The Kuhn Segment: Coarse-Graining a Real Chain into Freely-Jointed Links

What the Kuhn segment actually is

Real polymer backbones are stiff on short length scales. In polyethylene the C–C bonds are fixed at the tetrahedral angle, so a bond cannot point in an arbitrary direction relative to its neighbor — its orientation is strongly correlated with the previous few bonds. The freely-jointed chain (FJC), by contrast, is the idealization in which every link points in a random direction independent of all others, giving the clean result ⟨R²⟩ = N·b² for a chain of N links each of length b. The problem: no real bond is freely jointed, so N and b are not the number and length of real bonds.

Kuhn's insight (1934) was that you don't need the microscopic bonds to be independent — you need only to find a larger length over which orientational memory has effectively died. Group enough real bonds together and the summed vector connecting the ends of that group points in a direction nearly uncorrelated with the next group. Kuhn defined the Kuhn segment as exactly that coarse-grained link, and its length b as the Kuhn length. The chain of these segments is, to a very good approximation, an ideal freely-jointed chain.

Crucially, b and N are not free to choose — they are fixed by demanding that the equivalent chain reproduce two real quantities: the fully-extended contour length R_max and the unperturbed mean-square end-to-end distance ⟨R²⟩₀. Two equations, N·b = R_max and N·b² = ⟨R²⟩₀, pin down both unknowns uniquely.

The derivation: two constraints fix b and N

Start from the two matching conditions. Dividing them gives the Kuhn length directly:

• N·b = R_max  (match the maximum extension)
• N·b² = ⟨R²⟩₀  (match the coil size)

Divide the second by the first: b = ⟨R²⟩₀ / R_max. Substitute back: N = R_max² / ⟨R²⟩₀ = R_max / b. So everything reduces to two measurable numbers. The contour length for a vinyl backbone of n bonds of length l₀ = 0.154 nm, all-trans, is R_max = n·l₀·cos(θ/2), where the half-angle accounts for the zig-zag (θ ≈ 68° is the supplement of the 112° bond angle, so cos(θ/2) ≈ 0.83).

The mean-square size is written compactly through Flory's characteristic ratio C_∞: ⟨R²⟩₀ = C_∞·n·l₀². C_∞ measures how much stiffer the real chain is than a hypothetical freely-jointed chain of the same bonds — it is 1 for a truly free chain and rises with backbone rigidity (≈ 6.7–7.4 for PE depending on temperature and measurement, ≈ 9.5 for atactic polystyrene, ≈ 4.7–5.0 for cis-1,4-polyisoprene). Substituting both expressions gives the working formula b = C_∞·l₀ / cos(θ/2), tying the Kuhn length to a single dimensionless stiffness parameter and the geometry.

For semiflexible chains it is cleaner to route through the worm-like chain (WLC) model of Kratky and Porod (1949), where stiffness is captured by the persistence length l_p — the decay length of the bond-orientation correlation, ⟨cos θ(s)⟩ = exp(−s/l_p). In the long-chain limit the WLC gives ⟨R²⟩₀ = 2·l_p·R_max, and comparing to N·b² = b·R_max yields the celebrated identity b = 2·l_p. The Kuhn length is exactly twice the persistence length.

Worked example: polyethylene and DNA by the numbers

Polyethylene. Take a chain of molar mass 140 kg/mol. Each –CH₂– unit is 14 g/mol, so there are n ≈ 10,000 backbone carbons, i.e. ≈ 10,000 C–C bonds of l₀ = 0.154 nm. The all-trans contour length is R_max = n·l₀·cos(θ/2) = 10,000 × 0.154 nm × 0.83 ≈ 1280 nm. With C_∞ ≈ 7.4 (the melt value from neutron scattering), ⟨R²⟩₀ = C_∞·n·l₀² = 7.4 × 10,000 × (0.154 nm)² ≈ 1755 nm². Then b = ⟨R²⟩₀ / R_max ≈ 1755 / 1280 ≈ 1.37 nm, and N = R_max / b ≈ 935 Kuhn segments. So ~10 real bonds fold into each Kuhn link, and a chain that looks like 10,000 constrained bonds behaves like ~900 free ones. The coil radius of gyration follows as R_g = (⟨R²⟩₀/6)^(1/2) ≈ 17 nm.

Double-stranded DNA. Here the numbers are wildly different because dsDNA is a stiff rod on the nanoscale. Its persistence length in physiological salt is l_p ≈ 50 nm (≈ 150 base pairs at 0.34 nm/bp), so the Kuhn length is b = 2·l_p ≈ 100 nm. A 50 kbp fragment has R_max ≈ 50,000 × 0.34 nm = 17,000 nm = 17 μm and N = R_max/b ≈ 170 Kuhn segments. Its ideal size is ⟨R²⟩₀^(1/2) = (N·b²)^(1/2) = (170 × (100 nm)²)^(1/2) ≈ 1300 nm ≈ 1.3 μm — a coil a thousandth of its stretched length. The same two equations that tamed floppy polyethylene handle a molecule 70× stiffer without modification.

