Oxygen, Ozone, and Peroxide Chemistry
Liquid oxygen is pale blue. It also clings, faintly, to the poles of a magnet. This party trick cannot be explained by any Lewis structure of O2. The textbook double bond O=O predicts a diamagnetic, colorless molecule. It assumes all electrons are paired. The real explanation lives one level down, in molecular orbital theory. It is one of the cleanest, most quoted triumphs of MO theory. This triumph stands out over simple Lewis-dot reasoning in all of introductory inorganic chemistry. Get the electronic structure of O2 right. Then a whole family of species snaps into place: ozone O3, the peroxide ion O2^2−, the superoxide ion O2^−, and the ordinary oxide ion O^2−. They form a single coherent bond-order ladder. This ladder predicts bond lengths, bond strengths, magnetism, and reactivity all at once. This chapter matters because oxygen chemistry is not a museum piece. It is the chemistry of combustion, respiration, rocket propellants, bleaches, and atmospheric ozone. That ozone shields the biosphere from ultraviolet radiation. The paramagnetism of O2 explains why singlet oxygen matters. Singlet oxygen is an excited, diamagnetic form. It is a dangerously reactive species in photodynamic therapy and photo-oxidation of plastics. The instability of ozone relative to O2 has consequences too. It is the reason the ozone layer is a delicate steady-state balance. It is not a permanent deposit. The superoxide ion is a genuine biological hazard. It is generated as a byproduct of aerobic respiration. It is neutralized by the enzyme superoxide dismutase. None of this can be understood from "oxygen has six valence electrons" alone. It requires seeing how those electrons fill orbitals. The big picture organizes itself around a single quantity. That quantity is the number of electrons occupying the antibonding π orbitals. These orbitals belong to a diatomic O–O unit. For ozone, an analogous three-center π system applies instead. Adding electrons to π systematically changes the bond. This means going from O2^2+ conceptually down to O2^2−. It lengthens and weakens the O–O bond. It also converts magnetism from doubly paramagnetic to eventually diamagnetic again. This single idea, once mastered, lets you predict without memorization. You can predict bond order, bond length, and bond dissociation energy. You can also predict color, magnetic susceptibility, and characteristic reactivity. This works for every common oxygen species you will meet. This article builds the molecular orbital diagram of O2 from first principles. It extends this diagram to the O2^n− series. It also extends it to ozone's three-center π system. Then it surveys the oxide/peroxide/superoxide classification of solid-state oxygen compounds. This is the classification chemists actually use. It predicts what happens when an alkali or alkaline-earth metal burns in air. It closes with the reactivity chemistry that follows directly from bonding. This includes oxidizing power, disproportionation, and the role of oxygen species. These roles appear especially in catalysis and biology.
- O2's ground state is a triplet (paramagnetic, unpaired electrons in degenerate π* orbitals) — a direct, testable prediction of MO theory that Lewis structures miss entirely.
- Bond order in the O2^n− series runs 2 → 1.5 → 1 → (formally 0, i.e., no O–O bond) as electrons fill π*, tracking bond length (121 → 128 → 149 pm) and stretching frequency in lockstep.
- Ozone (O3) is bent (117°), resonance-stabilized by a delocalized 4-electron/3-center π system, and thermodynamically unstable relative to (3/2)O2 by about +143 kJ/mol.
- Oxide (O^2−), peroxide (O2^2−), and superoxide (O2^−) are classified by which metals form them: electropositive metals (alkali, heavy alkaline earth) stabilize peroxide/superoxide via lattice-energy and charge-density arguments.
- Peroxide and superoxide undergo characteristic disproportionation with water/acid, releasing H2O2 or O2 — reactions with quantitative stoichiometry worth memorizing.
- Singlet oxygen, an electronically excited, diamagnetic form of O2, is far more reactive toward organic substrates than the ground triplet state because spin restrictions on cycloaddition are lifted.
