The Covalent Bond and Electron Sharing

The covalent bond is the literal chemical infrastructure of biology and materials. Every peptide bond linking amino acids in a protein, every phosphodiester bond in the DNA backbone, and every C–C and C–H bond in the drugs sitting in a pharmacy are covalent bonds whose strength and geometry were set by exactly the overlap and electronegativity arguments developed here. Drug designers routinely tune a molecule's polarity by swapping a C–H for a C–F bond ($\Delta\chi_{\text{C–F}} = 1.5$, versus $\Delta\chi_{\text{C–H}} = 0.4$) to change lipophilicity, metabolic stability, and dipole-driven binding to a protein target — a direct, practical application of the electronegativity-to-bond-character reasoning above. In materials science, the covalent-versus-ionic spectrum explains why diamond (purely covalent C–C network) is an insulator and extremely hard, while sodium chloride (highly ionic) is brittle, water-soluble, and conducts electricity only when molten or dissolved — properties traceable directly to whether the "bonding electrons" are localized in directional shared pairs (covalent) or delocalized as free ions in a lattice (ionic). Semiconductors like silicon and gallium arsenide sit in between: Si–Si bonds are purely covalent ($\Delta\chi=0$), while Ga–As bonds ($\Delta\chi \approx 0.4$) have enough ionic character to shift the band gap and enable the specific optoelectronic properties exploited in LEDs and laser diodes. In instrumentation, infrared and Raman spectroscopy directly probe covalent bond force constants (essentially the curvature of the $E(r)$ well derived above) to identify functional groups, while X-ray crystallography measures bond lengths $r_e$ to sub-picometer precision, letting chemists directly verify predictions about bond order and hybridization. Computational chemistry (Hartree-Fock and density functional theory) builds numerically on exactly the bonding/antibonding molecular-orbital construction worked through in the derivation above, scaled up to molecules with thousands of orbitals — the same in-phase/out-of-phase overlap argument that explains H₂ also underlies modern drug-design software predicting how a candidate molecule's orbitals will overlap with a target enzyme's active site.

Ask the tutor: How do hybridized orbitals (sp³, sp², sp) change the strength and geometry of the covalent bonds described here?

Start with two isolated hydrogen atoms, far apart, each with one electron in a $1s$ orbital. Each atom's electron only feels the pull of its own proton. Bring the atoms closer. Once the $1s$ orbitals begin to overlap, quantum mechanics allows the two electrons — provided they have opposite spin — to occupy a single molecular orbital that is spread over both nuclei. In that new orbital, each electron spends part of its time near the other proton as well as its own. It is now attracted by two positive charges instead of one, and that extra attraction is the entire origin of the covalent bond in H$_2$. This is not free, though. As the nuclei approach, three things happen simultaneously: (1) each electron gains extra attraction to the "foreign" nucleus, (2) the two electrons, now confined to a smaller shared region, repel each other more, and (3) the two nuclei themselves begin to repel each other directly. At long range, the attraction term dominates (bringing the atoms together lowers energy). At very short range, nuclear-nuclear repulsion diverges and dominates. Somewhere in between, there is an energy minimum — this is the equilibrium bond length $r_e$, and the depth of that minimum (relative to separated atoms) is the bond dissociation energy. For H–H, $r_e = 0.74$ Å and the bond energy is $436\ \text{kJ/mol}$. Constructive overlap matters enormously. If the two $1s$ orbitals overlap in-phase (same sign of the wavefunction lobe pointing at each other), the wavefunctions add constructively between the nuclei, building up electron density exactly where it does the most good — in the internuclear region, pulling both nuclei toward it. This is the bonding molecular orbital, lower in energy than either isolated atomic orbital. If instead the orbitals overlap out-of-phase, the wavefunctions cancel between the nuclei, creating a node there; this is the antibonding orbital, higher in energy, and putting electrons into it pushes atoms apart rather than together. Covalent bond formation in H$_2$ is precisely: two electrons go into the low-energy bonding MO, and (crucially) none are left over to occupy the destabilizing antibonding MO. Why does the octet matter for heavier atoms? For second-row elements (Li–Ne), the valence shell consists of one $2s$ and three $2p$ orbitals — four orbitals total, and by the Pauli principle each can hold two electrons, for a maximum of eight. When all four orbitals are filled (as in Ne, or as in the bonding+lone-pair count around C, N, O, F in typical molecules), every valence orbital on that atom is either engaged in a bonding pair (stabilized by two nuclei) or a nonbonding lone pair (stabilized by its own nucleus and shielded from repulsion by being kept in a low-energy filled orbital). There is no room left for an additional electron to be forced into a higher-energy orbital, and no unfilled low-energy orbital left dangerously exposed to attack. Eight is not sacred; it is simply "four orbitals, completely used," which is why elements with different numbers of valence orbitals (transition metals, or main-group elements using $d$-orbitals) are not bound by an octet at all.

