Hydrogen Bond¶

An evidence-supported attractive X–H···Y interaction in which a polarized, covalently bound hydrogen couples a donor fragment to an acceptor site with characteristic directionality and electronic response.

Version: v1 · 2026-08-30 · History
Domain-specific #: 2027
Origin domain: chemistry
Subdomain: intermolecular and intramolecular interactions
Aliases: H-bond

Core Idea¶

A hydrogen bond is an attractive interaction involving a hydrogen atom in a molecular fragment X–H, where X is more electronegative than hydrogen, and an atom or group Y in the same or a different molecule, with evidence that a bond-like interaction has formed. It is conventionally written X–H···Y. X–H is the hydrogen-bond donor; Y is the acceptor. This follows the inclusive, evidence-demanding IUPAC recommendation developed to replace definitions restricted to a short list of elements or a single geometric cutoff.^[1]^[2]

The structural signature is polar covalent donor X–H + accessible acceptor Y + favorable relative geometry and environment → attractive, directional coupling supported by experimental or theoretical evidence → altered energy, geometry, spectroscopy, electron distribution, conformation, association, or assembly. The hydrogen remains covalently attached to X in the ordinary limiting description while interacting with Y; in very strong bonds the proton can become extensively shared, and proton transfer forms a neighboring limit rather than a different universe.

Hydrogen bonding is not explained adequately as a rigid “N/O/F rule,” a purely electrostatic attraction, or a fixed bond-energy interval. Electrostatics often contributes strongly, but polarization, charge transfer, exchange repulsion, and dispersion contribute in proportions that vary with donor, acceptor, geometry, and environment. Directionality is characteristic, not perfect linearity in every constrained structure. IUPAC therefore couples a broad definition to multiple criteria: no single short contact, red shift, or computed orbital interaction universally proves every case.^[2]

The candidate is accepted as a domain-specific abstraction at 0.99. It is a reusable explanatory and diagnostic object across gas-phase complexes, liquids, crystals, intramolecular conformations, proteins, nucleic acids, polymers, and designed supramolecular assemblies. No live or accepted-workspace node owns the donor–hydrogen–acceptor role system and evidence discipline.

Structural Signature¶

Seven roles make the abstraction operational:

donor fragment X–H: hydrogen is covalently bound to an atom or group X that polarizes the bond; classical donors include O–H and N–H, while weaker C–H and other donors can qualify when evidence supports the interaction;
acceptor Y: an atom, ion, bond, or group with electron density able to interact attractively with the donor hydrogen; it need not always be a textbook lone pair on O, N, or F;
interaction geometry: the X–H···Y arrangement is usually directional and often tends toward linearity as the interaction strengthens, but molecular constraints and competing interactions can bend it;
environment: solvent, phase, competing donors/acceptors, protonation, dielectric response, temperature, and confinement can change strength, population, and even which site acts as donor or acceptor;
mixed physical contributions: electrostatics and polarization are accompanied by varying charge-transfer/orbital, dispersion, and exchange components; the interaction should not be classified from one component alone;
evidence portfolio: geometry, favorable formation thermodynamics, vibrational or NMR changes, isotope effects, diffraction, electron-density topology, and credible computation can converge;
consequence: association, conformer stabilization, proton positioning, molecular recognition, crystal packing, network formation, or altered material properties.

The invariant is hydrogen-mediated directional attraction with evidence of bond formation. A donor and acceptor label without an observed or supported interaction is only a capability statement. A close H···Y distance in a crowded structure is suggestive but not decisive. Conversely, a genuine bond need not meet one universal distance, energy, or frequency-shift threshold; strong, moderate, weak, red-shifting, and some blue-shifting cases occupy a continuum.^[1]

Hydrogen bonds may be intermolecular, joining distinct molecules, or intramolecular, joining remote parts of one molecule. They may be isolated, bifurcated, cooperative, water-mediated, or organized into chains, rings, sheets, and three-dimensional networks. Those topologies change collective behavior while preserving the local donor–H···acceptor relation.

