How a long chain fits into a small region
Place a piece of string on a table. Its length measured along the string stays the same, while the straight-line distance between its ends can vary enormously. A polymer has the same distinction: a large number of connected units does not mean that the molecule remains stretched into a long rod in solution. Many positions along its backbone can rotate, and thermal motion continually changes the pattern of bending. Shapes obtained without changing the molecular connectivity are called conformations. Understanding them is a starting point for explaining molecular dimensions in solution, the restoring force of rubber and the role of flexible linkers between biological molecules.
A flexible linker does more than prevent two binding partners from separating completely. Its length and flexibility determine their typical separation, the number of arrangements that bring them close, and the force needed to hold them apart. Likewise, linear, cyclic and branched molecules with the same molecular mass can occupy different amounts of space. Their dimensions affect movement through pores and encounters in solution. Comparing such cases requires measures that represent many conformations, rather than a single molecular drawing chosen because it appears typical.
Kuhn's 1934 study of thread-like molecules helped establish a statistical approach to molecular shape. Many local rotations collectively determine the overall dimensions, and a chain can be represented as a sequence of steps in a random walk. Randomness here does not mean an absence of physical rules. Thermal fluctuations make individual outcomes variable, so we calculate which outcomes occur more frequently. Individual molecular snapshots differ, yet their size distribution and average dimensions can be reproducible and experimentally meaningful.
We begin with an ideal chain. This model retains connectivity and the necessary local geometry while neglecting the influence of long-range attraction or repulsion between chain segments on the statistics at the scale of interest. Chemical bonds remain present. The model provides a baseline against which excluded volume, charge, specific binding and crowding can be recognized. Real segments do not freely pass through one another, and real solutions can depart from ideal statistics. Establishing the baseline first makes an increase or decrease in size easier to connect to a particular physical cause.
Separating direction, length and averaging
Fix the beginning of a chain at the coordinate origin and follow the first segment, then the second, until reaching the last. Each displacement has both magnitude and direction, so it is represented by a vector. A vector is an arrow whose three components record displacement along mutually perpendicular spatial directions. Adding the small arrows head to tail gives the arrow directly connecting the starting point to the endpoint: the end-to-end vector. Its magnitude is the straight-line distance between the ends. The total distance traveled along all the segments is the contour length, which answers a different geometric question.
| Symbol | Meaning | Unit |
|---|---|---|
| R (vector) | End-to-end vector | nm |
| bi (vector) | Displacement vector of segment i, each of magnitude b | nm |
| R | Magnitude of the end-to-end vector | nm |
| N, i | Total segment count and segment index | 1 |
| b, L | Segment length and contour length | nm |
| Σ, | | | Summation; vector magnitude | — |
Reading the relationshipHead-to-tail vector addition gives net displacement. Contour length adds the positive lengths of all segments.
A worked exampleTwo 1 nm segments, one rightward and one upward, have contour length 2 nm and end-to-end distance √(1²+1²)=1.414 nm.
ConditionsThe contour expression assumes N equal-length model segments; unequal lengths must be added individually.
Now imagine many chains with no imposed force or preferred direction. An endpoint lying to the right in one configuration can be balanced by one lying to the left in another, so the mean end-to-end vector is zero. The chains have not collapsed into points. People arranged around the center of a square have an average position at the center while each remains some distance away. To characterize size, we can square the end-to-end distance, average the squared values and then take the square root. Squaring makes every contribution nonnegative and also allows correlations between segments to be separated mathematically.
