Soft matter
Reptation: How a Polymer Chain Snakes Through Its Own Melt
Double the length of a polymer chain and its melt becomes roughly ten times more viscous — a molten-plastic honey that flows about 2500 times slower for a tenfold longer chain. The reason is not friction in the ordinary sense but topology: a long chain cannot cross its neighbors, so it is trapped in a virtual "tube" of surrounding molecules and can only escape by wriggling along its own contour like a snake through tall grass. Pierre-Gilles de Gennes named this motion reptation in 1971, from the Latin reptare, to creep.
Reptation is the dominant relaxation mechanism for entangled polymer melts and concentrated solutions above a critical chain length. It predicts that the center-of-mass diffusion coefficient scales as D ∝ N⁻² and the terminal relaxation time and viscosity as τ ∝ N³ — a set of power laws that turned polymer rheology from empirical art into predictive physics.
- RegimeEntangled polymer melts / concentrated solutions, N > N_e
- Key relationsD ∝ N⁻², τ_rep ∝ N³, η ∝ M^3.4 (measured)
- Proposedde Gennes 1971; tube model formalized by Doi & Edwards 1978-79
- Characteristic scaleTube diameter a ≈ 4-6 nm; entanglement mass M_e ~ 1-2 kg/mol
- Realized/measured inNeutron spin echo on polyethylene/PEP; PFG-NMR diffusion; melt rheology
- Matters forPlastics processing, rubber elasticity, DNA electrophoresis, self-healing
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What reptation is and why it matters
A polymer melt is a dense tangle of long chain molecules — think of a bowl of cooked spaghetti with no sauce. Above a critical molecular weight, chains cannot pass through one another; these topological constraints (entanglements) dominate the dynamics and give melts their remarkable viscoelasticity: they flow like liquids on long timescales yet snap back elastically on short ones.
Reptation is the idealized answer to a deceptively simple question: how does one such trapped chain move? De Gennes' insight was that the surrounding, uncrossable chains confine the test chain to an effective tube running along its own backbone. The chain cannot move sideways past its neighbors, so its only large-scale escape route is to slither back and forth along the tube axis, gradually abandoning the old tube and creating new one at its ends.
This picture matters because it converts a hopeless many-body entanglement problem into a tractable single-chain diffusion problem, and it predicts — correctly, to a good approximation — the sharp molecular-weight dependence of melt viscosity that governs how every plastic is extruded, molded, and processed.
The mechanism, step by step
Coarse-grain the tube to its center line, the primitive path, of contour length L ≈ Nb²/a, where b is the monomer size and a the tube diameter. The chain performs one-dimensional curvilinear diffusion along this path with a curvilinear diffusion coefficient D_c = k_BT/(Nζ), where ζ is the monomeric friction — this is just Rouse (unentangled) friction acting along the tube, since local motion is unconstrained.
The chain fully forgets its original configuration only when it has diffused a curvilinear distance of order L, escaping the entire tube. The time for this is the reptation time τ_rep ≈ L²/D_c. Substituting L ∝ N and D_c ∝ N⁻¹ gives τ_rep ∝ N³. Because the chain travels a large curvilinear distance L to achieve only a modest real-space displacement (its end-to-end size R ∝ N^½), the center-of-mass diffusion is strongly suppressed: D ≈ R²/τ_rep ∝ N/N³ = N⁻². What balances what: entropic elasticity drives configurational renewal, monomeric friction resists it, and topology forces the escape to happen one-dimensionally.
Key equations, scales, and the four dynamical regimes
The central results are compact:
D ≈ (1/3)·(k_BT/ζ)·(a²/b²)·N⁻² and τ_rep ≈ (ζ b²/k_BT)·(N³ b²/a²) ∝ N³.
Three timescales order the motion, set by the entanglement segment N_e = a²/b²: the entanglement time τ_e (Rouse time of one tube segment), the Rouse time τ_R ≈ τ_e (N/N_e)², and τ_rep ≈ 3 τ_R (N/N_e). The monomer mean-square displacement g₁(t) therefore passes through four regimes: t^½ (free Rouse for t < τ_e), t^¼ (constrained Rouse inside the tube for τ_e < t < τ_R), t^½ (reptation, curvilinear diffusion along a random-walk tube for τ_R < t < τ_rep), and finally t¹ (free Fickian diffusion) beyond τ_rep. Characteristic numbers: the tube diameter a ≈ 4-6 nm, entanglement molar mass M_e ~ 1-2 kg/mol, and the plateau modulus G_N⁰ = (4/5)·ρRT/M_e, of order 0.1-2 MPa for common melts.
How it is measured: neutron spin echo and the tube
The tube is a hypothesis about geometry, so the decisive test is a microscopic one. Neutron spin echo (NSE) spectroscopy, pioneered for this purpose at the Jülich reactor by Dieter Richter and collaborators (from the 1990s onward), measures the single-chain dynamic structure factor S(q,t)/S(q,0) on nanometer length scales and nanosecond-to-hundred-nanosecond timescales, using deuterium labeling to isolate one chain in a melt.
Reptation predicts a striking signature: at intermediate times S(q,t) does not decay to zero but levels off at a q-dependent plateau, because transverse motion is confined to the tube. Fitting the de Gennes local-reptation expression yields the tube diameter directly — for polyethylene and poly(ethylene-propylene), a ≈ 4-6 nm, in quantitative agreement with the value inferred independently from the plateau modulus. Pulsed-field-gradient NMR and forced-Rayleigh-scattering diffusion measurements confirm D ∝ N⁻² over decades in molecular weight, closing the loop between microscopic confinement and macroscopic transport.
