Seven Clues to Ligand Noninnocence: The Metallocorrole Paradigm.

Noninnocent ligands do not allow an unambiguous definition of the oxidation state of a coordinated atom. When coordinated, the ligands also cannot be adequately represented by a classic Lewis structure. A noninnocent system thus harbors oxidizing (holes) or reducing equivalents (electrons) that are delocalized over both the ligand and the coordinated atom. To a certain degree, that is true of all complexes, but the phenomenon is arguably most conspicuous in complexes involving ligands with extended π-systems. The electronic structures of such systems have often been mischaracterized, thereby muddying the chemical literature to the detriment of students and newcomers to the field. In recent years, we have investigated the electronic structures of several metallocorrole families, several of which have turned out to be noninnocent. Our goal here, however, is not to present a systematic account of the different classes of metallocorroles, but rather to focus on seven major tools (in a nod to A. G. Cairns-Smith's ) that led us to recognize noninnocent behavior and subsequently to characterize the phenomenon in depth. (1) : For a series of noninnocent -triarylcorrole derivatives with different substituents X, the Soret maxima are typically exquisitely sensitive to the nature of X, red-shifting with increasing electron-donating character of the group. No such substituent sensitivity is observed for the Soret maxima of innocent triarylcorrole derivatives. (2) : Spin-unrestricted density functional theory calculations permit a simple and quick visualization of ligand noninnocence in terms of the spin density profile. Even for an = 0 complex, the broken-symmetry method often affords a spin density profile that, its fictitious character notwithstanding, helps visualize the intramolecular spin couplings. (3) : In principle, these two techniques afford experimental probes of the electronic spin density. (4) . Ligand noninnocence in metallocorroles is often reflected in small but distinct skeletal bond length alternations in and around the bipyrrole part of the macrocycle. In addition, for Cu and some Ag corroles, ligand noninnocence manifests itself via a strong saddling of the macrocycle. (5) . Unsurprisingly, the aforementioned bond length alternations translate to structure-sensitive vibrational marker bands. (6) . Noninnocent metallocorroles exhibit characteristically high reduction potentials, but caution should be exercised in turning the logic around. A high reduction potential does not necessarily signify a noninnocent metallocorrole; certain high-valent metal centers also undergo metal-centered reduction at quite high potentials. (7) . By focusing on a given element, typically the central atom in a coordination complex, X-ray absorption near-edge spectroscopy (XANES) can provide uniquely detailed local information on oxidation and spin states, ligand field strength, and degree of centrosymmetry. For metallocorroles, some of the most clear-cut distinctions between innocent and noninnocent systems have come from the K-edge XANES of Mn and Fe corroles. For researchers faced with a new, potentially noninnocent system, the take-home message is to employ a good majority (i.e., at least four) of the above methods to arrive at a reliable conclusion vis-à-vis noninnocence.