Welcome, Nietzschewoman!
Solids are complicated and you deserve all our encouragements on your path. While questions like "what makes bonds" are more or less answerable, "what makes semiconductors" can't be told from simple reasons. I regret it, but blame for that whoever or whatever created our Universe, not me.
==========
Strictly speaking, and
in solid state theory's wording, metals have no bandgap at all. In case you were influenced by models with atoms put on a line and a bandgap appearing necessarily then: remember that semiconductors use to be 3D objects, where the energy-versus-momentum relation depends on the direction, so the energy of the bands in varied directions can easily (and they do) overlap. And even in one direction, different bands can overlap their energies at different momenta.
You can admire true band diagrams (... though I'm not sure how much was measured in detail - more probably computed in detail and checked at a few strategic points) there
http://www.ioffe.ru/SVA/NSM/Semicond/http://www.ioffe.ru/SVA/NSM/Semicond/Si/bandstr.htmlhttp://www.ioffe.ru/SVA/NSM/Semicond/AlN/bandstr.html second diagram
https://www.researchgate.net/figure/269775630_fig1_Figure-2-Calculated-band-structure-of-aluminium-along-the-symmetry-axes-Reprinted-frommetals tend to be more complicated. And these are simplified diagrams along a few directions: only E versus simultaneous k
x, k
y and k
z would make sense, which is impossible to draw.
As a sidenote, components use presently "wide-gap semiconducting" materials that were defined as "insulators" three decades ago.
==========
A few solids have a
bandgap small enough that even solid state scientists may call them "metal", in the sense that room temperature suffices to put many mobile carriers in the conduction and valence bands and achieve metal-like conductivity. Grey tin is one example
https://en.wikipedia.org/wiki/List_of_semiconductor_materialsFor chemists, a metal has at least one basic oxide. The chemists' definition results from single atom properties, while for solid state, it's a collective property. Both definitions overlap almost. Grey tin is one exception.
==========
You wrote "two atoms". I misuse it as an excuse to put that two atoms don't make a solid nor a semiconductor. This needs millions of atoms.
Just to tell that solid properties don't result easily from atom properties:
- Solid hydrogen becomes a solid if putting enough pressure on it.
- Tin is a metal in one crystal form and a semi-metal in an other (grey).
- The band diagrams depend fundamentally on the crystal structure, SiC being one example
http://www.ioffe.ru/SVA/NSM/Semicond/SiC/bandstr.html - Si and Ge have different band structures, with conduction band minima in <100> and <111> directions, despite having both s2p2 orbital filling and diamond crystal structure, so the lower (non-binding) atomic orbitals matter too.
There are few general trends, supposedly with exceptions: lighter atoms tend to make bigger bandgaps, and combining atoms of different electronegativity tends to increase the bandgap.