Ginzburg Criterion and Mean-Field Theory Validity

Welcome back. In our previous lesson, we systematically derived the set of "mean-field" critical exponents from the analytical structure of the Landau free energy. This exercise revealed a profound aspect of theoretical physics: the universality of physical laws near a critical point, where exponents like $β = 1/2$ and $γ = 1$ emerge regardless of microscopic details, at least within this theoretical framework.

However, a core principle of the research methodology we are studying is to always question the validity of our approximations. Landau theory, by focusing on the most probable value of the order parameter, explicitly neglects the role of thermal fluctuations. Near a critical point, where the correlation length diverges, these fluctuations become long-ranged and are expected to dominate the physics. This lesson directly confronts this issue. We will develop a quantitative tool, the Ginzburg criterion, to determine precisely when these fluctuations invalidate the mean-field picture. This will not only define the limits of our theory but also resolve a fascinating experimental puzzle.

The Puzzle: When Does Mean-Field Theory Work?

The mean-field exponents we derived are not just a theoretical curiosity; they are a concrete prediction to be tested against experiment. The results are mixed, presenting a puzzle.

On one hand, many systems show clear deviations. For example, the specific heat of liquid helium at its superfluid transition does not show the finite jump ( $α = 0$ ) predicted by mean-field theory. Instead, it exhibits a sharp, logarithmic-like divergence known as the "lambda transition," as you can see in the experimental data below.

The specific heat of Helium-4 near the lambda transition at T ≈ 2.17 K. The shape of the curve, resembling the Greek letter $\lambda$, shows a divergence, in stark contrast to the simple discontinuity predicted by mean-field theory.

On the other hand, some systems agree with mean-field predictions with remarkable accuracy. This is most famously the case for conventional superconductors. The video lecture you'll see next presents a table summarizing experimental exponents for various systems, highlighting this discrepancy.

4. The Landau-Ginzburg Approach Part 3

This segment from an MIT OpenCourseWare lecture by Prof. Mehran Kardar compares the experimentally measured critical exponents for several systems with the predictions of mean-field theory. This comparison sets the stage for our main question.

Please watch the portion from the discussion of exponent identities. Pay close attention to the table comparing exponents for ferromagnets, superfluids, liquid-gas systems, and superconductors. The key takeaway is the stark contrast: most systems in d=3 disagree with mean-field theory, while superconductors agree almost perfectly. This is the puzzle we need to explain.

Our goal is to understand the physical principle that separates these two classes of systems. The answer lies in quantifying the importance of fluctuations.

The Ginzburg Criterion: A Heuristic Derivation

The essence of the Ginzburg criterion is a self-consistency check of the mean-field approximation. The approximation is valid if the thermal fluctuations of the order parameter are small compared to the value of the order parameter itself.

Let's formalize this. We consider the order parameter to be composed of its spatial average (the mean-field value, $m_{0}$ ) and a spatially varying fluctuation, $δ m (r)$ .

m (r) = m_{0} + δ m (r)

Mean-field theory is valid if the fluctuations are "small." But small compared to what? And on what length scale? Near a transition, fluctuations are correlated over a distance $ξ$ , the correlation length. It is therefore natural to compare the average fluctuation within a correlation volume, $ξ^{d}$ , to the mean-field order parameter $m_{0}$ .

The criterion can be stated as:
The average squared fluctuation of the order parameter within a correlation volume must be much smaller than the square of the mean-field order parameter.

⟨(δ m)^{2} ⟩_{ξ^{d}} ≪ m_{0}^{2}

This condition ensures that the system is, on average, well-described by its mean value $m_{0}$ .

The following video provides a clear introduction to the Landau-Ginzburg theory, which incorporates these spatial fluctuations, and lays out the derivation of the criterion based on this core idea.

Landau Ginzburg theory of Phase Transitions

This lecture by Jos Thijssen introduces Landau-Ginzburg theory and derives the Ginzburg criterion by assessing the importance of fluctuations.