Angle brackets denote a conformational average, with configurations weighted by their probabilities under the chosen conditions. Samples drawn according to these probabilities can approximate the average by ordinary arithmetic. Averaging before squaring differs from squaring before averaging. For three distances of 1, 2 and 3 nanometers, the mean distance is 2 nanometers, while the root-mean-square distance is about 2.16 nanometers because squaring gives greater weight to the larger values. A nanometer is one billionth of a meter, a convenient unit for molecular dimensions.
| Symbol | Meaning | Unit |
|---|---|---|
| R (vector), R | End-to-end vector and its magnitude | nm |
| bi (vector), bi, b | Segment-i vector, its magnitude and common segment length | nm |
| N, i, j | Segment count and indices of two segments | 1 |
| 〈 〉 | Average weighted by conformational probabilities | — |
| Σi, Σi<j | Sum over segments; sum over each distinct pair once | — |
| · | Dot product measuring directional alignment | — |
| Rrms | Root-mean-square end-to-end distance | nm |
| √, 0 (vector) | Square root; zero vector | — |
Reading the relationshipExpanding the square gives self-terms and cross-terms. Independent isotropic directions make the mean cross-terms zero, leaving b² per segment.
A worked exampleFor N=100 and b=1 nm, 〈R²〉=100 nm², the rms distance is 10 nm and the contour is 100 nm. At N=400 these lengths become 20 and 400 nm.
ConditionsNo imposed force, independent equal steps and isotropic directions. Cross-terms are generally nonzero in an individual configuration.
In a freely jointed chain, every segment has the same length, and its direction is independent and uniformly distributed in space. Expanding the squared total displacement gives self-terms and cross-terms. Each segment's dot product with itself contributes its squared length. Cross-terms between different segments have positive and negative values that cancel in the average. The dot product measures directional alignment: it is positive for parallel arrows, negative for opposite arrows and zero for perpendicular ones. The mean-square end-to-end distance therefore grows in proportion to segment number, while its square root grows only as the square root of that number. Four times as many segments give four times the contour length but only twice the typical end-to-end distance.
Beyond the two ends: radius of gyration
The end-to-end distance examines only two points. It captures an important dimension of a linear chain but does not describe how all the mass is distributed. A closed ring has coincident starting and ending points even while occupying a substantial region of space. A branched molecule has many ends, so selecting a particular pair introduces an arbitrary choice. The radius of gyration instead measures the distribution of all parts around the center of mass. This center is a mass-weighted mean position, analogous to the balance point of an object supported against gravity.
For equal-mass model beads, the center of mass is simply the average of their coordinates. Averaging the squared distances of the beads from this center gives the squared radius of gyration for one configuration. A second average can then be taken over configurations. The first operation describes the mass distribution in one snapshot; the second describes a collection of snapshots. This measure works for linear, cyclic and branched chains and is widely used in interpreting static scattering. It does not define a hard molecular boundary, and parts of the chain can lie outside the corresponding radius.
| Symbol | Meaning | Unit |
|---|---|---|
| ri, rj (vector) | Position of bead i or j | nm |
| rcm (vector) | Center-of-mass position | nm |
| Rg² | Squared radius of gyration of one configuration | nm² |
| N, N+1 | Number of segments and equal-mass beads | 1 |
| b | Freely jointed segment length | nm |
| i, j, Σ | Bead indices and summation over those indices | — |
| 〈 〉, | | | Conformational average; vector magnitude | — |
| ≃ | Long-chain approximation | — |
Reading the relationshipExpanding pairwise squared distances and using the center definition gives the pair identity. In a freely jointed chain, each pair has mean-square separation equal to its contour segment count times b²; summing gives the last line.
A worked exampleFor N=100 and b=1 nm, 〈R_g²〉=100×102/(6×101)=16.8317 nm², whose square root is about 4.103 nm. The long-chain approximation gives 4.082 nm.
ConditionsThe first three lines apply to any equal-mass bead configuration. The final line applies to a freely jointed linear chain, with the approximation requiring large N.
An equivalent calculation sums squared distances between all pairs of beads. The double sum counts each distinct pair in both orders, which explains the factor of two in the denominator. This identity is useful because the mean-square separation of two points on an ideal chain follows from the number of independent segments connecting them. For a long ideal linear chain, the mean-square radius of gyration is approximately one sixth of the mean-square end-to-end distance. These are squared sizes; the ratio of radii requires taking a square root. A finite chain uses the more precise expression that retains the endpoint counting.
