Section 43–2 The mean free path

Mean free path / Collision cross section / l-σ relation  

 

In this section, Feynman discusses the mean free path, the collision cross section, and the l-σ relation that connects them. These ideas resolve a paradox in kinetic theory of gases: if gas molecules move at high speeds, why do gases mix and diffuse slowly? The answer is that molecules do not travel unimpeded in straight lines, but they undergo incessant collisions in a chaotic random walk. The concept of mean free path was introduced by Clausius in his 1858 paper On the mean lengths of the paths described by the separate molecules of gaseous bodies, where he provided the first quantitative description of the average distance traveled by a molecule between successive collisions.

 

1. Mean free path

“If we say that the average time between collisions is τ, and that the molecules have a mean velocity v, we can expect that the average distance between collisions, which we shall call l, is just the product of τ and v. This distance between collisions is usually called the mean free path:

Mean free path l = τv (Feynman et al., 1963).”

 

A Comprehensive Definition of Mean Free Path

The mean free path (l) is the average distance a particle travels between successive collisions in a dilute gas. It is a statistical quantity that links the microscopic collisions to macroscopic transport phenomena such as diffusion, viscosity, and thermal conductivity.

Kinematic Perspective: It is the product of the particle's mean speed (v) and the mean collision time (t), l(v) = v t(v), which is directly related to the Maxwellian speeds. 

Geometric Perspective: For identical hard-spheres with diameter (d) and number density (n), the collision cross section is s = pd^2. Under the stationary target assumption, the mean free path is l =1/ns—it is inversely proportional to both number density and collision cross-section.

Statistical Perspective: The collision-free distances follow an exponential distribution, P(x)=(1/l)e^{-x/l], where P(x) is the probability that a particle travels a distance x. Thus, most particles travel less than l, while relatively fewer travel longer without any collision.

Thermodynamic Perspective: the mean free path also depends on the macroscopic states, e.g., at constant pressure, l = kT/(Ö2)sp, so increasing the temperature increases the mean free path.

       Thus, the mean free path is more than a simple average distance between collisions. It is a unifying parameter that connects the kinematics of particle motion, the geometry of molecular encounters, the statistics of random collisions, and the thermodynamic state of the gas.

 

The Stationary-Target model

Feynman’s formula is based on a simplifying assumption: all scattered molecules (targets) are treated as if they were stationary relative to the moving molecule. In reality, every molecule is moving with a high speed. What determines the collision rate is therefore not the speed of a single molecule relative to the laboratory, but the relative speed between pairs of molecules.

Suppose two identical molecules have velocity vectors v₁ and v₂, the relative velocity is

v_rel = v₁ - v₂

The mean-square relative speed is

<v_rel²> = <(v₁ - v₂)²>

Expanding the squares gives,

<v_rel²> = <v₁²> + <v₂²> - 2<v_1·v₂>

In thermal equilibrium, the velocities of different molecules are uncorrelated, so <v_1·v₂> = 0
Since the molecules are identical,

<v₁²> = <v₂²> = <v²> Þ  <v_rel²> = 2<v²>.

Taking the square root gives the mean relative speed

v_rel = (Ö2)v

where (v) denotes the corresponding root-mean-square molecular speed.

 

Physically, this means that collisions occur Ö2 times more frequently than predicted by the stationary-target model because every moving molecule contributes to the relative motion.

 

Note: In the audio recording (15 min: 45 sec) of this lecture, Feynman says something like: “Let’s say the nuclides are standing still…”, but this assumption is not included by the editors. The stationary-target model is intended as a simplification rather than an accurate derivation.

 

2. Collision cross section

“By ‘collision cross section’ we mean the area within which the center of our particle must be located if it is to collide with a particular molecule. (Feynman et al., 1963).”

The collision cross section, usually denoted by s, is a measure of the probability that two particles will collide or interact when they approach one another. Although it has the dimensions of an area, it should not always be interpreted as the particle's actual physical size. In the classical hard-sphere model, the collision cross-section is a fixed geometric area, which is independent of direction of approach. However, in Maxwell’s (1867) kinetic theory, the force between molecules is assumed to vary inversely as the fifth power of their separation. Under this assumption, the collision cross-section is no longer constant but varies inversely with the relative speed of the molecules. In general, the collision cross‑section is not a single, immutable number but a dynamical quantity—an effective interaction area that depends on the forces, speeds, and nature of the particles involved. It is the fundamental parameter that governs how frequently particles scatter, how far they travel between collisions, and ultimately how gases conduct heat, diffuse, and flow.

