IntroductionThe design of specific magnetic interactions and properties in bulk molecular solids is the ambitious goal of much research at the present time. In order to do this, reliable structure–property relationships must be determined. This is complicated in purely organic systems: exchange interactions are weak, and overall magnetic moments are limited. McConnell described a structure–property model suggesting that overlap between sites with opposite signs of spin density (McConnell I model1) is a plausible, basic paradigm for designing molecular solids with extended magnetic behavior. However, a McConnell I-type analysis of over 50 organic systems showed that experimental magnetic behavior often eludes simple analysis.2Modest changes in inter-orbital distance, alignment, and relative orientation can lead to sizable changes in exchange strength, and even to reversal from antiferromagnetic (AFM) toferromagnetic (FM) behavior.3As a result, over-simplified McConnell I-type analyses are not necessarily useful for structure–property analysis. Exchange behavior in systems that have multiple close contacts between larger spin density sites can be difficult to interpret. For this reason, systems with a limited number of significant spin site interactions are worth studying, because they allow more straightforward structure–property analysis.In 1994, Hosokoshiet al.described the magnetic behavior of 2-(3′,5′-difluorophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazol-1-oxyl 3-oxide (F2PhNN).4Half of the F2PhNN spins were antiferromagnetically exchange coupled as dimers, while the others acted in an isolated, paramagnetic fashion. They rationalized the behavior in terms of a close approach between two of the four molecules of the F2PhNN unit cell. Details of the crystal structure and of the AFM exchange mechanism were not given.As part of our efforts to test structure–property models of exchange in organic radicals, we re-determined the crystal structure of F2PhNN and carried out a detailed analysis of the close contacts between the major spin density sites. The results presented below exemplify the utility of spin density overlap models – such as those of McConnell,1Hoffmann,3aand Yamaguchi3c,d– in cases where only a few strong exchange interactions are present.