Atomic Orbitals: Shape and Energy

Orbital shape and energy are not a historical curiosity confined to introductory chapters — they are the daily working vocabulary of modern chemistry and its instrumentation. Every prediction of nucleophile/electrophile reactivity in drug design (which carbon of a scaffold is electrophilic enough for a nucleophilic amine to attack) starts from LUMO shape and energy computed, in practice, by density functional theory built entirely on atomic-orbital basis sets (Gaussian-type orbitals mimicking the $s/p/d$ shapes derived here). X-ray photoelectron spectroscopy (XPS), used industrially to characterize catalyst surfaces and semiconductor interfaces, directly measures the binding energies that split $ns$ from $np$ from $nd$ exactly as described above, reading off elemental composition and oxidation state from those splittings. In materials science, the band structure of silicon photovoltaics and the color of transition-metal pigments both trace back to $d$-orbital energy splitting in a ligand field, a direct descendant of the penetration/shielding argument given here. And at the most basic pedagogical level, every mechanism you will draw for the rest of your organic chemistry career — nucleophilic addition to a carbonyl, electrophilic addition to an alkene, SN2 backside attack — is literally a picture of same-phase orbital lobes overlapping while an antibonding orbital simultaneously grows into a new bond, the exact process worked out step-by-step in the second derivation above.

Ask the tutor: How does mixing 2s and 2p atomic orbitals actually generate the four equivalent sp³ lobes that give carbon its tetrahedral bond angles?

An atomic orbital is not a orbit in the planetary sense — an electron does not trace a path around the nucleus. It is a standing wave, a solution $\psi_{n,l,m_l}(r,\theta,\phi)$ of the time-independent Schrödinger equation for an electron in a central Coulomb potential. The wavefunction factors cleanly:

The radial function $R_{n,l}(r)$ tells you how the amplitude changes as you move outward from the nucleus — this controls size and radial nodes. The angular function $Y_l^{m_l}(\theta,\phi)$ tells you how the amplitude changes as you rotate around the nucleus at fixed distance — this controls shape, lobe count, and phase pattern. This factorization is the single most useful fact in the whole subject: shape lives in the angular part; size and energy trends live mostly in the radial part.

$s$ orbitals ($l=0$): $Y_0^0$ is a constant — no angular dependence at all. The orbital is perfectly spherical; density depends only on $r$. There are no angular nodes. A $1s$ orbital has zero radial nodes; a $2s$ orbital has one radial node (a spherical shell of zero density partway out); a $3s$ has two, and so on. The general rule for total nodes is:

$p$ orbitals ($l=1$): now $Y_1^{m_l}$ depends on direction. In the real (rather than complex) basis chemists use, the three $p$ orbitals are labeled $p_x$, $p_y$, $p_z$, each a two-lobed dumbbell aligned along one Cartesian axis, with a nodal plane perpendicular to that axis passing through the nucleus. A $2p_z$ orbital, for instance, has a node in the $xy$-plane: density is zero everywhere in that plane, positive phase above it, negative phase below it (or vice versa — the absolute sign is arbitrary, only relative sign matters). This is $l=1$, so there is exactly one angular node, consistent with the node-counting formula above. The phase of $\psi$ (its algebraic sign) is the feature students most often ignore and organic mechanism most desperately needs. It is not electric charge — both lobes of a $p$ orbital hold an electron of charge $-e$; the sign is a property of the wave amplitude, exactly like the crest and trough of a water wave carry no net charge but do carry a definite relative phase. When two orbitals overlap with matching phase (lobe $+$ meeting lobe $+$), the amplitudes add constructively, density builds up between the nuclei, and energy drops — a bonding interaction. When they overlap with mismatched phase ($+$ meeting $-$), amplitudes cancel, a node appears between the nuclei, and energy rises — an antibonding interaction. Every curved arrow in a mechanism is shorthand for exactly this: a filled, phase-matched lobe (HOMO of the nucleophile) pushing into an empty, phase-matched lobe (LUMO of the electrophile).

Now consider what happens across a whole shell. Within the same principal quantum number $n$, in a hydrogen atom (one electron, one proton) the energy depends only on $n$:

so $2s$ and $2p$ are exactly degenerate. But the moment you add a second electron — any multi-electron atom, carbon included — this degeneracy breaks. The reason is penetration: an $s$ electron's radial function has significant amplitude very close to the nucleus (it has no angular node forcing it away from $r=0$), so it spends more time inside the shielding cloud of core electrons and feels a larger effective nuclear charge on average. A $p$ electron, constrained by its angular node to avoid one whole plane through the nucleus, penetrates less and is shielded more effectively. The ordering that results, for the same $n$, is:

This is why, in a carbon atom's ground state ($1s^2 2s^2 2p^2$), the $2s$ electrons sit lower in energy than the $2p$ electrons, even though both belong to shell $n=2$. That energy gap is small enough, though, that mixing $2s$ and $2p$ character (hybridization) costs little and buys enormous geometric flexibility — the reason carbon can adopt sp, sp², and sp³ geometries at similar energetic cost.

Three governing relationships carry essentially all of the qualitative chemistry that follows from orbital theory.

Node counting, from the quantum numbers alone:

where: $n$ is the principal quantum number (shell), $l$ is the angular momentum quantum number ($0=s,1=p,2=d,\dots$), and each equation is exact for a one-electron hydrogenic wavefunction (and is used as a bookkeeping rule for real multi-electron atoms too). Reading it: bump $n$ up by one at fixed $l$ and you add exactly one radial node; bump $l$ up by one at fixed $n$ you trade one radial node for one angular node — total stays the same.

