Efros-Shklovskii VRH

In the pristine, ordered world of a perfect crystal, electrons move freely, giving rise to electrical conductivity. But what happens in the chaotic landscape of a disordered material—like doped silicon or an amorphous film—especially when cooled to near absolute zero? At these low temperatures, thermal energy is scarce, and electrons become trapped, or "localized," in a random potential landscape, unable to roam. This raises a fundamental question: how can such materials conduct electricity at all? The answer lies in a quantum mechanical leap of faith known as variable-range hopping, where electrons tunnel between localized states.

This article addresses the evolution of our understanding of this phenomenon, moving beyond early models to a more complete picture. We will explore how the crucial, yet often overlooked, long-range Coulomb interaction between electrons themselves dramatically reshapes the transport physics. You will learn not just what variable-range hopping is, but how different physical assumptions lead to distinct and testable predictions.

The journey begins in the "Principles and Mechanisms" chapter, where we will uncover the electron's dilemma in choosing a hop and contrast the foundational Mott VRH model with the more refined Efros-Shklovskii (ES) theory, introducing the pivotal concept of the Coulomb gap. Following this, the "Applications and Interdisciplinary Connections" chapter will demonstrate the remarkable power of the ES model as a practical tool, explaining its use in characterizing real-world systems from disordered semiconductors to granular superconductors and exploring the frontiers of its application.

Imagine you are an electron, and you find yourself in a rather unfortunate neighborhood. Instead of the pristine, repeating lattice of a perfect crystal, you're stuck in the messy, chaotic landscape of a disordered material—think of a glass, or a semiconductor that has been deliberately "roughed up" with impurities. At low temperatures, you don't have enough energy to roam freely. You are ​​localized​​, trapped in a small region of space, like being stuck in a valley in a rugged mountain range. How do you get anywhere? You can't just roll; you have to ​​hop​​.

To move from your current home (a localized state) to another, you need a little help. The atoms of the material are constantly vibrating, creating tiny packets of energy called ​​phonons​​. By absorbing a phonon, you can get the energy boost, $\Delta E$, needed to jump to a new state a distance $R$ away. But this is not a simple transaction. You face a fundamental dilemma, a trade-off that is at the very heart of how disordered materials conduct electricity.

The problem is twofold. First, quantum mechanics tells us that tunneling over a large distance $R$ is exponentially unlikely. The probability of making the leap is suppressed by a factor of roughly $\exp(-2R/\xi)$, where $\xi$ is your ​​localization length​​—a measure of how spread out your quantum wavefunction is. Hopping to a neighbor next door is easy; hopping across the street is nearly impossible.

Second, you need to find an empty state to hop into, and you need a phonon with just the right energy, $\Delta E$, to make the jump. The probability of finding such a phonon is governed by a Boltzmann factor, $\exp(-\Delta E / (k_B T))$, where $T$ is the temperature and $k_B$ is the Boltzmann constant. High-energy hops are rare, especially when it's cold.

So, what is a poor electron to do? You want to hop to a nearby site (small $R$) to make the tunneling easy, but the closest sites might be at very different energies (large $\Delta E$), making the energy cost prohibitive. Conversely, there might be a site far away with almost the same energy (small $\Delta E$), but the tunneling distance is too great. The path of least resistance—or rather, the hop with the highest probability—will be one that strikes a perfect compromise. We are looking for the "optimal" hop that minimizes the total difficulty, which is captured by the exponent $S = \frac{2R}{\xi} + \frac{\Delta E}{k_B T}$. Finding this optimum is the key to understanding conduction.

It turns out that the solution to this optimization problem depends critically on the landscape of available energy states. What does the ​​density of states (DOS)​​—the number of available energy levels per unit energy—look like near your home energy, the Fermi level? Two major scenarios emerge.

Mott's Democratic Landscape

Let's first take the simplest view, as Sir Nevill Mott did. He imagined a "democratic" energy landscape where the density of states, $g(E)$, is more or less constant near the Fermi level. It’s as if possible landing spots are scattered uniformly in energy.

In this world, if you are willing to hop a distance $R$, the typical energy difference you'll have to overcome to find any state is inversely related to the number of states in your search volume. In a $d$-dimensional world, this volume goes as $R^d$. So, the energy spacing is roughly $\Delta E \propto 1 / (g(E_F) R^d)$. A bigger search volume (larger $R$) means you're more likely to find a state with a conveniently small energy difference.

When we plug this into our optimization problem and find the best trade-off, we arrive at the celebrated ​​Mott variable-range hopping​​ (VRH) law. The conductivity $\sigma$ behaves as:

The peculiar exponent, $1/(d+1)$, comes directly from this trade-off in a world with a constant DOS. As it gets colder, the electron prefers to make longer, more ambitious hops to find sites with ever-smaller energy differences, hence the name "variable-range" hopping.

