Section 43–6 Thermal conductivity

Idealizations / Approximations / Limitations

 

In this section, Feynman discusses the idealized conditions needed to derive the thermal conductivity of a gas, presents approximate formulas, and examines their limitations. The section could therefore be appropriately titled “Thermal conductivity of a gas.” Although we can use Fourier’s law of heat conduction for solids, liquids, and gases, the underlying mechanisms and assumptions differ among these states of matter. The distinction is particularly important for gases, where the behavior depends on the molecular mean free path. In essence, modeling the thermal conductivity of a dilute gas requires distinct frameworks depending on how the mean free path compares to the physical dimensions of the system.

 

1. Idealizations

“The transfer of heat from the hotter gas to the colder gas is by the diffusion of the “hot” molecules—those with more energy—downward and the diffusion of the “cold” molecules upward. To compute the flow of thermal energy we can ask about the energy carried downward across an element of area by the downward-moving molecules, and about the energy carried upward across the surface by the upward-moving molecules. The difference will give us the net downward flow of energy (Feynman et al, 1963).”

Feynman’s explanation of heat transfer in terms of the diffusion of “hot molecules” and “cold molecules” can be misleading if taken literally. Strictly speaking, heat conduction does not involve a net transport of mass: molecules move in all directions, and in a stationary medium the upward and downward molecular flux are equal. The terms “hot molecules” and “cold molecules” are also imprecise: temperature is a macroscopic statistical property of a local population of molecules, not a permanent property of individual molecules. Molecules possess molecules in both regions have a broad energy distribution, and thus, molecules crossing the plane in opposite directions do not necessarily belong to two sharply defined regions of “hot” and “cold” molecules. More importantly, heat conduction occurs through the transfer of energy during the collisions of molecules moving in opposite directions instead of simply a net diffusion of molecules or mass. From the perspective of conventional elementary theory of heat conduction, the macroscopic heat flux is associated with a temperature gradient, whereas the microscopic mechanism is the diffusion of thermal energy through molecular collisions.

 

Fourier’s law of heat conduction states that the heat flux Q, which is the flow of thermal energy per unit area and per unit time, is directly proportional to the negative temperature gradient: Q = -kA(dT/dx), where k is the thermal conductivity. The minus sign means that heat flows in the direction of decreasing temperature. In the context of a gas, this law is strictly valid only under a set of idealized conditions: (1) Homogeneous and Isotropic: The thermal conductivity is independent of spatial position (homogeneous) and independent of the direction of heat flow (isotropic). (2) Linear response: The temperature gradient must be sufficiently small so that the transport coefficient can be regarded as approximately constant. (3) Continuum or local-equilibrium: There should be no convectional current and the molecular mean free path must be shorter than the characteristic length of the system. When these conditions are satisfied, Fourier’s law provides a reliable description of conductive heat transport in gases.

 

2. Approximations

“Since the details of the calculations are quite similar to those we have done above in considering molecular diffusion, we shall leave it as an exercise for the reader to show that κ=knlv/(γ−1), where (γ−1)kT is the average energy of a molecule at the temperature T (Feynman et al., 1963).”

 

Deriving the Thermal Conductivity of a Dilute Gas

Feynman’s problem on the thermal conductivity of a dilute gas is closely analogous to his earlier derivation of the diffusion coefficient. Mathematically, the transport equations for mass transport and thermal energy share the same underlying structure. The primary difference lies in the physical quantity being transported: diffusion transport number density of molecules (or mass), whereas heat conduction transports energy.

 

The derivation can be summarized in four-steps:

Step 1: Relate Energy Flux to Molecular Transport

Consider an imaginary horizontal plane at position z. The net thermal energy flux q_z (energy per unit area per unit time) transported across this plane can be estimated as:

q_z = nv[E(z - l) - E(z + l)]

where n is the number density, v is the mean molecular speed, E(z) is the average energy per molecule at z, and l is the mean free path.

 

Step 2: Use Taylor’s Expansion

Assuming the average molecular energy varies slowly over a mean free path l, we can approximate E(z ± l) using a first-order Taylor expansion:

E(z ± l) » E(z) ± (l)(dE/dz)

Substitution gives the one-dimensional estimate:

q_z = (nv){[E(z) - (l)(dE/dz)] - [E(z) + (l)(dE/dZ)]} = -2(nvl)(dE/dZ)

A more complete three-dimensional treatment introduces a geometric factor would yield:

q_z = -(1/3)(nvl)(dE/dZ)

 

Step 3: Relate Energy Gradient to Temperature Gradient and Heat Capacity

The average molecular energy depends on the local temperature. Using the chain rule:

dE/dZ = (dE/dT)(dT/dZ) = c_v (dT/dZ),

where c_v is the thermal capacity per molecule at constant volume.

