Evaluating Resonance Structures

Ranking resonance contributors correctly is not an academic exercise confined to textbook problem sets — it is the everyday reasoning tool bench chemists use to predict where a nucleophile or electrophile will attack. Medicinal chemists designing enolate alkylations or Michael additions must know which resonance contributor (α or γ carbon, or oxygen) actually carries the reactive charge density, directly informing regioselectivity in the synthesis of complex natural products and drug candidates. In physical organic chemistry, the ranking explains substituent effects on solvolysis rates (Example 2's methoxybenzyl cation logic underlies the design of protecting groups and the interpretation of Hammett $\sigma^+$ constants). In materials science, understanding which resonance structure dominates in extended conjugated polymers (polyanilines, push–pull chromophores for nonlinear optics) determines whether a material behaves more like a neutral aromatic form or a charge-separated quinoidal form — a distinction that controls conductivity and color. In biochemistry, the reduced basicity of the amide nitrogen (rank amide resonance: neutral contributor major, charge-separated minor but essential) is the entire structural reason peptide bonds are planar and resistant to rotation, underlying protein secondary structure (the Ramachandran plot's restricted geometries trace directly back to this resonance argument). Even spectroscopists rely on this ranking: NMR chemical shifts and IR carbonyl stretching frequencies are routinely rationalized by asking which resonance contributor dominates a given substrate — a shifted carbonyl stretch in an amide versus a ketone is a direct fingerprint of how much weight the charge-separated contributor carries.

Ask the tutor: How would the ranking change if we compared resonance structures of a carbanion stabilized by an adjacent sulfur versus an adjacent oxygen — does electronegativity alone predict which stabilizes the negative charge better?

Start with what a resonance structure is not: it is not a distinct chemical species, and the molecule does not oscillate between structures. A single Schrödinger equation describes one molecular electronic wavefunction, $\Psi$. Lewis structures are a bookkeeping device — localized two-electron, two-center bonds and lone pairs — invented before quantum mechanics, and many delocalized electron distributions simply cannot be captured by any single Lewis diagram. When more than one reasonable Lewis structure can be drawn for the same arrangement of nuclei (same connectivity, same positions of atoms, differing only in the placement of $\pi$ electrons and lone pairs), the true wavefunction is better approximated as a weighted superposition:

where each $\Psi_i$ is the wavefunction implied by one Lewis structure and $c_i$ is a mixing coefficient. This is the variational, resonance-theory (valence-bond) picture, and it is mathematically analogous to how atomic orbitals mix to form molecular orbitals in Conjugation: Overlapping p Orbitals. The physically meaningful, experimentally observed molecule corresponds to the coefficients $c_i$ that minimize the total energy — the variational principle guarantees that mixing in more structures (or better structures) can only lower the energy, never raise it, relative to the single best contributor. The question this article answers is: how do we estimate the relative sizes of $|c_i|^2$ without solving the Schrödinger equation? We cannot compute $c_i$ exactly by hand, but we can rank them, because each defect in a Lewis structure corresponds to a known, large or small energetic penalty. A structure that is a poor approximation to the true bonding (badly missing electrons where orbital overlap wants them, or badly misplacing charge relative to what electronegativity wants) mixes in weakly — small $c_i$ — and is called a "minor contributor." A structure that already resembles good bonding mixes in strongly — large $c_i$ — and is a "major contributor." The four criteria, in strict priority order, are:

1. Complete octets beat incomplete octets. An atom with only six electrons (a species like a carbocation, or boron in $\text{BF}_3$) has an empty, high-energy orbital. Any resonance structure that could instead place a full octet there — by forming an additional bond — represents dramatically better orbital overlap and a lower-energy configuration. The energy cost of an incomplete octet is on the order of an entire covalent bond strength (60–100 kcal/mol), because that structure is missing an actual bonding interaction, not merely mis-locating a charge.

2. Minimum number of formal charges beats more formal charges. Separating charge onto two different atoms costs Coulombic energy — you have pulled electron density away from where it would otherwise neutralize a nucleus, creating a charge-separated (essentially dipolar) configuration. This costs real energy (typically 20–40 kcal/mol per unit of separated charge in a small molecule), but it is much less than a missing bond, so this criterion is checked only after octets are satisfied.

