Radical Polymerization

Radical chain polymerization is not a historical curiosity; it is the manufacturing backbone of the modern plastics industry. Low-density polyethylene (LDPE) — the stretchy, branched polyethylene used in shrink wrap and squeeze bottles — is made industrially by exactly this mechanism, at extreme pressures (1000–3000 atm) and temperatures around 150–300 °C with organic peroxide or oxygen initiators; the high pressure favors the intermolecular and intramolecular chain-transfer events (backbiting) that give LDPE its characteristic short- and long-chain branching, distinguishing it mechanistically and physically from the coordination-catalyzed linear HDPE made by Ziegler–Natta-type methods. Poly(methyl methacrylate) (Plexiglas/Lucite), polystyrene (from disposable cutlery to high-impact packaging), poly(vinyl chloride) (PVC piping), poly(vinyl acetate) (adhesives, paint binders), and polyacrylonitrile (precursor to carbon fiber) are all produced industrially by free-radical chain mechanisms, often in emulsion (for latex paints and synthetic rubber, using water-soluble persulfate initiators and surfactant micelles as tiny reaction vessels) or suspension polymerization (for PVC and expandable polystyrene beads). The chain-transfer chemistry developed here is also the conceptual seed for the modern "living"/controlled radical techniques — atom transfer radical polymerization (ATRP), reversible addition–fragmentation chain transfer (RAFT), and nitroxide-mediated polymerization (NMP) — all of which suppress irreversible termination by keeping the vast majority of chains in a temporarily "dormant," capped state, allowing near-uniform (low-dispersity) molecular weights and block copolymer architectures used in modern drug-delivery nanoparticles, thermoplastic elastomers, and surface coatings. Radical initiators from this same mechanistic family (AIBN, benzoyl peroxide, ammonium persulfate) are also everyday laboratory tools for crosslinking dental resins, curing fiberglass composites, and initiating radical thiol–ene "click" reactions in materials science. Understanding this chain-reaction competition — the same tug-of-war between propagation, transfer, and termination that governs a plastic bag's molecular weight — is also excellent training for thinking about competing pathways more generally, a habit of mind carried forward explicitly when reasoning through bond disconnections in Retrosynthetic Analysis: Disconnections and Synthons.

Ask the tutor: How does controlled/living radical polymerization (ATRP, RAFT, NMP) suppress irreversible termination while still using this same radical chain machinery?

Think of the reacting system as a vast, dilute population of chain-carrying radicals swimming in a sea of monomer. Each radical is a single unpaired electron sitting in a singly occupied molecular orbital (SOMO) on a carbon, adjacent to whatever stabilizing groups (phenyl in styrene, ester in acrylates, nitrile in acrylonitrile, chlorine in vinyl chloride) happen to be on that monomer. The alkene it attacks presents a $\pi$ and $\pi^*$ pair of orbitals. Radical addition is fundamentally a three-electron-donating, orbital-controlled event: the SOMO of the radical interacts with the LUMO ($\pi^*$) of the alkene (dominant when the radical is nucleophilic and the alkene is electron-poor, as in acrylates) or with the alkene HOMO ($\pi$) donating into the radical's singly-occupied orbital (when the radical is electrophilic, as with chlorine or ester-stabilized radicals attacking electron-rich alkenes). Either way, the new $\sigma$ bond forms at the less substituted carbon, dumping the new radical onto the carbon that best stabilizes it — this is why polymer chains are overwhelmingly "head-to-tail" (1,3-relationship of substituents) rather than "head-to-head." The propagating radical is a moving target: every ~$10^{-3}$–$10^{-2}$ seconds (on average, for typical $k_p[\mathrm M]$ products) it adds one more monomer, extending the chain by one repeat unit and regenerating an identical-in-kind radical further down the chain. This self-similarity — the radical at carbon 5000 behaves kinetically almost exactly like the radical at carbon 5 — is why a single rate constant $k_p$, independent of chain length beyond the first few additions, adequately describes propagation for a polymer of any size. Meanwhile a much smaller population of radicals, whenever two chain ends randomly encounter each other by diffusion, terminates: either by direct C–C bond formation (combination) or by one radical abstracting a $\beta$-hydrogen from the other (disproportionation), leaving one saturated and one vinyl-terminated chain.

The whole quantitative structure of radical chain polymerization rests on four rate expressions, one for each mechanistic step, combined through the steady-state approximation on radical concentration.

