Fisher Zeros

How can a single mathematical tool describe the dramatic, collective transformations of matter, like water boiling or a metal becoming magnetic? In statistical mechanics, this power lies in the partition function, a master key that encodes all thermodynamic properties of a system. Yet, this function presents a puzzle: for any real-world physical system, it is always a positive number, which mathematically seems to forbid the abrupt, non-analytic changes characteristic of phase transitions. This article addresses this paradox by embarking on a journey into a mathematical "looking-glass world." We ask a simple but profound question: what happens if we allow physical parameters like temperature to become complex numbers? This inquiry reveals the existence of Fisher zeros—specific complex values where the partition function vanishes. In the following chapters, we will first delve into the "Principles and Mechanisms" of this theory, exploring how these non-physical zeros act as sentinels that signal real-world phase transitions. Subsequently, under "Applications and Interdisciplinary Connections," we will witness the astonishing breadth of this concept, seeing how it unifies phenomena in condensed matter, quantum field theory, and even pure mathematics.

Imagine you are a physicist trying to understand a block of iron. It contains a staggering number of atoms, each with a tiny magnetic personality, a "spin" that can point up or down. These spins talk to each other, preferring to align with their neighbors. As you change the temperature, this society of spins can behave in dramatically different ways—a chaotic mess at high temperatures, a disciplined magnetic army at low temperatures. How can we possibly capture this complex behavior in a single mathematical object?

The answer, one of the crown jewels of statistical mechanics, is the ​​partition function​​, usually denoted by the letter $Z$. You can think of $Z$ as a grand census of all possible states the system can be in. For each possible arrangement of spins (a configuration), we calculate its energy $E$, and assign it a statistical weight, the ​​Boltzmann factor​​ $e^{-\beta E}$, where $\beta$ is inversely related to temperature ($\beta = 1/(k_B T)$). The partition function is simply the sum of these weights over all possible configurations:

This humble-looking sum is astonishingly powerful. It is the master key from which we can derive every single thermodynamic property you can think of: the energy, the heat capacity, the magnetization, and so on. The entire secret life of our block of iron is encoded within $Z$.

Now, let's notice something simple but profound. For any real-world temperature $T \gt 0$, the inverse temperature $\beta$ is a positive real number. Since energy $E$ is real, the Boltzmann factor $e^{-\beta E}$ is always a positive number. The partition function $Z$ is a giant sum of these positive numbers. And what can we say about a sum of positive numbers? It can never, ever be zero. This means that for any physical situation you can create in a lab, $Z$ is always greater than zero. This seems like a trivial observation, but it is the calm before a theoretical storm.

Physics often progresses by asking "what if?". What if we could travel at the speed of light? What if we could be in two places at once? The physicists Chen Ning Yang, Tsung-Dao Lee, and Michael Fisher asked a similarly audacious question: what if temperature wasn't just a positive, real number? What if we allowed it to be a complex number?

This is a journey into a mathematical looking-glass world. There is no thermometer that can measure a temperature of, say, $(200 + 50i)$ Kelvin. This is purely a playground for the mind. But in this playground, our rule that $Z$ can never be zero suddenly breaks down. When $\beta$ becomes complex, the Boltzmann factors $e^{-\beta E}$ are no longer simple positive numbers; they are complex numbers, swirling around in the complex plane. And when you sum up a collection of complex numbers, they can conspire to cancel each other out perfectly, yielding zero.

The specific complex values of temperature (or a related variable) where the partition function vanishes, $Z=0$, are called ​​Fisher zeros​​.

Let's see this in action with a toy model, a tiny magnet made of just three spins on a triangle, a system so simple we can analyze it completely. The spins interact with their neighbors, and there are only $2^3 = 8$ possible ways to arrange them. After we list all the energies and sum up the Boltzmann factors, the partition function turns out to be:

where $J$ is the strength of the interaction. To make things neater, we define a new variable, let's call it $z = e^{-2\beta J}$. In terms of $z$, the partition function becomes a simple polynomial expression:

Now, finding the Fisher zeros is as easy as solving the high-school algebra problem $Z=0$. A little bit of manipulation leads to the startlingly simple equation $2 + 6z^2 = 0$. The solutions are not real numbers; they are purely imaginary: $z = \pm i/\sqrt{3}$. These are our Fisher zeros! We've found specific "complex temperatures" where our system's partition function vanishes. Similar calculations for a small chain of four spins give a similar result. The concept is not even limited to these simple magnetic models; it applies beautifully to quantum systems as well, like a ring of three interacting quantum spins.