Experimental quantities called radii must also be distinguished by definition. Dynamic light scattering analyzes time-dependent fluctuations of scattered light, usually yielding translational diffusion information that can be converted into a hydrodynamic radius. This is an equivalent size associated with resistance to motion through the solvent. The radius of gyration describes the spatial distribution of mass or scattering contrast. Solvent can move through a flexible coil, so the two measures need not agree. Before comparing numerical values, identify the physical quantity that the measurement actually reports.
How local chemistry changes overall dimensions
A chemical backbone is not a sequence of joints that can point freely in every direction. Common tetrahedral bonding geometries constrain the angle between neighboring bonds. Rotation about a bond can also be affected by side-group crowding and electronic interactions. The freely rotating chain retains a fixed bond angle while allowing unrestricted azimuthal rotation. Neighboring displacement directions are then correlated, with directional memory usually weakening along the chain. In the following convention, the angle is measured between successive forward-pointing displacement vectors. It is supplementary to the angle between two bonds drawn outward from their common atom, so distinguishing the conventions prevents a sign error.
| Symbol | Meaning | Unit |
|---|---|---|
| α | Angle between successive forward displacement vectors | radians (dimensionless) |
| q, cos | Adjacent-direction cosine and cosine function | 1 |
| li, li+s (vector) | Bond displacement vectors separated by s steps | nm |
| l | Fixed bond length | nm |
| N, i, s | Bond count, index and separation in steps | 1 |
| R | End-to-end distance | nm |
| 〈 〉, ·, Σ | Conformational average, dot product and summation | — |
Reading the relationshipFree azimuthal rotation multiplies directional memory by q per step. There are N−s pairs at separation s. Adding these correlations gives the finite-chain result and its long-chain limit.
A worked exampleFor q=1/3, the long-chain factor is (1+q)/(1−q)=2. The squared size doubles and the rms size increases by √2.
ConditionsFixed lengths and angles with uniform free azimuthal rotation; the asymptotic expression requires a chain much longer than its correlation scale and |q|<1.
A positive correlation between successive forward displacements makes a short portion of the chain more likely to continue in a similar direction. The same number of bonds of the same length then gives a larger mean-square end-to-end distance than the freely jointed baseline. For a forward-angle cosine of one third, the long-chain squared size is twice the freely jointed result; the size itself increases by the square root of two. The result sums directional correlations over every separation along the chain. It accounts for fixed-angle geometry, while actual torsional preferences introduce further changes.
A more detailed local model can distinguish a trans state and two gauche states around a rotatable bond. These describe relative orientations of neighboring backbone portions without changing chemical connectivity. Different energies give different probabilities, and the number of states at each energy matters as well. Boltzmann weights convert the energies into relative statistical weights. The exponential suppresses higher-energy states while preserving their finite probability under thermal motion. This makes it possible to examine how temperature changes the local choices available to the backbone.
Chemistry often reports the combined energy of many molecules, so the mole needs a definition here. One mole contains 6.02214076×10²³ specified entities; this count per mole is the Avogadro constant N_A. The entities may be molecules or the torsional units being compared in a model. An energy difference of 3 kilojoules per mole is the sum for that many equivalent units, corresponding to about 4.98×10⁻²¹ joules per unit. The joule is an energy unit, and one kilojoule equals 1000 joules. The thermal energy scale per unit is the Boltzmann constant k_B multiplied by absolute temperature T, with k_B=1.380649×10⁻²³ J/K. Multiplication by N_A gives the molar gas constant R, approximately 8.314 J/(mol·K). Absolute temperature is measured in kelvin, K; adding 273.15 to a Celsius temperature gives its kelvin value, so 300 K is approximately 26.85 °C. Dividing the molar energy difference by RT compares the state cost with the thermal scale on the same counting basis. Their units cancel, giving the dimensionless argument required by the exponential instead of placing an energy with units directly in its exponent.
| Symbol | Meaning | Unit |
|---|---|---|
| Δε | Molar energy difference of each gauche state relative to trans | kJ mol⁻¹ |
| R | Gas constant, 0.008314 | kJ mol⁻¹ K⁻¹ |
| T | Absolute temperature | K |
| w | Weight of one gauche state relative to trans | 1 |
| pt, pg+, pg− | Probabilities of trans and each of the two gauche states | 1 |
| exp | Natural exponential function | 1 |
Reading the relationshipTaking trans as zero energy gives weight 1. Two equivalent gauche states each contribute w, so normalization by 1+2w makes the probabilities sum to one.
