# Why is the distance between the atoms shorter in the dicarbon anion compared to the cation?

In general, the anions have bigger ionic radii than their corresponding neutral atoms, and the cations have smaller ionic radii then their corresponding neutral atoms. How come that does not apply to $\ce{C2-}$ and $\ce{C2+}$?

• What other anion/cation are you talking about? And are they mono-atomic or diatomic? – SteffX May 29 '17 at 15:23
• The distance between two C atoms has nothing to do with the ionic radii in the first place. It is a covalent bond. – Ivan Neretin May 29 '17 at 15:23
• @IvanNeretin how do I then determine the distance between the C-atoms? or at least which on is bigger? – DUDEofDK May 29 '17 at 17:38
• You consider the molecular orbitals and deduce the bond order. The higher it is, the shorter is the bond. – Ivan Neretin May 29 '17 at 18:51
• I assume you are concerned with the diatomic molecules/ions; could you provide some data for your claim? – Martin - マーチン Jun 1 '17 at 11:28

Looking up in a current compilation of bond lengths in the International Tables for Crystallography (IT), apparently there is indeed not so much deviation in the distance between two adjacent carbon atoms in a $\ce{C(sp)#C(sp)}$ pattern:
Some hints to render this excerpt easier to access. The IT indeed discern different bond types (13 for all combinations of $\ce{C-C}$, and 625 for all organic / organometallic compounds), and consequently, the source offers separate listings for, for example, $\ce{C(sp^3)-C(sp^2)}$ or $\ce{C(sp^2)-C_{ar}}$, too (first column). The second column is about the structure around the pattern in question highlighted in bold. For the $IT$, $\ce{C_{ar}}$ explicitly represents an aryl C in an six-membered ring; $\ce{C^{\#}}$ any $\ce{C(sp^3)}$ - linearly substituted, or not.
The third, fourth and fifth column are the unweighted sample mean $d$ in [Å], sample median $m$, and sample standard deviation $\sigma$, followed by the lower and upper quartile, $q_l$ and $q_u$, as well as the number of observations $n$. Overall, the compilation is based on a subset of 10 324 X-ray and neutron diffraction data found in the Cambridge Crystal Database by September 1985, judged to be "good enough" (the source offers additional insight about the criteria applied).