New Insights into the Chemistry of Imidodiphosphinates from Investigations of Tellurium-Centered Systems

Dichalcogenido-imidodiphosphinates, [N(PR(2)E)(2)](-) (R = alkyl, aryl), are chelating ligands that readily form cyclic complexes with main group metals, transition metals, lanthanides, and actinides. Since their discovery in the early 1960s, researchers have studied the structural chemistry of the resulting metal complexes (where E = O, S, Se) extensively and identified a variety of potential applications, including as NMR shift reagents, luminescent complexes in photonic devices, or single-source precursors for metal sulfides or selenides. In 2002, a suitable synthesis of the tellurium analogs [N(PR(2)Te)(2)](-) was developed. In this Account, we describe comprehensive investigations of the chemistry of these tellurium-centered anions, and related mixed chalcogen systems, which have revealed unanticipated features of their fundamental structure and reactivity. An exhaustive examination of previously unrecognized redox behavior has uncovered a variety of novel dimeric arrangements of these ligands, as well as an extensive series of cyclic cations. In combination with calculations using density functional theory, these new structural frameworks have provided new insights into the nature of chalcogen-chalcogen bonding. Studies of metal complexes of the ditellurido ligands [N(PR(2)Te)(2)](-) have revealed unprecedented structural and reaction chemistry. The large tellurium donor sites confer greater flexibility, which can lead to unique structures in which the tellurium-centered ligand bridges two metal centers. The relatively weak P-Te bonds facilitate metal-insertion reactions (intramolecular oxidative-addition) to give new metal-tellurium ring systems for some group 11 and 13 metals. Metal tellurides have potential applications as low band gap semiconductor materials in solar cells, thermoelectric devices, and in telecommunications. Practically, some of these telluride ligands could be applied in these industries. For example, certain metal complexes of the isopropyl-substituted anion [N(P(i)Pr(2)Te)(2)](-) serve as suitable single-source precursors for pure metal telluride thin films or novel nanomaterials, for example, CdTe, PbTe, In(2)Te(3), and Sb(2)Te(3).


CONSPECTUS
Unprecedented structural and reaction chemistry has also been revealed in studies of metal complexes of the ditellurido ligands [N(PR2Te)2] − . The large tellurium donor sites confer greater flexibility leading, in some cases, to unique structures in which the tellurium-centered ligand bridges two metal centers. The relatively weak P-Te bonds facilitate metal-insertion reactions (intramolecular oxidative-addition) to give new metaltellurium ring systems for some group 11 and 13 metals. From a practical perspective, certain metal complexes of the isopropyl-substituted anion [N(P i Pr2Te)2] − serve as suitable single-source precursors for pure metal telluride thin films or novel nanomaterials, e.g. CdTe, PbTe, In2Te3, Sb2Te3, which have potential applications in solar cells, thermoelectric devices and telecommunications.

Introduction
Chalcogen-centered chelating ligands of the type [N(PR2E)2] − (E = O, S, Se; R = alkyl, aryl) [i.e. dichalcogenido(imidodiphosphinates) (PNP)] have a long and venerable history that dates back to the 1960s. 1 In the intervening years many main group and transitionmetal complexes have been characterized. 2 Some early potential applications of these complexes included their use as lanthanide shift reagents, in luminescent materials, or in metal-extraction processes. 2a More recently, renewed interest in this class of compounds has been invigorated by the findings of O'Brien and co-workers that certain metal complexes of the isopropyl derivative [N(P i Pr2Se)2] − are suitable single-source precursors for the production of thin semiconducting films of binary metal selenides, e.g. MSe (M = Zn, Cd, Hg), M2Se3 (M = Ga, In, Bi) and PbSe, by using LP-(low-pressure) or AA-(aerosol-assisted) CVD (chemical vapor deposition) techniques. 3 The solvothermal generation of CdSe quantum dots has also been accomplished. 4 In addition to the synthetic challenge of making analogous Te-containing ligands that might display novel chemistry, the prospect of generating novel single-source precursors to metal tellurides excited our interest in this field.

