Schrödinger Equation for Atoms

Why the atomic Schrödinger equation matters

• It explains the periodic table from first principles. The hierarchy n > l > m_l > m_s with the Pauli exclusion principle generates exactly 2, 8, 8, 18, 18, 32, 32 elements per row — the structure Mendeleev tabulated empirically in 1869 emerges automatically once you solve the equation.
• The hydrogen atom is exactly solvable. Energies E_n = −13.6 eV · Z²/n², exactly. Bohr radius a₀ = 0.529 Å. Spectral lines reproduce Rydberg's 1888 formula 1/λ = R_H(1/n₁² − 1/n₂²) with R_H = 109737.32 cm⁻¹. The Lyman, Balmer, and Paschen series are all consequences.
• Atomic orbitals are squared wavefunctions. The familiar s, p, d, f shapes — spherical, dumbbell, cloverleaf, complex — are |Y_l^m|², the squared spherical harmonics. p_x, p_y, p_z are real linear combinations of m_l = −1, 0, +1.
• SCF launched computational chemistry. Hartree's 1928 mean-field iteration and Fock's 1930 antisymmetrization for fermions remain the foundation of every quantum-chemistry package. CCSD(T), MP2, CASSCF, MRCI, and DFT all start from a Hartree-Fock or Kohn-Sham reference.
• Bond formation is just orbital overlap. Heitler-London 1927 wrote the first quantum-mechanical wavefunction for H₂ and recovered the bond. The covalent bond is now understood as constructive overlap of one-electron wavefunctions; molecular orbital theory generalizes this to all molecules.
• Spectroscopy is energy-difference probing. Photons promote electrons between Schrödinger eigenstates. UV-Vis sees valence excitations; X-ray absorption probes core levels; rotational spectra (microwave) and vibrational (IR) read out the molecular Schrödinger equation. The whole hierarchy of spectroscopies is the atomic Schrödinger equation made experimental.
• Quantum chemistry now uses ~80% of supercomputing time across multiple national labs. Drug design (docking, ligand binding), catalyst design, battery electrode prediction, photovoltaic absorber screening — they all run Schrödinger-equation-based methods (DFT, post-HF) at scale.

Common misconceptions

• Electrons orbit the nucleus. They don't. The wavefunction has an instantaneous probability distribution; the electron does not follow a deterministic trajectory. Bohr's 1913 planetary model was superseded by Schrödinger 1926, even though the energy formula came out the same.
• Schrödinger's equation explains spin. No. Spin is an entirely external addition to non-relativistic Schrödinger; Pauli put it in by hand in 1925 with two-component spinors. Dirac's 1928 relativistic equation does derive spin self-consistently.
• The wavefunction is observable. ψ itself is not measurable; it is a complex amplitude. Born's 1926 interpretation states |ψ|² dV is the probability of finding the electron in volume dV. Phase information matters in interference and bonding but never appears as a direct observable for a single electron.
• Helium can be solved with another year of math. No. Even three-body atoms (He, H−) have no closed-form solution. Hylleraas (1929) achieved seven-decimal accuracy variationally using explicit r₁₂ terms, but no analytical formula exists. For He, MCSCF + r₁₂ methods now give microhartree precision but only numerically.
• Higher Z means simpler chemistry because the Coulomb pull dominates. The opposite. Heavier atoms have more electrons, more relativistic corrections, and stronger spin-orbit coupling. Lawrencium (Z = 103) chemistry differs visibly from lutetium because of relativistic 7s contraction.
• Atomic units make the equation prettier but optional. They are essentially mandatory in actual computation. In atomic units (e = m_e = ℏ = 4πε₀ = 1), energies come out in Hartrees (1 Ha = 27.211 eV) and lengths in bohrs (1 a₀ = 0.529 Å) — without them the floating-point conditioning at picometer scale is brutal.

From hydrogen to multi-electron atoms

For a one-electron atom of nuclear charge Z, the time-independent Schrödinger equation in atomic units reads −(1/2)∇²ψ − (Z/r)ψ = Eψ. Spherical symmetry of the Coulomb potential lets you separate ψ(r,θ,φ) = R(r) · Y_l^m(θ,φ). The angular factor Y_l^m is a spherical harmonic — well-known from any textbook on Laplace's equation — and gives quantum numbers l = 0, 1, ... and m_l = −l, ..., +l. The radial equation reduces to a Laguerre-polynomial differential equation; its normalizable solutions exist only for n = 1, 2, 3, ... with l < n. The eigenvalues are E_n = −Z²/(2n²) Hartree = −13.6 Z²/n² eV. The 1s orbital is R₁₀(r) = 2(Z/a₀)^3/2 exp(−Zr/a₀) · 1/√(4π).