Substituting c_v into the heat flux equation gives:

q_z = -(1/3)(nvl)c_v(dT/dZ)

Comparing this with Fourier’s Law q_z = -k(dT/dZ), we identify

k = (1/3)nvlc_v

 

Step 4: Express the Molecular Heat Capacity in terms of Adiabatic Index

For an ideal gas, Mayer’s relation for a single molecule is:

c_p - c_v = k,  where k is Boltzmann’s constant.

Since g = c_p/c_v

Þ g - 1 = (c_p - c_v)/c_v = k/c_v

Þ c_v = k/(g - 1)

Substituting c_v into k = (1/3)(nvl)c_v, completes the derivation: k = (1/3)(nvl)k/(g - 1)

 

Feynman’s expression omits the geometric factor, so it is simply k = (nvl)k/(g - 1). This derivation is an order-of-magnitude estimate rather than a rigorous result. It captures the essential dependencies on number density, mean free path, and molecular speed, but it relies on Talor’s expansion and simplified assumptions (such as an ideal gas, a single mean molecular speed, and local equilibrium).

 

3. Limitations

“The formula (43.43) was derived, as were all the others in this chapter, under the assumption that the mean free path between collisions is much smaller than any of the dimensions of the container. Whenever the gas density is so low that a molecule has a fair chance of crossing from one wall of its container to the other without having a collision, none of the calculations of this chapter apply (Feynman et al., 1963).”

 

Limitations of Fourier’s Law: The Knudsen Regime

Fourier’s law applies to heat conduction in solids, liquids, and gases, provided that a well-defined local temperature field exists and the heat flux responds approximately locally and linearly to the temperature gradient. In gases, however, this continuum description has its limitations when the molecular mean free path (l) is comparable, or larger than the characteristic length scale (L) of the system. This breakdown is most pronounced in a Knudsen gas—a gas at such low density or confined to such a small space where l >> L. In this regime, intermolecular collisions are negligible, and the transport is governed almost entirely by collisions between molecules and the container walls. In short, Fourier’s law fails because heat transfer can no longer be represented by a local temperature gradient and thermal conductivity; instead, the shape and size of the container must be considered in detail. Thus, while Fourier’s law remains a good continuum approximation when l << L, it breaks down in the Knudsen regime, where heat transfer becomes a free-molecular, geometry-dependent process.

 

Thermal conductivity is commonly defined as a material’s ability to conduct thermal energy via lattice vibrations, molecular collisions or free-electron motion. However, the phrase “thermal conductivity of a dilute gas” can be misleading when a confined gas is at sufficiently low pressure to approach vacuum conditions. At ordinary dilute-gas pressures, a bulk thermal conductivity still provides a useful macroscopic description. But when the molecular mean free path becomes comparable to or exceeds the container size, the assumptions underlying a local, continuum description break down. In this regime, molecules tend to travel from one boundary to another with few or no molecular collisions, so heat transfer is governed primarily by molecule-wall collisions rather than molecular collisions. Thus, the process becomes geometry-dependent, requiring a different formula and geometrical factor. In the limiting near-vacuum case, heat transfer must be described by physical principles different from the conventional heat conduction used for solids and liquids.

 

Key Takeaways (This Section):

Universal Forms, Distinct Physical Realities: Thermal Conductivity

The mathematical form of a transport law can be broadly universal, while its physical realization and range of validity are not. Fourier’s law of heat conduction can describe heat transport in solids, liquids, and gases when its idealized assumptions are satisfied. However, the microscopic mechanisms that determine the thermal conductivity (k), as well as the conditions required for the equation to hold, differ among materials and physical regimes. In an ordinary gas, thermal conductivity arises from molecular motion and intermolecular collisions, with the mean free path determining the transport of thermal energy. In the Knudsen regime, where the mean free path becomes comparable to or larger than the system’s characteristic dimension, Fourier’s law breaks down, and the geometry of the system must be taken into account. Thus, the term “thermal conductivity” should not be taken to imply a single microscopic mechanism with a wide range of applicability; the same term does not imply the same underlying physics.

 

Key Takeaways (This Chapter):

Shared Mathematical Structure across Transport Phenomena

Ohm’s law of electrical conduction, Fick’s law of diffusion, and Fourier's law of heat conduction are examples of a broader class of linear transport laws that share the same mathematical structure:

[Flux] = [Transport Coefficient] ´ [Driving Force / Gradient].

Although the mathematical equations may appear different, each law describes the transport of a distinct physical quantity:

• Fick's law (mass transport): J = -D(dn/dx) or J = -DÑn

where n is particle number density and D is the diffusion coefficient.