A worked exampleFor Δε=3 kJ/mol and T=300 K, w≈0.300, the trans probability is about 0.625 and each gauche probability is about 0.188. At Δε=0 all three states have probability 1/3.
ConditionsThree discrete states, equivalent gauche energies and independent local choices; neighboring coupling requires joint-state weights.
Treating every torsion independently already adds chemical information, but neighboring torsions can influence each other. A preceding gauche state may change crowding around the next bond, making a particular pair of states unfavorable. Rotational-isomeric-state models connect discrete local choices through such conditional weights. Protein backbones, polymers with bulky side groups and different polysaccharides require local descriptions appropriate to their chemistry and geometry. Flexibility emerges from these microscopic options together. The presence of a rotatable bond therefore does not imply that every rotation is equally easy or statistically independent.
Kuhn length: replacing chemical detail with effective steps
Following every chemical bond can be unnecessarily elaborate when the question concerns the overall dimensions of a long chain. Coarse-graining replaces local groups of bonds with effective units so that chemical details enter through fewer statistical parameters. A map can similarly replace a winding road with a smaller set of turning points while preserving the distances relevant to a particular journey. The effective unit is not an arbitrary short fragment, and every microscopic configuration of that fragment need not resemble a rigid rod. It is defined by the properties that the coarse model is required to match.
A common construction matches both contour length and the long-chain mean-square end-to-end distance. The Kuhn length is the segment length of the equivalent freely jointed chain; its effective segment count is the contour length divided by that value. Both matching conditions are needed. A fixed rule such as grouping every few monomers cannot be transferred unchanged between different materials. The contour definition must also be specified: an atom-by-atom zigzag and a smoothed backbone centerline can require a geometric conversion. Comparing parameters is meaningful when the associated definitions are consistent.
| Symbol | Meaning | Unit |
|---|---|---|
| bK | Kuhn length | nm |
| NK | Effective Kuhn-segment count, which may be noninteger | 1 |
| L | Contour length under a specified backbone-path convention | nm |
| 〈R²〉₀ | Unperturbed mean-square end-to-end distance; subscript 0 denotes unperturbed statistics | nm² |
Reading the relationshipThe equivalent freely jointed chain matches both contour length N_Kb_K and mean-square size N_Kb_K², determining the two parameters.
A worked exampleFor L=200 nm and 〈R²〉₀=400 nm², b_K=2 nm, N_K=100 and the rms end-to-end distance is 20 nm.
ConditionsA long chain with local correlations coarse-grained and unperturbed statistics at the relevant scale; retain the contour convention with the parameter.
At fixed contour length, a larger Kuhn length corresponds to fewer, longer effective steps and a more extended overall coil. At fixed Kuhn length, increasing the contour length adds steps and increases the typical size according to the square-root law. Both changes enlarge the coil, but for different reasons: one changes local stiffness, while the other adds chain length. When only a few effective segments fit within the chain, long-chain approximations become less accurate. A finite-chain or semiflexible description then retains directional correlations that still matter for the requested observable.
Grouping several independent steps produces a group whose mean-square displacement is the sum of the individual contributions. Fewer groups with larger typical displacements preserve the full chain's mean-square end-to-end distance. This statistical self-similarity underlies ideal-chain descriptions across scales. It does not mean that every short section has exactly the same shape as the entire chain, or that chemical and whole-molecule scales are interchangeable. The useful range begins beyond the local correlation scale and ends where larger-scale interactions or constraints change the statistics. Keeping this range explicit explains why a coarse model can be accurate for one question and insufficient for another.