Initiation. An initiator decomposes homolytically (unimolecularly) with rate constant $k_d$; only a fraction $f$ (initiator efficiency, typically 0.3–0.8) of the radicals produced actually add to monomer before being lost to cage recombination or side reactions.

where: $R_i$ = rate of production of chain-initiating radicals (M s$^{-1}$), $f$ = initiator efficiency, $k_d$ = unimolecular decomposition rate constant of initiator (s$^{-1}$), $[\mathrm I]$ = initiator concentration. The factor of 2 counts both radicals from each homolysis event, discounted by $f$ for the ones that actually escape the solvent cage and add to monomer.

Propagation. Every addition of monomer to a growing chain radical is treated as kinetically identical regardless of chain length (the "long-chain hypothesis"):

where: $R_p$ = rate of monomer consumption by propagation, $k_p$ = propagation rate constant (M$^{-1}$s$^{-1}$), $[\mathrm M]$ = monomer concentration, $[\mathrm M\cdot]$ = total concentration of all propagating radicals (summed over all chain lengths).

Termination. Two radicals meet and die, either by combination or disproportionation, with a single combined rate constant $k_t$:

where: $R_t$ = rate of radical destruction, $k_t = k_{tc} + k_{td}$ combines combination ($k_{tc}$) and disproportionation ($k_{td}$) pathways, and the factor 2 reflects that each termination event destroys two radicals.

Steady state. Because radical concentrations are minuscule and radicals are far more reactive than they are stable, the system reaches a quasi-equilibrium almost instantly: radicals are produced and destroyed at equal rates, $R_i = R_t$. Solving for the steady-state radical concentration:

Substituting into the propagation rate law gives the single most important equation in radical polymerization kinetics, the rate of polymerization:

This says the polymerization rate is first order in monomer but only half order in initiator — a signature, testable prediction of the radical chain mechanism (a chain mechanism where two radicals must meet to die shows this characteristic square-root dependence, exactly as in ordinary Radical Chain Reactions: Initiation, Propagation, Termination kinetics for small-molecule chains). The kinetic chain length $\nu$ — average number of monomers consumed per radical generated — is simply the ratio of propagation rate to the rate of radical-destroying events:

and the number-average degree of polymerization $X_n$ (average number of monomer units per dead polymer chain) depends on the termination mode:

Finally, chain transfer — a radical abstracting an atom (almost always H) from some other species $XY$, quenching the growing chain but generating a new radical $X\cdot$ that reinitiates — modifies $X_n$ without touching $R_p$ appreciably (transfer agents are typically present at low enough concentration, and regenerate radicals fast enough, that $[\mathrm M\cdot]$ is barely perturbed). This gives the Mayo equation:

where: $X_{n,0}$ is the degree of polymerization in the absence of any transfer, and each $C_X = k_{tr,X}/k_p$ is a chain-transfer constant — the rate constant for transfer to species $X$ divided by the propagation rate constant, for monomer ($C_M$), solvent ($C_S$), and initiator ($C_I$) respectively.

Example 1 (foundational) — computing $R_p$ and $[\mathrm M\cdot]_{ss}$ for a styrene/BPO system. Bulk styrene polymerization at 60 °C initiated by benzoyl peroxide has typical literature values $k_p = 176\ \mathrm{M^{-1}s^{-1}}$, $k_t = 3.7\times10^{7}\ \mathrm{M^{-1}s^{-1}}$, $k_d = 2.0\times10^{-6}\ \mathrm{s^{-1}}$, initiator efficiency $f = 0.60$, with $[\mathrm I] = 0.0084$ M and $[\mathrm M] = 8.35$ M (neat styrene). First find the steady-state radical concentration:

This confirms the claim that radical concentrations are minuscule — about eight orders of magnitude below monomer concentration, which is exactly why $[\mathrm M]$ barely changes on the timescale that a single radical lives. Now the rate of polymerization:

As a sanity/limiting check: if $[\mathrm I]\to 0$ (no initiator), $[\mathrm M\cdot]_{ss}\to 0$ and $R_p \to 0$ — no radicals, no chain reaction, exactly as required, and the half-order dependence means doubling the initiator concentration only increases $R_p$ by a factor of $\sqrt2 \approx 1.41$, not 2 — a distinctly non-intuitive, testable feature of the mechanism.