At this point, you should be asking: this is a fun mathematical game, but what does it have to do with the real world? Why should a physicist studying a real block of iron care about zeros that exist only at imaginary temperatures?

The answer is the magic word: ​​phase transitions​​. A phase transition, like water boiling into steam or a block of iron suddenly becoming magnetic below the Curie temperature, is a moment of dramatic, collective transformation. In the language of mathematics, it is a point of ​​non-analyticity​​. An analytic function is smooth and well-behaved, like a gently rolling hill. A non-analytic function has a sharp point, a cliff, or a sudden jump, reflecting the abrupt change in the physical system.

The central thermodynamic potential is the ​​free energy​​, $F$, which is derived from the logarithm of the partition function, $F = -k_B T \ln Z$. For the free energy to be non-analytic, something must go wrong with $\ln Z$. The only place the logarithm function "goes wrong" is at zero. So, a phase transition can only happen at a real, physical temperature $T_c$ if the partition function $Z$ becomes zero at that temperature.

But we've just hit a paradox! We argued earlier that for any real temperature, $Z$ is always positive and can never be zero. This implies that no phase transition can ever occur! And indeed, for any system with a finite number of particles—be it our 3-spin triangle or a trillion-atom diamond—the partition function is a finite sum of smooth exponential functions. This sum is always smooth and analytic. The startling conclusion is that ​​phase transitions do not happen in finite systems​​. A single molecule of $\text{H}_2\text{O}$ does not boil.

The resolution lies in the power of the infinite. A phase transition is an ​​emergent phenomenon​​ of a truly macroscopic system. We must take the ​​thermodynamic limit​​, letting the number of particles $N$ go to infinity.

Here is where the Fisher zeros re-enter the stage, not as curiosities, but as the main actors. For a finite system, the zeros are scattered in the complex plane, safely away from the real axis where the physical world lives. But as we increase the size of the system ($N \to \infty$), the number of zeros also increases. They begin to multiply and organize, marching like an army towards the real axis. In the infinite limit, these discrete zeros can coalesce into continuous lines or curves.

If one of these curves of zeros touches or crosses the real axis at a specific point, say $T_c$, then at that very point, the free energy becomes non-analytic. And voilà, a phase transition is born! The Fisher zeros, living in the complex plane, act as sentinels, mapping out the boundaries of stability. Their approach to the real world signals an impending transformation. This is why the two-dimensional Ising model has a phase transition—its zeros pinch the real axis in the thermodynamic limit—while the one-dimensional version does not; its zeros remain timidly in the complex plane.

The patterns these zeros form in the complex plane are not random; they possess a deep and often beautiful geometric structure. For many exactly solvable models, like the Ising model on certain two-dimensional lattices, the zeros in the thermodynamic limit lie perfectly on a circle in the complex plane of the variable $z=e^{-2\beta J}$. Imagine, the messy, collective behavior of countless interacting spins is governed by an unseen, perfect geometric object in a hidden mathematical space!

The elegance of the theory shines when we face seemingly intractable problems. The Ising model on a ​​Kagome lattice​​, a complex and beautiful web of corner-sharing triangles, seems daunting. Yet, through a clever mathematical trick called a ​​star-triangle transformation​​, physicists showed that the partition function of this complicated model is directly related to that of a much simpler honeycomb lattice. This means the unknown pattern of zeros for the Kagome lattice could be derived from the known pattern for the honeycomb lattice, revealing a profound and unexpected unity between different physical systems.

Even more, the way the locus of zeros approaches the real axis tells us about the nature of the phase transition. For the mean-field Curie-Weiss model, a simplified but powerful model of magnetism, we can calculate precisely how the line of zeros peels away from the real axis at the critical point. It does so at a sharp angle of $45$ degrees, a universal feature of this class of models. The geometry of the unseen dictates the physics of the seen.

This way of thinking—connecting phase transitions to the zeros of a partition function—is an incredibly general and powerful idea. It provides a common language to describe a vast menagerie of physical phenomena. While Fisher zeros arise from letting temperature become complex, a parallel theory from Yang and Lee explored what happens when the magnetic field becomes complex, leading to ​​Yang-Lee zeros​​. Together, they provide a multi-dimensional map of a system's analytic structure, revealing its critical points from different vantage points.