Why end-to-end distances form a probability curve
An average size compresses many configurations into one number. Binding, capture and stretching problems often require more: how frequently does a particular separation occur? Adding many independent small displacements makes the total displacement in each coordinate approach a bell-shaped Gaussian distribution. This is a consequence of the central limit theorem when there are enough effective steps and no few extreme contributions dominate. The three-dimensional end-to-end vector combines three coordinate components. With no preferred direction, its probability density depends only on the squared distance from the origin.
A probability density is not the probability of one exact point. Multiplying it by a small volume approximately gives the probability of finding the endpoint in that volume, and integrating over all space gives one. For a single fixed-length step, the endpoint is confined to a sphere. A Dirac delta distribution can represent this constraint mathematically, but it is a concentrated weight under an integral, not an infinite probability assigned to a point. Starting directly from the smooth distribution of many steps keeps the spatial counting clear without requiring that more specialized notation.
| Symbol | Meaning | Unit |
|---|---|---|
| Q | Mean-square end-to-end distance | nm² |
| N, b | Independent segment count and segment length | N: 1; b: nm |
| R (vector), R | End-to-end vector and its nonnegative magnitude | nm |
| p(R (vector)) | Endpoint probability density per spatial volume | nm⁻³ |
| w(R) | Radial probability density per distance interval | nm⁻¹ |
| Rmode, Rrms | Most probable distance and root-mean-square distance | nm |
| π, exp, √ | Pi, natural exponential and square root | — |
Reading the relationshipMultiplying the three coordinate Gaussian densities gives vector density. Summing directions in a thin shell multiplies by spherical area 4πR² to give radial density.
A worked exampleFor N=100 and b=1 nm, Q=100 nm², the radial peak is near 8.165 nm and the rms distance is 10 nm, while vector density remains maximal at the origin.
ConditionsThree-dimensional long-chain approximation with many independent effective segments, appropriate to typical coiled distances; an actual fixed-segment chain cannot extend beyond its contour.
The vector density is largest at the origin, yet the distribution of distance alone peaks at a nonzero value. A thin spherical shell at a larger radius contains more positions in different directions. A city provides an analogy: its central district can have a high population density while a broad surrounding annulus contains more residents in total. The radial probability density adds contributions from all directions and therefore includes the shell-area factor. This also explains why a zero mean vector, a nonzero most probable distance and a different root-mean-square distance can all describe the same ensemble.
Conformational samples of three-dimensional freely jointed chains. Each calculation uses 2000 chains and the same random sequence for comparison. The curve is a long-chain approximation; numerical feedback also gives the finite-chain gyration result.
Every chain in the interaction is generated in three dimensions. The displayed individual chain is a two-dimensional projection, while all distances are calculated using three coordinates. Vary segment number and compare contour length, root-mean-square end-to-end distance and radius of gyration. Then hold the segment count fixed and change segment length: all spatial distances scale by the same factor. The histogram represents many independent configurations, and the smooth curve is the long-chain Gaussian approximation. Finite sampling produces fluctuations, while shorter chains can differ systematically in the tails. A single snapshot, a finite-sample average and a theoretical limiting result are distinct levels of description.
The Gaussian approximation has a clear geometric boundary. A chain of fixed-length segments reaches its greatest end-to-end distance when every segment aligns, giving the contour length. A Gaussian distribution nevertheless has a very small tail at arbitrarily large distances. For a long chain near its usual coiled size, this discrepancy is small. Near full extension it becomes decisive. A probability model should therefore be applied within the range that supports its derivation. The same distinction determines whether its predicted force remains useful when the chain is strongly stretched.
A restoring force without stretching chemical bonds
Gently separating the ends of a flexible chain initially changes segment orientations, often without substantially stretching individual chemical bonds. An unconstrained chain can adopt many coiled arrangements. Holding its ends far apart leaves fewer compatible arrangements. Entropy measures the availability of microscopic configurations, and reducing that availability raises the free energy under the relevant conditions. The chain consequently exerts a mean restoring force on its anchors. This contribution is called entropic elasticity. In a rubber network, many connected chain segments transmit such tendencies into a macroscopic restoring force.
