Kelvin Relations

The coupling of thermal and electrical phenomena in materials gives rise to fascinating effects: a temperature difference can generate a voltage, and an electric current can transport heat. While the Seebeck, Peltier, and Thomson effects describe these behaviors, a crucial question remains: are they merely disconnected observations, or do they obey a deeper, unifying order? This article addresses this fundamental gap by exploring the Kelvin relations, the elegant set of laws that govern the world of thermoelectricity. By treating thermoelectric devices as heat engines, we will uncover the profound connection between these seemingly disparate effects. This exploration is structured to first reveal the core physical principles and then demonstrate their powerful real-world applications. The first chapter, "Principles and Mechanisms," derives the Kelvin relations from the fundamental laws of thermodynamics and statistical mechanics. Following this, the "Applications and Interdisciplinary Connections" chapter will demonstrate how these theoretical principles are indispensable tools for experimentalists and engineers in materials science and device design.

So, we have these curious effects where heat can generate electricity and electricity can move heat. But are they just a loose collection of phenomena? Or is there a deeper, more elegant order governing them? Nature, it turns out, is not a messy accountant; she is a master of profound and simple rules. Our mission in this chapter is to uncover these rules, the Kelvin relations, not as dry formulas to be memorized, but as inevitable consequences of the most fundamental laws of physics.

Let's begin with a thought experiment, one of the most powerful tools in a physicist's arsenal. Imagine we have a thermocouple—our simple device with two junctions, one hot and one cold—and we treat it as a tiny heat engine. What does a heat engine do? It takes in some heat from a hot place, turns some of it into useful work, and dumps the rest into a cold place.

Now, let's run this engine in a perfectly efficient, reversible way, just like the idealized Carnot engine you might have studied. We let a tiny amount of charge $dq$ creep around the circuit.

1. At the hot junction, at temperature $T_H$, it absorbs an amount of heat $Q_H = \Pi_{AB}(T_H) dq$. The term $\Pi$ is the Peltier coefficient, which is nothing more than the heat absorbed per unit of charge that crosses the junction.
2. As the charge moves through the circuit, it does work. The voltage from the Seebeck effect is $dV = S_{AB} dT$, so the work done is $dW = dV \cdot dq$.
3. At the cold junction, at temperature $T_C$, it releases heat $Q_C = \Pi_{AB}(T_C) dq$.

The First Law of Thermodynamics, the grand principle of energy conservation, tells us that the work we get out must be the net heat we put in: $dW = Q_H - Q_C$. Nothing surprising there.

But now comes the magic, courtesy of the Second Law of Thermodynamics. The Second Law puts a strict limit on the efficiency of any engine. It states that for a reversible engine, the ratio of heat to temperature must balance out: $\frac{Q_H}{T_H} = \frac{Q_C}{T_C}$. For our thermocouple engine operating between two infinitesimally different temperatures, $T$ and $T+dT$, this law, combined with the first law, leads to a startlingly simple conclusion about its efficiency. The efficiency must be the Carnot efficiency, $\eta = \frac{dW}{Q_H} = \frac{dT}{T}$.

When we substitute our expressions for $dW$ and $Q_H$ and do the algebra, something wonderful happens. We find a direct, unshakeable relationship between the Peltier coefficient and the Seebeck coefficient:

This is the first great Kelvin relation. Stop and appreciate what this means. On one side, we have $\Pi$, a purely thermal property—the heat absorbed at a junction. On the other side, we have $S$, a purely electrical property—the voltage generated by a temperature difference. And they are bound together by the absolute temperature, $T$. This isn't a coincidence or an empirical rule-of-thumb. It is a bargain enforced by the fundamental laws of thermodynamics. If this relation didn't hold, we could, in principle, build a device that violates the Second Law, and the universe simply does not allow that.

This relation also gives us a profound insight into what the Seebeck coefficient is. Since $\Pi$ is the heat energy per charge, $\Pi/T$ is the entropy per charge. The relation $\Pi = ST$ means that ​​the Seebeck coefficient, $S$, is nothing but the entropy carried by each charge carrier​​ in the material. A material with a high Seebeck coefficient is one where the charge carriers are very effective at transporting not just charge, but also disorder.

The Peltier effect happens at the junction between two different materials. But what if we have a single, continuous wire sitting in a temperature gradient? If a current flows through this wire from a hot spot to a cold spot, something else must happen. This is the ​​Thomson effect​​.

