High spin–low spin crossover and antiferromagnetic interactions in tris(1-pyrrolidinecarbodithioato)iron (III) and the 4-morpholine (FeM) and dibutyl analogs, effect of recrystallization solvent, and crystal structure of FeM⋅nitrobenzene

Abstract

High sensitivity magnetic susceptibility determinations, especially in the range 1.2–4.2 K on pure and dilute tris(pyrrolidinecarbodithioato)iron(III) (FeP) in its high spin form, show that a maximum at about 2 K is caused by antiferromagnetic interactions. The analogous chromium(III) complex does not exhibit significant antiferromagnetism compared to that of the iron complex, and it is likely that the upper e electrons possessed by the iron and not by the chromium are responsible for the bulk of the antiferromagnetism. As the iron atoms are about 9 Å apart in discrete molecules, the antiferromagnetic interactions presumably occur between unpaired spins delocalized on to the ligands of adjacent molecules. This is in keeping with NMR evidence that spin delocalization is greater in the iron(III) than in the chromium(III) complex. When diluted with large amounts of the cobalt(III) analog (CoP), FeP exhibits a spin state equilibrium. Thus, the structure of the FeP molecule is modified slightly (presumably with shortening of the Fe–S bond) to approach that of the CoP host lattice, which has a shorter metal–sulfur bond. The previous history of the samples of ferric dithiocarbamate complexes is shown to be far more important than had previously been suspected: When crystallized from benzene, FeP exhibits a high spin–low spin equilibrium, in constrast with the pure high spin behavior of the complex when not crystallized from benzene. The effect of adding 7% of benzene to the lattice is much greater than that of adding 50% of CoP. The dibutyl analog shows similar effects. The tris(4-morpholinecarbodithioato-S,S’)iron(III) complex FeM is shown, by single crystal x-ray data, to contain short Fe–S bond lengths (average 2.353 Å) when recrystallized from nitrobenzene. This indicates that the complex is principally low spin, in keeping with the observed magnetism and with the general strong solvent effect on the spin state. It is now proposed that the difference in Fe–S bond lengths between FeP crystallized from chloroform and FeP from benzene (the reverse of the expected differences) is due to experimental error. Crystal data for FeM⋅nitrobenzene: space group P21/c, Z=4, a=9.713(3) Å, b=31.419(8) Å, c =9.718(2) Å, β=105.04(2) °, V=2864 Å3, R=3.3%, 2712 reflections.