Computational Chemistry with MATLAB

Learn MATLAB through molecular evidence: structures, surfaces, kinetics, quantum states, spectra, dynamics, electronic structure, thermal ensembles, solvation, and design under uncertainty.

11 projects, 275 hands-on levels, run in your browser.

Syllabus

• Foundations: Molecules as Matrices: Learn MATLAB through atom identities and Cartesian coordinate matrices, then build a fail-closed molecular intake and evidence dossier. The project teaches functions, cell arrays, matrices, logical selection, loops, structures, validation, broadcasting, and invariants without detaching those programming ideas from molecular work.
• Are These Really the Same Conformer?: Build an identity-safe conformer comparison from local internal coordinates through proper rigid alignment, labelled distance audits, inertia evidence, and tolerance sensitivity. The final dossier distinguishes row permutation and rigid placement from genuine internal deformation.
• Where Does the Molecule Prefer to Sit?: Build and compare potential-energy models, verify analytic derivatives independently, search with complete trajectory evidence, classify stationary-point curvature, expose initial-condition and step sensitivity, and publish a bounded potential-surface dossier.
• Which Reaction Mechanism Fits the Evidence?: Build mass-action mechanisms, integrate them with solver evidence, fit kinetic and Arrhenius parameters, compare competing integrated laws, expose stiffness and identifiability, and publish a bounded mechanism dossier.
• Why Are Molecular Energies Quantized?: Discretize one-dimensional quantum operators, solve and audit eigenstates, measure observable evidence, challenge grid and domain convergence, and construct a bounded Hückel molecular-orbital dossier.
• What Is Hiding in the Spectrum?: Build mass-weighted normal modes, turn sampled motion into a traceable FFT spectrum, resolve candidate bands, fit overlapping signals and mixtures, and publish uncertainty-aware vibrational assignments.
• Will This Molecular Simulation Behave?: Build periodic systems and Lennard-Jones forces, integrate velocity-Verlet trajectories, challenge energy and ensemble evidence, measure radial structure, and publish a finite-size-aware simulation dossier.
• Can a Molecule Solve Its Own Electrons?: Validate supplied electronic integrals, solve generalized orbitals, construct density and Fock matrices, retain a complete damped SCF history, challenge guess and basis sensitivity, and publish a bounded RHF dossier.
• Can a Molecular Ensemble Support Its Free Energy?: Build a statistical-thermodynamics investigation from energy levels and Boltzmann populations through thermal observables, Metropolis sampling, autocorrelation and effective sample size, potentials of mean force, and a bounded free-energy dossier.
• What Does the Solvent Change?: Build and challenge a solvation calculation from continuum electrostatics and nonpolar terms through thermodynamic integration, exponential reweighting, uncertainty, thermodynamic-cycle closure, model sensitivity, and a complete solvation dossier.
• Which Molecules Should We Test Next?: Build a molecular-design workflow from candidate and descriptor contracts through transparent surrogate validation, Pareto tradeoffs, predictive uncertainty, applicability domains, diverse batch selection, and a decision dossier that proposes experiments without giving away a winning design.