Think about it this way. We just discovered that the Seebeck coefficient $S$ can be thought of as the entropy (and thus the "heat-carrying capacity") of the charge carriers. If $S$ changes with temperature—and for real materials, it always does—then as a charge carrier moves from a hot region to a cold region, its capacity to carry heat changes. If its capacity goes down, where does the extra heat go? It must be released into the wire. If its capacity goes up, it must absorb heat from the wire to make up the difference.

This is the essence of the Thomson effect: heat is absorbed or released along a current-carrying conductor in a temperature gradient. The amount of heat is described by the ​​Thomson coefficient​​, $\tau$. And, as you might now guess, thermodynamics again provides a direct link. The second Kelvin relation is:

This beautiful equation tells us that the Thomson effect is directly proportional to how fast the Seebeck coefficient changes with temperature ($\frac{dS}{dT}$). If a material had a Seebeck coefficient that didn't change with temperature, its Thomson coefficient would be zero. If we have a material where, say, $S(T) = \alpha T + \beta T^2$, we can immediately calculate its Thomson coefficient as $\tau(T) = T(\alpha + 2\beta T) = \alpha T + 2\beta T^2$ and from that, find the total heat absorbed by a wire segment in a temperature gradient. The relationship is precise and predictive. It doesn't matter why $S$ depends on $T$—whether it's due to simple electron diffusion or more complex effects like phonons dragging electrons along—the thermodynamic connection remains inviolate.

Why are these relations so robust? So universal? We derived them from thermodynamics, but there is an even deeper principle at play, one that comes from statistical mechanics: ​​Onsager's reciprocal relations​​.

Let's step back and look at the transport of heat and charge in a more general way. We have two "fluxes": a flow of charge ($J_q$) and a flow of heat ($J_h$). And we have two "forces" that can cause these flows: an electric field ($\mathcal{E}$) and a temperature gradient ($\nabla T$). In the linear regime (for small forces), we can write that each flux is a combination of both forces:

Here, $K_{11}$ is just the electrical conductivity $\sigma$, and $K_{22}$ is related to the thermal conductivity. But what about the "cross-coefficients," $K_{12}$ and $K_{21}$?

• $K_{12}$ describes how a temperature gradient causes a charge current (the Seebeck effect).
• $K_{21}$ describes how an electric field causes a heat current (the Peltier effect).

Lars Onsager showed, from the principle of microscopic reversibility (the idea that on a microscopic level, physical processes look the same whether you run the movie forwards or backwards in time), that these cross-coefficients must be related. In the absence of a magnetic field, the relation is stunningly simple: the matrix of coefficients is symmetric. For our case, this means:

By simply looking at the definitions of the Seebeck coefficient ($S = K_{12}/K_{11}$) and the Peltier coefficient ($\Pi = K_{21}/K_{11}$), this symmetry immediately gives us $\Pi = TS$. The Kelvin relation is a direct macroscopic manifestation of time-reversal symmetry at the microscopic level! This is a profound and beautiful piece of physics. It unifies disparate phenomena by tracing them to a single, fundamental symmetry of nature.

This symmetry has another incredible consequence. When we calculate the rate of ​​entropy production​​—the measure of true, irreversible waste—we find it is composed of two parts: waste from electrical resistance (Joule heating) and waste from heat conduction down a temperature gradient.

Where have the thermoelectric terms gone? They've vanished! The terms involving the cross-effects cancelled out perfectly, and they did so precisely because of the Kelvin relation $\Pi=ST$. This tells us something remarkable: the Seebeck, Peltier, and Thomson effects are, in themselves, ​​thermodynamically reversible​​. They don't generate waste; they merely orchestrate a delicate and balanced dance between heat and electricity. The only true sources of irreversibility and waste in our thermoelectric device are the familiar villains: electrical resistance and thermal conduction.

The power of fundamental laws lies not only in what they explain, but also in what they forbid. The Kelvin relations, rooted in the laws of thermodynamics, draw hard lines in the sand for materials science.

Could you ever invent a material with a perfectly constant, non-zero Seebeck coefficient, say $S(T)=S_0$? It sounds like a great engineering goal. However, we've learned that $S$ represents the entropy per charge. The ​​Third Law of Thermodynamics​​ demands that the entropy of any system must approach zero as the temperature approaches absolute zero. Therefore, we must have $\lim_{T \to 0} S(T) = 0$. A constant, non-zero $S_0$ would be a flagrant violation of this deep principle of nature. Thus, such a material cannot exist. Any real material's Seebeck coefficient must fade to nothing at absolute zero.

These laws are universal. But like all physical laws, they have a domain of applicability. They are built on the assumption of ​​local equilibrium​​—that even though the material as a whole is not in equilibrium (it has a temperature gradient), any tiny piece of it is. In most bulk materials, this is an excellent approximation. But in some exotic, very small (mesoscopic) systems or under very high-frequency conditions, this assumption can break down. In such cases, the beautiful simplicity of the Kelvin relations can be violated, and measuring this deviation becomes a powerful diagnostic tool for exploring physics far from equilibrium.

