Design and synthesis of multifunctional noninnocent ligands: electronic characterization and redox analysis with transition metals

Kyle D Spielvogel

doi:10.17077/etd.006232

Rising societal demand for fine and commodity chemicals routinely used in medicinal, agricultural and petroleum research has led to the proliferation and use of highly active metal catalysts that can offer unmatched rates of reaction and selective bond formation imperative to chemical design. However, many catalytic processes rely on the use of rare and expensive platinum group metals because they can more easily access oxidation states required for the two-electron exchange of oxidative addition and reductive elimination. The less expensive earth abundant metals prefer to undergo one electron exchanges that often impede their viability for catalytic applications. Recent efforts to engender platinum group type reactivity with abundant base metals have led to the development of reactive ligands that act noninnocently in catalytic reactions by assisting the metal in a chemical transformation. Noninnocent ligands can assist a metal either by acting as an electron reservoir and aiding in the transfer of electrons (redox noninnocent) or by actively participating in bond activation through collective coordination of small molecules with the metal (metal-ligand cooperativity). Previous designs of noninnocent ligands have primarily focused on installing only onenoninnocent function (chemical or redox) for a particular chemical transformation. An emerging area of interest for catalytic applications has been the combination of both functions into one ligand to synergistically address particularly challenging chemistries with small molecules like CO2. However, a fundamental understanding of how the multifunctional properties of a noninnocent ligand are affected when both functions are engaged simultaneously needs to be established to fully comprehend their potential use in catalysis. The research covered in this thesis was aimed at establishing a foundational understanding of the interplay between metal-ligand cooperativity (MLC) and redox activity in multifunctional noninnocent ligands bound to both precious and nonprecious metals. Initial investigations of the redox noninnocent character of our targeted multifunctionalligands were conducted by the synthesis, characterization, and electronic structure of nickel complexes containing triaryl MeNNNNMe and MeSNNSMe ligands derived from ophenylenediamine. The tetradentate ligands were investigated and compared to those in metal complexes with compositionally similar ligands to determine how ligand-centered redox properties change when redox-active flanking groups are replaced with redox-innocent NMe2 and SMe. A derivative of the MeSNNSMe ligand in which the phenylene backbone was replaced with ethylene was also prepared to interrogate the importance of o-phenylenediamine for ligandcentered redox events. Remarkably, several quasi-reversible ligand-centered redox waves were also observed despite the absence of the o-phenylenediamine subunit. Oxidizing the nickel complexes containing the o-phenylenediamine derived ligands with silver salts containing different counter anions (BF4-, OTf- and NTf2-) allowed the electrochemically generated complexes to be analyzed as a function of different oxidation states using single-crystal X-ray diffraction (XRD), EPR spectroscopy, and sulfur K-edge X-ray absorption spectroscopy. The experimental data are corroborated by DFT calculations and, together, they reveal how the location of unpaired spin density and electronic structure in the singly and doubly oxidized salts varies depending on the coordinating ability of the counteranions and exogenous ligands such as pyridine. The coordination chemistry of the o-phenylenediamine derived tetradentate ligandMeSNNSMe was then investigated with ruthenium to interrogate the potential added function of MLC and to understand how increased metal-ligand orbital mixing with a 4d metal affects the electronic structure of our ligand. Ruthenium complexes were initially targeted to investigate how MLC affected the redox properties of our ligand as square planar Ni2+ is electronically saturated and lacks available metal orbitals in the axial coordination sites to form additional interactions with small molecules. Deprotonation of H2(MeSNNSMe) with RuCl2(PPh3)3 resulted in a five coordinate, square pyramidal complex Ru(MeSNNSMe)(PPh3) where, upon addition of BH3·tetrahydrofuran (THF), resulted in the isolation of Ru[(μ-H)BH2](MeSNNSMe)(PPh3) with BH3 bound across a Ru-N bond in a MLC fashion. Exchanging the Ru starting material for RuCl3∙xH2O with H2(MeSNNSMe) in ethanol resulted in the six coordinate Ru(MeSNNSMe)Cl2 where single-crystal XRD revealed the ligand was fully oxidized to the o-diiminoquinone state. Subsequent reduction of