Bonding

Hydrogen Bonding

~5-30 kJ/mol — F/O/N donor + lone-pair acceptor; structures water, DNA, secondary protein structure

A hydrogen bond is a directional electrostatic-plus-partly-covalent attraction between a hydrogen atom covalently bonded to an electronegative donor (F, O, or N) and a lone pair on a neighboring electronegative acceptor. Energies typically span 5-30 kJ/mol per bond — far weaker than a covalent bond (~400 kJ/mol) but ten times stronger than London dispersion. The bifluoride ion FHF− is an exceptional case at ~163 kJ/mol. Linus Pauling extensively studied hydrogen bonding in The Nature of the Chemical Bond (1939). Hydrogen bonding explains water's high boiling point, ice's lower density, the alpha helix and beta sheet of proteins, and the Watson-Crick base pairing of DNA.

  • Typical strength5-30 kJ/mol
  • FHF− (extreme)163 kJ/mol
  • DonorsF, O, N (with H)
  • AcceptorsF, O, N lone pairs
  • D-H...A geometry~2.5-3.2 Å, near-linear
  • Studied byPauling 1939; Pauling-Corey 1951

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.

Why hydrogen bonding matters

  • Sets liquid water's properties. Without H-bonding, water would extrapolate to a boiling point near −90°C; the actual 100°C is a ~190 K shift. Each water molecule has two O-H donors and two lone-pair acceptors, allowing four H-bonds simultaneously and producing the tetrahedral network responsible for high heat capacity (4.18 J/g·K), high surface tension (72 mN/m), and high heat of vaporization (~40.7 kJ/mol).
  • Density anomaly. Ice (0.917 g/cm3) is less dense than liquid water (0.999 g/cm3) because tetrahedral H-bonding locks ice into an open hexagonal lattice. Liquid water reaches maximum density at 4°C (1.000 g/cm3). Lakes freeze top-down, allowing aquatic life to survive winter — a fact of life that depends entirely on hydrogen bonding.
  • Holds proteins in shape. Pauling and Corey predicted the alpha helix and beta sheet in 1951 from hydrogen-bond geometry alone. The alpha helix has 3.6 residues per turn with backbone N-H...O=C bonds bridging residues i and i+4; antiparallel beta sheets pair adjacent strands at ~7 Å spacing. Each backbone H-bond contributes ~5-10 kJ/mol of stabilization.
  • Drives DNA base pairing specificity. Watson-Crick pairing — A pairs with T via 2 H-bonds, G pairs with C via 3 H-bonds — is the molecular basis of genetic information storage and replication. The 2-vs-3 H-bond difference makes G-C 2-3 kJ/mol more stable per pair; PCR primers are designed with 40-60% GC content for balanced melting.
  • Determines solvent miscibility. Methanol, ethanol, and acetone are all infinitely miscible with water because they accept H-bonds; long-chain alcohols become immiscible above C5-C6 because the hydrocarbon tail outweighs the polar headgroup. Hexane is immiscible because it offers no H-bond donor or acceptor.
  • Encodes molecular recognition in biology. Antibody-antigen, enzyme-substrate, and drug-receptor binding all rely on patterned arrays of H-bond donors and acceptors. The active site of trypsin, the binding pocket of HIV protease, and the catalytic triad of chymotrypsin all use H-bonds to position substrates and stabilize transition states.
  • Stabilizes ionic liquids and deep eutectics. Modern ionic liquids and deep eutectic solvents (e.g., choline chloride + urea) achieve room-temperature melting via networks of weak hydrogen bonds. These solvents are increasingly used in green chemistry, electrolytes, and CO2 capture.

Common misconceptions

  • Hydrogen bonds are just strong dipole-dipole. H-bonds have a partial covalent character — modern AIM and NBO analyses show ~10-30% covalent contribution from the donor lone pair into the H sigma-star anti-bonding orbital. They are also strongly directional (preferring 180° D-H...A angles), unlike isotropic dipole-dipole.
  • Any X-H...Y interaction is a hydrogen bond. The strict IUPAC 2011 definition requires evidence of net attractive force between the H and the acceptor, with directionality and donor electronegativity. C-H...O contacts in crystals are sometimes called weak H-bonds (~4 kJ/mol), but accidentally close contacts in molecular dynamics shouldn't all be tagged as H-bonds.
  • FHF− is a typical hydrogen bond. No — at 163 kJ/mol it's a "low-barrier" or "symmetric" hydrogen bond where the H sits in a single potential well between the two F atoms. Most H-bonds in aqueous solution are 10-30 kJ/mol, and the H is much closer to one heavy atom than the other.
  • Stronger H-bonds always mean higher boiling point. HF (μ = 1.82 D, very strong individual H-bonds) boils at +20°C; H2O (slightly weaker per bond) boils at 100°C. Water wins because it forms four H-bonds per molecule (2 donors + 2 acceptors) versus HF's two-per-molecule chains.
  • H-bonds are too weak to matter for biology. Each is ~10 kJ/mol, but a typical protein has 200-500 backbone H-bonds. Total stabilization is hundreds of kJ/mol, comparable to a few kT per residue — exactly the right energy scale to fold reversibly.
  • The H-bond breaks at room temperature because kT ≈ 2.5 kJ/mol. Average breakage time of an O-H...O bond in liquid water at 25°C is ~1-3 ps; at any instant ~85% of bonds are intact. The collective network has lifetime far greater than any single bond.

