The Polymer Chain: Conformation and the Random Coil

One long chain, not a tiled motif

Every rung so far built structure the same way: take a rigid motif and stamp it, over and over, onto the points of a lattice. Even the amorphous rung kept that rigid motif — the SiO4 tetrahedron of a glass is exactly the tetrahedron of quartz, just linked at random angles. A polymer throws the whole tile away. Its fundamental object is a single molecule, but an absurdly long and floppy one: a polymer chain, thousands of small repeat units — the monomers — strung end to end into one continuous backbone by strong covalent bonds. Polyethylene, the plastic of a shopping bag, is just the two-carbon unit -(CH2-CH2)- repeated. The number of repeats is the degree of polymerization, N, and it is huge — commonly 10^3 to 10^5.

Put in real numbers and the strangeness jumps out. The backbone C-C bond is l = 0.154 nm and the tetrahedral angle is 109.5 degrees, so a fully-stretched all-trans zigzag advances only about 0.25 nm per two-carbon monomer. Take N = 5000 monomers (10^4 backbone carbons, a molecular weight near 140 kg/mol — an ordinary polyethylene): stretched out straight, its contour length L is roughly 5000 times 0.25 nm ~ 1260 nm, about 1.3 micrometres. A backbone half a nanometre thick, yet over a micrometre long if you could pull it taut. This is a genuinely new length scale of structure — the interesting arrangement lives not in the 0.5 nm bonding, and not at the naked-eye scale, but in the tens of nanometres in between.

Configuration you are born with; conformation you can change

Two words that sound alike but mean nearly opposite things, and getting them straight is the whole foundation of this rung. The configuration of a chain is the arrangement locked in by its chemical bonds — you cannot alter it without actually breaking and remaking bonds. The cis-or-trans of a carbon-carbon double bond, the head-to-tail versus head-to-head hookup of monomers, and above all the left-or-right handedness at each substituted carbon (which becomes tacticity) are all configuration. It is decided at the instant the chain is polymerized and is then permanent, like the picture printed on a stamp.

The conformation is the shape the chain actually takes at a given moment by rotating about its single backbone bonds — no bond is broken, the chain just twists. Every C-C single bond can sit in a trans state (backbone locally straight) or one of two gauche states (locally kinked), and thermal energy flips restlessly among them billions of times a second. A single chain therefore explores an astronomical number of shapes, and the random coil is simply the overwhelmingly most probable one. Link this in your mind to the crystallography you already own: configuration is like the fixed identity of the motif, while conformation is like the orientation the crystal happens to be lying in — a transient pose, not an identity.

The division cleaves this whole rung neatly in two. Configuration is what the next guide is about: a regular configuration — a chain in which every stereocentre points the same way, a regular tacticity — is precisely what lets stretches of chain register together and crystallize, while an irregular one can never fit and is doomed to stay a disordered glass. Conformation is what THIS guide is about: given whatever configuration a chain was born with, what shape does it fall into when left alone? For a flexible chain in a melt or solution, the answer is the random coil, and its geometry is beautifully simple.

The random coil: a drunkard's walk in three dimensions

To capture the coil's shape, physicists start from the boldest possible idealization, the freely-jointed chain: pretend the molecule is N rigid segments each of length l, and pretend each segment points in a completely random direction, utterly independent of its neighbours. That is not quite a real chain — we will pay back the difference later — but it is exactly a random walk, the same drunkard's walk a stumbling person takes away from a lamppost, one random step at a time. The most probable outcome of such a walk is the random coil: a loose, roughly spherical tangle with no preferred direction and no repeating pattern anywhere inside it.

The payoff is a clean formula. The end-to-end vector R is the sum of N little step-vectors pointing every which way, so on average R is zero — a coil is equally likely to end up in any direction from its start. But its size is not zero. Squaring and averaging, all the cross-terms cancel because the steps are independent, and you are left with the mean-square end-to-end distance <R^2> = N l^2. Take the square root and the typical size of the coil is R_rms = sqrt(N) times l. That square root is the fingerprint of every random walk: double the chain and its coil grows by only about 40 percent, not double.

RANDOM WALK  vs  STRETCHED CHAIN     ( N = 10^4 bonds,  l = 0.154 nm )

fully extended (all-trans zigzag),  R = contour length L :

  *-*-*-*-*-*-*-*- ... -*-*-*     L ~ N x 0.126 nm ~ 1260 nm  (1.3 um)


random coil (each bond points at random),  R_rms = sqrt(N) x l :

          *   *
         / \ / \_*         the chain wanders like a drunkard:
     *--*   *     \         after N steps of length l it strays
      \      *-*   *        only about sqrt(N) x l from home.
       *-*      \ /
          *      *          R_rms = 100 x 0.154 nm ~ 15 nm
         /
        *                   the coil is ~ 80x more compact than L

The same 10,000-bond polyethylene chain drawn two ways. Stretched straight it is 1.3 micrometres long; balled into its most probable random-coil shape it spans only about 15 nm, because a random walk of N steps strays only sqrt(N) steps from its start.