Hydrogenic energy (single electron, nucleus of charge $+Ze$):

where: $13.6\ \text{eV}$ is the Rydberg energy (the ionization energy of ground-state hydrogen, $Z=1,n=1$), $Z$ is the nuclear charge, and $n$ is the principal quantum number. Reading it: energy depends only on $n$ for one electron — this is the "unperturbed" reference state that multi-electron shielding then distorts.

Effective nuclear charge and the $ns<np$ split (Slater-type qualitative form):

where: $Z_{\text{eff}}$ is the effective nuclear charge felt by the electron, $Z$ is the actual nuclear charge, and $S$ is the shielding constant contributed by other electrons — crucially $S$ is smaller for more-penetrating orbitals (lower $l$ at fixed $n$), so $Z_{\text{eff}}(ns) > Z_{\text{eff}}(np)$, which pushes $E(ns)$ below $E(np)$.

Example 1 (foundational): node-counting for the $3p$ orbital of a hydrogen atom. Take $n=3$, $l=1$ (a $p$ orbital in the third shell — asymmetric on purpose, not the more commonly quoted $2p$). Using

So a $3p$ orbital has one angular node (the usual nodal plane perpendicular to its axis, giving the two-lobe dumbbell) and one radial node (a spherical shell of zero density cutting through each lobe), for two nodes total. This is testable directly against the general formula and is why $3p$ radial distribution plots show two humps of density along each lobe rather than one — a detail invisible in oversimplified "just a bigger dumbbell" pictures.

Example 2 (applied): ranking orbital energies in a ground-state phosphorus atom, and predicting its Lewis structure. Phosphorus, $Z=15$, ground configuration $1s^2\,2s^2\,2p^6\,3s^2\,3p^3$. We want the qualitative energy ordering of the valence orbitals, since this determines which orbitals are available for bonding. Using the penetration argument from the derivation above (extended down each shell): for the same $n=3$ shell,

because $3s$ has $l=0$ (maximal penetration, $Z_{\text{eff}}$ highest), $3p$ has $l=1$ (intermediate), $3d$ has $l=2$ (least penetration, $Z_{\text{eff}}$ lowest, and in P the $3d$ is unoccupied and much higher still). With $3s^2$ filled and $3p^3$ half-filled (one electron each in $3p_x,3p_y,3p_z$ by Hund's rule, maximizing unpaired spins to minimize electron-electron repulsion), phosphorus presents three singly-occupied $p$ orbitals, mutually perpendicular, each capable of forming one covalent bond by pairing with an electron from another atom (e.g., three Cl atoms in PCl₃, before any $s$–$p$ mixing/hybridization is invoked). The near-90° angles predicted by unhybridized, mutually perpendicular $p$ orbitals are a first approximation — real PCl₃ has bond angles nearer 100°, the deviation being exactly the kind of correction that hybridization arguments in Hybridization: The sp3 Carbon are built to explain.

Example 3 (exam/competition level): predicting which orbital overlap is symmetry-allowed for a $\sigma$ bond versus forbidden by cancellation. Consider an $s$ orbital on atom A approaching a $p_x$ orbital on atom B, with the approach axis chosen to be the $z$-axis (i.e., A approaches B from directly "above," along $z$, while B's $p$ orbital lobes point along $x$, perpendicular to the approach direction). Does bonding overlap occur? Set up the geometry: the $s$ orbital is uniformly $+$ in all directions. The $p_x$ orbital has a $+$ lobe on the $+x$ side of B and a $-$ lobe on the $-x$ side, with its node being the $yz$-plane. Because the incoming $s$ orbital sits along the $z$-axis (in the $yz$-plane, B's nodal plane), it overlaps equally with the $+x$ lobe and the $-x$ lobe of the $p_x$ orbital by the mirror symmetry of the arrangement:

These two contributions are equal in magnitude and opposite in sign (one is overlap with a $+$ lobe, the other with a $-$ lobe, at mirror-symmetric points), so the net overlap integral is exactly zero:

This is a symmetry-forced cancellation, not a coincidence of magnitudes — it holds regardless of the orbitals' radial functions, sizes, or the exact internuclear distance, because it follows purely from the mirror-plane argument. The general rule this illustrates: net bonding requires that the two orbitals' phase patterns be compatible with at least one common axis or plane of symmetry along the approach direction. An $s$ orbital does bond productively with a $p_z$ orbital oriented directly along the approach axis (head-on, same phase on the near lobe) — that combination has no such cancellation and gives a proper $\sigma$ bond. This exact symmetry-cancellation argument is the microscopic justification later used to explain why, in $\pi$-bond formation, only $p$ orbitals aligned parallel to each other (not perpendicular ones) give nonzero overlap.

The modern orbital picture is the joint product of quantum mechanics and its immediate chemical translation, but if one figure must be singled out for turning the abstract wavefunction into the chemist's working orbital diagram, it is Linus Pauling, who — building on Schrödinger's equation and Heitler and London's 1927 quantum treatment of the H₂ bond — synthesized valence bond theory, electronegativity, and hybridization into a single coherent, teachable framework in his 1939 book The Nature of the Chemical Bond. Pauling took orbitals that were, in the hands of physicists, purely mathematical eigenfunctions of an operator, and showed practicing chemists exactly how to combine, mix, and reshape them to predict real molecular geometries — precisely the move that makes an "atomic orbital" article the direct ancestor of a hybridization article.