Furthermore, the theory of zeros provides a microscopic ruler to measure the character of a phase transition through a concept known as ​​finite-size scaling​​. The distance of the closest zero to the real axis shrinks as the system size ($L$) grows, but how it shrinks depends on the type of transition:

• For ​​first-order transitions​​, like the boiling of water where there is a latent heat, the change is very abrupt. The gap to the nearest zero closes very quickly, scaling as $L^{-d}$, where $d$ is the dimension of the system.

• For ​​continuous transitions​​, like the gradual onset of magnetism at the Curie point, the gap closes much more slowly, scaling as $L^{-1/\nu}$. Here, $\nu$ is a famous ​​critical exponent​​ that describes how the correlation length diverges.

This is a remarkable connection. By studying the precise location of these non-physical zeros in finite-sized computer simulations, we can extract fundamental constants of nature like $\nu$ that describe the universal behavior of infinitely large systems at a critical point. The abstract world of complex zeros provides one of the most powerful and precise tools we have for understanding the concrete world of critical phenomena. From a simple "what if" question about complex numbers, a deep and beautiful theory has emerged, unifying the behavior of magnets, fluids, and quantum matter through the elegant dance of zeros in the complex plane.

In our previous discussion, we saw that the universe, when described by the laws of statistical mechanics, seems to have a peculiar aversion to certain "complex temperatures." At these specific complex values, the partition function—the grand sum over all possible states—vanishes. These points, the Fisher zeros, act as sentinels, mapping out the forbidden zones in the abstract plane of temperature. On the real axis, they pinch down towards a single point, heralding the dramatic onset of a phase transition. This idea is powerful, but its true beauty lies in its universality. It is not just a story about water boiling or magnets aligning; it is a story that echoes through an astonishing variety of scientific disciplines. Let us now embark on a journey to see how this one concept serves as a unifying thread, weaving together the physics of materials, the fundamental forces of nature, and even the abstract world of pure mathematics.

Our journey begins in the natural habitat of the partition function: the world of statistical and condensed matter physics. Here, we study systems made of countless interacting particles—spins in a magnet, atoms in a crystal. Even for the simplest "toy" systems, which we can solve exactly with paper and pencil, the Fisher zeros tell a profound story.

Imagine a tiny magnetic grid, perhaps a mere two-by-two square of spins with periodic boundaries, forming a miniature torus. Or consider a small, oddly shaped molecule whose atoms interact according to the rules of the Potts model. For such systems, we can write down the partition function, not as an infinite sum, but as a finite polynomial in a variable related to temperature, like $v = e^{\beta J} - 1$. The task of finding the Fisher zeros then becomes a familiar one from high-school algebra: find the roots of the polynomial. These roots, which are typically complex numbers, lie scattered in the complex plane, and as we make our system larger, we would see them march towards the real axis, ready to signal a phase transition. Even for a single triangular arrangement of spins in a so-called clock model, the zeros can be found with elegant simplicity, revealing locations like $x^2 = -3$ in the plane of the temperature variable $x=e^{\beta J}$. These simple examples are the training grounds where we build our intuition.

But the real world is quantum mechanical. The concept of Fisher zeros extends beautifully into this more complex domain. Consider the Hubbard model, a cornerstone for understanding the behavior of electrons in materials that are not simple metals or insulators. It describes a delicate dance between electrons hopping between atomic sites and repelling each other when they land on the same site. The strength of this repulsion is governed by a parameter, $U$. We can ask a new kind of question: instead of complex temperature, what if we imagine a complex interaction strength $U$? The partition function, now a function of complex $U$, will again have zeros. These zeros tell us about quantum phase transitions driven not by temperature, but by the very strength of the interactions between electrons. The landscape of zeros in the complex $U$-plane provides a map of the rich physics of strongly correlated electrons.

This line of inquiry pushes right to the frontiers of modern physics. Consider the Sachdev-Ye-Kitaev (SYK) model, a deceptively simple model of interacting quantum particles (Majorana fermions) that has created a storm of excitement. Why? Because its strange, chaotic behavior appears to be mathematically equivalent to the physics of a certain type of black hole in a universe with anti-de Sitter geometry. It's a "black hole in a bottle." The partition function of this model is an object of intense study, and analyzing the location of its zeros gives us crucial clues about its thermal properties and its connection to quantum chaos and gravity.

The power of statistical mechanics is that its methods can be translated to solve problems that seem, at first glance, completely unrelated. One of the most spectacular translations is to the realm of quantum field theory, the language we use to describe the fundamental forces of nature.