What It Is Not¶

Not the covalent X–H bond. The solid line in X–H···Y identifies the covalent donor bond; the dotted relation identifies the hydrogen bond to Y.
Not every dipole–dipole attraction. Hydrogen bonding requires the specific X–H···Y role structure and evidence of bond formation, not merely two polar molecules attracting.
Not a purely dispersive van der Waals contact. Dispersion can contribute, but an interaction dominated only by nondirectional dispersion does not satisfy the modern criteria.
Not merely “H close to O or N.” Distance and angle are evidence, not a complete definition. Packing, steric constraint, thermal motion, and refinement uncertainty can generate close contacts without a defensible bond claim.
Not restricted absolutely to O–H, N–H, F–H donors and O/N/F acceptors. Those are common strong textbook cases. Weaker C–H donors, sulfur acceptors, π systems, and other groups can qualify when the evidence portfolio supports them.^[3]
Not a universal energy band. Hydrogen bonds span a broad continuum and interact with solvent and geometry. IUPAC deliberately avoids a single energy cutoff.
Not proton transfer. Proton transfer changes which heavy atom owns the covalent H bond. Strong hydrogen bonding can approach a shared-proton limit and facilitate transfer, but the processes and states should still be distinguished.
Not a salt bridge or ionic bond. Ionic charge can strengthen or participate in hydrogen bonding, yet the ionic relation and the directional X–H···Y relation are analytically separable.
Not hydrophobic association or base stacking. These can cooperate with hydrogen bonds in proteins, membranes, and nucleic acids but have different driving forces and geometries.
Not adsorption. Adsorption is net accumulation at an interface. Hydrogen bonding can contribute to adsorption, but it can occur in bulk solution or within one molecule, and adsorption can arise without it.

Scope of Application¶

Small-molecule physical chemistry. Gas-phase dimers and solution complexes permit controlled measurement of geometry, association thermodynamics, vibrational shifts, and isotope effects. These cases test definitions without the interpretive density of a macromolecule.

Liquids and solvation. Water and alcohols form dynamic networks in which donor and acceptor partners continually exchange. Hydrogen-bond topology and lifetime affect local structure and transport, but a fluctuating network should not be mistaken for permanent molecule pairs.

Intramolecular conformation and reactivity. A donor and acceptor within one molecule can favor a ring-like conformation, change tautomer populations, mask polarity, organize a transition geometry, or prepare a proton-transfer pathway. Competing solvation may weaken the internal bond.

Structural biology. Repeated backbone N–H···O=C interactions organize α helices and β sheets; the Molecular Biology of the Cell account explicitly links local peptide-bond hydrogen bonding to these secondary structures.^[4] Side-chain and water-mediated bonds also participate in ligand recognition and conformational specificity.

Nucleic acids. Complementary donor/acceptor patterns allow efficient A–T and G–C base pairing, helping specify strand complementarity.^[5] Base stacking, hydration, ions, and backbone forces also matter; hydrogen bonds must not be credited as the sole source of duplex stability.

Crystal engineering and supramolecular chemistry. Directional donor/acceptor patterns act as supramolecular synthons for reasoning about recurring packing motifs and target assemblies. Desiraju's crystal-engineering program treats noncovalent interactions as design-bearing links while recognizing competition and polymorphism.^[6]

Materials and polymers. Reversible networks influence cellulose, polyamides, hydrogels, adhesives, self-healing materials, and responsive assemblies. Multiple individually reversible bonds can generate cooperative strength while retaining rearrangeability.

Clarity¶

A disciplined hydrogen-bond claim answers four questions. Who is the donor? Name the covalent fragment X–H. Who is the acceptor? Name Y and the relevant electron-rich site. What is the geometry and environment? Report distance and angle with phase, protonation, temperature, solvent, or structural constraints. What evidence shows bond formation? Give at least one credible experimental or theoretical consequence and, for marginal cases, seek convergent evidence.

The evidence should be interpreted as a portfolio. A shorter-than-expected donor–acceptor separation and an angle tending toward linearity are useful structural signs. Changes in the X–H stretching band, NMR shielding, isotope response, association free energy, or electron density can strengthen the case. But red shifting has blue-shifting exceptions, crystallographic hydrogen positions may be uncertain, and a bond critical point alone is not an all-purpose yes/no oracle. The IUPAC framework intentionally makes claims evidence-sensitive rather than cutoff-sensitive.^[2]

This also clarifies causality. Observing a hydrogen bond does not prove it is the dominant stabilizer of an entire protein, crystal, or DNA duplex. Compare the state with and without the interaction while accounting for replacement bonds to solvent, conformational entropy, stacking, electrostatics, and cooperative networks. A local directional interaction can be decisive for specificity or geometry even when its isolated energetic contribution is modest.

