Fluctuations, Dimensionality, and Mean-Field Theory Breakdown

Hello! In our previous lesson, we saw a key feature—and a key weakness—of mean-field theory (MFT). By applying it to both the Ising and classical Heisenberg models, we found that MFT predicts the same set of critical exponents for both, despite their different order parameter symmetries. This universality is a powerful result, but it hints that MFT is missing crucial physics, as real-world experiments show that exponents do depend on properties like symmetry and spatial dimension. The conclusion was that MFT fails because it neglects fluctuations.

Today, we will delve into that failure. Our goal is to qualitatively discuss the role of fluctuations and dimensionality in the breakdown of mean-field theory. We will explore what fluctuations are, how they behave near a critical point, and how their influence is mediated by the dimensionality of the system. This will lead us to a formal criterion for the validity of MFT, which introduces the important concept of an upper critical dimension.

1. The Mean-Field Approximation and What It Misses

Let's briefly recall the central assumption of MFT. Whether we use the molecular field picture or the variational free energy, the core idea is to replace the complex, fluctuating interactions of a given spin with its neighbors by an interaction with a single, static, average field.

Mean-Field Approximation:

$$\sum _{⟨i,j⟩}{\vec{S}}_i\cdot {\vec{S}}_j⟶\sum _i{\vec{S}}_i\cdot ⟨{\vec{h}}_{\text{eff}}⟩$$

This approximation linearizes a many-body problem, making it solvable. However, it completely erases the possibility that the neighbors of spin ${\vec{S}}_i$ might be pointing in directions that deviate from the global average. These local deviations from the average order parameter are called fluctuations. By design, MFT averages them away. Near a phase transition, this omission becomes catastrophic.

2. Visualizing Fluctuations: Critical Opalescence

Fluctuations are not just a theoretical construct; they have dramatic, observable consequences. One of the most famous examples is critical opalescence, which occurs in a fluid at its liquid-gas critical point.

To see this phenomenon and understand its origin, please watch the following segment from an MIT OpenCourseWare lecture by Prof. Mehran Kardar.

As you saw, at the critical point, the fluid is teeming with patches of higher and lower density that are comparable in size to the wavelength of visible light. These density fluctuations scatter light so strongly that the fluid becomes opaque. This is a direct visualization of the collective behavior that MFT ignores. Away from the critical point, fluctuations are small and localized. But as we approach $T_c$, they not only grow in amplitude but also become correlated over increasingly large distances.

3. From Fluctuations to Diverging Correlations

We can formalize the connection between fluctuations and macroscopic properties. The response of a system to an external field, such as magnetic susceptibility ($\chi$) or compressibility (${\kappa }_T$), is directly related to the fluctuations of its order parameter.

For a magnetic system with local magnetization $m(\vec{r})$, the susceptibility is related to the two-point correlation function $G(\vec{r})=⟨m(\vec{r})m(0)⟩-⟨m{⟩}^2$:

$$\chi \propto \int d^d\vec{r}G(\vec{r})$$

This is a version of the fluctuation-dissipation theorem. It states that the system's response to an external probe ($\chi$) is determined by its internal, spontaneous fluctuations ($G(\vec{r})$).

The correlation function $G(\vec{r})$ typically decays exponentially with distance, characterized by the correlation length, $\xi$:

$$G(\vec{r})\sim e^{-|\vec{r}|/\xi }$$

As we approach the critical point, we know that the susceptibility diverges ($\chi \to \infty$). For the integral over $G(\vec{r})$ to diverge, the range of the integral must also diverge. This implies that the correlation length must go to infinity: $\xi \to \infty$.

The following clip makes this crucial connection explicit.

This is the heart of critical phenomena: at $T_c$, fluctuations are correlated across the entire system. Every spin "knows about" every other spin, no matter how far apart they are. MFT, which assumes each spin only feels a local average field, cannot possibly capture this physics.

4. The Crucial Role of Dimensionality

The fact that $\xi \to \infty$ at $T_c$ is universal, but the consequences of these large-scale fluctuations are profoundly affected by the spatial dimension $d$ of the system.

Imagine creating a very long-wavelength fluctuation, like a slow twist in the direction of magnetization across the system. The energy cost of such a fluctuation is related to the gradient of the order parameter, $(\nabla m{)}^2$.

• In high dimensions, there are many neighbors to enforce conformity, making it energetically costly to deviate from the average order. Fluctuations are more easily suppressed.
• In low dimensions, a spin has fewer neighbors. It's "easier" for thermal energy to create large-scale fluctuations that can disrupt or even destroy long-range order.

This effect is so strong that it leads to a rigorous theorem. As mentioned in the previous lesson, the Mermin-Wagner theorem states that for systems with a continuous symmetry (like the Heisenberg model), spontaneous symmetry breaking is impossible at any non-zero temperature in dimensions $d\le 2$. This means there is no phase transition ($T_c=0$), because long-wavelength fluctuations are always sufficient to destroy any long-range order.

This defines a lower critical dimension, $d_l$, below which fluctuations are so powerful they prevent ordering at any finite temperature.

