Opening the workbench…
Two hydrogens and one oxygen snap to water at 104.5°. That angle is measured, not guessed. A bent 2p cartoon says 90°. A tetrahedral lone-pair cartoon says 109.5°. The molecule then sits between them. The geometry is a PDB string parsed by Three.js’s PDBLoader — atoms as spheres, CONECT records as cylinders.
A recipe book, not a calculator
Click or drag H, C, N, O, Na, or Cl onto the box. Known mixes (water, carbon dioxide, ammonia, methane, the diatomics, a gas-phase NaCl pair) build the classroom model. Anything else stays a pile. This page will not invent a mystery compound. It will also not run a quantum-chemistry engine.
The NaCl pair is not table salt. Solid salt is a lattice of many ions. The crystal page is where packings tile space. Space-filling mode uses van der Waals radii. Ball-and-stick then shrinks the atoms so the bond angle is the thing you see.
Questions the bench answers
Can I mix any two elements and get a molecule?
No. This bench is a recipe book of common classroom molecules. It parses them as PDB and then builds them with the real bond angles. It is not a quantum-chemistry engine. Unknown mixes therefore stay as a pile of atoms.
Is the NaCl here table salt?
It is the gas-phase Na–Cl pair (about 2.36 Å). Solid salt, however, is a 3D lattice of many ions. That is the FCC-related rock-salt structure. See the crystal-lattice page for how packings tile space.
Why 104.5° for water, not 90° or 109°?
A bent 2p picture suggests 90°. A tetrahedral lone-pair picture suggests 109.5°. The measured H–O–H angle is 104.5°, between those cartoons, because the lone pairs occupy more space than the O–H bonds.