First, watch the brief introduction from the introduction, where the goal of deriving the Ginzburg criterion is stated. Next, watch the segment from the introduction of spatial fluctuations. This part explains how we move from a uniform order parameter m (Landau theory) to a spatially varying field m(\mathbf{r}) (Landau-Ginzburg theory) by introducing a gradient term, K(\nabla m)^2, into the free energy. This term penalizes variations in the order parameter. Finally, watch the crucial segment from the derivation of the Ginzburg criterion. This part explains that the criterion arises from demanding that the fluctuation terms we neglected in the mean-field approximation are indeed small compared to the mean-field solution. This leads directly to the condition \langle (\delta m)^2 \rangle \ll m_0^2.

The Upper Critical Dimension

We now have a formal criterion, but to use it, we need to express both sides of the inequality $⟨(δ m)^{2} ⟩ ≪ m_{0}^{2}$ in terms of the reduced temperature $t = (T - T_{c}) / T_{c}$ . This will reveal how the validity of mean-field theory depends on the dimensionality of space, $d$ .

From our previous lesson on mean-field exponents, we know:

Mean-field order parameter: $m_{0}^{2} \sim ∣ t ∣^{2 β} = ∣ t ∣$ for $T < T_{c}$ .
Correlation length: $ξ \sim ∣ t ∣^{- ν} = ∣ t ∣^{- 1/2}$ .

Now we need the left side: the fluctuations. The calculation involves relating the local fluctuations to the correlation function, which in turn depends on the correlation length $ξ$ . The result of this calculation, which is detailed in the video below, is that the mean-squared fluctuation within a correlation volume scales as:

⟨(δ m)^{2} ⟩_{ξ^{d}} \sim ξ^{2 - d}

This is a non-trivial result from statistical field theory, but it is physically intuitive: as the dimension $d$ increases, a fluctuation at a point is "felt" less by its neighbors, so local fluctuations become less correlated and smaller in magnitude for a fixed $ξ$ .

Let's see how these pieces combine to give the final result.

Landau Ginzburg theory of Phase Transitions

The second half of Jos Thijssen's lecture calculates the scaling of the fluctuation term and combines it with the mean-field results to determine the upper critical dimension.

First, watch the derivation of the fluctuation scaling from the analysis of fluctuations. The key result to focus on is how the local fluctuation \langle (\delta m_r)^2 \rangle is related to an integral over the correlation function, ultimately yielding a scaling with the correlation length \xi. The video finds that the fluctuations in a local region scale as \xi^{2-d}. Now, watch the final step from the conclusion of the derivation. This segment substitutes the temperature dependencies of both the fluctuations (\sim |t|^{(d-2)/2}) and the mean-field order parameter (m_0^2 \sim |t|) into the Ginzburg criterion. The analysis shows that the inequality can only be satisfied as t \to 0 if the dimensionality d > 4.

This derivation leads to a remarkable conclusion. The Ginzburg criterion, $ξ^{2 - d} ≪ ∣ t ∣$ , becomes:

(∣ t ∣^{- 1/2})^{2 - d} ≪ ∣ t ∣^{1} ⟹ ∣ t ∣^{(d - 2) /2} ≪ ∣ t ∣^{1}

As we approach the critical point ( $t \to 0$ ), both sides of the inequality vanish. For the left side (fluctuations) to vanish faster than the right side (mean-field value), its exponent must be larger:

\frac{d - 2}{2} > 1 ⟹ d - 2 > 2 ⟹ d > 4

This defines $d_{u} = 4$ as the upper critical dimension for this universality class.

For $d > 4$ , fluctuations become irrelevant near the critical point, and mean-field theory is exact.
For $d < 4$ , fluctuations dominate the mean-field behavior, and mean-field theory breaks down. The critical exponents will deviate from the mean-field values.
At $d = 4$ , mean-field theory is mostly correct but with logarithmic corrections.

This explains why experiments in our three-dimensional world ( $d = 3$ ) generally do not match mean-field exponents. But it deepens the puzzle of superconductors: why do they follow mean-field behavior in $d = 3$ ?

The Ginzburg Criterion for Real Materials

The derivation above seems to give a simple binary answer based on dimension. However, the full Ginzburg criterion includes non-universal prefactors that depend on the material's microscopic properties. A more careful analysis, which compares the fluctuation contribution to the specific heat with the mean-field jump, allows us to quantify the temperature range where fluctuations are important.