“If molecules were little spheres (a classical picture) we would expect that σ_c=π(r₁+r₂)², where r₁ and r₂ are the radii of the two colliding objects (Feynman et al., 1963).”

A Feynman-Style Explanation of the Collision Cross Section

Imagine a large male dancer standing in the middle of a crowded dance floor. Let us say he has a radius R. Moving through the crowd is a smaller female dancer with radius r. Both are free to move, but to keep things simple, let us freeze the large dancer and let only the smaller one move.

Now the question is: When do they “collide”?

 

A casual observer might say, "When her elbow touches his belly." But for a physicist, that’s sloppy thinking. We can ignore the complicated shapes of bodies and keep track of their centers. At the instant the two dancers just graze one another—when her shoulder just brushes his arm—the distance between their centers is neither R nor r. It is the sum: R + r. Why? Her center is one r away from her edge, and his center is one R away from his edge. When the two surfaces just touch, the centers must be separated by the sum of their radii.

 

Now comes the clever trick.

Keeping track of two moving objects with finite size is inconvenient. So we cheat. We shrink the smaller dancer down to a mathematical point. Of course, doing that would change the physics—however, we can transfer the smaller dancer’s radius to the larger dancer.

Mathematically, the larger dancer acquires an effective radius R+r.

 

In other words, we can idealize the female dancer as a point zipping around, and the male dancer becomes am inflated stationary target. She will collide with him if her center ever reaches a circle of radius R+r centered on him.

 

What's the collision cross-section then? 

It is simply the area of that inflated target—the area of a circle with radius (R+r):

s = p(R+r)²

What does it mean?

The collision cross section is determined not by the size of either particle alone, but by the combined radius of both particles. Collisions depend on how close the centers of the particles can approach, not simply on the so-called physical size of one particle.

 

However, this is only the beginning. One may modify the explanation or analogy to include, for example, the interaction forces, relative speeds, and quantum effects.

 

3. l-σ relation 

“… if we write it as σ_cn₀l=1. (43.12). This formula can be thought of as saying that there should be one collision, on the average, when the particle goes through a distance l in which the scattering molecules could just cover the total area. In a cylindrical volume of length l and a base of unit area, there are n₀l scatterers; if each one has an area σ_c the total area covered is n₀lσ_c, which is just one unit of area (Feynman et al., 1963).”

 

According to Feynman, the formula σnl = 1 can be visualized as follows: a particle traveling a distance l sweeps out a cylindrical volume of cross-sectional area σ. On the average, this cylinder contains exactly one scattering target, so one collision is expected over that distance. However, this formula rests on a hidden assumption: the scattering molecules are treated as stationary targets, which is not realistic. For a gas of identical molecules obeying the Maxwell-Boltzmann distribution, the average relative speed of two colliding molecules exceeds the mean molecular speed by a factor of Ö2. This correction results in l = 1/(Ö2)nσ, which implies: nlσ = 1/Ö2 » 0.707, not 1. Thus, the relation nlσ = 1 is an approximation—a pedagogical device for grasping the scaling behavior, but not the exact statistical result.  It is also known as the l-s relation, which expresses the inverse proportionality between the mean free path and the collision cross‑section: larger cross‑section, shorter path; smaller cross‑section, longer path. Its true value lies not in numerical precision but in the physical intuition it provides.

 

Real-World Nuance: It's Not Just Physical Size!

The l-σ relation provides a simple geometric picture of molecular collisions, allowing the concept of mean free path intuitively accessible. However, its simplicity comes at the cost of accuracy. Here are some key limitations.

1. Charged Particles: Attractive forces can pull particles together from distances far larger than their physical dimensions, effectively increasing the collision cross section, whereas repulsive forces can deflect particles before they come close, shrinking it.

