Expansion of Free Energy

Phase transitions, the sudden reorganizations of matter from disordered states to ordered ones are among the most dramatic events in nature. From a magnet gaining its power to a liquid freezing into a crystal, these transformations pose a fundamental question: how can we describe the collective behavior of countless particles without tracking each one? This article addresses this challenge by delving into the elegant framework of Landau theory, which uses the expansion of free energy as a conceptual tool of breathtaking power. We will first explore the core principles and mechanisms, revealing how the abstract concept of symmetry dictates the behavior of a system near a critical point. Following this, we will journey through its diverse applications, demonstrating how this single idea unifies our understanding of phenomena ranging from ferromagnetism and superconductivity to the structural changes in advanced materials.

Imagine you are watching a pot of water come to a boil. At one moment, it's a placid liquid; the next, it's a turbulent chaos of steam bubbles. Or think of a piece of iron, mundane and non-magnetic at room temperature, which, when cooled, suddenly gains the power to snap up paperclips. These are phase transitions, and they represent some of the most dramatic and profound events in nature. They are not just about a substance changing its state, but about a system spontaneously reorganizing itself, switching from a state of high disorder to one of remarkable collective order.

How can we hope to describe such a sudden and collective rearrangement of countless atoms? We can’t possibly track every single particle. The genius of physics often lies in finding a simpler, macroscopic language to describe complex microscopic behavior. For phase transitions, this language is the expansion of free energy, a conceptual tool of breathtaking power and elegance developed by the great Soviet physicist Lev Landau. It doesn't rely on the messy details of atomic interactions, but on a principle of almost aesthetic purity: ​​symmetry​​.

To talk about organization, we first need a way to measure it. We need a quantity that is zero in the disorganized, high-temperature phase and takes on a non-zero value in the organized, low-temperature phase. This quantity is called the ​​order parameter​​, typically denoted by the Greek letter eta, $\eta$. For a magnet, the order parameter is the net magnetization, $M$. Above the critical ​​Curie temperature​​, $T_c$, the individual atomic magnets (spins) point in random directions, so their effects cancel out, and the total magnetization is zero ($M=0$). Below $T_c$, the spins spontaneously align, creating a net magnetization ($M \neq 0$). For the transition from liquid to solid, the order parameter could describe the periodic density variation that defines the crystal lattice. It is the hero of our story.

Now, what governs the behavior of this order parameter? Nature is fundamentally lazy, in a very specific sense. Any physical system will spontaneously arrange itself to minimize a quantity called the ​​free energy​​, which we'll call $F$. Think of it as the ultimate decider. The value of the order parameter we actually observe in a system at a given temperature is simply the one that makes the free energy as low as possible.

So, the whole problem of a phase transition boils down to this: how does the shape of the function $F(\eta)$ change with temperature, such that the location of its lowest point suddenly jumps from $\eta=0$ to some $\eta \neq 0$?

This is where the magic begins. We might not know the exact, complicated formula for $F(\eta)$, but we can figure out its general shape using a simple and profound idea. The free energy function must have the same symmetries as the physical system it describes.

Let's stick with our ferromagnet. In the absence of an external magnetic field, there is no physical difference between a state where all spins point "up" (magnetization $+M$) and a state where all spins point "down" (magnetization $-M$). The underlying laws of physics don't have a preference for "up" or "down". Therefore, the free energy of these two states must be identical. Mathematically, this means:

$F(M) = F(-M)$

The function must be even! This simple statement is a tremendously powerful constraint. Landau's brilliant insight was to expand the free energy as a power series in the order parameter around $\eta=0$. A general expansion would look like:

$F(\eta) = F_0 + A\eta + B\eta^2 + C\eta^3 + D\eta^4 + \dots$

But if the function must be even, all the terms with odd powers of $\eta$ must vanish! The $A\eta$ term must be gone, the $C\eta^3$ term must be gone, and so on. Why? Because if $C$ were not zero, for instance, then $F(\eta) - F(-\eta) = 2C\eta^3$, which would not be zero. The only way for $F(\eta) = F(-\eta)$ to hold for any value of $\eta$ is if all the odd coefficients ($A, C, \dots$) are exactly zero.