Manages Complexity¶

Hydrogen bond compresses diverse molecular observations into a small role map. Instead of treating each anomalous boiling point, conformational preference, spectral shift, crystal contact, base pair, or protein motif as unrelated, the analyst checks for a donor, hydrogen, acceptor, geometry, environment, and evidence. The map is rich enough to generate tests but limited enough to remain reusable.

It also decomposes network complexity. A liquid or macromolecule may contain thousands of transient contacts, yet each local edge can be labeled by donor, acceptor, strength proxy, lifetime, and topology. Graphs of those edges expose chains, rings, cooperative clusters, solvent bridges, and unsatisfied sites. The same local abstraction thus scales from a dimer to a network without treating the network as one giant bond.

For design, donors and acceptors become manipulable features. Chemists can block a donor, protonate an acceptor, change solvent, preorganize geometry, or add competing sites, then predict which contacts weaken, which conformers shift, and which assemblies may change. The abstraction does not make crystal or protein design deterministic—competition and entropy remain—but it narrows the variables worth testing.

Finally, the evidence portfolio prevents false precision. A single arbitrary distance cutoff would classify every structure quickly but unreliably. Requiring convergent geometric, energetic, spectroscopic, and electronic consequences directs attention to borderline cases and explains why different measurement regimes may legitimately resolve an interaction differently.

Abstract Reasoning¶

Donor–acceptor inference. If protonation removes the acceptor's available electron density, an X–H···Y interaction should weaken or disappear unless a different site assumes the acceptor role. Resulting conformer, spectrum, or association changes test the assignment.

Geometry inference. Holding chemistry comparable, a donor–H···acceptor arrangement closer to its preferred direction generally supports stronger coupling than a severely bent one. Steric or ring constraints can override the tendency, so angle is evidence rather than a universal law.

Spectroscopic inference. Formation often lengthens and weakens X–H and shifts its vibrational signature while increasing intensity; NMR may show hydrogen deshielding. Exceptions such as blue-shifting bonds mean the inference must be combined with other evidence.^[1]

Competition inference. Adding a solvent that strongly solvates the donor or acceptor can disrupt an intramolecular bond or replace an intermolecular partner. The observed population reflects competition among all accessible partners, not intrinsic donor/acceptor labels alone.

Cooperativity inference. In a chain or ring, formation of one hydrogen bond can polarize a donor/acceptor and alter neighboring bond strengths. Network behavior is therefore not always the sum of independent identical edges.

Mutation inference. Replacing a protein side chain that donates or accepts a proposed bond may change affinity or conformation. A change supports involvement only after steric, charge, solvation, and global-fold effects are controlled; mutation is not a clean deletion of one abstract edge by default.

Boundary inference. A short contact with no favorable orientation, no association consequence, and no spectroscopic or electronic support should remain a contact or hypothesis, not be promoted automatically to hydrogen bond.

Knowledge Transfer¶

The abstraction transfers literally across chemistry, biochemistry, and materials science because the same molecular roles recur. A water dimer, enzyme active site, DNA base pair, crystal synthon, and polymer network each contain an X–H donor, an acceptor, a directional interaction, environment-dependent strength, and evidence of structural or energetic consequence. Experimental techniques differ, but the recognition test remains molecularly identical.

Transfer across phase and scale requires calibration. Gas-phase association energies do not transplant numerically into water, where solvent competes for both partners. A contact found in a static crystal may be transient in solution. A repeated network can exhibit cooperativity absent from an isolated dimer. The abstraction transfers; its parameter values and causal weight do not transfer without environment.

Prime promotion fails. Directed coupling, complementarity, network formation, and evidence convergence recur outside chemistry, but hydrogen, electronegativity, covalent X–H, electron density, vibrational shifts, proton transfer, and donor/acceptor chemistry do not. Their portable residue belongs to coupling, compatibility, and related primes. “Hydrogen bond” used for a social or organizational tie is metaphor, not a literal cross-substrate instance.

Examples¶

Water dimer and liquid network. One water molecule donates an O–H hydrogen toward another molecule's oxygen acceptor. In liquid water, each molecule can donate and accept through changing neighbors, producing a dynamic network rather than permanent dimers. The local X–H···Y identity survives continual partner exchange.