2. Speed Dependence: At high relative speeds, particles zip past one another so quickly that they barely have time to feel the force—the effective cross‑section shrinks. At low speeds, they interact for longer periods, allowing attractive or repulsive forces to exert a greater influence and thereby modifying the collision probability.

3. Quantum Effects: At atomic and subatomic scales, the de Broglie wavelength of a particle can become effectively larger, diffraction and interference effects then alter the effective cross‑section in ways that have nothing to do with physical size.

Thus, while the geometric picture provides a useful starting point, the collision cross section is essentially a measure of interaction probability, shaped by interaction forces, relative motion, and, in many cases, quantum mechanics. Thus, the mean free path is ultimately determined not simply by geometry, but by the underlying physics governing particle interactions.

Key Takeaways:

1. Mean Free Path: The Average Distance Between Collisions

The mean free path is the average distance a particle travels between successive collisions. It is not the distance traveled by every particle, but rather the statistical average of many randomly distributed free-flight distances.

 

2. Collision Cross Section: The Effective Interaction Area

The collision cross section measures the probability that two particles will collide or interact. Although it has units of area, it should generally be interpreted as an effective interaction area rather than the particle’s “actual” geometric size. For ideal hard spheres, the cross-section is simply a geometric area, but for real molecules or ions, it depends on interaction force, relatively velocity, and quantum effects.

 

3. The l-σ Relation: Geometry Meets Probability

The formula nlσ = 1 means that, on average, a particle traveling a distance   sweeps out a cylindrical volume containing exactly one scattering target, so one collision is expected over that distance. This relation provides a geometric interpretation of the mean free path, l = 1/nσ, but it is an approximation that assumes stationary scatterers and neglects the relative motion of all particles in a real gas.

 

In short, the collision cross section tells us how large a target a particle presents, while the mean free path tells us how far a particle typically moves before hitting another. Increasing the number density of particles or enlarging their effective collision cross section increases the collision frequency and shortens the mean free path. This simple idea forms the microscopic foundation of diffusion, electrical conduction, viscosity, and many other transport phenomena.

 

A Real-Life Application: The Science of Fart Odor Transport The Source: Chemistry and Transport Physics The characteristic odor flatulence arises from volatile sulfur compounds, such as hydrogen sulfide and methanethiol. Although these compounds make up only a tiny fraction of the gas mixture, the human nose is remarkably sensitive to them. Once release, the odor molecules undergo a multi-stage process: Jet Advection – The gas is initially expelled as a turbulent jet, rapidly mixing with the surrounding air. Thermal convection –Released at about body temperature, the flatus may initially rise due to buoyancy. As it cools, this effect gradually diminishes.  Molecular diffusion: On a microscopic scale, individual odor molecules undergo random thermal motion, colliding with air molecules and slowly spreading the smell outward. Together, advection, convection, and diffusion result in how quickly the odor spreads.   Why Distance Matters There is no sharp boundary beyond which an odor suddenly vanishes—the concentration decreases continuously as the molecules disperse. The farther one stands from the source, the more the odor has been diluted by mixing with the air. Consequently, fewer odor molecules reach the olfactory receptors in the nose, reducing the probability that the smell exceeds the human detection threshold. Ventilation, air currents, room geometry, humidity, and temperature all affect the diffusion rate. In a poorly ventilated room, an odor may linger for several minutes; with good airflow, it dissipates in seconds.   The Probabilistic Nature of Smell Detecting an odor is fundamentally a probabilistic process. At high concentrations, a large number of odor molecules reach the nose every second, making detection almost certain. As the concentration falls, fewer molecules arrive, and the likelihood of detecting the smell decreases. Eventually the concentration drops below the detection threshold, and the odor becomes imperceptible, even though a small number of molecules may still be present.   Practical Takeaways To minimize significant olfactory exposure: Distance helps. Greater separation generally reduces odor concentration and therefore the probability of detection. Ventilation is even better. Fresh air rapidly dilutes and removes odor molecules from the room. Activated carbon masks are good at adsorbing sulfur compounds. Standard cloth, surgical, and N95 masks, however, are particulate filters—they capture aerosols, not gaseous molecules—and lack the necessary carbon bed to reduce odor in any meaningful way. If feasible, a small hand-held fan can be used to redirect the local airflow, spreading the odor away from your breathing zone and back toward its source.   Conclusion The next time a social situation requires a scientific description of “safety distance,” you can respond with a blend of statistical mechanics and social grace:  “Please maintain a radius of about three meters—or roughly two exponential decay lengths from the source—to guarantee negligible odor.” However, Feynman might have concluded it as follows: The molecules don’t know whether they’re carrying perfumes or unhealthy fumes—they simply obey the laws of physics.