How hydrogen bonds form and behave

Start with a polar covalent bond X-H, where X is F, O, or N. The high electronegativity of X (4.0 for F, 3.4 for O, 3.0 for N on the Pauling scale) shifts electron density toward X and leaves H with a substantial partial positive charge (δ+ ≈ 0.3-0.4e). Because H has no inner-shell electrons to provide repulsion, this exposed positive nucleus can approach a second electronegative atom Y (also F, O, or N) much more closely than a normal van der Waals contact. The lone pair on Y donates electron density toward H, producing a partly electrostatic, partly covalent attraction. Modern Bader (AIM) and NBO analyses show ~70-90% electrostatic + 10-30% covalent character; the covalent component grows as the bond shortens below ~2.7 Å.

Geometric signatures: D-H...A distances of 2.5-3.2 Å (heavy-atom to heavy-atom), with the D-H bond elongated by 1-3 pm versus the gas-phase value, and a strong preference for D-H...A angles within 30° of linear. Spectroscopic signatures: the X-H stretching mode in IR drops by 100-1000 cm−1 and broadens dramatically (the "nu(O-H) bridge band" in alcohols at 3200-3550 cm−1). NMR signatures: H-bonded protons appear far downfield, often at 9-15 ppm for amides and 12-16 ppm for carboxylic acid dimers.

Cooperativity is critical. In water and ice, H-bonds reinforce each other: the H-bond from molecule A to molecule B polarizes B further, strengthening B's H-bond to molecule C. This explains why the per-bond enthalpy in liquid water (~21 kJ/mol) exceeds that of an isolated water dimer in vacuum (~13 kJ/mol). DFT calculations on (H2O)n clusters reproduce the cooperative reinforcement: the 5th and 6th molecules added to a chain bind ~2-3 kJ/mol more strongly than the 2nd. The same cooperativity stabilizes alpha-helices through transmitted dipoles along the helix axis.

Comparison: hydrogen bond strengths

Bond typeStrength (kJ/mol)ExampleNotes
F-H...F (low barrier)163[HF2]−Symmetric, partly covalent
O-H...O (charged)40-60H5O2+ (Zundel cation)Proton transfer regime
N-H...N (charged)30-50NH4+...NH3Salt-bridge analogues
O-H...O (neutral, water)21Liquid H2OCooperatively reinforced
N-H...O (amide-amide)5-10Protein backbone~96 per 100-residue helix
O-H...N8-15Methanol-pyridineO is better donor than N
N-H...N5-15DNA base pair (G-C inner)Specificity in genetic code
C-H...O (weak)2-5CHCl3...acetoneRecognized as H-bond since 1990s
O-H...π (aromatic)4-8Water...benzene faceLone-pair-aromatic ring contact
S-H...S3-8H2S liquidWhy H2S boils at −60°C

Applications and examples

  • Watson-Crick base pairing. A=T (2 H-bonds at ~10 Å centroid spacing) and G≡C (3 H-bonds) produce nearly isosteric pairs that fit a uniform double helix with ~10.5 bp/turn. The 2-vs-3 H-bond difference is the molecular origin of GC content's effect on DNA melting temperature: Tm ≈ 81.5°C + 16.6·log[Na+] + 0.41·%GC − 600/N for short oligonucleotides.
  • Alpha helix and beta sheet. The 1951 Pauling-Corey predictions identified H-bonded backbone structures with rise per residue 1.5 Å (helix) and ~7 Å strand spacing (sheet). Confirmed by Kendrew's 1958 myoglobin structure and Perutz's 1959 hemoglobin (joint Nobel 1962). Most globular proteins are 30-50% helix + 10-30% sheet.
  • Solvent miscibility tables. Water (donor + acceptor), methanol, ethanol, acetone, DMSO (acceptor only) are all water-miscible. Hexane, benzene, CCl4 (no donor or acceptor) are immiscible. A drug's logP partition coefficient — the standard pharmacological measure of lipophilicity — depends primarily on its H-bond donor/acceptor count.
  • Boiling-point ladder. Compare hydrides at MW ≈ 18: H2O (100°C, H-bonded), CH4 (−161°C, no H-bond) — a difference of 261 K from a single donor/acceptor pair. The same effect explains why ethanol (78°C) boils ~50 K higher than ethane (−89°C) at near-equal MW.
  • Cellulose and silk strength. Cellulose I has chains aligned by inter-strand O-H...O bonds at ~2.7 Å; spider silk's beta-sheet crystallites depend on N-H...O backbone bonds. Both materials achieve tensile strength of GPa (greater than mild steel by mass) entirely from the cooperative H-bond network.

