Atomic orbitals are wavefunctions describing electron probability densities. s-orbitals are spherically symmetric, p-orbitals are dumbbell-shaped, d-orbitals are cloverleaf-shaped. Nodes—regions where the wavefunction vanishes—emerge from quantization and angular momentum constraints. Radial nodes correspond to standing waves in the radial direction; angular nodes define the shape. Higher energy orbitals contain more nodes.
From solving the hydrogen atom, you know that the wavefunction separates into a radial part and an angular part: ψ_nlm(r,θ,φ) = R_nl(r) · Y_l^m(θ,φ). The three quantum numbers n, l, m each govern a distinct aspect of the orbital's geometry. The principal quantum number n determines the overall energy and size scale; the angular momentum quantum number l (ranging from 0 to n−1) determines the shape; and the magnetic quantum number m (ranging from −l to +l) determines the orientation in space. The orbital shapes you see visualized — the dumbbells, cloverleaves, and spheres — are contour surfaces of |ψ|², the probability density.
Nodes are the key to understanding why different orbitals have the shapes they do. A node is any surface where the wavefunction ψ = 0, which means the electron has zero probability of being found there. Think of a standing wave on a guitar string: the allowed vibration modes have 0, 1, 2, ... fixed points (nodes) where the string never moves, and more nodes correspond to higher frequency (higher energy). Orbitals obey the same principle in three dimensions. There are two types: radial nodes are spherical shells at specific radii where ψ vanishes as you move outward from the nucleus, arising from the radial standing wave condition in R_nl(r). Angular nodes are planes or cones defined by the angular part Y_l^m(θ,φ), independent of r. The total number of nodes is always n − 1, split as: (n − l − 1) radial nodes and l angular nodes.
For l = 0 (s-orbitals), there are no angular nodes — the wavefunction depends only on r, giving spherical symmetry. The 1s orbital has no nodes at all; the 2s has one spherical radial node at a specific radius where ψ changes sign; the 3s has two. For l = 1 (p-orbitals), there is one angular nodal plane. The p_z orbital (m = 0) has a nodal plane at z = 0 (the x-y plane); electrons are found above and below but never in that plane, giving the characteristic dumbbell shape. The 2p orbital has no radial nodes; the 3p has one additional radial shell. For l = 2 (d-orbitals), two angular nodes create the cloverleaf patterns: the d_z² orbital has two nodal cones, while d_xy, d_xz, d_yz each have two perpendicular nodal planes rotated relative to each other.
The connection to chemistry is direct: orbital shapes determine how atoms bond. Two orbitals can form a bond only if they have significant spatial overlap where their wavefunctions are both nonzero and have the same sign (constructive overlap). The nodal planes and lobes of p and d orbitals dictate the geometry of sigma and pi bonds, the directionality of molecular shapes (VSEPR theory is a coarser approximation of this), and the selection rules for electronic transitions. When you memorize orbital shapes, you are not memorizing an arbitrary classification — you are reading off the angular momentum content of the quantum state from its visible geometry.