The Coulomb Gap: A Valley of Impossibility

But wait a minute. We forgot something crucial. Electrons are not neutral; they are charged. And they repel each other with the long-range Coulomb force. This simple fact, as Boris Shklovskii and Alexei Efros realized, dramatically reshapes the energy landscape.

Imagine you hop from site $i$ to an empty site $j$ a distance $R$ away. You've left behind a positively charged "hole" at site $i$ and created a negative charge at site $j$. The energy of the system has increased by at least the Coulomb interaction energy between this new electron-hole pair, which is approximately $E_C = e^2/(\kappa R)$, where $\kappa$ is the material's dielectric constant.

For the system to be in a stable ground state at zero temperature, any such hop must not release energy. This means the single-particle energy difference, $\epsilon_j - \epsilon_i$, must be greater than the Coulomb energy you get back, or $\epsilon_j - \epsilon_i > e^2/(\kappa R)$. This simple stability requirement has a profound consequence: it forbids states of different energies from being too close to each other. It carves out a "soft" gap in the density of states right at the Fermi level—the ​​Efros-Shklovskii Coulomb gap​​.

Unlike a "hard" gap in a typical semiconductor, this isn't a region with strictly zero states. Instead, the DOS smoothly goes to zero, forming a valley whose width is determined by the Coulomb interaction itself. For a $d$-dimensional system, the theory predicts the DOS should follow the form $g(E) \propto |E-E_F|^{d-1}$. The uniform, democratic landscape of Mott is gone, replaced by a deep valley right where we need to find states to hop between!

How does our electron's dilemma resolve in this new, more realistic landscape? The situation is paradoxically simpler. The energy cost of a hop is no longer an independent variable we need to average over; it is determined by the hopping distance itself. The minimum energy to create an electron-hole pair separated by $R$ is precisely the Coulomb energy, $\Delta E(R) \approx e^2/(\kappa R)$.

Our optimization problem is now much more constrained. We simply need to minimize:

This is a classic problem of finding the minimum of a function that is the sum of a term that increases with $R$ and a term that decreases with $R$. The minimum occurs when the two terms are of comparable magnitude. Taking the derivative and setting it to zero reveals the optimal strategy for the hopping electron.

The optimal hopping distance is found to be $R_{\text{opt}} \propto T^{-1/2}$, and the corresponding optimal hopping energy is $\Delta E_{\text{opt}} \propto T^{1/2}$. As the temperature drops, the electron can't afford the high energy cost of short hops across the Coulomb valley. It is forced to look for longer, more improbable tunnels to sites whose Coulomb energy cost is lower.

Plugging these back into the exponent gives the magnificent ​​Efros-Shklovskii (ES) VRH​​ law:

The characteristic temperature, $T_{ES}$, is proportional to $e^2/(\kappa \xi k_B)$, a beautiful combination of the fundamental constants governing the Coulomb interaction and quantum tunneling.

Notice the exponent: $1/2$. Unlike Mott's law, this exponent is ​​universal​​—it does not depend on the dimensionality $d$ of the system. This universality is a direct consequence of the $1/R$ form of the Coulomb potential, a law that is the same in any dimension greater than one. The physics of long-range interaction imposes its own geometry, overwhelming the dimensionality of the space the electrons live in. It's a stunning example of how a fundamental interaction dictates macroscopic behavior in a simple and powerful way.

This elegant picture of electrons hopping across a Coulomb-carved landscape is not just a theoretical curiosity. It is a robust framework that we can test by poking and prodding the system in various ways.

• ​​The Crossover:​​ At high temperatures, an electron has so much thermal energy that the subtle dip of the Coulomb gap is just a minor bump in the road. The system behaves according to Mott's law. As the temperature is lowered, the electron's energy budget shrinks, and it begins to "see" the steep walls of the Coulomb valley. At a specific crossover temperature, $T_c$, the behavior switches from Mott-like to ES-like. This crossover is routinely observed in experiments and is a smoking gun for the presence of a Coulomb gap.

• ​​Screening the Interaction:​​ What if we could turn off the long-range part of the Coulomb force? We can! By placing a metal gate near our disordered material, we introduce screening. The interaction between two charges is now only $1/R$ for short distances; for distances larger than the gate separation $r_g$, it falls off much faster. This effectively "fills in" the Coulomb gap at very low energies. What happens? As we cool the system, it follows the ES law. The optimal hop distance $R_{\text{opt}}$ grows as $T^{-1/2}$. But once $R_{\text{opt}}$ becomes comparable to the gate distance $r_g$, the electron is forced to make hops where the interaction is screened. The Coulomb gap is no longer relevant for these long hops, and the system remarkably reverts to Mott-like behavior!.