- Key idea: every property of an oxygen species — bond length, magnetism, color, reactivity — is a direct readout of how many electrons occupy the π* (or bridging π) orbitals of the O–O framework.
Where this lives today
Superoxide dismutase (SOD), an enzyme found in essentially all aerobic organisms, exists precisely because mitochondrial respiration leaks electrons to O2, generating superoxide as an unavoidable byproduct; SOD catalyzes the disproportionation of two O2^− radicals into O2 and H2O2 (which catalase then further disproportionates to water and oxygen), and failures or mutations in this defense system are implicated in oxidative-stress pathologies and neurodegenerative disease. Photodynamic therapy for certain cancers deliberately generates singlet oxygen (the excited, spin-paired form of O2) using a photosensitizer drug plus light, exploiting singlet oxygen's much higher reactivity toward cellular lipids and proteins compared to ordinary ground-state triplet O2 — the same spin-state argument that explains why triplet O2 reacts only sluggishly with most organic molecules despite being thermodynamically a powerful oxidant (the "kinetic stability of dioxygen" that keeps the biosphere from spontaneously combusting). Hydrogen peroxide, industrially manufactured by the anthraquinone auto-oxidation process on a multi-million-tonne annual scale, is a peroxide-classified oxidant used in bleaching (paper pulp, textiles), disinfection, and as a "green" oxidant in fine-chemical synthesis because its only byproduct is water. Sodium percarbonate and sodium perborate — both peroxide-adduct solids — are the active bleaching agents in modern non-chlorine laundry detergents, releasing H2O2 on dissolution exactly as Na2O2 releases H2O2 on hydrolysis. Potassium superoxide, KO2, is used in industrial and military self-contained breathing apparatus and submarine air-regeneration systems precisely because its disproportionation with exhaled moisture and CO2 simultaneously scrubs carbon dioxide and releases breathable oxygen. Stratospheric ozone remains the single most important atmospheric shield against biologically damaging UV-B and UV-C radiation, and understanding its thermodynamic instability relative to O2 (it does not simply "sit there" — it is continuously photochemically produced and destroyed in the Chapman cycle) is central to modeling ozone depletion and recovery following the Montreal Protocol's phase-out of chlorofluorocarbons. The same bonding logic that classifies O2^n− species extends naturally into the broader catenation chemistry of Sulfur Allotropes and Catenation, where sulfur's willingness to form long homonuclear chains and rings contrasts instructively with oxygen's preference for small, highly π-bonded units — a contrast rooted in the same p-block periodic trends that govern orbital overlap efficiency down a group.
:::faq[Why do alkali metals form different oxygen compounds (oxide, peroxide, superoxide) down the group instead of all forming the same one]? The controlling factor is lattice energy, which scales with the product of ionic charges divided by the sum of ionic radii. Small cations like Li^+ pair very favorably with the small, doubly-charged, compact O^2− ion, generating enough lattice energy to pay the substantial energetic cost of forming O^2− from O2 in the first place. As cations get larger down the group (Na^+, K^+, Rb^+, Cs^+), that lattice-energy advantage for the small O^2− ion shrinks faster than it does for the larger, more diffuse peroxide and superoxide ions, so the thermodynamically preferred product shifts toward the anion whose size and charge density better match the cation's — peroxide for sodium, superoxide for potassium through cesium.
Ask the tutor: How would the MO diagram and bond-order predictions change for a heteronuclear analogue like NO or CO, where the two atoms are no longer equivalent?