The Gaussian vector density gives the free-energy cost of fixing an endpoint at a specified position. A natural logarithm converts statistical weights into an additive free-energy contribution. Differentiating with respect to displacement gives the mean external force needed to maintain that position. The constraint concerns the endpoint's position in space, rather than only its distance while it remains free to move over a spherical shell. Using the radial density instead would introduce shell-position counting into a different constrained problem. Identifying the constraint first determines which probability belongs in the calculation.
The force on one chain is small and is commonly expressed in piconewtons, pN: one piconewton equals 10⁻¹² N. Force multiplied by displacement along its direction has units of energy, so 1 pN·nm equals 10⁻²¹ J. These units keep molecular forces and distances within convenient numerical ranges. At 300 K, the thermal scale k_BT is about 4.142 pN·nm. The free-energy difference below uses this energy unit; taking its rate of change with displacement in nanometers leaves piconewtons, giving the force required to hold the endpoint in place.
| Symbol | Meaning | Unit |
|---|---|---|
| ΔF | Free-energy difference relative to an endpoint at the origin | pN nm |
| kB | Boltzmann constant, 0.01380649 pN·nm/K | pN nm K⁻¹ |
| T | Absolute temperature | K |
| R (vector), R, 0 (vector) | End-to-end vector, magnitude and zero vector | nm |
| p | Probability density of endpoint-vector position | nm⁻³ |
| N, b | Effective segment count and length | N: 1; b: nm |
| ln | Natural logarithm; its density-ratio argument is dimensionless | 1 |
| fhold | Magnitude of the mean external holding force | pN |
| dΔF/dR | Rate of free-energy change with distance | pN |
Reading the relationshipFixing the endpoint reduces configuration weight; −k_BT times its logarithm gives the free-energy cost. Its displacement derivative gives holding force; chain restoring force points oppositely.
A worked exampleAt T=300 K, N=100, b=1 nm and R=5 nm, k_BT=4.142 pN·nm, the spring constant is 0.1243 pN/nm and the holding force is about 0.621 pN.
ConditionsAn equilibrium Gaussian chain with fixed vector endpoint and extension well below contour length; Nb² is defined under the specified conditions.
Within its range of validity, the relation resembles an ordinary spring: greater displacement requires greater holding force. Its spring constant, however, comes from temperature and the chain's statistical dimensions. Holding other parameters fixed, raising temperature increases the free-energy cost of losing configurational entropy. A longer chain offers more rearrangement options and generally requires less force for the same endpoint displacement. Actual materials can also change Kuhn length, interactions or network structure when heated, so these contributions should accompany the entropic comparison when interpreting temperature-dependent experiments.
Near the contour length, the remaining orientation choices become scarce and finite extensibility becomes essential. A linear Gaussian spring must then give way to a more suitable description. Stronger forces may stretch bonds, unfold a local structure or rupture an association, each with its own energetic and kinetic behavior. Single-molecule force–extension measurements are useful because they can distinguish successive regimes: conformational adjustment, suppression of bending and changes within the molecular structure. Connecting a curve's change in slope to the relevant process gives a physical meaning to a measured increase in stiffness.
From flexible coils to semiflexible rods
A short thin strip can appear almost straight, while a much longer strip of the same material bends appreciably. Whether a polymer appears flexible also depends on observation length. Persistence length describes how directional memory decays along the contour. Starting from one point and following the chain, the local tangent gradually loses correlation with its initial direction. In the simple wormlike-chain model, the correlation falls to about 37% after one persistence length. The separation is contour distance traveled along the chain, rather than straight-line spatial distance. Larger persistence length means that local orientation persists farther.