Example 2 (applied) — molecular weight control via a thiol chain-transfer agent. Suppose the same bulk styrene system (60 °C, $k_p = 176\ \mathrm{M^{-1}s^{-1}}$) has an intrinsic (transfer-free) degree of polymerization $X_{n,0} = 1900$ from termination/propagation alone, and a chemist adds $n$-dodecanethiol as CTA at $[S]/[\mathrm M] = 2.0\times10^{-3}$ to deliberately lower the molecular weight for a coatings application. The chain-transfer constant for thiols to styrene is $C_S \approx 15$ at 60 °C (thiol S–H bonds are weak and easily abstracted). Ignoring transfer to monomer and initiator (both much smaller, $C_M\approx 6\times10^{-5}$ for styrene) for this estimate:

Notice how completely the thiol dominates: a transfer agent present at only 0.2 mol % relative to monomer collapses the degree of polymerization from ~1900 down to ~33 — this is precisely why CTAs are the workhorse tool for tuning molecular weight in industrial free-radical polymerization without needing to touch initiator concentration (which would also change $R_p$). As a limiting check: setting $[S]\to 0$ recovers $X_n \to X_{n,0} = 1900$ exactly, confirming the equation correctly reduces to the transfer-free case.

Example 3 (exam/competition level) — extracting $k_p^2/k_t$ from combined rate and average-lifetime data, and predicting the termination mode from $X_n$ vs. $\nu$. A methyl methacrylate (MMA) bulk polymerization at 50 °C is initiated thermally by AIBN with $R_i = 4.20\times10^{-8}\ \mathrm{M\,s^{-1}}$ (measured independently by a radical-scavenger method) and gives an observed $R_p = 1.19\times10^{-3}\ \mathrm{M\,s^{-1}}$ at $[\mathrm M] = 9.04$ M. Independent gel permeation chromatography gives a number-average degree of polymerization $X_n = 3550$. Step A — find $\nu$. Since $\nu = R_p/R_i$:

Step B — diagnose the termination mode. Compare $X_n$ to $\nu$:

If termination were purely by disproportionation we would expect $X_n = \nu$ (ratio $=1$); if purely by combination, $X_n = 2\nu$ (ratio $=2$). A ratio near 0.125 is far below either limit, which signals that chain transfer — not the termination mode — is the dominant chain-limiting event here (recall the Mayo equation: transfer lowers $X_n$ below the "intrinsic" value set by termination alone). This is a common trap: a low $X_n/\nu$ ratio does not mean some exotic termination pathway; it means transfer agents (often adventitious solvent, or even MMA's own transfer constant, which is unusually significant at higher temperatures) are consuming most chain ends before termination ever gets the chance. Step C — back out $[\mathrm M\cdot]_{ss}$ and $k_p/k_t^{1/2}$. From $R_p = k_p[\mathrm M][\mathrm M\cdot]_{ss}$ and known $R_i = R_t = 2k_t[\mathrm M\cdot]_{ss}^2$:

so eliminating $[\mathrm M\cdot]_{ss}$ between the two rate laws gives the classic combined ratio used to extract $k_p^2/k_t$ from experiment:

Plugging in numbers:

This composite parameter ($k_p^2/k_t$, units M$^{-1}$s$^{-1}$) is exactly the quantity historically measured by rotating-sector and, later, pulsed-laser-polymerization experiments — because $R_p$ and $R_i$ (or $[\mathrm M\cdot]$ via ESR) are directly measurable, while $k_p$ and $k_t$ individually require a second, independent experiment (e.g., pulsed-laser polymerization–size exclusion chromatography, PLP-SEC) to separate.

Radical chain polymerization as a quantitative science owes its clearest early articulation to Paul Flory, whose 1930s–1950s work at DuPont and Cornell built the statistical and kinetic framework — steady-state kinetics, the Mayo equation's logical ancestor, and the molecular-weight-distribution theory for step-growth and chain-growth polymers alike — that this entire article rests on. Flory arrived at DuPont's central research department in 1934, joining a group under Wallace Carothers that was already producing nylon and neoprene by empirical trial; Flory's distinctive contribution was to insist that the statistics of random bond-forming events, not just the chemistry of individual bonds, determined the measurable properties (viscosity, molecular weight distribution, gel point) of a polymer sample. He derived the now-standard "most probable distribution" for step-growth polymers and, in parallel work with collaborators including Frank Mayo (for whom the Mayo equation of chain transfer is named), extended rigorous kinetic chain-length analysis to radical vinyl polymerization. Flory received the 1974 Nobel Prize in Chemistry "for his fundamental achievements, both theoretical and experimental, in the physical chemistry of macromolecules" — the first Nobel awarded explicitly for polymer science, marking the field's arrival as a rigorous branch of physical organic chemistry rather than an empirical industrial art.