Intramolecular bond. In an enol capable of bringing O–H and a carbonyl oxygen into a favorable ring-like geometry, an intramolecular O–H···O bond can stabilize one conformation or tautomer. Polar solvent may compete for both sites and shift the equilibrium, illustrating environmental dependence.

Protein α helix. A backbone carbonyl oxygen accepts from a later backbone N–H in a repeated pattern; the common α-helical arrangement uses the carbonyl of residue i and N–H of residue i+4.^[7] Repetition organizes local secondary structure, but hydrophobic packing and other interactions govern the whole fold.

β sheet. Hydrogen bonds connect peptide groups on neighboring strands, allowing parallel or antiparallel sheet organization.^[4] The donor/acceptor role is local while the sheet is the emergent network topology.

DNA base pairing. A–T and G–C expose complementary donor/acceptor patterns that form efficient cross-strand hydrogen bonds, contributing to pairing specificity and complementarity.^[5] Stacking and the solvent/ion environment remain essential to duplex thermodynamics.

Carboxylic-acid dimer in a crystal. Two acid groups can form a cyclic pair of reciprocal O–H···O bonds. Crystal engineers treat such recurring patterns as synthons, while checking whether other donors, acceptors, solvent, and packing forces redirect assembly.^[6]

Strong-bond boundary. In a very strong, nearly symmetric donor–H–acceptor arrangement, the proton can become extensively shared. The case still belongs to the hydrogen-bond continuum but warns against picturing every bond as a weak, localized electrostatic contact.

Negative case. A crystallographic model places H near O because crowded geometry leaves little space, but the angle is unfavorable and no energetic, spectroscopic, or electronic evidence supports association. Report a close contact or proposed bond, not a confirmed hydrogen bond.

Structural Tensions¶

Inclusive definition versus false positives. Recognizing weak and unconventional bonds prevents an O/N/F-only blind spot, but makes every close contact tempting. Diagnostic: require evidence of bond formation and preferably multiple independent criteria for marginal cases.

Geometric simplicity versus electronic complexity. Distance and angle are easy to inspect, while the interaction combines several physical contributions. Diagnostic: use geometry for screening, then test energetics, spectroscopy, or electron distribution before assigning mechanism or strength.

Local specificity versus global stability. A hydrogen bond may determine which partner or orientation is selected without dominating total free energy. Diagnostic: separate specificity, local geometry, and whole-system stability claims.

Isolated bond versus cooperative network. Pairwise models are tractable, but linked bonds can strengthen, weaken, or reorganize one another. Diagnostic: compare isolated-edge and network-aware calculations or perturbations.

Structural persistence versus dynamic exchange. Crystals show relatively fixed contacts; liquids and proteins exchange partners. Diagnostic: distinguish occupancy, lifetime, and ensemble probability from the existence of a permanent bond.

Hydrogen bonding versus proton transfer. Stronger attraction can lower transfer barriers until donor/acceptor identity changes. Diagnostic: locate the proton distribution and timescale rather than forcing one static Lewis structure.

Structural–Framed Character¶

Hydrogen bond is structural with an aggregate score of 0.05 and no framing boundary. The interaction is value-neutral and exists in molecules regardless of whether anyone names it. Donor, hydrogen, acceptor, geometry, energy, and electronic redistribution are measurable or calculable; historical changes in definition reflect improved observation of one physical family rather than institutional creation of it.

The vocabulary is technical and domain-specific, but that alone does not make the object framed. “Donor” and “acceptor” are role names for asymmetries in electron distribution, not social judgments. Borderline classification is epistemic—how much evidence is sufficient—not evidence that the phenomenon is conventional.

Structural Core vs. Domain Accent¶

The structural core is a directional coupling relation mediated by a constrained intermediate site, recognized through converging effects rather than a single surface cue. This skeleton supports network and competition reasoning beyond chemistry.

The domain accent is decisive: hydrogen covalently bound to more-electronegative X, acceptor electron density at Y, geometry and polarization, spectroscopy, charge transfer, proton sharing, solvent competition, and molecular thermodynamics. Remove those roles and one obtains generic coupling or affinity, not hydrogen bonding. Conversely, retaining only “attraction involving H” is too broad; the X–H donor relation and evidence of bond formation are load-bearing.