The following text provides an excellent and concise derivation of this more practical form of the criterion.

[PDF] Chapter 2 Ginzburg-Landau Phenomenology

This chapter from a textbook by B.D. Simons provides a clear, alternative derivation of the upper critical dimension by analyzing fluctuation corrections to the specific heat. It then uses this to explain the experimental puzzle of superconductors.

First, read the section "Fluctuation Corrections". This part calculates the correction \delta C to the mean-field specific heat due to fluctuations and shows how its divergence for d<4 signals the breakdown of the approximation. Next, read the crucial section "Ginzburg Criterion". This is the heart of the matter. It establishes a "Ginzburg temperature" t_G, which defines the crossover region where fluctuations become dominant. The expression for t_G depends on microscopic parameters, particularly the zero-temperature correlation length \xi_0. Follow the argument for why t_G is experimentally accessible for superfluids but astronomically small for conventional superconductors.

The key insight is that fluctuations become important when the reduced temperature $∣ t ∣$ is smaller than a characteristic Ginzburg temperature, $t_{G}$ :

∣ t ∣ ≪ t_{G} \approx [\frac{a ^{d}}{( ξ _{0} ) ^{d} ( Δ C _{s p} / k _{B} )}]^{\frac{2}{4 - d}}

where $Δ C_{s p}$ is the mean-field jump in specific heat, $a$ is the interatomic spacing, and $ξ_{0}$ is the correlation length far from the transition (a measure of the "size" of the fundamental constituents).

This resolves our puzzle:

For superfluids and liquid-gas systems, $ξ_{0}$ is on the order of the interatomic distance ( $\sim$ angstroms). This makes $t_{G}$ relatively large (e.g., $1 0^{- 1}$ to $1 0^{- 2}$ ). Experiments can easily access the temperature regime $∣ t ∣ < t_{G}$ , where fluctuations dominate and mean-field theory fails.
For conventional superconductors, the order parameter describes Cooper pairs of electrons. Due to long-range Coulomb repulsion, these pairs are very large. The correlation length $ξ_{0}$ can be thousands of angstroms. This enormous value of $ξ_{0}$ makes $t_{G}$ incredibly small (e.g., $1 0^{- 16}$ ). No experiment can resolve temperatures that close to $T_{c}$ . Therefore, all experiments on conventional superconductors are performed in the regime $∣ t ∣ > t_{G}$ , where mean-field theory is perfectly valid.

The Ginzburg criterion is thus a beautiful example of how a theoretical tool can explain why a simple model works in some contexts and fails in others, based on the underlying physical scales of the system.

Conclusion

In this lesson, we have taken a crucial step in the methodology of theoretical physics: assessing the domain of validity of an approximation. By developing the Ginzburg criterion, we have established a quantitative measure for the importance of fluctuations, which were neglected in the simple Landau theory.

Key Takeaways:

The Ginzburg criterion provides a self-consistency check for mean-field theory by comparing the size of order-parameter fluctuations to the mean-field value itself.
This criterion leads to the concept of an upper critical dimension, $d_{u} = 4$ , above which fluctuations are irrelevant and mean-field theory becomes exact at the critical point.
For real materials in $d < 4$ , the criterion defines a crossover temperature $t_{G}$ . Mean-field theory is valid for $∣ t ∣ ≫ t_{G}$ , while fluctuation-dominated behavior appears for $∣ t ∣ ≪ t_{G}$ .
The value of $t_{G}$ depends on microscopic length scales. This explains why conventional superconductors ( $d = 3$ , large $ξ_{0}$ , tiny $t_{G}$ ) obey mean-field theory, while superfluids ( $d = 3$ , small $ξ_{0}$ , large $t_{G}$ ) do not.

The failure of mean-field theory in most low-dimensional systems is not an end, but a beginning. It tells us that a more powerful theoretical framework is needed to handle the scale-invariant physics of strong fluctuations at a critical point. This motivates our next topic: the Renormalization Group, a revolutionary theoretical tool designed specifically for this purpose.

Recommended resources

The Puzzle: When Does Mean-Field Theory Work?

The Ginzburg Criterion: A Heuristic Derivation

The Upper Critical Dimension

The Ginzburg Criterion for Real Materials

Conclusion