The physical picture
Start with the Lewis structure everyone draws for O2: :O=O:, two lone pairs on each oxygen, one σ bond and one π bond, all electrons paired. This structure is wrong in a specific, falsifiable way — it predicts a diamagnetic molecule, and liquid O2 is famously paramagnetic, poured between the poles of a magnet in countless lecture demonstrations where it visibly clings to the field. The resolution requires molecular orbital theory: build MOs from the 2s and 2p atomic orbitals of the two oxygen atoms, respecting symmetry, and fill them with the 12 valence electrons (6 from each oxygen) in order of increasing energy. For homonuclear diatomics of the second period, 2s–2p mixing is small once you reach O2 (unlike for the earlier elements B2, C2, N2 where 2s–2pz mixing pushes the σ_2p above π_2p). By oxygen, the ordering reverts to the "textbook" sequence: σ_2s, σ_2s, σ_2p (lowest), π_2p (doubly degenerate, next), π_2p (doubly degenerate, next), σ_2p (highest). Filling 12 valence electrons: σ_2s² σ_2s² σ_2p² π_2p⁴ π_2p² — and it is that last pair, distributed one-each into the two degenerate π orbitals by Hund's rule, that makes O2 a ground-state triplet with two unpaired electrons. Bond order is computed from bonding minus antibonding electron pairs:
For O2: bonding electrons in σ_2s, σ_2p, π_2p total 2+2+4 = 8; antibonding in σ_2s and π_2p total 2+2 = 4; net (8−4)/2 = 2. A bond order of 2 recovers the "double bond" of the Lewis picture, but now correctly decorated with two unpaired π electrons — magnetism the dot structure could never produce. Removing or adding electrons from this π pair generates the entire O2^n− family with a completely predictable trend. Removing one electron (dioxygenyl cation, O2^+, found in exotic salts like O2^+PtF6^−) raises bond order to 2.5 and shortens the bond. Adding one electron gives superoxide, O2^−, bond order 1.5, one unpaired electron (paramagnetic, radical anion). Adding two electrons gives peroxide, O2^2−, bond order 1, all electrons paired (diamagnetic), a genuine single O–O bond exactly as in H2O2. There is no stable "O2^2−-plus-more" species with a bonding O–O interaction; adding further electrons past peroxide would have to occupy σ*_2p, which is so strongly antibonding that the O–O bond simply breaks, giving two separate O^2− oxide ions instead of a diatomic unit at all. Ozone extends the same logic to a bent triatomic. Each terminal oxygen and the central oxygen contribute p orbitals perpendicular to the molecular plane; these combine into a three-center π system holding 4 electrons distributed over 3 MOs (one bonding, one nonbonding, one antibonding), with the 4 electrons filling the bonding and nonbonding levels. The result is a delocalized 4-electron/3-center π bond exactly analogous to the allyl anion in organic chemistry, giving each O–O linkage a fractional π bond order of 0.5 on top of the σ framework — consistent with the experimental O–O bond length of 127.8 pm, intermediate between a single bond (~148 pm) and a double bond (~121 pm), and with the equivalence of the two O–O bonds required by the C2v-symmetric resonance structures.
The core ideas
The single governing relationship for diatomic oxygen species is the bond-order formula already introduced, restated with explicit reference to the π* population that actually varies across the series:
where:
- is the number of electrons occupying the two degenerate π*_2p orbitals (ranges from 0 in O2^2+ to 4 in peroxide),
- the leading 2 comes from the fixed σ_2p and π_2p bonding contributions (one σ bond order unit + one full π bond order unit) once σ*_2s and σ_2s are fully occupied and cancel as they do throughout this series,
- bond order decreases by 0.5 for every electron added to π*, since two electrons (one full orbital) lower bond order by exactly 1.
Reading it: every additional electron placed in the antibonding π* system directly and linearly weakens and lengthens the O–O bond; this is why the bond order ladder in the figure steps down in units of 0.5 as you move O2^+ → O2 → O2^− → O2^2−, and why bond length increases correspondingly (112 → 121 → 128 → 149 pm). A second governing relationship connects bond order to both stretching frequency and bond dissociation energy through the empirical (but robust) correlation that stiffer, shorter bonds vibrate faster and require more energy to break:
where:
- is the effective force constant of the O–O bond (N/m),
- is the O–O stretching wavenumber (cm⁻¹), directly measurable by IR/Raman spectroscopy,
- is the reduced mass of the two oxygen atoms (a constant across the whole isotopically-uniform series).
