Homolytic versus Heterolytic Bond Cleavage

Radical chain chemistry launched by homolysis underlies industrial-scale radical polymerization (polyethylene, polystyrene, acrylics), where peroxide or azo initiators are chosen specifically for their tunable, weak bonds — the same logic as Mechanism 1 above, scaled to megaton production. Autoxidation of edible oils, rubber, and plastics (the slow degradation that antioxidants like BHT are added to prevent) proceeds through homolytic O–H and C–H abstraction steps, and understanding BDEs lets chemists design better antioxidants that intercept the chain by donating a weakly-bound H atom preferentially. Atmospheric photochemistry — ozone layer chemistry, smog formation, the reactions that break down CFCs and pollutants — is governed almost entirely by photochemical homolysis of weak bonds (O–O, C–Cl) driven by solar UV. On the heterolytic side, essentially all of acid-base catalysis and much of pharmaceutical synthesis relies on controlled heterolysis: protonation of a carbonyl oxygen polarizes and weakens the adjacent C=O π-bond heterolytically to generate an electrophilic, resonance-stabilized oxocarbenium-like center; Lewis acid catalysts (BF₃, TiCl₄) assist heterolysis of C–X or C–O bonds in Friedel–Crafts and glycosylation chemistry; and drug metabolism in the liver (cytochrome P450 oxidations) often proceeds through a delicate mixture of both modes — some steps homolytic (hydrogen-atom abstraction by an iron-oxo radical-like species) and some heterolytic (subsequent ionic rebound or elimination) within the same catalytic cycle. Modern physical organic chemistry routinely uses computed and measured BDEs and heterolytic bond dissociation energies (via mass spectrometry and photoionization) to predict, in advance, which pathway a new substrate under new conditions will prefer — the same reasoning built from first principles here.

Ask the tutor: If a C–H bond has both meaningful polarity (C is slightly more electronegative than H) and could in principle form a stabilized carbanion, why do most C–H bonds still cleave homolytically rather than heterolytically under radical halogenation conditions?

Picture a single covalent σ-bond, A–B, as two atomic orbitals overlapping constructively to make a bonding molecular orbital, populated by two electrons of paired spin. This bonding MO is lower in energy than the two separate atomic orbitals; that energy gap is (roughly) why the bond exists at all. Breaking the bond means promoting those two electrons out of the stabilized bonding MO and back onto A and B as separate species. The chemistry hinges entirely on how the two electrons are partitioned between the resulting fragments.

Homolysis (homo- = same): each fragment keeps exactly one electron. A• and •B are formed, each a radical — a species with an unpaired electron sitting in a singly-occupied molecular orbital (SOMO). No charge is created; the total electron count is simply redistributed one-for-one. This is the symmetric outcome, and it is favored when A and B are similar in electronegativity (so neither atom has a strong pull on both electrons) and when the environment cannot stabilize full charges (gas phase, nonpolar solvent, high temperature, or photons providing energy directly to the σ* or via excitation).

Heterolysis (hetero- = different): one fragment keeps both electrons, the other keeps none. A⁺ and :B⁻ (or A:⁻ and B⁺, depending on which atom is more electronegative/stabilizes charge better) are formed — a cation and an anion. This is the asymmetric outcome, favored when A and B differ substantially in electronegativity or in ability to stabilize charge, and when a polar or protic solvent (or a nearby Lewis acid/base) can solvate or stabilize the resulting ions. Curved-arrow notation is built to make this fork visually explicit. A full-headed, double-barbed arrow ⤻ always represents the movement of an electron pair — the signature of heterolysis (or of any elementary step involving two-electron flow, including ionic bond formation). A half-headed, single-barbed fishhook arrow ↿ represents the movement of a single electron — the signature of homolysis and of all subsequent radical propagation steps. Mixing these up is not a stylistic error; it describes a chemically different, often impossible, process (e.g., using a full arrow to "form" a radical would imply a pair went one place and left an odd electron unaccounted for).