 

The Moral of the Lesson: Mean Free Path and COVID‑19 The concept of mean free path provides a useful way of thinking about the spread of airborne diseases such as COVID-19. In this context, however, the idea appears in two different forms (Fernández-Terán et al., 2020). At the microscopic level, it describes how far virus-laden aerosols travel before colliding with another air molecule. At the macroscopic level, it serves as an analogy for the distance traveled by individuals in a crowd before encountering an infected person.   1. Safe Distancing: Dilution of Aerosol Concentration An infected individual continuously emits viral-laden aerosols through breathing, speaking, or sneezing. These aerosols are initially most concentrated near the source. As they spread through diffusion, gravitational settling, and ventilation, their concentration decreases (sometimes exponentially) with distance from the source. By increasing the physical separation between individuals, the aerosol concentration is effectively lowered. Statistically, a lower concentration means fewer virus-containing aerosols enter the breathing zone of an individual and therefore a lower probability of inhaling an infectious dose.   2. Wearing a Mask: Increasing the Probability of Collisions A mask is not a simple sieve. It is a complex three‑dimensional network of microscopic fibers that acts as a highly efficient particle filter. From the perspective of an incoming viral aerosol, the fibers form a dense array of collision targets. Without a mask: The air is equivalent to a low density of solid obstacles, allowing aerosols to move relatively unobstructed into the respiratory tract. With a mask: The probability of an aerosol encountering and adhering to a fiber increases dramatically. The result is a substantial reduction in the number of virus-laden aerosols that penetrate the filter and reach the lungs.   3. The Selection of Masks The effectiveness of a mask can be intuitively understood through the formula σnl = 1:   σ is the “effective” collision cross-section. n is the fiber number density (or density of collision targets). l is the mean free path. A useful analogy: increasing the number density of collision targets and their effective cross section reduces the distance a particle can move before an interaction occurs. This is why effective respirators (e.g., N95) fulfil the formula by combining a dense fiber network with suitable material to enhance aerosol capture. The goal is not to eliminate all aerosol motion but to maximize the chance that viral aerosols are intercepted before reaching the respiratory system.   Conclusion: COVID-19 transmission is not solely a problem for epidemiologists—it is also a problem in transport physics. By understanding the mean free path and collision cross-section, we can design effective interventions: safe distancing lengthens the total path particles must travel and masks reduce mean free path. Together, they lower the probability that viral aerosols complete their journey from one person to another.   A Note on COVID‑19 and Long-Term Health Risks Researchers continue to investigate the long-term consequences of SARS‑CoV‑2 infection, which extend beyond the acute respiratory phase to encompass potential effects on the cardiovascular, neurological, and respiratory systems (Tanrıverdi, 2024). Emerging evidence has also raised questions about possible links between COVID‑19 and cancer development. However, it is important to emphasize that current evidence remains preliminary, and no definitive causal relationship between COVID-19 and can cancer has been established. The field is still emerging, and further research is needed to clarify the underlying mechanisms and quantify long-term risks. This reinforces the importance of understanding aerosol transport: reducing exposure to viral droplets is not just about acute infection, but potentially about long‑term health consequences.