Reading it: because is fixed, stretching frequency alone tracks bond order — O2 (triplet) stretches near 1556 cm⁻¹, superoxide near 1145 cm⁻¹, and peroxide near 770–880 cm⁻¹ (values depend somewhat on the counter-cation/lattice), a spectroscopic fingerprint chemists use routinely to identify which oxygen species is present in an unknown solid. For ozone, the third governing relationship is thermodynamic rather than orbital, quantifying why ozone is a high-energy, metastable allotrope relative to O2:
where:
- is the standard Gibbs energy of formation of ozone from O2 (positive, confirming instability),
- the large positive value reflects that converting three strong O2 double bonds (bond order 2) into two ozone units (effective O–O bond order 1.5, weaker per linkage) costs more energy than the modest entropy/enthalpy gained,
- this is why ozone slowly reverts to O2 at room temperature — kinetically slow (a symmetry-forbidden, spin-allowed but activation-barrier-limited decomposition) but thermodynamically strongly favorable.
Goal. Show, without invoking any empirical magnetic data, that filling the correct MO diagram for O2 with 12 valence electrons necessarily produces two unpaired electrons — i.e., that paramagnetism is a prediction, not a patch, of MO theory.
Step 1 — Count valence electrons and identify symmetry-adapted orbitals. Each oxygen atom contributes 6 valence electrons (2s²2p⁴), so O2 has 12 valence electrons to place. The relevant atomic orbitals are 2s and 2p on each center; by symmetry (the molecular axis defines z), these combine into σ_2s, σ_2s (from 2s + 2s), σ_2p, σ_2p (from 2pz + 2pz, along the bond axis), and two degenerate pairs π_2p, π*_2p (from 2px/2px and 2py/2py, perpendicular to the axis).
Step 2 — Fix the energy ordering appropriate for O2 (not B2/C2/N2). For Z ≥ 8 (O, F, Ne), the 2s and 2p atomic orbitals are far enough apart in energy that s–p mixing is weak, so σ_2p sits below π_2p in energy — the "normal" ordering. (For earlier diatomics, strong 2s–2pz mixing inverts this, putting π_2p below σ_2p; this inversion is why N2's diagram looks different, but by O2 the crossover has already happened.) The full ordering, lowest to highest: σ_2s < σ_2s < σ_2p < π_2p (×2) < π_2p (×2) < σ*_2p.
Step 3 — Fill electrons using the Aufbau principle and Hund's rule. Placing 12 electrons: σ_2s² (2), σ_2s² (4), σ_2p² (6), π_2p⁴ (10, two electrons in each of the two degenerate π orbitals), leaving 2 electrons for the doubly-degenerate π pair (12 total). Hund's rule dictates that these last two electrons occupy the two different π* orbitals singly, with parallel spins, rather than pairing up in one — exactly as electrons fill degenerate p or d orbitals in atoms.
Step 4 — Read off bond order and spin state. Bonding electrons: σ_2s(2) + σ_2p(2) + π_2p(4) = 8. Antibonding electrons: σ_2s(2) + π_2p(2) = 4. Bond order = (8 − 4)/2 = 2. Because the two π* electrons are unpaired (parallel spins in different orbitals), the total spin quantum number S = 1, giving a triplet ground state (²S+1 = 3), with a paramagnetic moment expected near = ≈ 2.83 Bohr magnetons for n = 2 unpaired electrons — matching the experimentally measured effective moment of O2 gas closely (spin-only value ≈ 2.83 μB, observed ≈ 2.8 μB after minor orbital contribution corrections). This is a genuine quantitative prediction, not a retrofit: nothing in Steps 1–3 assumed magnetism, yet the diagram forces it.
Go deeper · Derivation B · 2 minOzone's bent geometry and the 4-electron/3-center π bond
Goal. Derive why ozone is bent (not linear) and why its two O–O bonds are equivalent with fractional bond order, using VSEPR for the σ framework and a Hückel-style treatment for the π system.