• ​​Hopping in an Electric Field:​​ What happens at absolute zero temperature, when there are no phonons to help? Can conduction stop entirely? Not if you apply a strong electric field, $F$. The field can provide the energy for a hop. An electron moving a distance $R$ along the field gains energy $eFR$. At $T=0$, the optimal hop will be one where this energy gain exactly pays the Coulomb cost: $eFR \approx e^2/(\kappa R)$. This leads to a unique, non-thermal conductivity law $\sigma(F) \sim \exp[-(F_0/F)^{1/2}]$.

The ES-VRH model is so powerful that it can also explain how these materials respond to a temperature gradient, giving rise to ​​thermopower​​, and how their resistance changes in a magnetic field—a phenomenon known as ​​magnetoresistance​​. Each of these effects provides another window into the intricate dance of electrons hopping across a landscape shaped by their own mutual repulsion. The journey of the reluctant electron, it seems, is governed by principles of remarkable beauty and unity.

Now that we have explored the elegant argument that leads to the Efros-Shklovskii (ES) law, $\sigma(T) \propto \exp[-(T_{ES}/T)^{1/2}]$, you might be tempted to think of it as a neat but perhaps esoteric piece of theoretical physics. Nothing could be further from the truth. The real power and beauty of this law lie not in its derivation, but in its application. It is a bridge connecting the microscopic quantum world of localized electrons to the macroscopic, measurable world of electrical resistance. By touching a material with a probe and measuring its conductivity as we cool it down, we can learn a surprising amount about the dance of electrons within. Let's explore some of the places where this dance is performed.

The most natural home for Efros-Shklovskii hopping is in a disordered semiconductor at very low temperatures. Imagine a crystal of silicon, not perfectly pure, but "compensated" – meaning it has been doped with both donor atoms (which want to give up an electron) and acceptor atoms (which want to grab one). At room temperature, this is a bustling city of charge carriers. But as we cool it down to just a few kelvins, the thermal energy vanishes and the electrons "freeze" onto the donor atoms. They are no longer free to roam. The semiconductor has become an insulator.

But is it a perfect insulator? Not quite. An electron trapped on one donor can still, with the help of a tiny vibration from the crystal lattice (a phonon), "hop" to a nearby empty donor site. This is our variable-range hopping. Because the carriers are charged electrons, their long-range Coulomb interactions are inescapable, opening the characteristic soft gap in the density of states. And so, the electrical conductivity is beautifully described by the ES law.

This is more than just a formula that fits the data. It's a powerful diagnostic tool. By measuring the conductivity of a silicon sample at cryogenic temperatures, an experimentalist can create a plot of $\ln \sigma$ versus $T^{-1/2}$. The slope of the resulting straight line directly yields the characteristic temperature, $T_{ES}$. The formula for $T_{ES}$, which we've seen contains fundamental constants and material properties, is approximately $T_{ES} \propto e^2 / (4\pi\epsilon_0\epsilon_r k_B \xi)$. Since everything else in this expression is known—the charge of the electron $e$, the dielectric constant of silicon $\epsilon_r$, etc.—the measurement of $T_{ES}$ gives us a direct experimental value for the localization length, $\xi$! This is remarkable: by making a simple resistance measurement, we gain profound insight into the quantum-mechanical extent of the electron's wavefunction, a quantity we could never hope to "see" with a microscope.

Furthermore, the theory makes concrete predictions. If we increase the compensation—that is, the number of charged donor and acceptor impurities—the random electrical landscape becomes more rugged. This enhanced disorder traps the electrons more tightly, causing their localization length $\xi$ to shrink. Our formula for $T_{ES}$ then predicts that $T_{ES}$ should increase, making the material even more insulating. This is precisely what is observed in experiments, giving us great confidence in the physical picture.

You might think this hopping game is peculiar to electrons in semiconductors, but the principles of physics are rarely so parochial. The same logic applies whenever charged particles are localized and interact via the Coulomb force. Consider a completely different system: a thin film of a granular superconductor. This material is composed of tiny, microscopic islands of a superconductor, separated by a thin insulating layer.

Above a certain transition temperature, the material is a normal metal. Below this temperature, each island becomes superconducting, and the charge carriers within it are not electrons, but Cooper pairs—bound pairs of electrons with a charge of $q=2e$. However, because of the insulating barriers, these Cooper pairs are trapped on their islands. For current to flow, a Cooper pair must quantum-mechanically tunnel, or "hop," from one island to the next.

Here we have all the ingredients for ES hopping: localized charge carriers (the Cooper pairs), hopping transport, and the long-range Coulomb interaction. The physics is identical. And indeed, the conductivity of such systems at low temperatures follows the hallmark $T^{-1/2}$ law. The only change is that the charge $q$ in the derivation is now $2e$, which simply modifies the value of the characteristic temperature $T_{ES}$ we measure. This beautiful correspondence between a doped semiconductor and a granular superconductor is a testament to the unifying power of physical principles.