3. Given equal charge separation, negative charge on the more electronegative atom (and positive charge on the less electronegative atom) is better. This is a direct electronegativity argument: electronegative atoms (O, N, F, Cl) have lower-energy valence orbitals and "want" extra electron density; putting negative charge there is like letting the actual electron distribution relax toward where it already prefers to sit. The energetic benefit here is smaller than octet or charge-count effects — a few kcal/mol to ~10 kcal/mol — so it only breaks ties between structures that already tie on criteria 1 and 2.

4. Adjacent like charges are destabilizing; adjacent opposite charges are stabilizing. This is pure Coulomb's law, $E \propto q_1q_2/r$, applied at the length scale of a bond (~1.5 Å). Two positive charges one bond apart cost real energy; a positive charge adjacent to a negative charge is a small stabilization. This is the finest tie-breaker, invoked last.

The variational statement underlying everything above is:

where:

• $\hat{H}$ is the molecular Hamiltonian (kinetic + Coulomb potential energy operator for all electrons and nuclei),
• $\Psi_{\text{true}}$ is the actual (delocalized) electronic wavefunction,
• $E(\Psi_i)$ is the energy that would be computed by treating contributor $i$ as if it were the exact wavefunction.

Reading: mixing in additional Lewis structures always lowers the energy relative to any single contributor taken alone — this is why delocalization stabilizes molecules, and why "the more good resonance structures, the more stable." The mixing coefficients that weight each contributor's importance are governed, to leading approximation, by the energy of each localized structure relative to the others:

where:

• $E_i, E_j$ are the estimated "energies" of contributors $i$ and $j$ (higher when octets are broken or charge is poorly placed),
• $R$ is the gas constant, $T$ the temperature — included only to show that even a modest energy gap (tens of kJ/mol) makes $c_j$ negligible compared with $c_i$.

Reading: this is not a formula to plug numbers into (we rarely know $E_i$ precisely) — it explains qualitatively why one bad feature (say, an incomplete octet, worth tens of kcal/mol) is enough to make a contributor essentially irrelevant to the hybrid, even though it remains a "legal" Lewis structure. Finally, the electronegativity ordering that criterion 3 depends on is the ordinary Pauling scale relevant to organic atoms:

where $\chi$ denotes Pauling electronegativity (F 3.98, O 3.44, N 3.04, Cl 3.16, C 2.55, H 2.20). Any resonance structure placing negative charge on the leftmost feasible atom in this list (given the connectivity) is preferred by criterion 3.

Example 1 (foundational): Nitromethane, $\text{CH}_3\text{NO}_2$. Nitromethane's nitro group can be drawn two ways for each N–O bond. Structure (i): $N$ double-bonded to one O, single-bonded to the other O⁻, with formal charge +1 on N (since N here has 4 bonds: 1 to C, 1σ+1π to one O, 1σ to the other O, and 0 lone pairs: FC $=5-0-4=+1$), and −1 on the singly-bonded O (FC on that O $=6-6-1=-1$; it has 3 lone pairs + 1 bond). The doubly-bonded O has FC $=6-4-2=0$. Structure (ii) is identical but with the double bond swapped to the other oxygen — by symmetry it is chemically indistinguishable, just relabeled. Apply the criteria: both O's satisfy full octets (each O: either 1 bond+3 lone pairs = 8 e⁻ region, or 2 bonds (1σ+1π) + 2 lone pairs = 8 e⁻); N has 4 bonds and 0 lone pairs = 8 e⁻ (counting shared pairs twice). Criterion 1: tied, both perfect octets. Criterion 2: each structure has exactly two formal charges (+1 on N, −1 on one O) — tied. Criterion 3: negative charge sits on oxygen (the more electronegative atom, $\chi=3.44$ vs. N's 3.04) in both — tied. Criterion 4: the + and − charges are adjacent (N–O bond) and of opposite sign, which is Coulombically favorable in both — tied. Because structures (i) and (ii) are related by nothing more than swapping the labels on two chemically equivalent oxygens, they are true equivalent resonance contributors, mixing with equal coefficients $c_1 = c_2$. The experimental consequence: nitromethane's two N–O bond lengths are equal (both ≈ 122 pm, intermediate between typical N–O single ~140 pm and N=O double ~115 pm), confirming that the real molecule is the symmetric 50:50 hybrid, not a molecule flickering between two distinguishable tautomers. This example demonstrates the "no discrimination needed" limiting case: when all four criteria tie exactly, the structures are equal contributors, and this is the correct way to recognize true resonance (as opposed to different molecules or genuine tautomers, which differ in atom connectivity and are not resonance structures at all).