Step 1 — Assign the σ-framework geometry with VSEPR. The central oxygen in O3 has, formally, one lone pair, one lone pair used for the π system (see below), and two σ bonds to the terminal oxygens — but the simplest correct VSEPR count treats the central O as having 3 σ-type electron domains (2 bonding pairs to the terminal O's + 1 lone pair) around it, giving a trigonal-planar electron geometry and hence a bent molecular geometry, analogous to SO2 and NO2^−. Repulsion from the lone pair compresses the O–O–O angle from the idealized 120° to the observed 117°, closely tracking the pattern lone-pair repulsion > bonding-pair repulsion familiar from VSEPR (compare NH3's 107° from 109.5°, or SO2's 119° from 120°).
Step 2 — Identify the π system perpendicular to the molecular plane. Each of the three oxygens contributes one p orbital perpendicular to the O3 plane. These three atomic p orbitals combine (Hückel-style, exactly as for the allyl radical/anion in organic pi systems) into three molecular orbitals: one fully bonding (all three p orbitals in phase), one nonbonding (a node through the central atom, zero coefficient there), and one fully antibonding (alternating phases, two nodes).
Step 3 — Count π electrons and fill the MOs. Each terminal oxygen contributes one electron to this π system from a formally "extra" lone pair-derived p orbital, and the central oxygen contributes... the bookkeeping that matches experiment allocates 4 π electrons total to this 3-orbital system (this is the same electron count as the allyl anion, isoelectronic in π-framework terms). Filling from the bottom: 2 electrons in the bonding MO, 2 electrons in the nonbonding MO, 0 in the antibonding MO — analogous to how the allyl anion places 4 electrons in its 3-orbital π system.
Step 4 — Compute the resulting π bond order per O–O linkage. With the antibonding MO empty, there is no antibonding π contribution to either O–O bond; the bonding MO contributes π bonding character delocalized equally over both O–O linkages by the molecule's C2v symmetry (which relates the two terminal oxygens by a mirror plane and forces their bonds to be identical). Distributing the bonding pair (2 electrons) over 2 equivalent bonds gives a π bond order of 0.5 per O–O linkage; the nonbonding pair contributes no bond order at all (it's localized off the internuclear regions). Added to the σ bond order of 1 already present from the VSEPR σ-framework, each O–O linkage in ozone has:
This sits neatly between a pure single bond (order 1, ~148 pm as in H2O2 or F2O2) and a pure double bond (order 2, 121 pm as in O2), predicting an intermediate bond length — and the measured O–O distance in ozone, 127.8 pm, confirms exactly this intermediate character, essentially identical to superoxide's 128 pm (also bond order 1.5), a satisfying cross-check between two completely different bonding pictures converging on the same number.
Worked examples
Example 1 (foundational): predicting the magnetic behavior and bond order of the peroxide ion. Predict the bond order, O–O bond length trend, and magnetic behavior of O2^2− using the MO diagram, and compare to the known peroxide bond length in Na2O2 (149 pm) and in H2O2 (147.5 pm). Peroxide has 14 valence electrons (12 for neutral O2, plus 2 extra for the 2− charge). Filling the same MO ordering as before: σ_2s² σ_2s² σ_2p² π_2p⁴ π_2p⁴ (all four π electrons now pair up, filling both π orbitals completely) — σ_2p remains empty. Bonding electrons: 2+2+4 = 8 (unchanged). Antibonding electrons: σ_2s(2) + π*_2p(4) = 6. Bond order:
A bond order of exactly 1 predicts a single O–O bond, consistent with treating peroxide as isoelectronic (in bonding terms) with H2O2's central O–O single bond. All electrons are now paired (π_2p is completely filled, not half-filled), so peroxide is <strong>diamagnetic</strong> — a clean, testable prediction. The measured bond lengths, 149 pm in solid Na2O2 and 147.5 pm in gas-phase H2O2, both sit right at the typical O–O single-bond length (compare to 148 pm in F2O2 and other peroxo compounds), confirming the bond-order-1 prediction quantitatively. As a limiting-case check: if you (incorrectly) added electrons further, filling σ_2p, bond order would drop to (8−8)/2 = 0 — no bond at all — correctly predicting that O2^4− is not a stable diatomic species; instead the O–O bond simply breaks, giving two separate O^2− oxide ions, which is exactly what is observed in ionic oxides like MgO or Na2O (no O–O contacts at all, isolated O^2− ions in a rock-salt lattice).