Kratky and Porod's 1949 work connected local chain stiffness to scattering from molecules in solution. The wormlike chain represents the backbone as a continuous bendable line and describes the transition from locally rod-like behavior to an overall coil. A contour much shorter than the persistence length offers little opportunity to accumulate large bends. A much longer contour contains widely separated portions whose directions have largely decorrelated. One model thus includes both appearances without assigning a material a permanently rigid or permanently flexible identity.
| Symbol | Meaning | Unit |
|---|---|---|
| t(s), t(0) (vector) | Unit tangent vectors at contour positions s and 0 | 1 |
| s, L | Contour separation and total contour length | nm |
| ℓ_p | Persistence length | nm |
| R | End-to-end distance | nm |
| bK | Long-chain equivalent Kuhn length | nm |
| 〈 〉, ·, exp | Conformational average, dot product and natural exponential | — |
| ≫, ≃ | Much greater than; approximation in that limit | — |
Reading the relationshipTotal displacement accumulates tangent vectors along the contour. Integrating their exponential pair correlation gives the mean-square end-to-end distance; the end correction becomes relatively small for long chains.
A worked exampleFor L=100 nm and ℓ_p=10 nm, 〈R²〉≈1800.009 nm² and the rms distance is about 42.43 nm. The leading long-chain value 2000 nm² is still about 11.1% higher, showing the value of the finite-length correction.
ConditionsA three-dimensional, unforced, homogeneous inextensible wormlike chain with constant local bending stiffness and no long-range interactions.
In the long-chain limit of this continuous model, the Kuhn length is twice the persistence length. The factor follows from integrating tangent correlations and matching the contour length. For other discrete-angle models or short chains, the definitions and conversion should be checked explicitly. Another useful measure, the characteristic ratio, compares an unperturbed chain's mean-square end-to-end distance with the freely jointed result using the same chemical bond count and bond length. Monomer number, backbone-bond number and effective Kuhn-segment number consequently remain different quantities that should be recorded separately.
Double-stranded DNA provides a biological example. Its double helix has substantial local bending stiffness, yet a sufficiently long molecule forms a coil. Marko and Siggia's 1995 analysis of DNA stretching used wormlike-chain statistics to explain force–extension behavior and examined changes associated with electrostatic interactions and stronger forces. Salt modifies interactions between charged portions, and local molecular structure can also change with conditions. A persistence length therefore belongs together with its measurement environment. The same comparison of contour length with directional-memory length guides model choice for protein filaments, polysaccharides and nucleic acids.
Rings, stars and branches: connectivity changes size
The same total number of segments can produce different average dimensions when connected differently. A linear chain has free ends. Ring closure imposes a global constraint and changes the distance statistics between positions along the chain. A star joins several arms at one center. Topology here refers to connectivity, rather than the outline in a particular drawing. Bending a linear chain into a star-like silhouette does not create a branch point. Among natural molecules, glycogen is branched, whereas the basic cellulose chain is linear; membership in the polysaccharide family does not determine one common architecture.
Zimm and Stockmayer studied dimensions of branched and ring-containing chains in 1949. For ideal chains, pairwise separations can be calculated using connectivity and then combined into the radius of gyration. In a tree without loops, exactly one contour path connects any two points, allowing independent segment contributions along that path to be added. A closed loop couples the available paths through a closure constraint, so they cannot all be treated as unrelated free walks. Separating the connection graph from its spatial realization provides a useful route to more complex molecular architectures.
| Symbol | Meaning | Unit |
|---|---|---|
| Rg² | Squared radius of gyration | nm² |
| 〈 〉 | Conformational average | — |
| N, b | Total statistical segment count and common segment length | N: 1; b: nm |
| lin, ring, star, architecture | Subscripts for linear, ring, equal-arm star and compared architecture | — |
| g | Ratio of mean-square gyration radii, the contraction factor | 1 |
| f | Number of equal star arms | 1 |
Reading the relationshipThe radius follows from all pairwise mean-square distances. Ring closure and branch paths alter these distances; comparison at equal total segment count gives g.
A worked exampleFor N=120 and b=1 nm, the linear mean-square gyration radius is about 20 nm² and the ring value about 10 nm². A four-arm star has g=10/16=0.625 and a value about 12.5 nm².
ConditionsIdeal Gaussian long chains at equal total segment count and length, with equal star arms; excluded volume, knots and threading constraints are omitted.