• For continuous symmetries (Heisenberg, XY models), $d_l=2$.
• For discrete symmetries (Ising model), $d_l=1$.

The predictions of MFT, which always yield a finite $T_c$ regardless of dimension, are in stark contrast with these exact results.

5. The Ginzburg Criterion and the Upper Critical Dimension

We've established that MFT fails because it neglects fluctuations. This leads to a natural question: is there a regime where this neglect is justified? The Ginzburg criterion provides the answer.

The core idea is simple: MFT should be valid as long as the fluctuations of the order parameter within a "correlation volume" (${\xi }^d$) are small compared to the average value of the order parameter itself.

To see how this is formally derived and what it implies, please read the following section from a set of lecture notes.

As the text shows, the condition for MFT to be valid is:

$$d\cdot {\nu }_{\text{MF}}-2{\beta }_{\text{MF}}-{\gamma }_{\text{MF}}>0$$

Plugging in the mean-field exponents we've calculated (${\beta }_{\text{MF}}=1/2$, ${\gamma }_{\text{MF}}=1$, ${\nu }_{\text{MF}}=1/2$, where $\nu$ is the exponent for the correlation length $\xi \sim |T-T_c{|}^{-\nu }$), we get:

$$d\cdot \frac{1}{2}-2\cdot \frac{1}{2}-1>0⟹\frac{d}{2}-2>0⟹d>4$$

This defines the upper critical dimension, $d_c=4$.

• For $d>d_c=4$, fluctuations are not strong enough to affect the critical behavior, and the mean-field exponents are correct.
• For $d<d_c=4$, fluctuations dominate near the critical point, and MFT fails. The critical exponents will take on different, non-mean-field values.
• For $d=d_c=4$, MFT is almost correct, but there are subtle logarithmic corrections to the scaling laws.

Since we live in a three-dimensional world, MFT is not expected to give quantitatively correct exponents for most standard phase transitions, such as the liquid-gas or Ising ferromagnetic transitions.

Test your understanding!

Imagine you are studying a system where interactions are very long-ranged, for example, where every spin interacts with every other spin in the system with equal strength. How would this affect the validity of mean-field theory and the role of dimensionality?

Show answer

In such a system, each spin genuinely feels the average field produced by all other spins. The local environment of a spin is identical to the global average. In this case, the mean-field approximation becomes exact. Fluctuations are suppressed because any deviation by a single spin is overwhelmed by the collective field of all other N-1 spins. The concept of dimensionality becomes less relevant because the interaction is no longer local (i.e., with nearest neighbors). This is the physical scenario where MFT is expected to work perfectly, even for d < 4.

6. The Ginzburg "Window": Why MFT Can Still Be Useful

The Ginzburg criterion provides one final, crucial insight. It tells us not just if MFT will fail, but also when. The breakdown occurs within a specific temperature range around $T_c$, known as the Ginzburg or critical "window". The size of this window is given by the Ginzburg reduced temperature, $t_G$:

$$|t_G|=\left|\frac{T-T_c}{T_c}\right|\propto {\left(\frac{a}{{\xi }_0}\right)}^{\frac{2d}{4-d}}$$

Here, $a$ is the microscopic lattice spacing and ${\xi }_0$ is the "bare" correlation length far from $T_c$.

• If ${\xi }_0$ is small (on the order of the lattice spacing $a$), as in magnetic materials or simple fluids, then $t_G$ is large (e.g., $10^{-1}$ or $10^{-2}$). The critical window is wide and experimentally accessible. We will observe the breakdown of MFT and see true, non-classical critical exponents.
• If ${\xi }_0$ is very large compared to $a$, then $t_G$ becomes extremely small. This is the case for conventional superconductors, where the Cooper pairs have a size (${\xi }_0$) of hundreds of nanometers. For these systems, $t_G$ can be as small as $10^{-14}$.

This means that for a superconductor, you would have to get impossibly close to $T_c$ to see the breakdown of mean-field theory. In any practical experiment, you are always outside this tiny critical window, in a region where MFT provides an excellent description. This is why the BCS theory of superconductivity, which is a form of mean-field theory, is so successful.

Conclusion

Key Takeaways:

• Mean-field theory breaks down because it fundamentally ignores spatial fluctuations of the order parameter.
• Near a critical point, these fluctuations become correlated over a diverging length scale $\xi$, leading to observable phenomena like critical opalescence.
• The dimensionality of the system determines the impact of these fluctuations. In low dimensions ($d\le d_l$), they can completely destroy long-range order.
• The Ginzburg criterion quantifies the breakdown of MFT. It defines an upper critical dimension, $d_c=4$, above which MFT is exact.
• For $d<4$, MFT fails within a critical temperature window around $T_c$. The size of this window depends on the microscopic parameters of the system, explaining why MFT works well for some systems (like superconductors) but not others (like magnets).

Preview of the Next Lesson:
We have thoroughly explored the successes and, more importantly, the failures of mean-field theory. To move forward, we need a framework that embraces fluctuations from the start. This leads us to Landau Theory. In the next lesson, we will begin Module 5 by constructing the Landau free energy expansion based on symmetry arguments, providing a more general and powerful phenomenological language to describe phase transitions.