The same logic scales up to multi-electron atoms and heteronuclear bonds. Two chlorine atoms, each one electron short of argon's configuration, share a pair drawn from a $3p$ orbital on each; the bonding MO built from those two $3p$ orbitals is filled, no antibonding orbital is populated, and Cl$_2$ is bound by 242 kJ/mol. But now the two nuclei are identical, so the shared pair sits, on time-average, exactly midway — a nonpolar covalent bond. Replace one chlorine with sodium, and everything changes: sodium's $3s$ electron is barely held (ionization energy 496 kJ/mol) while chlorine's electron affinity is large (349 kJ/mol) and its nucleus, with a much higher effective nuclear charge felt at the valence shell, pulls far harder on any electron placed between them. Instead of an intermediate, partly-shared arrangement, the system reaches its lowest energy by transferring the electron completely: Na$^+$ and Cl$^-$, held together not by orbital overlap but by pure Coulombic attraction between full charges. This is the ionic limit of the same underlying competition.

The energy of two atoms as a function of separation $r$ can be written, in the crudest useful approximation, as a sum of an attractive term and a repulsive term:

where:

• $A$ encodes the net electrostatic attraction gained from sharing electrons between two nuclei (or, in the ionic limit, the attraction between point charges $+e$ and $-e$),
• $B$ and the exponent $n$ (typically $8$–$12$) encode short-range repulsion from overlapping electron clouds and internuclear Coulomb repulsion,
• $r$ is the internuclear separation.

This is the qualitative content of the classic Lennard-Jones / Born-type potential, and minimizing it with respect to $r$ (setting $dE/dr = 0$) locates the equilibrium bond length, exactly as shown by the minimum in the H$_2$ curve above. For predicting how covalent (versus ionic) a given bond is, the working tool is the electronegativity difference. Pauling's empirical relationship between excess bond energy and electronegativity difference is:

where:

• $\chi_A, \chi_B$ are the Pauling electronegativities of atoms A and B,
• $\Delta_{AB} = D(A\text{–}B) - \tfrac{1}{2}\left[D(A\text{–}A) + D(B\text{–}B)\right]$ is the "excess" bond energy (in kJ/mol) of the real A–B bond over the average of the two homonuclear bonds — the extra stabilization attributable to ionic character,
• $D(X\text{–}Y)$ is the bond dissociation energy of X–Y.

Reading this equation: if A–B were purely covalent (nonpolar, shared exactly like two identical atoms), its bond energy would simply be the average of the A–A and B–B bond energies (a geometric-mean-like additivity). Any bond energy stronger than that average reflects extra ionic (electrostatic) stabilization from unequal sharing, and the size of that excess energy — via $\Delta_{AB}$ — tells you $\Delta\chi$. A separate, more intuitive metric is percent ionic character, estimated from the dipole moment:

where:

• $\mu_{\text{observed}}$ is the experimentally measured dipole moment of the molecule,
• $\mu_{\text{ionic}}$ is the dipole moment the bond would have if the bonding pair sat 100% on the more electronegative atom (i.e., a full elementary charge $e$ separated by the actual bond length $r_e$),
• $e$ is the elementary charge, $1.602\times10^{-19}\,\text{C}$.

Reading this: a bond that is perfectly covalent (symmetric sharing) has $\mu_{\text{observed}} = 0$, giving 0% ionic character; a bond with electrons fully localized on one atom has $\mu_{\text{observed}} = \mu_{\text{ionic}}$, giving 100%. Every real polar covalent bond lies somewhere between, and this is the rigorous version of the qualitative "covalent vs. ionic" spectrum sketched above.