Geometry differs sharply between the two intermediate families. A carbon radical, with three σ-bonds and one unpaired electron in a roughly nonbonding orbital, sits close to planar (sp²-like) with the SOMO in a p-orbital perpendicular to that plane — shallow enough to invert rapidly, and stabilized by adjacent π-systems through spin delocalization. A carbocation, with only three bonds and an empty p-orbital, is essentially always planar sp² (to maximize s-character in the three σ-bonds and leave a pure, most-stabilizable empty p-orbital) and is stabilized by any adjacent electron density (hyperconjugation, resonance, alkyl induction). A carbanion, with three bonds plus a lone pair, is pyramidal sp³ (like an amine), with the lone pair stabilized by withdrawing the electron density it does not want, i.e., by adjacent electronegative atoms or π-acceptors.

The energetic cost of homolysis is the bond dissociation energy, BDE, defined for the gas-phase process

where:

• $\mathrm{BDE}$ is always positive (bond breaking costs energy) and is tabulated for common bonds (e.g., $\mathrm{BDE}(\mathrm{CH_3-H}) \approx 105\ \text{kcal/mol}$, $\mathrm{BDE}(\mathrm{Cl-Cl}) \approx 58\ \text{kcal/mol}$).
• No solvent or counterion terms appear — homolysis is intrinsically a gas-phase-like, low-polarity-friendly event; solvent barely changes BDE because no charge is created.

The energetic cost of heterolysis is different in kind, because forming isolated gas-phase ions is enormously endothermic — you must also pay for ionization and reclaim electron affinity, and in solution you recover most of that cost through solvation:

where:

• $\mathrm{DH}_{\text{het}}$ is the gas-phase heterolytic bond dissociation enthalpy, forming $A^+$ and $B^-$.
• $\mathrm{IE}(A^{\bullet})$ is the ionization energy of the radical $A^{\bullet}$ (energy to remove an electron and form $A^+$).
• $\mathrm{EA}({}^{\bullet}B)$ is the electron affinity of the radical ${}^{\bullet}B$ (energy released when it accepts an electron to form $B^-$).

Reading this equation: heterolysis is always more endothermic than homolysis in the gas phase (often by 100–200 kcal/mol), because $\mathrm{IE} \gg \mathrm{EA}$ for almost every organic fragment. The only reason heterolysis competes at all in real chemistry is that solvation energy of the resulting ions — especially in polar protic solvents like water or methanol — can recover 80–100+ kcal/mol, collapsing that gap. This is why heterolysis is essentially a solution-phase phenomenon, while homolysis is comfortable in the gas phase, in flames, in the upper atmosphere, and in nonpolar solvents. A second governing relationship is qualitative but load-bearing: the polarity of the bond itself predisposes it toward one cleavage mode. For a bond A–B with substantial electronegativity difference, the bonding MO is already polarized toward the more electronegative atom (it contributes more atomic-orbital character to the bonding MO). Heterolysis simply completes a partition the bond was already biased toward; homolysis of such a bond, forcing an electron away from the atom that "wants" it, is disfavored. Symmetric, nonpolar bonds (Cl–Cl, C–C between similar carbons, O–O in peroxides) have no such bias and homolyze far more readily than they heterolyze in typical (non-ionizing) media.

Example 1 (foundational) — Classify the cleavage mode and draw correct arrows for Cl₂ under UV light. Chlorine gas, $\mathrm{Cl-Cl}$, absorbs near-UV/visible light ($\lambda \approx 330\text{–}400\ \text{nm}$) and dissociates in the gas phase or in nonpolar solvent (CCl₄) during radical chlorination. Reasoning. Both atoms are identical chlorine, so there is zero electronegativity difference — the bonding MO is perfectly symmetric, giving no thermodynamic preference for sending both electrons to one chlorine over the other. The reaction is run in the gas phase or nonpolar solvent, with no polar medium available to stabilize $\mathrm{Cl^+}$ and $\mathrm{Cl^-}$ (and even if it could, $\mathrm{Cl^+}$ is a very high-energy species — chlorine's ionization energy is large, roughly $299\ \text{kcal/mol}$, $12.97\ \text{eV}$). Photon energy is delivered directly to an electron in the $\sigma^{*}$ antibonding orbital region (or more precisely promotes the molecule to a dissociative excited state), symmetrically depositing energy across the bond. Arrows. Two fishhook arrows, one from the bond to each chlorine:

Check against BDE. $\mathrm{BDE}(\mathrm{Cl-Cl}) \approx 58\ \text{kcal/mol}$, corresponding to a photon of wavelength

This falls in the visible/near-UV range, consistent with chlorine's pale yellow-green color and its known photodissociation onset — confirming homolysis is energetically accessible with ordinary lamps or sunlight, with no ionic pathway competing at all.

Example 2 (applied) — Predict the dominant cleavage mode for the same C–O bond under two different reaction setups. Consider the C–O bond in a tertiary alkyl tosylate, $R_3C\text{-OTs}$, where $\mathrm{OTs}^-$ (tosylate) is an excellent, resonance- and inductively-stabilized leaving group. Setup A: neat, heated in the gas phase at 300 °C. No solvent exists to solvate ions. Even though tosylate is a superb anion once formed (charge delocalized over the sulfonate and aromatic ring), the intrinsic gas-phase cost of separating $R_3C^+$ and $\mathrm{OTs}^-$ (dominated by the large ionization energy term) remains on the order of 130–150 kcal/mol — unrecoverable without solvation. At 300 °C, thermal energy ($RT \ln$-scale reaction barriers accessible are typically 40–60 kcal/mol, not 130+), homolysis of the weaker of the available bonds (often a C–H or C–C bond elsewhere in the molecule, or, if forced, the C–O bond itself) becomes the kinetically competitive channel, or, more realistically, a concerted pericyclic elimination pathway (E_i, syn-periplanar, no charged intermediate at all) dominates instead. The lesson: heterolysis needs solvent, and its absence forces the molecule toward homolytic or concerted alternatives. Setup B: dissolved in acetic acid (polar protic, ionizing solvent) at 25–60 °C. Now solvation of both the incipient tertiary cation (by dipole alignment of acetic acid carbonyls) and the tosylate anion (by hydrogen bonding of acetic acid O–H to the sulfonate oxygens) recovers well over 100 kcal/mol, collapsing the net barrier to a range readily crossed at mild heating. Heterolysis proceeds with a full-arrow mechanism:

followed by rapid capture of the planar cation by acetic acid (or by the conjugate base) to give the substitution product — the classic $S_N1$ solvolysis pathway used to measure relative carbocation stabilities (solvolysis rate studies). Take-away. Identical bond, identical intrinsic polarity — but the solvent environment alone flips the dominant cleavage mode from effectively unobservable heterolysis (gas phase) to facile heterolysis (ionizing solvent), while a nonpolar, non-ionizing medium at high temperature would instead favor homolytic or concerted routes.

