I’d appreciate ‘only’ the experts to answer.

I’m facing not only a literature full of discrepancies, and dealing with those discrepancies, in the quantitative characterization ‘covalency and ionicity’ in solids but also confusion as to the existence of covalency alongside ionicity.

Theoretical and experimental studies have reported conflicting quantitative results about how covalent and ionic a solid such as α-Al2O3 is, the main reason being lack of a unique quantitative definition, and therefore lack of precise measurements, of electronegativity. Understandable.

What confuses me is a lack of a clear distinction in the literature between a ‘distortion of geometry’ caused by covalent contribution in a mainly ionic solid and other types of distortion such as ‘symmetry requirement of the crystal, Jahn-Teller effect, magnetic interactions, metal-metal bonding’ etc.

As an example, consider the series: α-Al2O3, Cr2O3, α-Fe2O3. All three are isostructural (R−3c) metal oxides with only one octahedral metal site. We have ‘the same’ distortion of MO6 in all three cases but to various extents. In α-Al2O3, the small Al3+ ion located between the two O closed-packed layers is shifted from the center towards the upper 3 O2− ions creating two different Al–O distances (and different angles), with the smaller distance being 94% of the bigger. The same ratio is 97.7% for Cr2O3 and 92% for α-Fe2O3.

Now, let’s exclude what effect doesn’t belong in here. Jahn-Teller effect is out in all three cases since Al has no d orbitals, and the d orbitals in Cr3+ and Fe3+ are evenly filled with electrons as (t2g)3(eg)0 and (t2g)3(eg)2, respectively. Magnetic interaction doesn’t exist in alumina (we ignore the nuclei), neither does the metal-metal type. We do have magnetic interactions (and possibly metal-metal?) in the other two cases.

So, what really causes this significant distortion in α-Al2O3 which is the same but bigger than in Cr2O3 and smaller than in α-Fe2O3?

Sometimes the literature would speak about covalent contribution contracting the distance in one direction and other times it would discuss it simply in terms of crystal requirements lowering its symmetry even when there are no d-orbitals. Almost like covalency-caused distortion and purely symmetry requirements are equivalent, if I have not misunderstood it. Like chemist vs. physicist perceptions.

A clarification will be appreciated (if possible with further examples).