Example 2 (applied): quantitative disproportionation of sodium peroxide in water. Solid sodium peroxide, Na2O2, is added to water. Predict the products, balance the reaction, and calculate the mass of H2O2 produced (before any further decomposition) when 15.6 g of Na2O2 (M = 77.98 g/mol) is dissolved completely in excess cold water. Peroxide ion in water undergoes simple hydrolysis rather than true disproportionation (disproportionation of O2^2− would require unequal splitting into different oxidation states of oxygen, but with water simply protonating the ion, both oxygens stay at the same −1 oxidation state):
Moles of Na2O2:
By the 1:1 stoichiometry above, moles of H2O2 produced equal moles of Na2O2 consumed, 0.2000 mol. Mass of H2O2 (M = 34.02 g/mol):
This reaction is exactly why Na2O2 is used industrially as a bleaching agent and oxygen source: cold water hydrolysis is controllable and gives H2O2 cleanly, while hot water or acid promotes further decomposition of the H2O2 itself (2 H2O2 → 2 H2O + O2, a true disproportionation in which peroxide oxygen, at oxidation state −1, splits into oxide oxygen at −2 in water and O2 at 0). As a limiting-case check, note that if instead of Na2O2 you used the superoxide KO2, the analogous reaction with water is a genuine disproportionation, since superoxide's oxygen sits at oxidation state −1/2 and must split unevenly:
Here O2^− (oxidation state −1/2 per O) disproportionates into OH^− (oxidation state −2) and O2 (oxidation state 0) — a genuinely different reaction type from the peroxide hydrolysis above, and the reaction actually used in some self-contained breathing apparatus (KO2 absorbs exhaled CO2 and moisture, releasing O2).
Example 3 (exam/competition level): using thermochemical cycles to rank the thermal stability of alkali-metal oxide, peroxide, and superoxide. Explain, using lattice-energy arguments (Born–Haber-style reasoning), why lithium forms predominantly the simple oxide Li2O when burned in air, sodium forms predominantly the peroxide Na2O2, and potassium, rubidium, and cesium form predominantly the superoxide MO2 — even though all four species (O^2−, O2^2−, O2^−) are accessible in principle to every alkali metal. The controlling factor is the competition between the lattice energy gained by pairing a given metal cation with a given oxygen anion, and the energy cost of forming that anion in the first place (successive electron affinities of oxygen: O + e^− → O^−, ΔH ≈ −141 kJ/mol, favorable; O^− + e^− → O^2−, ΔH ≈ +744 kJ/mol, strongly unfavorable in the gas phase, because you are forcing a second negative charge onto an already-negative ion). Forming O^2− from O2 costs a great deal of energy overall and is only recovered if the lattice energy of the resulting M2O salt is large enough to compensate — and lattice energy scales inversely with the sum of ionic radii (Born–Landé/Kapustinskii reasoning):
where are the ionic charges and the ionic radii. The small Li^+ ion (76 pm) sits close to the small, doubly-charged O^2− ion, generating a very large lattice energy (Li2O lattice enthalpy ≈ −2800 kJ/mol) — large enough to pay the steep cost of forming O^2−, so oxide is thermodynamically preferred for lithium. As the alkali metal cation grows (Na^+ 102 pm, K^+ 138 pm, Rb^+ 152 pm, Cs^+ 167 pm), the lattice energy advantage of pairing with the small, highly-charged O^2− ion shrinks (the term drops rapidly, while the more diffuse, singly-charged, larger anions O2^2− and O2^− lose relatively less lattice energy as the cation grows, because their own greater ionic radius already dominates the denominator). This is the radius-ratio/charge-density rule: large, low-charge-density cations preferentially stabilize large, low-charge-density (weakly charged, diffuse) anions, and