This symmetry argument is the heart of the theory. It tells us that for any system where the states $+\eta$ and $-\eta$ are equivalent—be it a magnet, a ferroelectric crystal with inversion symmetry, or many others—the free energy expansion must take the form:

$F(\eta) = F_0 + B\eta^2 + D\eta^4 + \dots$

We've already simplified the problem immensely without knowing a single detail about the atoms! We also keep the $D\eta^4$ term for a crucial reason. As we'll see, the coefficient $B$ will become negative to favor ordering. If we only had the $B\eta^2$ term, the free energy would plummet to negative infinity, which is a physical disaster. The $D\eta^4$ term, with $D > 0$, acts like a safety wall, ensuring the free energy has a stable minimum at a finite value of $\eta$.

What if a system lacks this symmetry? For example, some crystal structures do not have a center of inversion. In such a case, the cubic term $C\eta^3$ might be allowed to exist. Its presence has dramatic consequences, often leading to a more abrupt, "first-order" transition (like water boiling) rather than the smooth, "second-order" onset of magnetization. The very structure of the free energy expansion is a fingerprint of the system's underlying symmetries.

Now we have our stage, $F(\eta) = F_0 + B\eta^2 + D\eta^4$. The coefficients $B$ and $D$ are not universal constants; they depend on temperature. This is where the action happens.

• ​​High Temperature ($T > T_c$)​​: In the hot, disordered phase, we know the system prefers $\eta=0$. For this to be the minimum of our free energy function, the curve must look like a simple bowl with its bottom at $\eta=0$. This requires the coefficient of the $\eta^2$ term to be positive, $B > 0$.

• ​​Low Temperature ($T T_c$)​​: In the cold, ordered phase, the system spontaneously chooses a non-zero order parameter, $\eta \neq 0$. For this to happen, our free energy curve must change shape. The bottom of the bowl at $\eta=0$ must pop up into a hump, and two new, lower minima must appear on either side. This "double-well" potential is the hallmark of a system that has undergone a spontaneous symmetry breaking. This shape change requires the coefficient of the $\eta^2$ term to become negative, $B 0$.

The phase transition occurs precisely at the critical temperature $T_c$ where the coefficient $B$ passes through zero. The simplest way to model this is to assume $B$ depends linearly on temperature near $T_c$:

$B(T) = \alpha(T-T_c)$

where $\alpha$ is some positive constant. This simple assumption is the engine of the whole theory. It connects the coefficients of our abstract expansion to the one thing we can easily control in an experiment: temperature.

With this, we can now make a concrete prediction! Let's find the new minimum for $T T_c$. We minimize the free energy by taking its derivative with respect to $\eta$ and setting it to zero:

$\frac{dF}{d\eta} = 2B\eta + 4D\eta^3 = 2\alpha(T-T_c)\eta + 4D\eta^3 = 0$

This equation has one solution $\eta=0$ (the unstable hump) and two symmetric solutions:

$\eta^2 = -\frac{B}{2D} = -\frac{\alpha(T-T_c)}{2D} = \frac{\alpha(T_c-T)}{2D}$

So, the equilibrium value of the order parameter is:

$\eta_{eq} = \pm \sqrt{\frac{\alpha}{2D}} \sqrt{T_c - T}$

The order parameter grows from zero as the square root of the distance in temperature from the critical point! This is a universal prediction. It doesn't matter if we are talking about a magnet, a superconductor, or a ferroelectric. If the transition is governed by the same symmetry, the behavior near the critical point should be the same. In fact, this is exactly the result one gets by doing a much more complicated calculation starting from a specific microscopic model of magnetism, a beautiful confirmation of the power of the symmetry-based approach.

The beauty of the Landau framework is its adaptability. The principle of symmetry is universal. What if the order parameter is more complicated than a simple number that can be positive or negative?