For our purposes, however, the message is clear. The thermoelectric effects are not a grab-bag of disconnected oddities. They are a tightly-knit family, governed by elegant and powerful rules derived from the deepest principles of thermodynamics and statistical mechanics. They reveal a beautiful, hidden symmetry in the way nature couples the flow of heat and the flow of charge.

We have learned the elegant rules of the game, the Kelvin relations that orchestrate the interplay of heat and electricity. But a set of rules is one thing; the game itself is another. What symphony does this score conduct? What can we do with this knowledge? As it turns out, these simple-looking equations are not just abstract thermodynamic statements. They are the working tools of physicists, chemists, material scientists, and engineers—a powerful bridge connecting the deepest theories of matter to the practical design of futuristic technologies. In this chapter, we will embark on a journey to see these relations in action, to appreciate their astonishing utility and the beautiful unity they reveal across the scientific disciplines.

Imagine you are a materials scientist and you have just synthesized a promising new compound. To understand its potential for converting heat into electricity, you need to know its Seebeck coefficient ($S$), its Peltier coefficient ($\Pi$), and its Thomson coefficient ($\tau$). Measuring each of these properties independently can be a complex and time-consuming task, each requiring a different experimental setup. Here is where the Kelvin relations first show their practical magic. They act as a "Rosetta Stone," allowing us to translate between the different languages of thermoelectric phenomena.

If you can measure just one of these properties as a function of temperature, the Kelvin relations empower you to deduce the others. For instance, the Seebeck coefficient is often the most accessible property to measure experimentally. Once you have a reliable measurement of $S(T)$, the second Kelvin relation hands you the Thomson coefficient on a silver platter:

This equation tells us that the Thomson effect is directly linked to how sensitively the Seebeck coefficient changes with temperature. For many materials, the Seebeck coefficient is not constant. At low temperatures, its behavior might be dominated by the simple diffusion of charge carriers, while at higher temperatures, a phenomenon called "phonon drag"—where vibrating crystal lattices give electrons an extra push—can cause $S$ to change more dramatically. By measuring $S(T)$ and calculating its derivative, we gain direct insight into the Thomson heat that would be generated or absorbed if we were to pass a current through the material in a temperature gradient.

But the translation service doesn't stop there. Once we have $S(T)$, the first Kelvin relation immediately gives us the Peltier coefficient:

This connection is remarkably direct. The Peltier heat exchanged at a junction is simply the Seebeck coefficient multiplied by the absolute temperature. This powerful cycle of deduction works in all directions. If, for instance, we were to measure the Thomson coefficient $\tau(T)$, we could perform an integration to find the Seebeck coefficient. And from that, we'd find the Peltier coefficient. This interconnectedness provides a crucial consistency check for experimentalists and dramatically simplifies the task of characterizing new materials. Instead of three difficult experiments, one carefully performed experiment, combined with the Kelvin relations, can reveal the entire thermoelectric profile of a material.

Of course, real science is never as clean as a perfect formula. Experimental data points are always flecked with uncertainty; they are not divine pronouncements but noisy messages from nature. The Kelvin relations, far from being fragile theoretical ideals, are robust enough to guide us through this messy reality.

In modern materials science, researchers will often measure the Seebeck coefficient at various temperatures and then fit this data to a mathematical model, perhaps a polynomial like $S(T) = \alpha T + \beta T^3$. The Kelvin relations then become a computational engine. They allow the scientist to take the best-fit parameters for $S(T)$ and "propagate" them to calculate the expected Peltier and Thomson coefficients. Crucially, this propagation also applies to the uncertainties. The statistical errors from the initial Seebeck measurement can be mathematically transformed to predict the range of uncertainty in the calculated values of $\Pi$ and $\tau$. This gives a realistic, honest assessment of the material's properties, which is essential for any serious engineering application.

You might ask, where does the profound certainty of these relations come from? Why should they hold in the messy, real world? The answer lies deep in the bedrock of statistical mechanics, in the work of Lars Onsager. He showed that for any system near thermodynamic equilibrium, the response of one flow (like charge) to the driving force of another (like a temperature gradient) is symmetrically related to the response of the second flow to the first force. This "reciprocity principle," grounded in the time-reversal symmetry of microscopic physical laws, is the ultimate source of the Kelvin relations. They are not arbitrary; they are a macroscopic echo of a fundamental symmetry of the universe.