Example 1 (foundational): Predicting bond character in H–F from electronegativities. Hydrogen has $\chi_H = 2.1$; fluorine has $\chi_F = 4.0$ (Pauling scale). First find $\Delta\chi$:

Apply the exponential relationship derived above:

So H–F is predicted to be roughly 59% ionic, 41% covalent — a strongly polar covalent bond, not a fully ionic one. Cross-check against the experimental dipole moment: $\mu_{\text{obs}}(\text{HF}) = 1.82\ \text{D}$, and the bond length is $r_e = 0.92$ Å. The fully-ionic reference dipole is

(using the convenient unit conversion $e = 4.803$ D per Å of separation). Then

Both methods agree that H–F is substantially, but far from completely, ionic — a genuinely intermediate case, exactly the regime where the covalent/ionic distinction is most informative and where the two independent estimation methods (bond-energy-based and dipole-based) should be expected to differ somewhat, since they rest on different approximations, while agreeing on the qualitative conclusion.

Example 2 (applied): Why NaCl crystallizes ionically but PCl₃ does not — a mechanistic comparison. Consider two P–Cl-type and Na–Cl-type interactions side by side. Sodium's valence electron sits in an unusually high-energy, poorly shielded $3s$ orbital (first ionization energy only $496\ \text{kJ/mol}$), while chlorine's incompletely-filled $3p$ orbital sits at a much lower energy and chlorine has a high electron affinity ($-349\ \text{kJ/mol}$, i.e., strongly favorable to gain an electron). The energy cost to fully ionize Na (496 kJ/mol) is more than repaid by the energy released forming Cl⁻ (349 kJ/mol) plus the enormous lattice/Coulombic stabilization of packing Na⁺ and Cl⁻ ions into an extended ionic lattice (lattice energy of NaCl $\approx 787\ \text{kJ/mol}$). Once summed, full electron transfer plus lattice formation is energetically far more favorable than any shared-pair arrangement between such mismatched orbitals — so NaCl is properly described as ionic ($\Delta\chi = 2.1$). Phosphorus, by contrast, has $\chi_P = 2.1$, giving $\Delta\chi(\text{P–Cl}) = |3.0-2.1| = 0.9$:

With only modest electronegativity mismatch, and no comparably large lattice-stabilization pathway available to a small discrete molecule like PCl₃ (there is no macroscopic ionic lattice to form — three separate Cl⁻ ions and a bare P³⁺ cation would cost enormous, unrecovered ionization energy), the lowest-energy structure keeps the electrons shared in three polar covalent P–Cl bonds rather than transferred. This example shows explicitly that "ionic vs. covalent" is decided by comparing the total energetics of each pathway (including lattice or solvation energies), not electronegativity difference in isolation — $\Delta\chi$ predicts the character of an individual bond, but whether a substance crystallizes as discrete covalent molecules or as an extended ionic lattice depends on whether that ionic pathway has anywhere to put the released lattice energy.

Example 3 (exam/competition-level): Excess bond energy and Pauling's original electronegativity derivation, applied to Si–Cl. Given the following bond dissociation energies — $D(\text{Si–Si}) = 226\ \text{kJ/mol}$, $D(\text{Cl–Cl}) = 242\ \text{kJ/mol}$, $D(\text{Si–Cl}) = 464\ \text{kJ/mol}$ — compute the electronegativity difference $\Delta\chi(\text{Si,Cl})$ predicted by Pauling's original method, and compare to the tabulated values ($\chi_{\text{Si}} = 1.9$, $\chi_{\text{Cl}} = 3.0$).

Step 1. Compute the "expected" purely-covalent bond energy as the average of the two homonuclear bond energies:

Step 2. Compute the excess bond energy $\Delta_{AB}$ — the amount by which the real, polar Si–Cl bond is stronger than this naive covalent average, attributable to ionic/electrostatic stabilization:

Step 3. Apply Pauling's relation:

Step 4. Compare with the tabulated electronegativity difference:

The bond-energy method gives 1.55, somewhat higher than the tabulated 1.1 (tabulated Pauling values are themselves fitted across many bonds simultaneously by least-squares, so any single-pair back-calculation will deviate somewhat — this is an expected and instructive discrepancy, not an error). What both numbers agree on is the qualitative conclusion: $\Delta\chi$ is well above the ~0 value for a purely covalent bond and well below the ~2 needed to call Si–Cl essentially ionic, correctly placing SiCl₄ as a polar covalent, molecular (not ionic-lattice) compound — consistent with its low melting point (−70 °C) and volatility, properties utterly unlike NaCl's high-melting ionic lattice.