Consider a magnet where the spins can point anywhere in a 2D plane, like the XY model. The order parameter is now a two-component vector, $\vec{M} = (M_x, M_y)$. The key symmetry of this system (in zero field) is rotational invariance in the plane. The physics, and therefore the free energy, cannot change if we rotate our coordinate system.

So, what kind of mathematical combinations of $M_x$ and $M_y$ are invariant under rotation? The answer is any combination that depends only on the length of the vector, not its direction. The most fundamental rotational invariant is the square of the vector's magnitude: $|\vec{M}|^2 = M_x^2 + M_y^2$. Any other rotational invariant (like $|\vec{M}|^4, |\vec{M}|^6, \dots$) is just a power of this fundamental one.

Therefore, the Landau free energy for this system must be an expansion in powers of $|\vec{M}|^2$:

$F(T, \vec{M}) = F_0 + a(T)|\vec{M}|^2 + b(T)(|\vec{M}|^2)^2 + \dots$

The symmetry principle, in one clean stroke, dictates the form of the free energy, sidestepping the immense complexity of building it from individual component terms. The logic remains the same: the transition happens when the coefficient $a(T)$ changes sign, causing the system to spontaneously pick a direction in the plane and acquire a non-zero magnetization of a specific length.

Landau's theory is a masterpiece of physical reasoning. But, like all great theories, it's important to understand its boundaries—to be honest about what it simplifies and what it neglects.

First, the very idea of expanding the free energy as a truncated power series ($F \approx F_0 + B\eta^2 + D\eta^4$) is an approximation. Such an expansion is only accurate when the order parameter $\eta$ is small. Where is $\eta$ small? Only in the immediate vicinity of the critical temperature $T_c$. As we cool the system far below $T_c$, the order parameter grows large, and the higher-order terms we ignored ($\eta^6$, $\eta^8$, etc.) become significant. This is why the Ginzburg-Landau theory for superconductors, for example, works beautifully near its $T_c$, but gives incorrect predictions at very low temperatures. The theory is a brilliant description of the onset of order.

Second, and more profoundly, the simple Landau theory assumes that the order parameter $\eta$ has the same value everywhere in the system. It describes a perfectly uniform state. This is what physicists call a ​​mean-field theory​​. In reality, especially near a critical point, the system is a roiling, fluctuating mess. There are temporary domains of order forming and dissolving within a sea of disorder. These spatial fluctuations are completely ignored in our simple model.

A more advanced version of the theory, known as ​​Ginzburg-Landau theory​​, rectifies this by adding a term to the free energy that represents the energy cost of these spatial variations, a term proportional to the square of the gradient of the order parameter, $(\nabla\eta)^2$. The neglect of this gradient term is the central approximation that makes the basic Landau theory "mean-field" and is the reason it fails to predict certain details (like critical exponents) with perfect accuracy.

Even so, the picture it paints is remarkably robust. By starting with nothing more than the abstract concept of symmetry and the idea that nature minimizes free energy, we have unveiled the universal mechanism behind a vast class of transformations in the world around us. We have built a framework that is not just a calculation, but a way of thinking—a testament to the profound idea that the most fundamental rules of the universe are often the most elegant.

In our previous discussion, we uncovered a remarkably powerful secret of nature: that near a critical point of change, the complex tapestry of interactions within a system often simplifies dramatically. We found that the free energy—the ultimate arbiter of the system's fate—can be expressed as a simple polynomial of an "order parameter," with the allowed terms dictated solely by symmetry. This idea, the heart of Landau theory, might seem like a convenient mathematical trick. But is it? Or is it a window into a deeper reality?

In this chapter, we embark on a journey to see this principle in action. We will discover that this single, elegant idea is a master key, unlocking the secrets of an astonishing variety of phenomena across physics, chemistry, and materials science. We are about to witness how this "universal language" of free energy expansions describes everything from the magnetism of iron and the structure of crystals to the quantum mystery of superconductivity.