• Fourier's law (thermal energy transport): q = -k(dT/dx) or q = -kÑT,

where T is temperature and k is thermal conductivity.

• Ohm's law (electrical charge transport): J = -s(dV/dx) or J = -sÑV

where V is electric potential and s is electrical conductivity.

The deeper lesson is that a common mathematical structure can describe radically different physical phenomena. What distinguishes these physical laws is not their mathematical equation alone, but the definitions of their variables—the quantity being transported, the driving force or gradient, and the transport coefficient. In short, we can use the same mathematical idea to model different physical phenomena, provided one interprets the symbols correctly.

 

The Moral of the Lesson: Maxwell’s Error and Scientific Humility

At the end of the audio recording of the lecture, Feynman recounts an anecdote about Maxwell that appears to serve a pedagogical purpose, but it was omitted from the edited text. After deriving electrical conductivity, diffusion, and thermal conductivity using simplified kinetic-theory arguments, Feynman was not simply making fun of Maxwell. Rather, the anecdote underscores an important lesson: having the right physical insight does not guarantee that every numerical result will be correct. Maxwell’s initial calculation of the ratio of the thermal conductivity of copper to that of air was wrong because of unit-conversion errors—specifically, failure to convert kilograms to pounds and hours to seconds. This is a striking reminder that even a great physicist can obtain a wildly incorrect numerical result through seemingly mundane mistakes. The story also illustrates scientific humility: Maxwell acknowledged his errors and recognized Clausius’s contribution to correcting his work.

 

There is, therefore, a broader moral to the chapter: physics is built not only from profound ideas, but also from rigorous attention to seemingly small details. Approximate reasoning may reveal the essential physics, but careful attention to the details is what turns physical insight into reliable quantitative knowledge. In science, there are many “little” things, but being meticulous about the little things is a very big thing.

 

Historical Note:

In his paper On the Dynamical Theory of Gases, Maxwell (1867) writes: “It is to Professor CLAUSIUS, of Zurich, that we owe the most complete dynamical theory of gases. His other researches on the general dynamical theory of heat are well known, and his memoirs “On the kind of Motion which we call Heat,” are a complete exposition of the molecular theory adopted in this paper. After reading his investigation of the distance described by each molecule between successive collisions, I published some propositions on the motions and collisions of perfectly elastic spheres, and deduced several properties of gases, especially the law of equivalent volumes, and the nature of gaseous friction. I also gave a theory of diffusion of gases, which I now know to be erroneous, and there were several errors in my theory of the conduction of heat in gases which M. CLAUSIUS has pointed out in an elaborate memoir on that subject.”

 

Clausius’s Comment on Maxwell’s Numerical Error

Feynman’s remark was based on a footnote (See below) in Rudolf Clausius’s (1862) paper on the thermal conductivity of gases.

Thus, the discrepancy in Maxwell’s numerical result arose from unit conversion mistakes instead of incorrect physical idea underlying his model of thermal conduction.

 

Review Questions

1. What idealized assumptions underlie the derivation of thermal conductivity for a dilute gas? In particular, explain why a well-defined local temperature, a sufficiently small temperature gradient, and a mean free path much smaller than the characteristic system length are each necessary for the derivation.

2. Feynman presents two approximate formulas for the thermal conductivity of a dilute gas—k = (nvl)k_B/(g - 1) and k = k_Bv/s_c(g - 1). What does each formula reveal about the microscopic physics of heat conduction? In particular, why does the second expression imply that thermal conductivity is independent of gas density, and under what conditions does that conclusion hold?

3. Under what conditions do these approximate formulas, and the Fourier-law description more generally, become unreliable for a dilute gas? Identify the relevant physical regimes—such as the Knudsen regime, high-density non-ideal gases, or convection-dominated transport—and explain what physical descriptions replace them.

 

References:

Clausius, R. J. E. (1862). Ueber die wärmeleitung gasförmiger körper. (On the Conduction of Heat in Gases). Annalen der Physik, 191(1), pp.1-56.

Feynman, R. P., Leighton, R. B., & Sands, M. (1963). The Feynman Lectures on Physics, Vol I: Mainly mechanics, radiation, and heat. Reading, MA: Addison-Wesley.

Fick, A. (1855). On Liquid Diffusion. Philosophical Magazine, Vol. 10, pp. 30–39.

Grattan-Guinness, I. (2005). Joseph Fourier, Théorie analytique de la chaleur (1822). In Landmark Writings in Western Mathematics 1640-1940 (pp. 354-365). Elsevier Science.

Maxwell, J. C. (1867). On the dynamical theory of gases. Philosophical transactions of the Royal Society of London, 157, 49-88.

Reif, F. (1965). Fundamentals of Statistical and Thermal Physics. McGraw-Hill.