The theory of the strong nuclear force, which binds quarks into protons and neutrons, is called Quantum Chromodynamics (QCD). Its equations are notoriously difficult to solve. One of the most powerful tools we have is Lattice Gauge Theory, where the continuous fabric of spacetime is replaced by a discrete grid, or lattice. In this framework, the quantum field theory problem miraculously transforms into a statistical mechanics problem! The partition function is calculated by summing over all possible field configurations on the lattice. The parameter $\beta$ in this context is no longer related to temperature, but to the fundamental coupling constant $g$ of the force itself, via a relation like $\beta = 4/g^2$.

What, then, is a "phase transition" here? It can correspond to something as profound as the liberation of quarks. At low energies (or strong coupling), quarks are permanently "confined" inside protons and neutrons. But at extreme temperatures and densities, like those in the early universe, they are predicted to exist in a "deconfined" plasma. This transition is governed by the Fisher zeros of the lattice gauge theory partition function in the plane of the complex coupling constant. For a simplified SU(2) gauge theory on a single square of the lattice (a "plaquette"), the zeros can be calculated exactly. One finds that the partition function is related to a Bessel function, and its zeros lie on the imaginary axis in the plane of $\beta$. This means the Fisher zeros for the squared coupling constant, $g^2$, also lie on the imaginary axis, providing a beautifully clean mathematical picture of where transitions in the nature of the vacuum might occur.

The ambition doesn't stop there. Some of the most advanced theories attempting to describe quantum gravity itself, like Group Field Theory (GFT), also borrow this language. In GFT, spacetime is not fundamental but is built up from discrete "atoms" of geometry, such as tetrahedra. The partition function is a sum over all ways of gluing these atoms together to form a quantum spacetime. And once again, the zeros of this partition function in the complex coupling plane can signal phase transitions between different phases of spacetime—for instance, a transition from a disconnected collection of geometric atoms to a smooth, extended universe like the one we live in. The search for the structure of reality itself can be guided by finding where the partition function vanishes.

Perhaps the most startling and beautiful application of Fisher zeros is not in physics at all, but in the abstract realm of pure mathematics. It reveals a hidden unity between the physical world and the world of patterns and structures that is nothing short of breathtaking.

Consider a classic mathematical puzzle: map coloring. How many ways can you color a map with $q$ colors such that no two adjacent countries share the same color? This question is codified in an object called the chromatic polynomial, $P_G(q)$, which tells you the number of valid colorings for any given number of colors $q$. Now for the leap: in 1968, it was discovered that this purely combinatorial polynomial is physically realized as the partition function of a specific statistical model—the antiferromagnetic Potts model—in the limit of zero temperature. The number of colors, $q$, plays the role of a parameter in the partition function.

This means that the complex roots of the chromatic polynomial—the values of $q$ for which the number of colorings is formally zero—are none other than the Fisher zeros of a physical system! A physical phase transition in this model corresponds to a change in the large-scale coloring properties of the graph. When we calculate the zeros for a simple graph like a five-vertex cycle, we find elegant roots like $q = 1 \pm i$. These "impossible" complex numbers of colors carry deep information about the structure of the graph, all through the lens of physics.

The story culminates in an even more profound connection: to Knot Theory. A knot is just a closed loop of string in three-dimensional space that cannot be untangled. A central goal of knot theory is to find ways to tell if two complicated-looking knots are actually the same. In the 1980s, a powerful new tool was discovered: the Jones polynomial, an algebraic expression that is a "fingerprint" of a knot. And in a stunning turn of events, it was realized that the Jones polynomial, too, can be formulated as a partition function of a statistical mechanics model.

The variable $t$ in the Jones polynomial corresponds to a complex temperature-like parameter. Therefore, the complex zeros of the Jones polynomial are Fisher zeros. These zeros, rooted in the principles of phase transitions, are now fundamental characteristics of a knot's topology. They are part of the knot's very identity. Finding the zeros of the polynomial for, say, the cinquefoil knot is equivalent to probing the analytic structure of a physical partition function. The bridge between the tangible physics of phase transitions and the ethereal mathematics of topology is complete.

From magnets to black holes, from the forces of nature to the coloring of maps and the twisting of knots, the principle of Fisher zeros provides a common language. It is a powerful reminder that the deep truths of the universe often reveal themselves in simple, elegant ideas that resonate far beyond their original context, connecting disparate fields of thought in a grand, unified tapestry.