 

Richard P. Feynman To Yetta Farber, MARCH 30, 1982 Ms.Yetta Farber wrote to remind Feynman that she had once dated him at Cornell. She also had a story to tell that made her laugh every time she saw Feynman’s name in the papers. Immediately after their enjoyable date, word had gone out that there was a rapist, described as wearing “a brown or brown leather-like jacket,” loose on the Cornell campus. “A-hah! I said—I went out with this nice fellow and he wore a brown leather-like jacket! Maybe it was he. When you called me for another date, I said, ‘No, I’m busy.’” Ms. Farber had thought that Feynman was too young to be an assistant professor (in fact, Feynman was a full professor), so she was rather suspicious.   Dear Yetta, Naturally I could never understand why the girls I went out with in Ithaca wouldn’t go out with me again. At last I find out—it was my brown leather jacket! So often, was I thus frustrated by pretty girls (like you) that I came out to California. Since the weather was so much better I threw away my leather jacket and at last found someone who would go out with me more than once—so I married her.   I always thought that the girls in California were more tolerant—but now I know the inner workings of the phenomenon. Physics is much easier to understand.   Your former date, Richard P. Feynman (Source: Feynman, 2005, p. 345)   Feynman’s humorous letter about his brown leather jacket provides a valuable lesson in transport phenomena. The tragedy of Feynman’s dating life in Ithaca, New York, is not a charming anecdote—it is a real-world lesson in transport theory, molecular diffusion, and the physics of dating.   The Moral of the Lesson: Feynman’s Jacket Odor and the Mean Free Path 1. The Source: A Reservoir of Odor Molecules A well-worn leather jacket can act as a reservoir of volatile organic molecules (sweat, bacteria, and skin oils…). These molecules continuously evaporate from the jacket, creating a region of elevated odor around the jacket.   2. The Mean Free Path In short, the diffusion is slow and the mean free path is tiny. Although an odor molecule moves at hundreds of meters per second, they do not travel directly from the jacket to another person’s nose. It constantly collides with air molecules in a random walk. A molecule’s mean free path is typically less than a micrometer. Because it bounces backward, forward, and sideways continually, the molecule undergoes random walk, causing the smell to diffuse outward slowly.   3. Why Proximity Matters The concentration of odor molecules decreases with distance from the source. At a distance: When Feynman first met someone, he was standing far enough away that the flux of odor molecules reaching the girl’s nose was negligibly small. The "collision cross-section" of her olfactory receptors didn’t get enough molecules to trigger a warning. During a date: Once in close proximity—sitting across a table or walking side by side—the separation distance decreased and the exposure time increased. Under these conditions, enough odor molecules could diffuse to the olfactory receptors to confirm the jacket as a source.   A Reflection: The Hidden Variable The most amusing aspect of Feynman’s story is that he initially searched for the explanation in the wrong place. He thought the difference lay between the women of Ithaca and those of California. Eventually, he discovered that the “unknown” variable was neither geography nor human psychology, but his jacket. This is a lesson familiar to every physicist: when observations disagree with expectations, look for a hidden variable before constructing a new theory.   The Ultimate Takeaway In physics, as in dating, proximity changes everything. Thanks to a short mean free path, diffusion is inherently slow. A localized hazard—whether odor molecules, viral aerosols, or any other unwanted transport—can remain completely undetectable at a distance. The flux across a short distance could be surprisingly fast; you cannot rely on the sluggishness of diffusion to protect you. To lower the hazard rate of an undesirable transport process, do not rely on slow diffusion to save you—change the boundary conditions by eliminating the source at the boundary. And, for heaven’s sake, throw away the jacket or at least give it a generous spray of perfume.

 

Review Questions:

1. Explain why the stationary-target assumption is introduced when deriving the classical expression for the mean free path.

2. How would you explain the expression for the collision cross‑section of two different molecules?

3. How would you explain the physical meaning of the l-s relation and its limitations?

References:

Clausius, R. (1858). On the mean lengths of the paths described by the separate molecules of gaseous bodies. Philosophical Magazine, 15(101), 417–424.

Fernández-Terán, R., Sucre-Rosales, E., Echevarría, L., & Hernández, F. E. (2020). Social distancing during the COVID-19 pandemic: an analogy to explain collision cross-sections in chemical kinetics. Journal of Chemical Education, 97(12), 4540-4544.

Feynman, R. P. (2005). Perfectly reasonable deviations from the Beaten track: The letters of Richard P. Feynman (M. Feynman, ed.). New York: Basic Books.

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.

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

Tanrıverdi, Ö., Alkan, A., Karaoglu, T., Kitaplı, S., & Yildiz, A. (2024). COVID-19 and carcinogenesis: exploring the hidden links. Cureus, 16(8), e68303.