Let's start with one of the most familiar phase transitions: a ferromagnet losing its spontaneous magnetization as it's heated past its Curie temperature, $T_c$. Below $T_c$, countless atomic spins are aligned; above $T_c$, they are in thermal chaos. The order parameter is the net magnetization, $M$. The theory we developed gives us the simplest possible free energy expansion that respects the up/down symmetry of magnetism: $F \approx \frac{a}{2}(T-T_c)M^2 + \frac{b}{4}M^4$.

From this humble starting point, a cascade of precise, testable predictions emerges. By simply asking the system to find its minimum energy state, we can calculate exactly how the spontaneous magnetization $M_s$ should grow as we cool just below the critical point. The theory predicts $M_s \propto (T_c - T)^{1/2}$. This isn't just some vague trend; it's a quantitative law with a specific "critical exponent" of $\beta = 1/2$.

But there's more. We can ask how the disordered phase above $T_c$ responds to a small external magnetic field. This response is the magnetic susceptibility, $\chi$. Our simple model predicts that as you approach the transition from above, this susceptibility should diverge, with the material becoming infinitely sensitive to the field right at the critical point. Again, the theory gives a precise law: $\chi \propto (T - T_c)^{-1}$, defining another critical exponent, $\gamma=1$.

Finally, we can ask what happens to the heat capacity, $C$, which tells us how much energy the material absorbs for a given change in temperature. The free energy expansion reveals that while the energy itself is continuous, its second derivative with respect to temperature is not. This leads to a stunning prediction: a finite, sudden jump in the heat capacity right at $T_c$. All three of these behaviors—the onset of order, the divergent response, and the jump in heat capacity—are hallmarks of second-order phase transitions, and they all fall out of our simple polynomial. The details of the atomic interactions seem to have vanished, replaced by the pure logic of symmetry and thermodynamics.

The world, however, is not always so gentle. Some transitions are abrupt and violent, like water boiling into steam. These are "first-order" transitions, characterized by a latent heat and a discontinuous jump in the order parameter itself. Can our theory handle this?

Absolutely. The key lies in the coefficients of our expansion. Consider a ferroelectric material, where the order parameter is electric polarization, $P$. If the material properties are such that the coefficient of the $P^4$ term is negative, the free energy landscape is dramatically altered. The system is now unstable until we add a stabilizing $P^6$ term. The resulting energy landscape has two distinct wells separated by a barrier, even right at the transition temperature. To change its state, the system must "jump" from the disordered state ($P=0$) to an ordered one ($P \neq 0$). This jump involves the release or absorption of latent heat, a quantity our theory can calculate precisely.

This raises a fascinating question: what if we could tune the material's properties to be right on the borderline between a continuous second-order transition and an abrupt first-order one? This special state of matter is called a ​​tricritical point​​. At this point, the coefficient of the fourth-order term vanishes precisely as the second-order term does. By carefully adjusting parameters like pressure, we can guide a system to this crossroads. At a tricritical point in an antiferromagnet, for instance, the free energy becomes dominated by the sixth-order term, leading to new, unique critical exponents, such as the order parameter appearing as $L \propto (T_c - T)^{1/4}$.

This tuning doesn't always have to be external. Internal couplings within the material can have the same effect. In some ferroelectrics, the polarization is coupled to the physical strain of the crystal lattice—a phenomenon called electrostriction. This coupling effectively "renormalizes" the coefficients of the free energy expansion. The strain accommodates the polarization in a way that can cancel out the positive $P^4$ term, driving the system from a second-order to a first-order transition via a tricritical point. The simple polynomial is now a dynamic stage where different physical effects compete to dictate the very nature of the transformation.

So far, our order parameters have been simple scalars. But many of the most important transitions in nature involve changes in structure and shape. Imagine a crystal transforming from a perfect cube into a rectangular (tetragonal) shape, a process known as a martensitic transformation, which is fundamental to shape-memory alloys. The order parameter is no longer a single number but a collection of strain components that describe the distortion.