Example 2 (applied): Ranking contributors of an aryl cation vs. a stabilized benzylic/allylic cation — why $\text{SN1}$ solvolysis rates differ by orders of magnitude. Consider the cation formed on ionization of $p$-methoxybenzyl chloride, $p\text{-CH}_3\text{O-C}_6\text{H}_4\text{-CH}_2^{+}$. The benzylic cation (empty $p$ orbital on the exocyclic $CH_2$) can delocalize into the ring, and further into the oxygen lone pair, generating a family of contributors. Contributor G (initial): positive charge localized on the exocyclic carbon, ring fully aromatic (three alternating double bonds), oxygen neutral with 2 lone pairs. Octet check: the cationic carbon has only 3 bonds (2 to H, 1 to ring) and 0 lone pairs = 6 electrons — incomplete octet, formal charge +1. Contributor H: push the arrow from a ring π bond into the exocyclic C, forming $CH_2=C(\text{ring carbon})$, and placing the resulting positive charge on the ring carbon ortho or para to the exocyclic position (a normal benzylic-type delocalization, breaking ring aromaticity temporarily in that resonance structure). This ring carbon, if it's the one para to the OMe group, has 3 bonds and empty p orbital: still an incomplete octet, formal charge +1, but now conjugated further. Contributor I: from oxygen's lone pair, push an arrow to form a new π bond O=C(ring carbon), placing the resulting formal positive charge on oxygen. Now check octets: oxygen has 2 bonds (1σ to CH₃, 1σ+1π to ring carbon = actually 2σ+1π = 3 bonds) and 1 remaining lone pair: FC on O $=6-2-3=+1$; the exocyclic $CH_2$ carbon now has 4 bonds (2 H, 1 to ring, plus the π bond completed through the chain) — full octet restored on every carbon, and oxygen, despite bearing +1 formal charge, still has a complete octet (3 bonds + 1 lone pair = 8 electrons around it). Ranking: Criterion 1 — contributor I is the only one with all atoms octet-complete; G and H both leave some carbon with only 6 electrons. So I dominates, even though I has a formal positive charge sitting on the highly electronegative oxygen (which criterion 3 would normally penalize) — but criterion 1 is checked first and overrides that concern completely, since G and H are disqualified before criterion 3 ever gets applied to compare them with I. Chemical payoff: because contributor I is a legitimate, low-energy, fully-octet-satisfied resonance structure (an oxocarbenium-like structure, $\text{O}^+=\text{C}$), the para-methoxybenzyl cation is dramatically stabilized relative to unsubstituted benzyl cation — solvolysis rate accelerations of $10^3$–$10^6$ are typical for a para-OMe substituent in $\text{S}_N1$ reactions, exactly because this contributor exists and is not merely a curiosity but a major, octet-complete resonance form. This is the mechanistic root of why methoxy (and other lone-pair-bearing substituents) are powerful cation-stabilizing groups, foreshadowing the broader treatment of Inductive Effects and Charge Distribution alongside resonance effects.