Armed with this reliable toolset, let's build something. How does one go about designing a thermoelectric generator to power a deep-space probe or a solid-state cooler for a microprocessor? The Kelvin relations are indispensable guides in this engineering endeavor.

First, to accurately predict how a device will behave, we must model the flow of energy within it. When an electric current $J$ flows through a material that also has a temperature gradient $\frac{dT}{dx}$, we must account for all sources of heat. There is the familiar irreversible Joule heating, which scales as $\rho J^2$. But the Kelvin relations revealed the Thomson effect, a reversible heating or cooling given by $\tau J \frac{dT}{dx}$. A full energy balance, as derived from the first law of thermodynamics, shows that the net volumetric heat generation is a competition between these two effects: $\rho J^2 - \tau J \frac{dT}{dx}$. Forgetting the Thomson term would be like trying to balance a budget while ignoring a major income stream or expense; your predictions for the device's temperature profile and overall performance would be wrong.

When designing a material for a device, "performance" can mean two different things: raw power or high efficiency. The Kelvin relations help us distinguish the material properties that govern each.

1. ​​The Power Factor ($S^2\sigma$):​​ If your goal is to extract the maximum possible electrical power from a given temperature difference, what matters most is the ​​power factor​​, defined as $S^2\sigma$, where $\sigma$ is the electrical conductivity. A simple analysis shows that the maximum power output scales directly with this quantity. Notice what's missing: the thermal conductivity, $k$. For sheer power, you want a large Seebeck coefficient and high electrical conductivity, and you don't (at first glance) care how leaky the material is to heat.

2. ​​The Figure of Merit ($ZT$):​​ If, however, your goal is maximum efficiency—converting the largest possible fraction of heat into useful electricity—then you can't ignore thermal leakage. It does no good to have a great thermoelectric engine if most of the heat simply flows through the material without being converted. To maximize efficiency, one must balance a high power factor with a low thermal conductivity, $k$. This balance is captured by the celebrated dimensionless ​​figure of merit​​:

   $$ZT = \frac{S^2 \sigma}{k} T$$

   This single number is the king of thermoelectric performance metrics. The higher the $ZT$ of a material, the closer its efficiency can get to the absolute thermodynamic limit set by Carnot. The entire field of modern thermoelectric materials research can be seen as a quest to find materials with ever-higher $ZT$ values.

Using the figure of merit, engineers can calculate a realistic maximum efficiency for a generator operating between a hot temperature $T_H$ and a cold temperature $T_C$. The formula, derived by combining the laws of thermodynamics with the thermoelectric transport equations, is:

where $T_m$ is the average temperature. This equation beautifully shows that the actual efficiency is the ideal Carnot efficiency, $\eta_{\text{Carnot}} = 1 - T_C/T_H$, multiplied by a factor that depends entirely on the material's quality, $ZT$.

Perhaps the most profound application of the Kelvin relations is their ability to serve as a bridge from our macroscopic world to the bizarre and beautiful quantum realm of electrons. The coefficients $S$, $\Pi$, and $\tau$ are not just arbitrary material parameters; they are determined by the collective quantum behavior of electrons moving through a crystal.

Consider, for example, exotic materials known as "heavy-fermion" metals. At very low temperatures, interactions between electrons cause them to behave as if they have enormous mass, leading to unusual thermodynamic properties. Condensed matter physicists have quantum theories that predict, for example, the electronic specific heat $C_e(T)$ of such a system. From the third law of thermodynamics, we know that the electronic entropy is related to the specific heat by $s_{el}(T) = \int_0^T \frac{C_e(T')}{T'} dT'$. Another thermodynamic relation connects this entropy to the Seebeck coefficient.

Here is the magic: a physicist can start with a quantum-mechanical prediction for $C_e(T)$, calculate the resulting entropy $s_{el}(T)$, use that to find the Seebeck coefficient $S(T)$, and then—using the Kelvin relation $\Pi = ST$—predict the Peltier coefficient. This creates a complete, testable chain of logic from the deepest quantum theories of matter to a macroscopic, measurable quantity. A simple voltage measurement in the lab can become a verdict on a complex quantum theory.

As we have seen, the Kelvin relations are far more than a textbook curiosity. They are a central pillar of thermoelectricity, connecting fundamental theory to practical application. They are the grammarian's rules for the language of heat and charge, the experimentalist's multi-tool, the engineer's design principle, and the theorist's window into the quantum world. The ongoing quest for clean energy and efficient cooling has placed thermoelectric materials at the forefront of modern research. In this quest, the elegant and powerful connections first laid down by Lord Kelvin over a century ago continue to light the way.