Here, the role of symmetry becomes paramount. The free energy expansion must be invariant not just under a simple sign flip, but under all the rotational symmetries of the original cubic crystal. Constructing this expansion is like solving a beautiful puzzle. Abstract mathematics, in the form of group theory, provides the exact rules. It tells us that we must build our polynomial not from the strain components directly, but from specific, invariant combinations of them. For the cubic-to-tetragonal case, we find there are two fundamental "building blocks"—a quadratic invariant $Q^2$ and a cubic one $I_3$—from which all other invariants can be constructed. The resulting free energy is a rich and complex function that correctly predicts which crystallographic directions the material will choose to distort along. This is a breathtaking demonstration of how abstract symmetry principles dictate the concrete, physical properties of a material.

At this point, you might still harbor a suspicion. This all works beautifully, but is it just a phenomenological model, a clever story we tell that happens to fit the facts? Or does it reflect a deeper, underlying truth? The answer lies in one of the most profound phenomena in physics: superconductivity.

Superconductivity, where electrical resistance vanishes completely, is a purely quantum mechanical effect. Its description requires the full machinery of many-body quantum field theory. Yet, miraculously, the onset of superconductivity can be described perfectly by a Ginzburg-Landau free energy expansion. The connection was made through the microscopic theory of Gor'kov. By taking the complex quantum equations and making a single, valid approximation—that we are very close to the critical temperature $T_c$—the entire theory collapses, simplifying into our familiar polynomial form.

This procedure is not just qualitative; it allows us to derive the Landau coefficients from first principles. When we do this for the quadratic term, we find that $\alpha(T)$ is not just like $a(T-T_c)$, it is proportional to $(T-T_c)$, with the proportionality constant determined by fundamental properties of the metal like its density of states. This is a triumph. The Landau expansion is not just a guess; it is the emergent, macroscopic manifestation of the underlying quantum reality near a critical point.

Our entire discussion has assumed the system is uniform in space. But what if the order parameter varies from point to point? The "Ginzburg" part of Ginzburg-Landau theory addresses this by adding terms involving spatial gradients, $(\nabla M)^2$, to the free energy. These terms represent an energy cost for creating variations, like the surface tension on a domain wall between north- and south-pointing magnetic regions.

Usually, these terms favor uniformity. But in certain materials that lack a center of symmetry, the crystal structure allows for a peculiar kind of gradient term, a "Lifshitz invariant," that is linear in the gradient, like $M_x \frac{\partial M_y}{\partial z} - M_y \frac{\partial M_x}{\partial z}$. This term, also known as the Dzyaloshinskii-Moriya interaction, doesn't just penalize any change; it actively favors a specific direction of twist. The lowest energy state is no longer a uniform alignment, but a continuous, swirling spiral. This competition between the exchange interaction favoring alignment and the DMI favoring twisting can lead to the formation of stable, particle-like magnetic vortices known as ​​skyrmions​​. These are topological objects, whose existence is dictated by the subtle symmetries of the crystal, as revealed by the gradient expansion of the free energy. Our simple polynomial has now led us to the cutting edge of modern physics, connecting thermodynamics to the fascinating world of topology and spintronics.

Finally, it is important to realize that the power of series expansions is a recurring theme throughout thermodynamics, extending beyond phase transitions. Consider a real gas, as opposed to an idealized one. The interactions between gas molecules cause its behavior to deviate from the simple ideal gas law. This deviation can be systematically captured by the virial expansion, which expresses the pressure as a power series in the gas density, $\rho$.

Starting from this expansion, we can construct the corresponding Helmholtz free energy for the real gas. This potential is a function of volume (or density). However, it is often more convenient to work with the Gibbs free energy, which is a function of pressure. Using the powerful mathematical tool of a Legendre transform, we can convert our Helmholtz free energy into the Gibbs free energy, which naturally appears as a power series in pressure. This exercise showcases a general and elegant principle: different thermodynamic potentials are simply different "views" of the same system, and the logic of series expansions provides a common language to translate between them.

From the universal behavior of magnets to the quantum origins of superconductivity, from the structural birth of shape-memory alloys to the topological dance of skyrmions, the principle of expanding the free energy based on symmetry has proven to be an astonishingly faithful guide. It is a testament to the idea that in the face of complex change, nature often chooses the simplest and most elegant path.