Atmospheric Chemistry

Why CO₂ and Methane Trap Heat: The Molecular Physics of the Greenhouse Effect

Nitrogen and oxygen make up 99% of the air, yet they are effectively invisible to the infrared glow rising off the planet's surface. The heat-trapping is done almost entirely by trace gases — CO₂ at ≈ 424 ppm, CH₄ at ≈ 1.92 ppm, water vapor, and N₂O — that together add roughly +3.0 W/m² of extra energy to every square meter of Earth's surface compared with 1750. That single number, an imbalance smaller than the heat of a Christmas-tree bulb, is why the last decade ran about 1.3 °C warmer than the preindustrial average.

The reason a molecule 1,800 times rarer than N₂ can dominate the energy budget comes down to one quantum-mechanical rule: a molecule can absorb infrared light only if the vibration it excites changes the molecule's electric dipole. O₂ and N₂ can't; CO₂, CH₄, and H₂O can — and they do so precisely in the 4–20 μm band where a 288 K planet radiates.

  • Absorption ruleVibration must change dipole moment
  • Key bandsCO₂ 667 & 2349 cm⁻¹; CH₄ 1306 & 3019 cm⁻¹
  • Atmospheric window8–13 μm (mostly transparent)
  • ConcentrationsCO₂ ≈ 424 ppm; CH₄ ≈ 1.92 ppm; N₂O ≈ 336 ppb
  • Total forcing≈ +3.0 W/m² vs 1750
  • CH₄ GWP≈ 28 (100 yr), ≈ 80 (20 yr)

Interactive visualization

Press play, or step through manually. The visualization is yours to drive — try it before reading on.

Open visualization fullscreen ↗

Watch the 60-second explainer

A condensed visual walkthrough — narrated, captioned, under a minute.

The one rule that decides everything: a changing dipole

A molecule absorbs an infrared photon only when the photon's energy matches a vibrational transition and that vibration changes the molecule's electric dipole moment. Light is an oscillating electric field; it can push and pull on a vibration only if that vibration has a fluctuating charge separation to grab onto. This is the gross selection rule for IR spectroscopy, and it is the entire reason the atmosphere behaves as it does.

Homonuclear diatomics fail the rule completely. N₂ (N≡N) and O₂ (O=O) have two identical atoms; stretching the bond keeps the charge distribution perfectly symmetric, so the dipole stays fixed at zero. No dipole change means no IR absorption at any intensity — which is why 99% of the atmosphere is essentially a transparent window to outgoing heat.

  • CO₂ (O=C=O) is linear and symmetric, so its symmetric stretch (both O's moving out together, 1388 cm⁻¹) is also IR-inactive. But two of its four vibrational modes do change the dipole: the bending mode at 667 cm⁻¹ (≈ 15 μm) and the asymmetric stretch at 2349 cm⁻¹ (≈ 4.3 μm).
  • CH₄ and H₂O are non-symmetric enough that several modes are active; water even has a permanent dipole to begin with.

So the heat-trapping club has a strict membership rule, and only polyatomic molecules with the right geometry get in.

The 15-micron trap: how CO₂ intercepts Earth's glow

Earth's surface, at about 288 K, radiates like a blackbody with a broad peak near 10 μm in the thermal infrared — this is the outgoing longwave radiation (OLR) that ultimately balances incoming sunlight. Crucially, CO₂'s bending mode sits right on the long-wavelength shoulder of that peak, at 15 μm (667 cm⁻¹), exactly where a warm planet emits copiously.

The sequence is a genuine energy-relay, not a mirror:

  • Absorption: a surface-emitted 15 μm photon excites the CO₂ bending vibration, promoting the molecule to a higher vibrational state.
  • Thermalization: at surface pressures a CO₂ molecule collides with N₂/O₂ roughly 10⁹ times per second — far faster than the ~1 s radiative lifetime — so the vibrational energy is handed off as translational heat before it can be re-emitted. The absorbed photon warms the surrounding air.
  • Re-emission: higher in the atmosphere, collisions become rare and excited CO₂ radiates 15 μm photons in all directions — half of them downward, back toward the surface.

The net effect is that adding CO₂ raises the altitude (the effective emission level) from which the planet's 15 μm radiation finally escapes to space. Because temperature falls with height in the troposphere (lapse rate ≈ 6.5 °C/km), radiating from a higher, colder layer means the planet emits less energy at those wavelengths than it receives — and it must warm the whole column until balance is restored. That altitude shift, not a literal reflecting blanket, is the physical core of the greenhouse effect.

