The Crystal Structures of α‐ and β‐F₂ Revisited

Abstract

The crystal structures of a-F 2 and b-F 2 have been reinvestigatedu sing neutron powder diffraction. For the low-temperature phase a-F 2 ,w hich is stable below circa 45.6 K, the monoclinic space group C2/c with lattice parameters a = 5.4780(12), b = 3.2701 (7), c = 7.2651(17) , b = 102.088(18)8, V = 127.26(5) 3 , mS8, Z = 4a t1 0Kcan now be confirmed.T he structure model was significantly improved, allowed forthe anisotropic refinement of the Fatom, anda n FÀFbond length of 1.404(12) was obtained, which is in ex-cellent agreement with spectroscopic data and high-level quantum chemical predictions.T he high-temperature phase b-F 2 ,s table between circa 45.6 Ka nd the melting point of 53.53 K, crystallizes in the cubic primitive space group Pm3 n with the lattice parameter a = 6.5314(15) , V = 278.62(11) 3 , cP16, Z = 8, at 48 K. b-F 2 is isotypic to g-O 2 and d-N 2 .T he centres of gravity of the F 2 molecules are arranged like the atoms in the Cr 3 Si structure type.[a] Dr.S.I.I vlev,Dr. M. Conrad, Prof.

The Crystal Structure of α‐F2: Solving a 50 Year Old Puzzle Computationally

“…For a-F 2 ,P BE predicts only av ery small bond elongation from 1.413 for af ree F 2 -molecule to 1.416 in the F 2 crystal, which agrees with newly reported data from neutron diffraction experimentsb yI vlev et al [7] Periodic singlepoint LMP2 calculations based on this refined structure yield a cohesive energy of À9.19 kJ mol À1 ,e xceedingt he experimental value by 10 %. Adding incremental corrections at the CCSD(T) level, extrapolated to the complete basis set limit (CBS), as well as considering long range electron correlation effects and zero point energy correction (see Supporting Information for details) yields our best estimate for the cohesive energy of a-F 2 of À8.72 kJ mol À1 .I vlev et al report av alue of À3.9 kJ mol À1 from DFT calculations at PBE0/TZVP level with D3(BJ + ABC) dispersion correction, based on their newly-obtained crystal structure for a-F 2 , [7] which appearst ob ei nt he right ballpark but too low. Figure 2i llustrates our assessment of different methods for the cohesive energies of solid-F 2 crystal structures in C2/c and Cmca space group symmetry.…”

“…After assessing a number of different DFT functionals, PBE was chosen since it describes the F 2 molecules best (see Table ). For α‐F 2 , PBE predicts only a very small bond elongation from 1.413 Å for a free F 2 ‐molecule to 1.416 Å in the F 2 crystal, which agrees with newly reported data from neutron diffraction experiments by Ivlev et al . Periodic single‐point LMP2 calculations based on this refined structure yield a cohesive energy of −9.19 kJ mol −1 , exceeding the experimental value by 10 %.…”

“…Ivlev et al. report a value of −3.9 kJ mol −1 from DFT calculations at PBE0/TZVP level with D3(BJ+ABC) dispersion correction, based on their newly‐obtained crystal structure for α‐F 2 , which appears to be in the right ballpark but too low.…”

“…Ivlev et al confirmed both assignments in very recent state-of-the-art neutron diffraction experiments. [7] Interestingly,t he crystal structures of fluorine differ from the orthorhombic Cmca structure, whichi so bserved for solid chlorine, bromine and iodine. Both, the a-form of fluorine and the crystal structures of chlorine, bromine and iodine feature layers of hexagonally close-packed X 2 molecules (see Figures S1 and S2 in the Supporting Information, or Figure 4below).…”

“…Ivlev et al. confirmed both assignments in very recent state‐of‐the‐art neutron diffraction experiments …”