Synthesis
The neutral precursors HN(PR2E)2 (E = S, Se; R = Ph, i Pr) are readily generated by the direct reaction of the corresponding P III /P III systems HN(PR2)2 with elemental sulfur or selenium. 5 In the case of tellurium, however, this oxidation is limited to the formation of the yellow monotellurido derivative Te i Pr2PNP(H) i Pr2 (1a), 6 which is isolated as the P-H tautomer in 81 % yield (Scheme 1); the phenyl derivative HN(PPh2)2 is unreactive towards tellurium. 7 In 2002 we demonstrated that this lack of reactivity can be circumvented by generating the anion [N(PPh2)2] − prior to the reaction with elemental tellurium. In this way sodium salts of the ditellurido ligands [N(PR2Te)2] − (2a, R = i Pr; 8 2a′, R =Ph; 7 2a′′, R = t Bu 9 ) are obtained in good yields (Scheme 1). This protocol can be adapted for the synthesis of the mixed chalcogen ligands [N(P i Pr2Te)(P i Pr2E)] − (E = S, Se) (Scheme 1). 10 The best procedure for obtaining the Te/S ligand, as the Li derivative 2c, in high purity (99 %) involves the in situ deprotonation of the monotelluride (1a) with n-butyllithium followed by reaction with sulfur in THF. A similar methodology is employed for the optimal synthesis of the Te/Se reagent 2b (97 % purity) from the monoselenide (1b). The mixed chalcogen anions [N(P i Pr2Te)(P i Pr2E)] − (E = S, Se) can also be obtained as ion-separated cobaltocenium salts by reduction of the corresponding cations (see Section 3.2) with cobaltocene. 11 The "metallation-first" approach to these novel tellurium-containing ligands opened the door to a comprehensive investigation of their fundamental chemistry, as well as potential applications of metal complexes as single-source precursors to metal tellurides in the form of thin films or nanomaterials.

Redox Behavior
3.1. One-electron Oxidation. In our initial studies we observed that yellow solutions of the monotelluride 1a become red upon exposure to air and a few crystals of the unusual ditelluride (TeP i Pr2N i Pr2PTe-)2 (4a) were isolated from the red solution and structurally characterized. 6 This intriguing transformation prompted us to undertake a systematic examination of the redox behavior of the monoanions [N(PR2E)2] − (E = S, Se, Te; R = i Pr, t Bu). The one-electron oxidation of their sodium salts with iodine produces the dimers (EPR2NR2PE-)2 either in the form of dichalcogenides (DCs) 4a, 4b, 4b′ and 4c′ or as the spirocyclic contact ion pair (CIP) 5a ′ . 9 The disulfide 4c (E = E′ = S; R = i Pr) could not be obtained owing to hydrogen-abstraction reactions. 9 The availability of the mixed chalcogen anions in 2b and 2c provoked the interesting question -which chalcogen-chalcogen bond will be formed preferentially upon one-electron oxidation, E- towards the more electropositive tellurium atom when E = S, Se resulting in stronger Te-Te overlap. 11 The elongation of the central chalcogen-chalcogen bond in the diselenides 4b and 4b ′ is ca. 6 % and that in the disulfide 4c′ is only 2 %, suggesting better overlap of the two radical SOMOs for the lighter chalcogens We note, however, that the markedly different conformations of acyclic DCs is likely to be a contributing factor. 9 DFT calculations of the relative energies of the two structural isomers observed for the symmetrical dimers (EPR2NR2PE-)2 (DC and CIP) as a function of (a) the chalcogen and (b) the R group reveal interesting trends. 9 For both the i Pr and t Bu series the stability of the CIP increases relative to that of the DC upon going from sulfur to tellurium. However, the CIP is predicted to be significantly more stable in only one case (R = t Bu, E = Te), consistent with the observed isolation of 5a ′ . However, the differences is coordinated in an E-monodentate fashion to the cyclic cation [N(P i Pr2Te)2] + . 12 DFT calculations indicate that the observed bidentate coordination mode is more stable than the monodentate isomer by 35-45 kJ mol −1 for the tert-butyl derivatives 5a ′ -c ′ , but preferred by only 5-10 kJ mol −1 for the iso-propyl analogues 5b and 5c; crystal packing forces may be responsible for the observed formation of monodentate arrangement in the latter case. 12 The trends in Te-Te bond lengths along the series 5a ′ (2.981 Ǻ), 5b ′ (2.922 Ǻ) and 5c ′ (2.868 Ǻ) indicate that the incipient cation in the CIP is more well-developed upon going from Te to S in the counter-anion. Comparison of the bond orders of the central chalcogen-chalcogen bonds reveals that this trend is determined by the strength of the anion-cation interaction. Thus, the Te-Te bond order increases as the E-Te bond order decreases. 12 Consistently, DFT calculations predict that the extent of electron transfer is 0.20 e − for 5a ′ , 0.39 e − for 5b ′ , and 0.50 e − for 5c ′ . Indeed, symmetrical cations of the type [N(P i Pr2E)2] + are readily generated as the surprisingly air-stable iodide salts 6a and 6b by oxidation of the corresponding anions with one equivalent of I2 (Scheme 2). 13 The mixed chalcogen systems 6c and 6d are prepared in a similar manner. 11 The hexafluoroantimonate salts 7a,b are produced by metathesis of the corresponding iodide salts with Ag[SbF6] (Scheme 2). 14 The salts 6a and 6b are comprised of a five-membered cyclic cation [N(P i Pr2E)2] + and an iodide counterion that interacts with one of the chalcogens to form an infinite chain structure (Figure 2). 13 By contrast, the tert-butyl derivatives [N(P t Bu2E)2]I (E = Se, Te) are dimeric with close Se-Se and Te-Te contacts, while the sulfur system [N(P t Bu2S)2]I3 is an ion-separated monomer with a triiodide counterion. 9 In the mixed chalcogen systems 6c and 6d the iodide counterion interacts preferentially with the tellurium center. 11 The Te-I interaction is much stronger with the mixed chalcogen cations (6c and 6d) than that in the ditellurido cation (6a) to the extent that the former are essentially monomeric in the solid state.
Simple electron-counting rules predict that the cyclic cations [N(PR2E)2] + are 6 πelectron systems. DFT calculations confirm this prediction, but reveal that the net π-bond order within the five-membered ring is close to zero. The three highest occupied orbitals are indeed π-type orbitals ( Figure 3). However, the bonding effect of the E-E π-bonding orbital (HOMO−2) is essentially cancelled by the double occupation of the HOMO, which is the E-E π * -antibonding orbital. The third occupied π-orbital (HOMO−1) is a primarily non-bonding nitrogen-centered orbital.
The stronger iodide-chalcogen interaction in the mixed chalcogen salts 6c and 6d produces a more pronounced elongation of the chalcogen-chalcogen bonds in the cyclic cations (6c, 8 %; 6d, 12 %) 11 than that observed for the Te-Te bond in 6a (4 %). 13a The latter is attributed to the donation of electron density from a lone pair on the iodide counterion into the Te-Te σ * orbital (LUMO) (Figure 3) of the symmetrical cation in The NMR spectra of tellurium-or selenium-containing imidodiphosphinates have provided an initial indication of unexpected structures. For example, the 31 P NMR spectra on the monochalcogenides 1a and 1b showed one resonance with 1 J(PE) values consistent with terminal a P=E bond and a second resonance which revealed 1 J(PH) = 440-445 Hz, signifying the formation of the P-H tautomer. 6 The NMR spectra of the dimeric structures shown in Chart 1 are also revealing. At room temperature the 31 P NMR spectra of the dichalcogenides (DCs), e.g. 4a (E = E′ = Te; R = i Pr) exhibit a broad, unresolved resonance indicative of a fluxional process. At low temperatures a pair of mutually coupled doublets is resolved with 1 J(PTe) = 1500 and 1026 Hz, consistent with the solid-state structure. 6 By contrast, the observation of four resonances in the low temperature 31 P NMR spectrum of the tert-butyl derivative 5a′ was a signal of a different structure, subsequently shown to be a contact ion pair (CIPs, Chart 1). 9 31 P NMR spectra are also diagnostic of the purity of the Li derivatives of mixed chalcogen ligands, 2b and 2c, which exhibit characteristic satellite peaks, since the presence of the corresponding symmetrical ligands is readily detected. 10 39 The reason for the preferential formation of the heavier chalcogenide from these SSPs has not been established.