Methane, water, and the atmospheric window

Between roughly 8 and 13 μm lies the atmospheric window — a band where none of the major gases absorb strongly and Earth radiates almost directly to space. This window is the planet's main cooling vent, and its importance is why the gases that do nibble at its edges punch above their weight.

  • Methane (CH₄) is a tetrahedral molecule with IR-active bending (deformation) modes near 1306 cm⁻¹ (7.7 μm) and C–H stretches near 3019 cm⁻¹ (3.3 μm). At only 1.92 ppm it is over 200 times rarer than CO₂, yet its bands fall on the edge of the window and are far from saturated, so each additional CH₄ molecule absorbs much more efficiently than each additional CO₂ molecule.
  • Water vapor (H₂O) is a bent molecule with a large permanent dipole and strong bending (1595 cm⁻¹) and stretching (3657, 3756 cm⁻¹) bands, plus a dense far-IR rotational forest. It supplies roughly half of the natural greenhouse effect — but because its concentration is set by temperature (Clausius–Clapeyron: vapor rises ~7% per °C), it acts as a powerful feedback that amplifies CO₂-driven warming, not as an independent driver.
  • N₂O and CFCs/HFCs absorb inside the window itself, which is why molecule-for-molecule they are potent: N₂O's GWP ≈ 273, and some fluorocarbons exceed 10,000.

The takeaway: what matters is not just whether a gas absorbs IR, but where in the spectrum, and whether that band is already saturated by more abundant gases.

Radiative forcing: turning spectroscopy into watts

Climate scientists collapse all of this molecular physics into one bookkeeping number: radiative forcing (RF), the change in Earth's net energy balance at the top of the atmosphere, in W/m², caused by a change in a gas. The reference is the year 1750.

  • CO₂: RF ≈ 5.35 × ln(C/C₀) W/m². With C₀ = 278 ppm and C ≈ 424 ppm, that gives RF ≈ 5.35 × ln(1.53) ≈ +2.3 W/m² — the single largest term.
  • CH₄:+0.5 W/m², more than expected from its tiny concentration precisely because its bands are unsaturated.
  • N₂O:+0.2 W/m²; halocarbons add another ≈ +0.4 W/m².

Note the logarithmic dependence for CO₂: because the center of the 15 μm band is already saturated (fully opaque), extra CO₂ only widens the absorbing band's wings. Each doubling of CO₂ adds a roughly constant ≈ 3.7–4.0 W/m² of forcing — the basis of the ~3 °C equilibrium climate sensitivity per doubling. The total anthropogenic forcing today is about +2.7 to +3.0 W/m²; that persistent imbalance is currently loading the ocean, ice, and atmosphere with roughly 0.9 W/m² of unrealized heat.

GWP: why one methane molecule ≠ one CO₂ molecule

Comparing gases fairly requires accounting for both how strongly a molecule absorbs and how long it survives. That is the job of Global Warming Potential (GWP): the time-integrated radiative forcing of 1 kg of a gas relative to 1 kg of CO₂, over a chosen horizon.

  • Methane's lifetime is short — about 12 years — because it is destroyed in the troposphere by the hydroxyl radical (HO•): CH₄ + HO• → CH₃• + H₂O, the first step of an oxidation chain that ends at CO₂ and H₂O. This short life is why its GWP depends so heavily on the time window: GWP₂₀ ≈ 80 but GWP₁₀₀ ≈ 28.
  • CO₂ has no single lifetime. A pulse of CO₂ is drawn down over decades to centuries by the ocean and land, but a substantial 15–40% fraction persists for over 1,000 years until slow silicate weathering and carbonate burial finish the job. This longevity is why CO₂, despite a modest per-molecule absorption, is the dominant long-term forcer.

Practically, this means cutting methane buys fast relief — because HO• clears it within a couple of decades — while CO₂ cuts are what actually stabilize the climate over centuries. Both are needed, on different clocks.

The carbon that stays: where CO₂ comes from and goes

The forcing numbers ultimately trace back to a mass balance. Humans emit roughly 37 gigatonnes of CO₂ per year from fossil-fuel combustion — the reverse of photosynthesis, e.g. CH₄ + 2 O₂ → CO₂ + 2 H₂O for natural gas or C + O₂ → CO₂ for coal — plus ~4 Gt from land-use change.