So far, we have imagined our materials to be isotropic—the same in all directions. But many modern materials, especially two-dimensional ones like black phosphorus, are anisotropic. Their crystal structure or bonding makes them behave differently along different axes. For such a material, the dielectric constant isn't just a number; it's a tensor. For a 2D material in the $x-y$ plane, we might have different values, $\kappa_x$ and $\kappa_y$.

This seemingly small complication has a fascinating consequence: the Coulomb interaction itself becomes anisotropic. The electrostatic force between two charges depends on the direction of the line connecting them. If we apply our hopping theory, we find that the characteristic temperature, $T_{ES}$, also becomes direction-dependent. A measurement of conductivity along the $x$-axis will be governed by one characteristic temperature, $T_{ES,x}$, while a measurement along the $y$-axis will be governed by another, $T_{ES,y}$. The physics of hopping provides a direct window into the anisotropic electronic structure of the material, a crucial aspect for designing novel electronic devices.

The environment of the hopping system is just as important as its internal structure. The entire ES framework rests on the long-range, $1/r$ nature of the Coulomb potential. What happens if we tamper with it? A clever way to do this is to place a metallic plate (a gate) near our 2D hopping system. The mobile charges in the metal will rearrange to screen the electric fields from the hopping electrons. This effectively cuts off the Coulomb interaction at a distance comparable to the separation between the system and the gate, $W$.

At high temperatures, hops are short, $r_{\text{opt}} \ll W$, and the electrons don't "see" the gate. The physics is pure Efros-Shklovskii. But as we lower the temperature, the optimal hop distance, $r_{\text{opt}} \propto T^{-1/2}$, grows. Eventually, we reach a temperature where $r_{\text{opt}}$ becomes larger than $W$. For these long hops, the interaction is no longer long-range. The physical justification for the Coulomb gap vanishes! The system reverts to a model where the density of states is constant, which is the assumption behind the older Mott variable-range hopping theory. The conductivity law beautifully crosses over from the ES form ($\sigma \propto \exp[-(T_{ES}/T)^{1/2}]$) to the 2D Mott form ($\sigma \propto \exp[-(T_M/T)^{1/3}]$). This ability to tune between two fundamental laws of physics simply by changing temperature or geometry is a stunning demonstration of the interplay between theory and experiment.

The theoretical framework for hopping is more than just a description of the $1/r$ world; it's a general-purpose toolkit. We can ask: what if the interaction between charges followed a different law? In some exotic physical systems, such as near a Weyl semimetal, the effective interaction between charges at long distances might be screened into a steeper $1/R^2$ form.

We can feed this new interaction law into our theoretical machine. The logic is the same: the interaction potential determines the form of the "gap" in the density of states; the density of states determines the relationship between hopping distance and energy; and that relationship determines the final temperature dependence. For a $1/R^2$ potential in 3D, the machine churns and produces a new result: the conductivity should follow a $T^{-1/3}$ law. This demonstrates the predictive power of the theory; it allows us to explore the consequences of new and undiscovered physical interactions.

What about a messy, real-world material that is a mixture of different regions—some where the Coulomb interaction is screened (Mott-like) and others where it is not (ES-like)? We can model such a composite system, for instance, as a checkerboard of the two types of domains. Using effective medium theory, we find that the overall conductivity is a blend of the two behaviors. We can even define an effective hopping exponent, $p_{\text{eff}}(T)$, which is no longer a fixed number like $1/2$ or $1/3$, but is itself a function of temperature, smoothly varying between the two limits as one mechanism or the other starts to dominate the overall resistance of the network.

As a final illustration of the richness of this topic, let's consider a one-dimensional wire. Here, electrons can only hop forward or backward along a line. If the charges in this wire interact with the standard 3D Coulomb $1/r$ potential (as they would if the wire is embedded in an insulator), one might expect the usual ES physics to apply. If you carry out the derivation for the hopping exponent, you find $p=1/2$, which seems to confirm this.

But here nature throws us a curveball. A more careful analysis of the stability conditions reveals that in one dimension, a long-range $1/r$ interaction is not sufficient to open a true Coulomb gap in the density of states; the density of states remains constant. The system should therefore obey Mott's law. But what exponent does Mott's law predict for one dimension ($d=1$)? It predicts $p = 1/(d+1) = 1/(1+1) = 1/2$. So we arrive at the same exponent, but for a completely different physical reason! It is a beautiful and subtle coincidence. It serves as a warning, a classic Feynman-esque lesson: do not be fooled by mere appearances. A true physical understanding requires that we look not just at the final answer, but at the integrity of the physical reasoning that gets us there.

From the heart of a silicon chip to the frontiers of materials science, the seemingly simple law of Efros-Shklovskii hopping provides a surprisingly versatile and insightful lens through which to view the quantum world.