Example 3 (exam / competition level) — Rank three C–X bonds by preference for heterolysis vs. homolysis, and justify using both the polarity argument and an estimated thermodynamic comparison. Rank these three bonds — (i) $\mathrm{CH_3-CH_3}$ (ethane C–C), (ii) $(\mathrm{C_6H_5})_3\mathrm{C-Cl}$ (trityl chloride), (iii) $\mathrm{CH_3-I}$ (methyl iodide) — by their tendency to undergo heterolysis in a polar protic solvent (e.g., aqueous acetone) versus homolysis under thermal/photochemical gas-phase conditions. Bond (i), C–C in ethane. Electronegativity difference is zero (carbon–carbon), so the bonding MO is symmetric — no polarity bias toward heterolysis at all. Additionally, forming either a primary carbocation or a primary carbanion is exceptionally destabilizing (no adjacent groups to donate density to the empty p-orbital of a cation, and no adjacent electron-withdrawing groups to stabilize a lone pair on the anion). Heterolysis of this bond in any solvent is essentially never observed; when this bond breaks at all (e.g., in cracking, pyrolysis, combustion), it is exclusively homolytic, consistent with $\mathrm{BDE}(\mathrm{C-C}) \approx 90\ \text{kcal/mol}$ being reachable thermally at several hundred °C while heterolysis would require an inaccessible 200+ kcal/mol even with generous solvation. Bond (ii), trityl chloride, $\mathrm{Ph_3C-Cl}$. Here the electronegativity gap (C vs. Cl) already biases the bond toward heterolysis in the "send both electrons to Cl" direction, and critically, the cation formed upon ionization is the trityl cation, $\mathrm{Ph_3C^+}$ — spectacularly stabilized by resonance delocalization of the positive charge into all three aromatic rings simultaneously (a textbook example of extended cation stabilization). This dramatically lowers the effective $\mathrm{IE}$-like term in the heterolytic bond-dissociation relationship (the radical $\mathrm{Ph_3C^{\bullet}}$ itself is already unusually stable and easy to further oxidize to the cation, because the same resonance stabilizes both species, but stabilizes the cation more per the Hammond-postulate-like reasoning applied to charge-bearing resonance structures). Trityl chloride ionizes so readily that it is a classroom standard for demonstrating spontaneous, visible heterolysis (a deep yellow/orange color develops from the trityl cation) merely upon dissolution in polar solvents such as liquid $\mathrm{SO_2}$ or acetone/water mixtures, with no external initiator needed. Bond (iii), methyl iodide, $\mathrm{CH_3-I}$. The C–I bond is polarized (I more electronegative than C, though only mildly: 2.66 vs. 2.55) but weak in absolute terms, $\mathrm{BDE}(\mathrm{C-I}) \approx 57\ \text{kcal/mol}$ — one of the weakest common C–halogen bonds. This bond is genuinely ambivalent: under photochemical or thermal gas-phase conditions (e.g., UV photolysis in the gas phase or nonpolar solvent), it readily homolyzes to give methyl radical and iodine atom, because the resulting primary methyl cation would be extremely unstable (no stabilization at all) making heterolysis unfavorable despite bond polarity. But methyl iodide is, famously, an excellent electrophile in heterolytic $S_N2$ reactions — not through unimolecular ionization (there is no stable methyl cation to form), but through concerted, nucleophile-assisted heterolysis, where the incoming nucleophile's electron pair pushes into carbon exactly as the C–I bond's electron pair is pushed onto iodide, all in one step, avoiding ever generating a free, high-energy methyl cation. Ranking (tendency toward spontaneous unimolecular heterolysis in polar protic solvent):

with methyl iodide reserved for bimolecular, concerted heterolytic pathways ($S_N2$) rather than free ionization, and ethane essentially confined to homolysis under any accessible conditions. This example demonstrates that "heterolysis vs. homolysis" is not merely about bond polarity, but crucially about whether the resulting ion (especially the cation) is stable enough to justify the enormous ionization penalty — precisely the logic embedded in the $\mathrm{IE} - \mathrm{EA}$ correction term.

The homolysis/heterolysis distinction crystallized through the convergence of two research traditions in the early-to-mid twentieth century: physical-organic mechanistic chemistry (Christopher Ingold, Robert Robinson, and coworkers in Britain, who built the language of curved arrows, nucleophiles, and electrophiles in the 1920s–30s) and free-radical chemistry (initiated experimentally much earlier but placed on a rigorous mechanistic footing by Morris S. Kharasch in the United States, who in 1933 demonstrated that "peroxide effects" reversed the regiochemistry of HBr addition to alkenes — the first clear proof that a different bond-cleavage mode (radical chain vs. ionic) could completely change a reaction's outcome).