Conclusions
In summary, comprehensive investigations of the redox behavior and coordination complexes of anionic tellurium-centered imidodiphosphinates have provided new insights into the chemistry of this well-established class of inorganic ligand. One-electron oxidation provides a variety of dimers whose structures are influenced by the nature of the chalcogen as well as the organic substituent on phosphorus; two-electron oxidation generates novel cyclic cations. In conjunction with DFT calculations, the structural data for these dimers and cations have enhanced our understanding of chalcogen-chalcogen bonding. Significant differences in the coordination behavior of the tellurium-centered ligands compared to that of their lighter chalcogen analogues was observed as a result of (a) the larger size of tellurium and (b) the weakness of P-Te bonds. The former is manifested in the tendency for the tellurium ligands to adopt a doubly bridging coordination mode leading to unprecedented structures and, in the case of coinage metals, the possibility of metallophilic interactions. The lability of P-Te bonds is demonstrated by the occurrence of tellurium-transfer processes (intramolecular oxidative additions) that generate novel metal-tellurium rings. It is also evident in the use of metal complexes as single-source precursors to thin films of binary metal tellurides. The majority of our work has been carried out on ligands with isopropyl (or tert-butyl) substitutents on phosphorus.
However, preliminary investigations presage that significantly different chemistry will be observed for the phenyl-substituted analogues. 13b