For a two-electron atom you add a second kinetic-energy term, a second Coulomb-attraction term, and crucially a 1/r₁₂ repulsion that depends on both electron coordinates simultaneously. This term blocks separability — there is no closed-form analytical solution. Hartree's 1928 fix was to replace the exact 1/r₁₂ with each electron seeing the time-averaged charge density of the other. You guess one-electron orbitals, build the average potential, solve a one-electron Schrödinger equation in that potential, get new orbitals, and iterate. Fock added antisymmetry — the wavefunction must change sign on swapping any two fermions — by writing the product as a Slater determinant. The combined Hartree-Fock procedure recovers about 99% of atomic and molecular total energies.

The remaining 1% is electron correlation — the deviation from the mean-field assumption. It dominates chemistry: bond energies, reaction barriers, intermolecular forces all live in that 1%. Modern approaches systematically improve on Hartree-Fock with Mller-Plesset perturbation theory (MP2, MP3, ...), configuration interaction, coupled cluster (CCSD(T)), and DFT. Each trades cost for accuracy. CCSD(T) — the "gold standard" — costs O(N⁷) but gets bond energies to within 4 kJ/mol. DFT runs at O(N³) and captures most of the correlation through an approximate exchange-correlation functional, which is why ~80% of computational chemistry papers now use it.

Hydrogen-atom exact solutions: orbital energies and radial extents

Quantum numbers	Notation	Energy (eV)	Most probable radius (a0)	Radial nodes	Angular nodes
n=1, l=0	1s	−13.6	1.0	0	0
n=2, l=0	2s	−3.40	5.24	1	0
n=2, l=1	2p	−3.40	4.0	0	1
n=3, l=0	3s	−1.51	13.07	2	0
n=3, l=1	3p	−1.51	12.0	1	1
n=3, l=2	3d	−1.51	9.0	0	2
n=4, l=0	4s	−0.85	23.7	3	0

Methods for multi-electron atoms — accuracy vs cost

Method	Treatment of correlation	Cost scaling	Typical error on small-atom binding energy	Where it shines
Hartree-Fock (HF)	None — mean field only	O(N4)	50-100 kJ/mol; misses dispersion entirely	Reference wavefunction; orbitals for post-HF
MP2 (Møller-Plesset 2nd order)	Perturbation on top of HF	O(N5)	~30 kJ/mol; problematic at stretched bonds	Affordable correlation for <100-atom systems
CCSD (coupled-cluster, single+double)	Iterative; near-exhaustive in single/double space	O(N6)	~10 kJ/mol	Closed-shell molecules and atoms
CCSD(T) — "gold standard"	CCSD plus perturbative triples	O(N7)	~4 kJ/mol; benchmark accuracy	Reaction thermochemistry <30 atoms
DFT (B3LYP, hybrid)	Approximate exchange-correlation functional	O(N3) (hybrid: O(N4))	~15-30 kJ/mol	~80% of computational papers; up to ~1000 atoms
Multireference (MCSCF, CASPT2)	Combines static + dynamic correlation	O(Nk), k variable	10-20 kJ/mol	Systems with bond breaking, transition metals
Quantum Monte Carlo (QMC)	Stochastic sampling of correlated wavefunction	O(N3) per sample, parallel	~5-10 kJ/mol	Benchmark accuracy for >100 atoms

Applications and examples

• Hydrogen atom radial functions. The exact analytical solutions R_nl(r) are products of an exponential and a generalized Laguerre polynomial. R₁₀ = 2(Z/a₀)^3/2 exp(−Zr/a₀) · 1/√(4π); R₂₀ has a node at r = 2a₀; R₂₁ has no radial node. The most-probable distance for the 1s orbital is the Bohr radius a₀ = 0.529 Å.
• Heitler-London 1927 H₂. The first quantum-mechanical molecular calculation. Heitler and London wrote a two-electron wavefunction with one electron on each H atom, antisymmetrized for fermions, and recovered the H₂ bond energy of 4.7 eV with about 60% accuracy. This launched valence-bond theory and proved that chemistry follows from the Schrödinger equation.
• Hartree-Fock self-consistent field. Hartree (1928) and Fock (1930) replaced the unsolvable N-body problem with iterative one-body calculations in a mean field. Modern packages (Gaussian, ORCA, NWChem, Q-Chem, PySCF) all start from HF or Kohn-Sham SCF and add correlation on top.
• Walter Kohn and John Pople 1998 Nobel. Kohn for density-functional theory (the alternative to wavefunction methods that dominates today); Pople for developing computational methods including Gaussian, the program. Together their work made quantum chemistry routinely usable.
• Drug discovery and catalysis at scale. Pharmaceutical molecule docking (computing binding affinities), homogeneous-catalyst design, photovoltaic absorber screening, and battery-electrode prediction all run on Schrödinger-derived methods. CCSD(T) is the benchmark, DFT the workhorse, and ML-corrected DFT (Δ-machine learning) is the new frontier in 2024-2026.