Now feel the numbers. For our N = 10^4 bonds with l = 0.154 nm, R_rms = 100 times 0.154 = 15.4 nm, against a contour length of 1260 nm. The very same molecule is about 80 times more compact balled up than stretched — the sqrt(N) coil versus the N-long backbone. And it is not a tight ball: a coil of this size is more than 95 percent empty space, so in a real melt many chains happily interpenetrate through one another like a bucket of wet spaghetti. That open, self-penetrating cloud is the resting structure of nearly all polymers above their softening point.

How big is a coil, really? End-to-end distance and radius of gyration

The end-to-end distance is conceptually clean but experimentally awkward. The two chain ends are nothing special — you cannot grab them, you cannot see them, and a ring polymer has no ends at all — yet the formula only works if you can find them. So the size that experiments actually report is the radius of gyration R_g: the root-mean-square distance of all the chain's segments from the chain's own centre of mass. It needs no special endpoints, applies to a ring or a branched chain just as well, and it is what a scattering instrument reads out.

For an ideal random coil the two measures are locked together by a tidy factor: R_g^2 = <R^2> / 6, so R_g = R_rms / sqrt(6). Our example coil, with R_rms = 15.4 nm, has R_g = 15.4 / 2.449 ~ 6.3 nm. And here is the lovely echo of the diffraction rung: you measure R_g by small-angle X-ray or neutron scattering (SAXS, SANS), which is exactly the low-angle cousin of the wide-angle diffraction you already learned — the shallow initial slope of the scattered intensity versus angle (the Guinier region) hands you R_g directly. Scattering read atomic spacings before; here the same physics reads the size of a whole molecular cloud.

Count the backbone bonds N and write down the bond length l (for a carbon backbone, l = 0.154 nm).
Contour (fully stretched) length: L ~ N times the projected bond length (~0.126 nm per C-C bond for an all-trans zigzag).
Ideal end-to-end size: R_rms = sqrt(N) times l — the drunkard's-walk result.
Radius of gyration: R_g = R_rms / sqrt(6), the size a scattering experiment actually measures.
Compare R_g against L: for our chain, ~6 nm against 1260 nm shows just how compact the coil is.
For a REAL chain, correct the naive R_rms upward by multiplying <R^2> by the characteristic ratio C-infinity (the next section explains why).

Real chains are stiffer — and, surprisingly, ideal in the melt

The freely-jointed chain lied to us in a useful way, and now we repay the difference. A real backbone bond does NOT point at random: the bond angle is pinned near 109.5 degrees, and rotation is hindered (trans sits lower in energy than gauche), so each bond's direction stays correlated with its neighbours over some distance. The chain is locally stiff, and a stiff walk wanders farther than a free one, so a real coil is bigger than the naive sqrt(N) l. Two honest bookkeeping devices fix this. The characteristic ratio C-infinity = <R^2> / (N l^2) is simply how many times bigger the real mean-square size is than the freely-jointed guess; for polyethylene C-infinity is about 6.7, so R_rms scales up by sqrt(6.7) ~ 2.6, turning our 15 nm coil into a truer ~40 nm.

The second device is more physical: the Kuhn segment. Replace the fussy real chain by an equivalent freely-jointed chain of N_K straight segments of length b (the Kuhn length), chosen so that it reproduces BOTH the true end-to-end size and the true contour length. Then <R^2> = N_K b^2 exactly, as if the chain really were freely jointed — you have just coarse-grained the stiffness away. The Kuhn length b (and its cousin the persistence length, l_p = b/2) is a direct measure of chain stiffness: it is the distance over which the chain forgets which way it was heading. Flexible polyethylene has b ~ 1.4 nm and l_p ~ 0.7 nm; stiff chains have far longer ones — double-stranded DNA has l_p ~ 50 nm, and rigid backbones like aramid (Kevlar) and cellulose are stiffer still, which is exactly why they can line up straight and reinforce a fibre instead of coiling.

Two subtleties worth stating plainly. First, the ideal Gaussian coil ignores that a chain cannot pass through itself — its excluded volume. In a good solvent this self-avoidance swells the coil, and the size grows as R ~ N^0.6 (Flory's exponent) rather than N^0.5. BUT — Flory's beautiful theorem — in a dense melt or an amorphous glassy solid a chain is surrounded by identical chains and cannot tell its own segments from a neighbour's, so the excluded volume is screened away and the chain relaxes back to IDEAL Gaussian statistics, R ~ N^0.5. That startling prediction was confirmed by neutron scattering off a few deuterium-labelled chains, and it is why the simple sqrt(N) picture is exactly the right one for solid plastics. Second, a random coil is disorder WITH sharp short-range order — every bond length and angle is fixed — but no long-range order at all. It is the same lesson as the glass, now living inside a single molecule.

Why should a soft materials person care about all this counting of steps? Because the coil is where structure meets property. A coil has astronomically more conformations than a stretched-out chain, so pulling it straight costs entropy and the chain fights back — the entropic spring behind rubber elasticity, a direct structure-property link running from a single bond's freedom to twist all the way up to why a rubber band snaps back (and warms slightly as it does). And the coil is only the disordered starting point of this rung. Give a chain a regular configuration and cool it, and neighbouring stretches can fold flat and lock into crystals. The next guide takes up tacticity — the configurational regularity that decides whether a chain can crystallize at all — and guide 3 builds the semicrystalline lamellae and spherulites that grow when it does.