interhalogen compounds (ClF3, IF7)
If halogens are so hungry for electrons, what happens when two different halogens meet? They form molecules of each other — interhalogen compounds, where a larger, less electronegative halogen sits at the centre and one or more smaller, more electronegative halogens (usually fluorine) surround it. The general formula is XYn, with n being 1, 3, 5, or 7, and the central atom carries an odd, positive oxidation state.
Take chlorine trifluoride, ClF3. Chlorine is the central atom, oxidized to +3; three fluorine atoms bond out, and chlorine keeps two lone pairs, giving five electron domains and the famous bent T-shape (the lone pairs take the equatorial positions). It is one of the most violently oxidizing substances known — it ignites glass, sand, and asbestos, and was once studied as rocket fuel. At the extreme stands iodine heptafluoride, IF7: only the large iodine atom is big enough to hold seven fluorines around it, in a pentagonal-bipyramidal cage, with iodine in its highest oxidation state of +7. The pattern is clear — only the bigger central halogens can support high coordination numbers, because they have the room and the available higher oxidation states.
Interhalogens matter as some of the most powerful fluorinating and oxidizing agents in inorganic chemistry, useful for making fluorides and, like BrF3, as non-aqueous ionising solvents. They also illustrate VSEPR beautifully (count the bonds and lone pairs, predict the shape). One honest correction: the older story that a chlorine atom 'expands its octet using empty d orbitals' to bond five or seven atoms is now regarded as largely wrong — the bonding in these hypervalent molecules is better described by molecular-orbital pictures with three-centre interactions and highly polar, partly ionic bonds, with the d orbitals playing at most a minor role.
ClF3 has chlorine surrounded by three bonding pairs and two lone pairs; VSEPR puts both lone pairs in the equatorial plane, leaving a distorted T-shaped molecule rather than a flat triangle.
Counting bonds and lone pairs predicts the T-shape of ClF3 by VSEPR.
The 'expanded octet via d orbitals' story is now considered largely incorrect; hypervalent interhalogens are better explained by three-centre/MO bonding with very polar bonds.