Example 3 (exam/competition level): Ranking five contributors of the diazonium-coupled azide anion, $\text{N}_3^-$, and computing formal charges explicitly. Azide ion, $N_3^-$ (atoms $N_1$–$N_2$–$N_3$, linear, total charge −1), is a classic multi-contributor system. Generate contributors systematically. Contributor J: $N_1{\equiv}N_2^+{=}N_3^{2-}$ — obtained by a poorly chosen arrow push that dumps two extra lone pairs on $N_3$. Check: $N_3$ would need 3 lone pairs (6 e⁻) plus a double bond (counts 2) $=8$ around it giving FC $=5-6-2=-3$? This overcounts — let's discard this as not even a legitimately balanced structure (total charge would not equal −1); it fails basic electron bookkeeping and is not a valid contributor at all. (Always verify total formal charge sums to the ion's actual charge before ranking — this is a mandatory checkpoint, separate from, and prior to, the four criteria.) Contributor K (valid): $N_1^{-}{=}N_2^{+}{=}N_3^{-}$, i.e., $:\overset{-}{N}=N^+=\overset{-}{N}:$ Wait — check total charge: $-1+1-1 = -1$. Good, matches overall charge. Octet check: $N_1$ has 1σ+1π (2 bonds) to $N_2$, plus 2 lone pairs (4 e⁻): total $=4+4=8$, FC $=5-4-2=-1$. $N_2$ has 2 double bonds (4 bonds total) and 0 lone pairs: FC$=5-0-4=+1$. $N_3$ symmetric to $N_1$: FC $=-1$. Total: $-1+1-1=-1$ ✓. All full octets. This is the symmetric linear structure, contributor K. Contributor L: $N_1\equiv N_2^{+}-N_3^{2-}$: $N_1$ triple-bonded to $N_2$ (3 bonds, 1 lone pair): FC $=5-2-3=0$. $N_2$: triple bond to $N_1$ (3 bonds) + single bond to $N_3$ (1 bond) = 4 bonds, 0 lone pairs: FC$=5-0-4=+1$. $N_3$: single bond (1) + 3 lone pairs (6 e⁻) = FC$=5-6-1=-2$. Total: $0+1-2=-1$ ✓ charge balances, and octets are all complete (each N has 8 electrons in its accounting), but this structure has a doubly-charged atom (−2 on $N_3$) — a highly unfavorable, large charge separation. Contributor M: the mirror image of L, $N_1^{2-}-N_2^{+}\equiv N_3$, equivalent by symmetry to L (charge −2 now on $N_1$ instead).

Rank all valid contributors (K, L, M; J discarded for failing charge bookkeeping): Criterion 1 (octets): K, L, M all have complete octets everywhere — tied. Criterion 2 (charge count/magnitude): K carries formal charges of magnitude {−1, +1, −1} — three unit charges total, but no charge exceeds magnitude 1. L and M carry {0, +1, −2} — also three units of charge in total, but concentrated so that one atom bears a doubly charged center (−2). Standard practice ranks structures with smaller maximum charge magnitude on any single atom as better — a −2 formal charge is considerably less realistic than two separate −1's, because concentrating two extra electrons on one nitrogen is a larger local distortion than spreading one extra electron each onto two nitrogens. K is preferred over L and M by this refinement of criterion 2. Criterion 3 (electronegativity placement): moot here — all atoms are nitrogen, so there is no electronegativity difference to invoke; this criterion only discriminates between different elements. Criterion 4 (adjacency): in K, the −1 charges are on $N_1$ and $N_3$, which are not adjacent (they are both bonded to $N_2$, but not to each other) — two like charges separated by an intervening +1 is a favorable, low-energy arrangement (alternating charges along the chain, each adjacent pair being +/− and therefore attractive). In L, the −2 sits directly bonded (adjacent) to the +1 — while opposite-charge adjacency is favorable, having a full −2 concentrated locally still costs more than K's spread-out −1/+1/−1 pattern, consistent with K also winning criterion 2.

Conclusion: $N_3^-$'s major contributor is the fully symmetric $:\overset{-}{N}=N^+=\overset{-}{N}:$ structure K, with L and M as real but minor contributors. This correctly predicts the experimentally observed structure: azide is linear (consistent with sp hybridization at $N_2$ demanded equally by K, L, and M — all agree on linearity) with two equal N–N bond lengths (116 pm each, essentially double-bond-like), which matches K's symmetric double-double bonding far better than L or M's asymmetric triple/single pattern (which would predict unequal bond lengths, ~110 pm and ~140 pm — not observed). The correct ranking is not just an abstract exercise; it directly predicts measurable bond lengths.

The systematic rules for ranking resonance structures trace to Linus Pauling, who built resonance theory into a quantitative (variational, valence-bond) framework in the 1930s, synthesizing quantum mechanics (Heitler–London bonding theory) with the empirical Lewis-structure language that organic chemists already used. Pauling's 1939 book The Nature of the Chemical Bond systematized exactly the hierarchy used here: octet completeness, charge separation, and electronegativity-based charge placement, derived from his own electronegativity scale (which he had introduced in 1932 specifically to quantify bond polarity). Before Pauling, chemists like G. N. Lewis had the electron-pair bond and the octet rule, and chemists intuitively understood that some Kekulé-type structures for benzene "felt" more important than others, but there was no principled ranking — Pauling supplied both the electronegativity numbers and the variational justification (mixing lowers energy) that make the ranking rigorous rather than a matter of taste.