small, high-charge-density cations preferentially stabilize small, high-charge-density (compact, highly charged) anions — a lattice-energy-driven matching principle sometimes summarized loosely as "big likes big, small likes small." Quantitatively: for cesium, the lattice energy gained by using O^2− over O2^− is no longer enough to offset the large second electron affinity cost of forming O^2− in the first place, so cesium is oxidized by O2 all the way to CsO2 (superoxide) as the thermodynamic product, and burning cesium metal in air gives the orange superoxide directly rather than the oxide. Sodium sits at an intermediate radius, favoring the peroxide Na2O2 as the dominant product of combustion in air (though NaO2 and Na2O are both known as minor or high-pressure products). This is a genuinely predictive periodic trend: as you descend Group 1, the thermodynamically preferred combustion product of the metal in air shifts systematically oxide → peroxide → superoxide, purely from ionic radius and lattice-energy bookkeeping, with no need for separate empirical rules for each element.
Limiting cases, scope & common traps
| Situation | What happens | Why |
|---|---|---|
| O2 in strong magnetic field (liquid) | Visibly attracted, clings between pole faces | Two unpaired π* electrons, S = 1, genuine paramagnetism |
| Singlet oxygen (¹Δg, excited state) | All electrons paired, diamagnetic, highly reactive toward alkenes (ene reaction, [4+2]) | π* electrons paired in one orbital rather than Hund's-rule unpaired; spin-forbidden barrier to reaction with normal (triplet) organics is lifted |
| O2^2+ (dioxygenyl, e.g. in O2^+PtF6^−) | Shortest, strongest O–O bond in the series (~112 pm, bond order 2.5) | One electron removed from π*, less antibonding population |
| Alkaline-earth peroxides (BaO2 stable, MgO2 unstable) | Peroxide stability increases down the group | Same lattice-energy/charge-density logic as alkali metals: larger cations favor larger, lower-charge-density anions |
| Ozone at high altitude vs. ground level | Stratospheric ozone is protective (absorbs UV-B/UV-C); tropospheric ozone is a pollutant | Same molecule, different context — reactivity/toxicity is about where, not what it is |
| Treating O2^− as if it always came from complete reduction of O2 | Wrong assumption in biology | O2^− (superoxide) in cells arises from one-electron reduction of O2 (e.g., leaked from the electron-transport chain), not from a controlled two-electron process |
| Assuming ozone's resonance structures mean "two single + one double bond, localized" | Wrong; bonds are equivalent, not localized | The 4e/3-center π system delocalizes charge and bond order equally over both O–O linkages, exactly like allyl anion, not like a rapidly interconverting pair of distinct isomers |
The chemist behind it
Origins · 1 minRobert S. Mulliken and how this was discovered
Robert Mulliken, working largely at the University of Chicago from the 1920s through the 1960s, was the chief architect of molecular orbital theory as a rigorous alternative to Heitler–London valence-bond pictures, and he was the one who worked out, with quantitative spectroscopic backing, that O2's paramagnetism was a direct consequence of degenerate π* orbitals half-filled according to Hund's rule — resolving a puzzle that simple bonding pictures (including early valence-bond treatments) could not explain. He received the Nobel Prize in Chemistry in 1966 "for his fundamental work concerning chemical bonds and the electronic structure of molecules by the molecular orbital method," and his systematic correlation of spectroscopic term symbols with orbital occupancy for diatomics remains the direct ancestor of every O2 MO diagram drawn in introductory chemistry today.