Frequently asked questions

What atoms can act as hydrogen-bond donors and acceptors?

Conventional hydrogen bonding requires that hydrogen be covalently attached to one of the three most electronegative second-row elements: fluorine, oxygen, or nitrogen. The F-H, O-H, or N-H bond polarizes hydrogen sufficiently positive (δ+ ~0.3-0.4e) to attract a lone pair on a neighboring F, O, or N (the acceptor). Sulfur, chlorine, and pi systems can act as weak acceptors, and C-H...O bonds with electron-poor carbons (e.g., CHCl3, alkynes) are recognized as 'weak hydrogen bonds' at ~4-5 kJ/mol. The F-H...F bond is by far the strongest at 163 kJ/mol in [HF2]-, exceeding even some single covalent bonds because the proton sits at the geometric midpoint.

Why is water's boiling point so anomalously high?

Compare hydrides going down Group 16: H2O boils at 100 degrees C, H2S at −60, H2Se at −41, H2Te at −2. Without hydrogen bonding the boiling point of water would extrapolate to about −90 degrees C — a 190-degree difference. Water has two O-H donors and two lone pairs (acceptors) per molecule, so it can form four hydrogen bonds simultaneously, giving rise to a tetrahedral local network. The cohesion energy of liquid water is about 44 kJ/mol of which roughly 40 kJ/mol comes from the broken hydrogen-bond network on vaporization. Without H-bonding, life at room temperature would be impossible; oceans would be steam.

Why is ice less dense than liquid water?

In ice (Ih structure), each water molecule donates two and accepts two hydrogen bonds, locking it into a tetrahedral geometry with O-O distances of 2.76 Angstroms. The result is an open hexagonal lattice with substantial empty space. When ice melts at 0 C, about 10-15% of hydrogen bonds break and fragments collapse into the cavities, increasing density by 8.3% — from 0.917 g/cm3 to 0.999 g/cm3. Density continues rising until water reaches its 4 C maximum at 1.000 g/cm3, then falls slightly with thermal expansion. This anomaly means lakes freeze top-down rather than bottom-up, allowing aquatic ecosystems to survive winter.

How does hydrogen bonding stabilize protein secondary structure?

The protein backbone has one N-H donor and one C=O acceptor per residue. Pauling and Corey predicted in 1951 that intramolecular hydrogen bonds between these groups would stabilize two regular structures: the alpha helix (H-bonds from residue i to residue i+4, ~3.6 residues per turn, ~1.5 Angstrom rise) and the beta sheet (H-bonds between adjacent strands, parallel or antiparallel). Each backbone hydrogen bond contributes ~5-10 kJ/mol; a 100-residue alpha helix contains ~96 such bonds for ~600-900 kJ/mol of stabilization. X-ray crystallography of myoglobin (Kendrew, 1958) and hemoglobin (Perutz, 1959) confirmed the prediction. Both shared the 1962 Nobel.

How does hydrogen bonding implement DNA base pairing?

Watson and Crick's 1953 double-helix model relies entirely on specific hydrogen-bond patterns. Adenine pairs with thymine via two H-bonds (N6-H...O4 and N1...H-N3); guanine pairs with cytosine via three H-bonds (O6...H-N4, N1-H...N3, N2-H...O2). The geometry of A-T and G-C is nearly identical (10.85 Angstrom across the helix), allowing a uniform double-helix with ~10 base pairs per turn. The two-versus-three H-bond difference makes G-C 2-3 kJ/mol more stable per pair, which is why GC-rich DNA has higher melting temperature and PCR primers are usually designed with target GC content of 40-60% to balance stability and specificity.

What is a low-barrier hydrogen bond?

When the donor and acceptor are nearly identical in proton affinity and the donor-acceptor distance is below 2.5 Angstroms, the proton sits in a single broad potential well rather than two distinct wells separated by a barrier. The result is a bond strength of 60-100 kJ/mol — three to ten times a normal hydrogen bond — and a proton position not biased toward either heavy atom. The bifluoride ion FHF- is the textbook example at 163 kJ/mol with the H exactly centered between the two F atoms (F-F = 2.27 Angstroms). Low-barrier hydrogen bonds have been proposed as catalytic intermediates in enzymes such as ketosteroid isomerase and serine proteases, though the size of the catalytic boost is debated.