The Radical Anions and Trianions of 1,8-Diphenylnaphthalene and of Some of its Derivatives Containing a Cyclophane Substructure

The radical anions of 1,8-diphenylnaphthalene ( 1 ) and its decadeuterio-(D₁₀- 1 ) and dimethyl-( 2 ) derivatives, as well as those of [2.0.0] (1,4)benzeno(1,8)naphthaleno(1,4)benzenophane ( 3 ) and its olefinic analogue ( 4 ) have been studied by ESR and ENDOR spectroscopy, At a variance with a previous report, the spin population in documentclass{article}pagestyle{empty}begin{document}$ rm {1}^{-kern-4pt {.}} $end{document} and documentclass{article}pagestyle{empty}begin{document}$ rm {2}^{-kern-4pt {.}} $end{document} is to a great extent localized in the naphthalene moiety. A similar spin distribution is found for documentclass{article}pagestyle{empty}begin{document}$ rm {3}^{-kern-4pt {.}} $end{document} and documentclass{article}pagestyle{empty}begin{document}$ rm {4}^{-kern-4pt {.}} $end{document}. The ground conformations of documentclass{article}pagestyle{empty}begin{document}$ rm {1}^{-kern-4pt {.}} $end{document}-documentclass{article}pagestyle{empty}begin{document}$ rm {4}^{-kern-4pt {.}} $end{document} are chiral of C₂ symmetry. For documentclass{article}pagestyle{empty}begin{document}$ rm {1}^{-kern-4pt {.}} $end{document}, an energy barrier between these conformations and the angle of twist about the bonds linking the naphthalene moiety with the phenyl substituents were estimated as ca. 50 kJ/mol and ca. 45°, respectively. The radical trianions of 1 , D₁₀- 1 , and 2 , have also been characterized by their hyperfine data. In documentclass{article}pagestyle{empty}begin{document}$ rm {1}^{3-kern-4pt {.}} $end{document} and documentclass{article}pagestyle{empty}begin{document}$ rm {2}^{3-kern-4pt {.}} $end{document}, the bulk of the spin population resides in the two benzene rings so that these radical trianions can be regarded as the radical anions of ‘open-chain cyclophanes’ with a fused naphthalene π-system bearing almost two negative charges. The main features of the spin distribution in both documentclass{article}pagestyle{empty}begin{document}$ rm {1}^{-kern-4pt {.}} $end{document} and documentclass{article}pagestyle{empty}begin{document}$ rm {1}^{3-kern-4pt {.}} $end{document} are correctly predicted by an HMO model of 1 .