Where it operates, and how it differs from Rouse and Zimm
Reptation is the correct picture only for linear, flexible chains that are long enough to be entangled (N ≫ N_e) in a melt or a concentrated/semidilute solution. Below N_e a chain feels no topological cage and relaxes by Rouse dynamics (D ∝ N⁻¹, τ ∝ N²); in dilute solution, hydrodynamic backflow dominates and one has Zimm dynamics instead. The distinction is sharp: reptation is a topological, not hydrodynamic, effect — solvent-mediated forces are screened out in a melt.
Architecture also breaks it. A star or branched polymer cannot reptate — its branch point cannot follow the arm down a tube — and relaxes instead by exponentially slow arm retraction, giving η growing exponentially rather than as a power of arm length. Ring polymers, having no free ends, cannot reptate at all and remain an active puzzle. The pure reptation exponent τ ∝ N³ is likewise an idealization: real melts show η ∝ M^3.4, discussed next.
Applications, corrections, and open questions
Reptation underpins the entire rheology of plastics processing — the η ∝ M^3.4 law sets extrusion pressures and injection-molding windows — and explains rubber elasticity, adhesion, polymer welding and self-healing, where interfacial strength grows as chains reptate across a crack. It also governs DNA gel electrophoresis: a charged DNA coil reptates head-first through the gel's pore network, the basis of the biased-reptation model of size separation.
The famous discrepancy — theory gives N³, experiment N^3.4 — is not a failure but a finite-length correction. Doi (1980s) and later Milner and McLeish showed that contour-length fluctuations (CLF) (the tube ends breathe in and out) and constraint release (the tube itself decays as neighbors move) accelerate relaxation for finite N, steepening the effective exponent; both effects vanish as N → ∞, recovering the pure reptation limit. NSE has directly imaged CLF loosening the confinement in shorter chains. Open frontiers include the dynamics of rings and unconcatenated loops, nonlinear flow and shear banding, and reconciling the tube's rheological and microscopic diameters — which NSE finds to differ by a factor of order two.
| Quantity | Rouse (unentangled, N < N_e) | Reptation / tube (N ≫ N_e) | Physical origin |
|---|---|---|---|
| Diffusion coefficient D | ∝ N⁻¹ | ∝ N⁻² | Curvilinear escape from a tube of length L ∝ N |
| Longest relaxation time τ | ∝ N² | ∝ N³ (reptation) | τ_rep = L²/D_c, both grow with N |
| Zero-shear viscosity η | ∝ N¹ | ∝ N³ (theory), N^3.4 (experiment) | Terminal stress relaxation; CLF correction |
| Stress relaxation G(t) | Broad power law, no plateau | Rubbery plateau G_N⁰ then tube renewal | Entanglement network of lifetime τ_rep |
| Mean-square displacement g₁(t) | t^½ (single crossover) | t^½ → t^¼ → t^½ → t¹ | Four regimes set by τ_e, τ_R, τ_rep |
Frequently asked questions
Why does the diffusion coefficient scale as N⁻² instead of the N⁻¹ of an ordinary chain?
Along its tube the chain diffuses with curvilinear coefficient D_c ∝ N⁻¹ (ordinary Rouse friction). But to renew its configuration it must travel a curvilinear distance L ∝ N, taking time τ_rep = L²/D_c ∝ N³, while achieving a real-space displacement of only its coil size R ∝ N^½. The center-of-mass diffusion D ≈ R²/τ_rep ∝ N⁻² therefore picks up an extra factor of N from the topological confinement that forces one-dimensional escape.
What exactly is the 'tube' — is it a real object?
No, the tube is a mean-field representation of the many uncrossable neighboring chains. Their collective effect on a test chain is to permit motion along its own contour while suppressing lateral excursions beyond a diameter a. The tube's center line is the primitive path. Its reality is operational: neutron spin echo measures a confinement length of 4-6 nm that matches the tube diameter inferred from the plateau modulus, so the geometry it encodes is physically observable even though no wall exists.
Why is the measured viscosity exponent 3.4 rather than the predicted 3?
Pure reptation gives η ∝ N³ only in the infinite-chain limit. For finite chains, contour-length fluctuations (the tube ends 'breathe') and constraint release (neighboring tubes decay) provide faster relaxation channels that add a molecular-weight-dependent correction, steepening the apparent slope to about 3.4 over the experimentally accessible range. As N → ∞ these corrections become negligible and the exponent crosses back toward 3.
How does reptation differ from Rouse and Zimm dynamics?
Rouse dynamics describes unentangled chains (N < N_e) with no hydrodynamic interactions: D ∝ N⁻¹, τ ∝ N². Zimm dynamics applies to dilute solutions where solvent backflow couples the monomers hydrodynamically. Reptation applies to entangled melts and concentrated solutions where hydrodynamics is screened but topology dominates, giving D ∝ N⁻² and τ ∝ N³. The three are distinguished by whether entanglements and hydrodynamics matter.
Can branched or ring polymers reptate?
Not in the standard sense. A star polymer's branch point cannot thread down a single tube, so its arms relax by activated 'arm retraction,' giving viscosity that grows exponentially with arm length rather than as a power law. Ring polymers have no chain ends to lead the reptative motion, so they relax by amoeba-like loop rearrangements; their dynamics remain an active and unresolved research area.
What is the experimental signature that most directly confirms the tube picture?
The single-chain dynamic structure factor S(q,t) measured by neutron spin echo. Instead of decaying to zero, it plateaus at intermediate times because transverse monomer motion is confined to the tube. Fitting de Gennes' local-reptation form to this plateau yields the tube diameter directly (a ≈ 4-6 nm for polyethylene), independently confirmed by the rubbery plateau modulus and by pulsed-field-gradient NMR showing D ∝ N⁻².