sp² bonding: how graphene actually holds together

sp² hybridization carbon chemistry

This is one of three common bonding states for carbon. In sp³ hybridization (diamond), all four valence orbitals mix, giving four equivalent bonds pointed at tetrahedral angles — a rigid 3D solid with no delocalized electrons. In sp² hybridization (graphene, graphite), only three orbitals hybridize, producing a flat, three-bonded geometry with one orbital left over. That leftover p_z orbital is what makes graphene chemically and electronically distinct from diamond.

σ bonds + a delocalized π system the lattice

Each carbon's three sp² orbitals overlap head-on with the sp² orbitals of its three neighbors, forming strong σ (sigma) bonds about 1.42 Å long — this is the covalent scaffolding that gives graphene its extraordinary in-plane stiffness and tensile strength. It's the same bonding pattern found in benzene, extended into an infinite two-dimensional sheet rather than a single six-membered ring.

The leftover, unhybridized p_z orbital on every carbon points perpendicular to the sheet. Instead of bonding to just one neighbor, these p_z orbitals overlap sideways with all neighboring p_z orbitals at once, merging into one continuous π (pi) system delocalized across the entire flake. This shared electron cloud — not the σ framework — is what makes graphene a nearly perfect electrical conductor and gives it its distinctive semimetallic band structure.

No covalent bonds between layers interlayer force

Every carbon atom in a graphene sheet has already used all three sp² orbitals for in-plane σ bonds, and its π electron is already committed to the delocalized in-plane cloud. There's nothing left over to bond to the next layer. Stacked sheets are held in place purely by van der Waals forces — fluctuating-dipole attraction between the π clouds of adjacent layers. It's why graphite cleaves so easily along its basal planes (the same reason a pencil "lead" leaves a mark, and why mechanical exfoliation with adhesive tape can peel off single layers), and why layers can rotate relative to one another with comparatively little energy cost.

That last point matters directly for HydroGraph's material: its graphene is turbostratic — layers are randomly rotated relative to each other rather than locked into Bernal (AB) registry — which XRD measures as an interlayer spacing of 0.344–0.350 nm, slightly expanded from ideal graphite's 0.335 nm. Because van der Waals bonding is the only thing holding turbostratic layers together, and there's no shared lattice registry to reinforce it, the layers interact even more weakly than in ordinary graphite — electronically, each behaves more like an isolated monolayer than part of a bonded stack. See the turbostratic stacking definition for the full ISO terminology.

Dangling bonds at the flake edge zigzag / armchair

The σ/π picture above assumes an infinite sheet where every carbon has exactly three neighbors. A real flake is finite — a HydroGraph monomer is only 20–50 nm across — so atoms sitting at its perimeter don't have a third neighbor to bond to. That leaves an unsatisfied, dangling bond at every edge carbon. Depending on which crystallographic direction the boundary cuts across the honeycomb lattice, edges fall into one of two canonical geometries — zigzag or armchair — which even carry different electronic signatures (zigzag edges can host localized edge states that armchair edges don't).

A dangling bond is reactive: left alone it will grab whatever is available to satisfy it, typically hydrogen or an oxygen-containing group (hydroxyl –OH, carbonyl C=O, carboxyl –COOH, or an epoxide bridge). That termination is chemically real bonding — usually pulling the terminated edge carbon part-way toward sp³ character — it's just confined to the boundary instead of spread across the basal plane.

Edge chemistry differs by production route why it matters

How a flake's edges form depends heavily on how the flake itself was made. Methods that fracture a larger sheet down to size tend to leave more reactive, more oxidized edges: sonication-based liquid-phase exfoliation generates hydroxyl radicals in solution that attack graphene edges directly; ball milling creates fresh reactive edge sites that oxidize rapidly the moment they're exposed to air; electrochemical exfoliation leaves edge oxidation as a byproduct of the process itself. HydroGraph's chamber-explosion route works differently — flakes crystallize directly out of a carbon aerosol as multilayer platelets, so edges form during nucleation and growth rather than by tearing a larger sheet apart in a separate destructive step.

This lines up with HydroGraph's own bulk composition data: FGA-1 is measured at 99.8% carbon, 0.2% oxygen, 100% sp²-bonded carbon, with no PAHs and no volatiles up to 700°C. With the bulk lattice reported as fully sp², that small residual oxygen fraction is a reasonable candidate for edge-terminating functional groups rather than basal-plane defects — though the source doesn't explicitly localize it, so take that as informed inference, not a directly measured result.

Worth noting: it isn't simply "smaller flake, more edge, more oxygen." FGA-2 monomers are larger (75–200 nm, meaning proportionally less edge per unit volume than FGA-1) yet carry more oxygen (1.35%) and less sp² character (80%). That's driven by precursor chemistry — a higher oxygen/acetylene ratio (O/C = 0.5, vs. 0.3 for FGA-1) incorporates oxygen throughout synthesis, not only at the edges. Precursor stoichiometry can dominate over the simple edge-fraction story.