Why is O2's Lewis structure "wrong" if it correctly shows a double bond?
The Lewis structure :O=O: does correctly capture the overall bond order (2) and the presence of two lone pairs per oxygen, so it is not wrong about connectivity or bond strength. It is wrong specifically about electron pairing: a simple Lewis structure implicitly assumes every bonding and nonbonding pair is spin-paired, predicting a diamagnetic molecule. Molecular orbital theory shows that the two highest-energy electrons occupy a degenerate pair of π* orbitals, and Hund's rule then requires them to remain unpaired with parallel spins, producing a triplet ground state. This is why liquid O2 is measurably paramagnetic — a fact no amount of resonance structures drawn from Lewis theory can reproduce, because Lewis theory has no mechanism for representing degenerate, half-filled orbital sets.
Is ozone a stronger oxidizer than O2, and how does that connect to its bonding?
Yes — ozone's standard reduction potential (O3 + 2H^+ + 2e^− → O2 + H2O, E° = +2.07 V) is considerably higher than O2's (O2 + 4H^+ + 4e^− → 2H2O, E° = +1.23 V), making ozone one of the strongest common oxidizers. This connects directly to bonding: ozone's weaker, more delocalized 3-center π system (effective bond order 1.5 per O–O linkage versus 2 for the O2 double bond) means less energy is released simply by breaking ozone's bonds, but more importantly, ozone is thermodynamically high in energy relative to (3/2)O2 to begin with (ΔG°f ≈ +163 kJ/mol), so any reaction that lets ozone collapse toward the more stable O2 releases that "stored" instability as oxidizing power delivered to the substrate being oxidized.
Why doesn't the O2^n− bond-order series continue past peroxide to give a stable "O2^4−"?
Adding two more electrons past peroxide would require filling the σ*_2p orbital, which is the single most strongly antibonding orbital in the whole diagram (it correlates with the strongly bonding σ_2p, so its antibonding partner is proportionally strongly destabilizing). The resulting bond order, by the formula (8 bonding − 8 antibonding)/2 = 0, means there is no net bonding interaction left at all. Physically, this is exactly why ionic oxides contain isolated O^2− ions with no O–O contact whatsoever — the "extra" two electrons beyond peroxide are far more stable as two separate, fully reduced oxide ions than as any kind of diatomic unit, however weakly bonded.
How is singlet oxygen different from ordinary (triplet) O2, and why does it matter chemically?
Singlet oxygen has the same 12 valence electrons and same overall MO filling pattern as ground-state O2, but the two electrons that occupy the π set are paired together in a single π orbital (leaving the other π* orbital empty) rather than spread across both orbitals with parallel spins. This is an excited electronic state, roughly 94–157 kJ/mol above the triplet ground state, but critically it is diamagnetic (all electrons paired), which removes the spin-selection-rule barrier that normally makes triplet O2 react sluggishly with closed-shell (singlet) organic molecules. Singlet oxygen therefore reacts rapidly with alkenes and other unsaturated systems via concerted cycloadditions, which is exploited deliberately in photodynamic cancer therapy and occurs unintentionally in the photo-oxidative degradation of dyes, plastics, and biological membranes.
Is it true that ozone smells, and is that related to its bonding?
Yes; ozone has a sharp, distinctive smell (its name derives from the Greek word for "to smell"), often noticed near electrical equipment, photocopiers, or after lightning strikes, all of which generate ozone from atmospheric O2 via electrical discharge. The smell is a direct consequence of ozone's reactivity: as a strong oxidizer with a bent, polar, resonance-delocalized structure quite different from nonpolar O2, ozone readily oxidizes olfactory receptor proteins and airborne organic compounds at extremely low concentrations (detectable by human smell at parts-per-billion levels), a sensitivity that is itself indirect evidence of how much more chemically reactive ozone's weaker, higher-energy bonding makes it compared to ordinary molecular oxygen.