable: for a ferromagnet, all the partition function zeros lie precisely on the unit circle in the complex fugacity plane. Now, think about what happens as we take a larger and larger piece of our magnet, approaching the thermodynamic limit. The discrete, scattered zeros on the circle begin to crowd together, eventually forming a continuous distribution. A phase transition—the spontaneous appearance of magnetization as we cool the system—is governed by the behavior of these zeros near $z=1$. For temperatures above the critical temperature, $T_c$, this distribution of zeros is dense along the entire unit circle. But as we cool the system to below $T_c$, a gap miraculously opens up in the circle of zeros right around $z=1$. The system has developed a stable magnetized phase, and this stability is reflected by the fact that no zeros are near the physical point $z=1$. The phase transition is, in this language, the process of the gap opening and closing.

This isn't just a theoretical curiosity; it's a powerful practical tool. In computer simulations, we can only ever study finite systems, which can't have a true phase transition. But we can calculate the locations of their Lee-Yang zeros. By searching for the temperature at which the zero closest to the real axis gets as close as possible to $z=1$, we can obtain remarkably accurate estimates of the true critical temperature of the infinite system. In this way, the "unphysical" complex zeros cast a "shadow" of the phase transition that is visible even in a small, finite system.

The power of this method extends far beyond simple magnets and fluids. Nature is filled with materials whose microscopic interactions are more complex. Consider magnetic systems where the atomic spins can point not just up or down, but can also be in a non-magnetic zero state. Models like the Blume-Capel model describe such systems. Even for a tiny two-site version, we can calculate the fugacity zeros and see how they are affected by the new interactions.

Going further, in models like the Blume-Emery-Griffiths (BEG) model, which can describe phenomena like mixtures of Helium-3 and Helium-4, we find something even more profound. Depending on the values of the interaction parameters, the locus of the Yang-Lee zeros can undergo a dramatic change in its geometry—for instance, switching from lying on the unit circle to lying on the negative real axis. This geometric transition in the complex plane corresponds directly to crossing a boundary in the real-world phase diagram of the material. The abstract geometry of zeros is a map of the concrete physical phases!

The idea is not even restricted to particles or spins on a lattice. Consider the "six-vertex model," which is a fundamental model for the structure of two-dimensional ice and certain ferroelectric materials. Here, the "state" is a configuration of arrows on the edges of a grid, subject to local "ice rules." Again, we can introduce a fugacity for, say, arrows pointing up. The mathematics gets more sophisticated—the zeros are now found by studying the eigenvalues of a "transfer matrix"—but the core idea holds. In the limit of an infinitely large system, the condition for a zero is that the dominant eigenvalues of this matrix become equal in magnitude, and this condition carves out curves in the complex fugacity plane where the system is critical.

Perhaps the most stunning demonstrations of this concept's power come when we venture into the quantum world. The partition function of a quantum system is defined by a trace over its Hilbert space, but once computed, it can often still be viewed as a function of a complex fugacity. Consider a cluster of four quantum spins interacting antiferromagnetically, as described by the Heisenberg model. Unlike the ferromagnet, the Lee-Yang circle theorem no longer applies, and the zeros can lie in more complex patterns. These patterns, however, are just as meaningful, encoding the physics of quantum fluctuations and geometric frustration inherent in the system.

The story culminates at one of the most exciting frontiers of modern physics: topological matter. In certain materials known as topological insulators, there exist special, robust states of matter protected by the underlying topology of the system's quantum wavefunctions. The Su-Schrieffer-Heeger (SSH) model is the simplest example, describing a 1D chain with special "edge states" that have exactly zero energy. For non-interacting fermions, the partition function zeros are directly related to the single-particle energy levels. A zero-energy state corresponds precisely to a Yang-Lee zero at $z=-1$.

Now, what happens if we break the topological protection? The two zero-energy states, one at each end of the chain, acquire small, opposite energies. This splitting of energy levels is directly mirrored in the complex fugacity plane: the single zero at $z=-1$ bifurcates into two distinct zeros that move apart along the real axis. The separation distance between these two zeros becomes a direct, measurable proxy for the degree to which the system's topology has been broken. This provides an incredible, direct link between the abstract geometry of complex zeros and the profound, physically consequential properties of topological quantum matter.

The unifying power of complex fugacity even transcends physics itself, building a beautiful bridge to the world of pure mathematics. Let's return to the simple hard-core lattice gas. The partition function, which counts the number of ways to place non-overlapping particles on a graph, turns out to be mathematically identical to an object known as the ​​independence polynomial​​ in the field of graph theory.

This means that the Lee-Yang zeros studied by physicists are none other than the roots of the independence polynomial studied by combinatorialists. Questions about the existence and nature of phase transitions in a physical system can be translated into questions about the location of roots of a graph polynomial, and vice-versa. Two communities, using different languages, were in fact exploring different faces of the same underlying mathematical structure.

From the boiling of a liquid, to the magnetization of a solid, to the quantum properties of topological insulators, and even to abstract problems in graph theory, the decision to let our variables roam into the complex plane has paid off handsomely. It has given us a unified, geometric viewpoint that reveals deep connections between seemingly disparate phenomena. It is a powerful testament to the fact that sometimes, the most profound insights into the real world are found by having the courage to explore the "unreal."