The contraction factor compares a nonlinear architecture's mean-square radius of gyration with that of a linear chain having the same total segment count and statistical segment length. An ideal long ring gives a factor of one half, corresponding to a radius ratio near 0.707 rather than a halved radius. In an equal-arm star, adding arms at fixed total length shortens each arm and concentrates the mass distribution toward the center. Unequal arms require more information than arm count alone because a long arm can carry mass farther outward. A comb also requires backbone length, side-chain lengths and grafting positions to make a meaningful comparison.
Real highly branched molecules can expand because of side-chain crowding. Rings can be affected by knots, threading and constraints imposed by other chains. Even an ideal-chain snapshot is usually not a perfect sphere: the mass distribution can extend differently along its principal directions. A gyration tensor records directional second moments and cross-correlations; in principal axes, its three nonnegative values add to the squared radius of gyration. Averaging all random orientations first can produce isotropy. Aligning each molecule with its own principal axes before comparing shapes retains the asymmetry that orientation averaging would conceal.
Different positions on the same linear chain also explore different regions relative to its center of mass. In a long Gaussian chain, an endpoint has a larger mean-square distance from the center than the chain midpoint does. The midpoint is connected to long chain portions on both sides, whereas the endpoint is connected on only one side. For a total mean-square end-to-end distance of 100 square nanometers, the continuous-chain model gives about 33.3 square nanometers for an endpoint and 8.33 for the midpoint. The squared quantities differ by a factor of four, so their root-mean-square distances differ by two. This is an ensemble comparison: an endpoint need not be farther away in every snapshot, and it is not confined to a fixed spherical shell.
Connecting single-chain statistics to materials and experiments
The chain models provide different resolutions of the same physical problem. A freely jointed chain establishes an independent-step baseline. Freely rotating and rotational-isomeric-state descriptions retain more local chemistry. A Kuhn representation compresses those effects into effective steps, a Gaussian chain describes the aggregate distribution of many steps, and a wormlike chain retains continuous bending stiffness. Models can agree on large-scale average dimensions while differing at short distances, under strong extension or in their dynamics. Begin with the observable—size, bending, encounter probability or mechanical response—and retain the details that determine it.
Rouse's 1953 bead–spring description replaced sufficiently long subchains with entropic connections and assigned coarse-grained solvent resistance to beads. A bead need not correspond to one monomer, and the springs are not literal objects inside the backbone. They carry the friction and conformational free energy retained by the model. This construction links static chain statistics to recovery after deformation. Adding time requires specifying thermal fluctuations, friction and coupling through the fluid or surrounding chains. A static mean dimension alone does not determine the complete motion.
Consider a flexible linker between binding sites. Its effective segment number and typical size can first be compared with the separation the sites must span. A distance far above the typical size is compatible with fewer configurations and carries a substantial entropic cost. At shorter distances, end-group orientation, local crowding and binding chemistry also matter. Ring formation similarly requires more than the ends merely approaching: suitable orientation and reaction time influence whether an encounter becomes a bond. Chain statistics describe spatial opportunities, while the specific chemistry determines how efficiently those opportunities produce the desired event.
A clear size comparison records connectivity, total contour length or molecular mass, the statistical-segment convention, solvent, temperature and the particular radius measured. Concentration-dependent changes suggest a role for interactions between chains. Salt-dependent changes invite examination of charge and local structure. Nonlinear stretching requires checking finite extensibility and internal transitions. In this way, ideal-chain theory becomes a sequence of useful questions: define the geometry, specify the probability rules, and identify the interactions that shift the real system away from the baseline.
Complementary measurements on the same sample help distinguish what a model captures. Static scattering probes spatial distribution, diffusion probes resistance to motion with the solvent, and endpoint labeling examines selected pair separations. Explaining these observables together is more informative than fitting one average dimension. Equal average sizes can still accompany different distribution widths, shapes or stretching responses, showing which information a single size parameter leaves unresolved. The useful procedure is to select the observable first, then seek the simplest model that retains its controlling physics. This keeps a calculation manageable while preserving the structural detail that the experiment is designed to reveal.