Why coarse-graining to Kuhn segments matters

The payoff is universality. Once you express a chain in Kuhn units, its large-scale statistics no longer remember the chemistry — a nylon, a polystyrene, and a strand of DNA all become the same N-link random walk. This is why de Gennes' scaling theory and Flory's mean-field arguments can quote a single exponent (R ~ N^ν, with ν = 1/2 ideal, ν ≈ 0.588 for a self-avoiding walk in good solvent) without specifying the monomer. The Kuhn segment is the natural "lattice spacing" that makes the renormalization-group and blob pictures apply to any polymer.

It also sets the correct counting unit for entropy and elasticity. Classical rubber elasticity treats a network strand as N freely-jointed Kuhn links; the tension to stretch it, f = (3k_BT/N·b²)·R at small extension, and the full inverse-Langevin law f = (k_BT/b)·ℒ⁻¹(R/N·b) at large extension, both depend on b and N, not on the count of chemical bonds. Get the Kuhn length wrong and you predict the wrong modulus and the wrong finite-extensibility limit. The same Kuhn-based tension law, the Marko–Siggia WLC interpolation formula, is what optical-tweezer experiments fit to extract l_p (hence b) from single DNA molecules.

Finally, the Kuhn length is a diagnostic of backbone chemistry. A stiff aromatic or helical backbone has a large b (poly(p-phenylene) or a peptide α-helix can reach many nm); a flexible aliphatic ether like poly(ethylene oxide) has a small b (≈ 1.1 nm). Reading b off a scattering curve tells you, quantitatively, how conjugation, hydrogen bonding, or bulky side groups stiffen a chain.

Limits, subtleties, and common mistakes

The Kuhn construction is exact only for an ideal (unperturbed) chain — a melt, a glass, or a dilute solution at the θ-temperature where two-body excluded-volume repulsion is exactly cancelled by solvent-mediated attraction (the θ-condition, Flory 1949–53). In a good solvent the chain swells, R ~ N^0.588·b, and ⟨R²⟩ ≠ N·b²; the Kuhn length remains a meaningful local stiffness parameter, but you must not plug the swollen size into N·b² = ⟨R²⟩ to extract it. This is the single most common error.

A few further cautions worth stating plainly:

• b = 2·l_p only in the long-chain WLC limit. For short, rod-like chains (contour length ≲ l_p, e.g. a short DNA oligomer) the chain never randomizes, ⟨R²⟩ deviates from 2·l_p·R_max, and neither the FJC nor the simple identity applies.
• The Kuhn segment is not a physical object. No 1.4 nm rigid rod exists in polyethylene; b is a statistical bookkeeping length. Cutting a chain into literal b-length pieces is meaningless.
• C_∞ is temperature- and tacticity-dependent. It rises as the trans population grows on cooling and differs for isotactic vs. atactic backbones, so b is not a fixed constant of a chemical species — quote the conditions.
• Charged chains complicate l_p. For polyelectrolytes the electrostatic persistence length (Odijk–Skolnick–Fixman) adds to the intrinsic one and depends strongly on ionic strength, so b for DNA shrinks toward ~80–90 nm at high salt.

History and the broader family of models

Werner Kuhn published the coarse-graining idea in 1934 ("Über die Gestalt fadenförmiger Moleküle in Lösungen," Kolloid-Zeitschrift 68, 2), essentially contemporaneously with Guth and Mark's independent 1934 random-flight treatments and well before the statistical-mechanics machinery matured. Kuhn's segment gave the field its first quantitative bridge between microscopic chemistry and the Gaussian coil, and it is the conceptual ancestor of Flory's rotational-isomeric-state (RIS) theory (1969, Statistical Mechanics of Chain Molecules), which computes C_∞ — and hence b — from the actual gauche/trans energetics of a given backbone.

Two neighboring models complete the picture. The freely-rotating chain fixes the bond angle but lets torsions rotate freely, giving ⟨R²⟩ = n·l₀²·(1+cos θ)/(1−cos θ) with θ the backbone (supplement) angle, cos θ ≈ 1/3 — consistent with the ≈ 68° convention used above; adding a hindrance factor from real torsional preferences upgrades it to the RIS result. The Kratky–Porod worm-like chain (1949) treats the backbone as a continuous elastic filament with bending stiffness κ, where l_p = κ/k_BT; it is the model of choice for stiff biopolymers (DNA, actin, collagen) and interpolates smoothly between rigid rod and Gaussian coil.

All three converge on the same large-scale statistics, and all three can be summarized by a single number: the Kuhn length. That reduction — collapse every chemical detail into b, then treat the polymer as a random walk of N free links — remains the first calculation any polymer physicist does when handed a new chain, ninety years after Kuhn wrote it down.

The real chain and its Kuhn-equivalent freely-jointed chain describe the same two large-scale observables while differing on everything local.

Property	Real chain (e.g. polyethylene)	Equivalent Kuhn chain
Link length	l₀ ≈ 0.154 nm (C–C bond)	b ≈ 1.4 nm (Kuhn segment)
Number of links	n ≈ 10,000 backbone bonds	N = R_max/b ≈ 900
Bond-angle constraint	fixed θ ≈ 68° supplement (112° tetrahedral)	none — links hinge freely
Orientational correlation	decays over ~l_p (short-range memory)	zero beyond one link by construction
Contour length	R_max = n·l₀·cos(θ/2) ≈ 1260 nm	R_max = N·b ≈ 1260 nm (matched)
Mean-square size	⟨R²⟩₀ = C_∞·n·l₀²	⟨R²⟩₀ = N·b² (matched)