  • About half stays airborne, which is why the Mauna Loa curve climbs ~2.5 ppm/yr.
  • The ocean absorbs ~25%, dissolving CO₂ into carbonic acid — CO₂ + H₂O ⇌ H₂CO₃ ⇌ H⁺ + HCO₃⁻ — which has lowered surface-ocean pH by ≈ 0.1 unit (from ~8.2 to ~8.1) since 1750, a ~30% rise in [H⁺].
  • The land biosphere takes up ~25% through faster plant growth and soil storage.

The interventions all target this budget. Carbon capture scrubs CO₂ from flue gas or air (e.g. amine sorbents: CO₂ + 2 R₂NH ⇌ R₂NCOO⁻ + R₂NH₂⁺). Enhanced weathering speeds the natural silicate reaction CaSiO₃ + 2 CO₂ + H₂O → Ca²⁺ + 2 HCO₃⁻ + SiO₂ that has regulated Earth's thermostat for eons. Cutting methane leaks and combusting them (CH₄ → CO₂, a 28-fold GWP reduction per molecule) is among the cheapest fast wins. But the arithmetic is unforgiving: because a large fraction of emitted CO₂ lingers for millennia, only net-zero emissions stops the warming — the atmosphere keeps the tab open until we stop adding to it.

Why some air molecules trap heat and others don't — the four molecules that matter, contrasted with the transparent bulk gases.
MoleculeIR-active?ConcentrationContribution / GWP-100
N₂ (78%)No — symmetric, no dipole change780,000 ppm0 (transparent)
O₂ (21%)No — symmetric, no dipole change209,000 ppm0 (transparent)
H₂O (vapor)Yes — bend + stretches, permanent dipole0–40,000 ppm (variable)~50% of natural effect; feedback, not forcing
CO₂Yes — bend (667) + asymmetric stretch (2349)≈ 424 ppm~+2.1 W/m²; GWP = 1 (reference)
CH₄Yes — deformation (1306) + stretch (3019)≈ 1.92 ppm~+0.5 W/m²; GWP ≈ 28
N₂OYes — asymmetric, permanent dipole≈ 336 ppb~+0.2 W/m²; GWP ≈ 273

Frequently asked questions

If CO₂ is only 0.04% of the air, how can it possibly matter?

Because the 99% that is N₂ and O₂ is completely transparent to infrared — those symmetric molecules can't change their dipole moment when they vibrate, so they absorb no outgoing heat. CO₂'s bending mode at 15 μm sits right where a 288 K planet radiates, so even 424 ppm is enough to raise the altitude from which heat escapes. Abundance is irrelevant if the abundant gases do nothing; a trace gas at the right wavelength controls the budget.

Isn't the CO₂ band already saturated, so more CO₂ can't do anything?

The very center of the 15 μm band is indeed saturated — it's fully opaque. But the band has wings, and adding CO₂ widens the wavelength range that becomes opaque and pushes the effective emission level higher and colder. That's why forcing scales as the logarithm of concentration, RF ≈ 5.35 × ln(C/C₀): each doubling adds a roughly constant ~3.7 W/m². Saturation slows the growth per molecule but does not stop it.

Why is methane called ~80 times worse than CO₂ but also only ~28 times worse?

Both numbers are correct — they're just measured over different horizons. Methane absorbs strongly in an unsaturated band, so per molecule it's very potent, but it's destroyed by hydroxyl radicals in about 12 years. Over 20 years its GWP is ~80; averaged over 100 years, its decay drags the figure down to ~28. CO₂, which lasts centuries to millennia, wins the long game.

Is water vapor the real greenhouse gas we should worry about?

Water vapor supplies roughly half the natural greenhouse effect, but its concentration is set by temperature, not by human emissions — warmer air holds ~7% more vapor per °C (Clausius–Clapeyron). So it acts as a feedback that amplifies CO₂-driven warming, roughly doubling it, rather than as an independent driver. We can't control water vapor directly; controlling CO₂ is what moves the whole system.

Does the greenhouse effect actually 'reflect' heat back down?

No — it's absorption and re-emission, not reflection. A greenhouse gas absorbs an upward infrared photon, thermalizes the energy through collisions, and higher in the atmosphere emits new photons in all directions, including downward. The net result is that heat escapes to space from a higher, colder altitude, forcing the surface to warm. Nothing bounces off a mirror.

Is the warming reversible if we stop emitting?

Partly, and slowly. Methane clears within a couple of decades once emissions stop, so cutting it delivers fast cooling. CO₂ is different: 15–40% of a pulse persists for over 1,000 years until silicate weathering and carbonate burial remove it. Temperatures roughly stabilize at net-zero rather than falling, so the practical answer is that we can stop the warming but cannot quickly undo it.