The smallest proposed direct parent is coupling. A hydrogen bond is a strict molecular specialization of a channel that makes the geometry, energy, electron distribution, and dynamics of donor and acceptor interdependent. Coupling exists without molecules or hydrogen bonds, while hydrogen bonding fixes the channel as X–H···Y and supplies chemical evidence criteria.

compatibility is related because donor/acceptor geometry and electronic complementarity determine whether interaction is favorable. network becomes relevant when local bonds form chains or sheets. environmental_coupling_strength can describe how solvent changes an association, but is not the bond itself. No additional direct parent is necessary.

Prospective DAG placement, proposal only:

parent: prime:coupling type: subsumption qualifier: strict

Relationships to Other Abstractions¶

Current abstraction Hydrogen Bond Domain-specific

Parents (1) — more general patterns this builds on

Hydrogen Bond is a kind of Coupling Prime

The smallest proposed direct parent is coupling.

Hierarchy path (1) — routes to 1 parentless root

Hydrogen Bond → Coupling

Neighborhood in Abstraction Space¶

Hydrogen Bond sits in a sparse region of the domain-specific corpus (91^st percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Pentagonal Planar Molecular Geometry — 0.79
Energy Level Splitting — 0.78
Proton Emission — 0.78
Periodic Trends — 0.78
Secondary Carbon — 0.78

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With¶

Hydrogen bonding: collective or process wording for formation and organization of hydrogen bonds; normally a grammatical variant, not a different force.
Hydrogen-bond donor: the X–H fragment, not the entire interaction.
Hydrogen-bond acceptor: the atom or group Y, not the entire interaction.
Covalent bond: electron-sharing bond such as X–H; hydrogen bonding is the additional X–H···Y interaction.
Dipole–dipole interaction: broader polar attraction lacking the full role/evidence requirement.
van der Waals contact: generic close-range noncovalent contact, often dominated by dispersion and repulsion.
Salt bridge: ionic interaction between charged groups; it can coexist with a hydrogen bond.
Proton transfer: change in covalent proton attachment, sometimes facilitated by a hydrogen bond.
Hydrophobic effect: solvent-mediated association of nonpolar surfaces, not donor–acceptor bonding.
Base stacking: aromatic stacking in nucleic acids; it cooperates with but is not hydrogen bonding.
Adsorption: surface accumulation that may use hydrogen bonds as one mechanism.
Dihydrogen bond or halogen bond: neighboring directional interaction classes with different role definitions.

References¶

[1] Arunan, E. et al. (2011). “Definition of the hydrogen bond (IUPAC Recommendations 2011).” Pure and Applied Chemistry, 83(8), 1637–1641. https://doi.org/10.1351/PAC-REC-10-01-02 registry ↩a ↩b ↩c

[2] Arunan, E. et al. (2011). “Defining the hydrogen bond: An account (IUPAC Technical Report).” Pure and Applied Chemistry, 83(8), 1619–1636. https://doi.org/10.1351/PAC-REP-10-01-01 registry ↩a ↩b ↩c

[3] Desiraju, G. R. (2011). “A Bond by Any Other Name.” Angewandte Chemie International Edition, 50, 52–59. https://doi.org/10.1002/anie.201002960 registry ↩

[4] Alberts, B. et al. (2002). “The Shape and Structure of Proteins,” in Molecular Biology of the Cell, 4^th ed. NCBI Bookshelf. https://www.ncbi.nlm.nih.gov/books/NBK26830/ registry ↩a ↩b

[5] Alberts, B. et al. (2002). “The Structure and Function of DNA,” in Molecular Biology of the Cell, 4^th ed. NCBI Bookshelf. https://www.ncbi.nlm.nih.gov/books/NBK26821/ registry ↩a ↩b

[6] Desiraju, G. R. (1995). “Supramolecular Synthons in Crystal Engineering—A New Organic Synthesis.” Angewandte Chemie International Edition in English, 34, 2311–2327. https://doi.org/10.1002/anie.199523111 registry ↩a ↩b

[7] Breda, A., Valadares, N. F., Norberto de Souza, O., et al. (2008). “Protein Structure, Modelling and Applications,” in Bioinformatics in Tropical Disease Research. NCBI Bookshelf. https://www.ncbi.nlm.nih.gov/books/NBK6824/ registry ↩