the complex with 2 equiv. NaBH4 resulted in the dual MLC complex Ru[(μ-H)BH3]2(MeSNNSMe) with two BH3 units bound across both Ru-N bonds. Electrochemical investigations revealed that stepwise chemical engagement of the o-phenylenediamine N atoms attenuates ligand-centered redox activity. However, the results demonstrate how incorporating more than one multifunctional reactive site could be an effective strategy for maintaining redox noninnocence in ligands that are also chemically reactive and competent for MLC. Electronic structure investigations of Ru(MeSNNSMe)(PPh3) were performed in a similarmanner to the previous Ni complexes. Addition of acetonitrile (MeCN) to Ru(MeSNNSMe)(PPh3) resulted in coordination of MeCN to the vacant coordination site identified by single-crystal XRD to result in Ru(MeCN)(MeSNNSMe)(PPh3). Electrochemical investigations in THF revealed two fully reversible redox features similar to those observed for the ligand-centered redox activity of the Ni complexes. Similar results were observed when electrochemical data were collected in MeCN, albeit with slightly shifted potentials similar to reported complexes featuring different ancillary ligand donor properties. Oxidation of the complex in MeCN generated products that were analyzed as a function of oxidation state using single-crystal XRD, NMR and EPR spectroscopy. Collectively the data revealed that the first oxidation is primarily a ligand-centered process and the second oxidation is a mixed metal and ligand process where resonance forms contain character of both. Electrochemical investigations of Ru[(μ-H)BH2](MeSNNSMe)(PPh3) revealed that theMLC-bound BH3 was in equilibrium. Monitoring the electrochemical response as a function of BH3 concentration resulted in the gradual elimination of one redox feature. Modeling the equilibrium data permitted the determination of an MLC binding affinity for Ru(MeSNNSMe)(PPh3) that was in excellent agreement with DFT-derived binding constants. As MLC binding affinity could be determined as a function of electrochemical response, we sought to determine if our electrochemical approach was sensitive enough to detect differences in MLC binding associated with changes to the redox active and ancillary supporting ligands in square pyramidal Ru(MeSNNSMe)(PR3) complexes. Structurally similar derivatives of H2(MeSNNSMe) were synthesized featuring extended π conjugation of the backbone or flanking arms by incorporation of 2,3-diaminonaphthalene or quinoline, respectively. The influence of the ancillary phosphine on MLC were investigated by swapping PPh3 for PCy3. Synthesis and subsequent isolation of the MLC-bound BH3 complexes allowed for single-crystal XRD and NMR and IR spectroscopy studies that showed the strongest MLC interaction with the 2,3-diaminonaphthalene derived ligand. Electrochemical binding studies corroborated this result and were sensitive enough to quantify subtle differences in the MLC binding affinity of the other complexes. MLC studies with the ruthenium complexes and BH3 provided a foundationalunderstanding of the interplay between MLC and redox noninnocence, and showed that the binding affinity of our complexes need to be increased to bind less electrophilic and more catalytically interesting MLC substrates like CO2. To address this need, the coordination chemistry of the noninnocent ligands was investigated with base metals (Al, Ti, and V). It was postulated that the increased oxophilicity of these metals would aid in the MLC binding of CO2. Dehydrohalogenation of TiCl4(THF)2 with H2(MeNNNNMe) resulted in isolation of monomeric Ti(MeNNNNMe)Cl2, but this complex was not amenable to further reduction needed for MLC studies. Protonolysis of H2(MeNNNNMe) with V[N(SiMe3)2]2Cl(THF) resulted in the dimeric complex [V(MeNNNNMe)Cl]2 with bridging chlorides and a highly distorted ligand coordination. Treating this compound with CO2 in noncoordinating solvents yielded dark purple crystals of [(MeNNNNMe)VCl(μ-O)2C]2. XRD studies confirmed our first successfully isolated CO2 complex, albeit with insertion between the o-phenylenediamine nitrogen and vanadium bond instead of the MLC-type binding observed with BH3 in our Ru complexes. As iron complexes with noninnocent ligands have shown moderate reactivity in theelectrocatalytic reduction of CO2, we explored the coordination chemistry of H2(MeSNNSMe) and H2(MeNNNNMe) with iron. Initial synthetic attempts modeled after the Ru work resulted in diverse complexes with structures that were dependent on the number of protonated N atoms, ligand identity and Fe starting materials. However poor yields and ligand degradation in the presence of stronger bases prevented further use of these complexes. Protonolysis of Fe[N(SiMe3)2]2 with H2(MeSNNSMe) and H2(MeNNNNMe) resulted in dimeric complexes of [(MeSNNSMe)Fe]2 and [(MeNNNNMe)Fe]2 displaying ligand supported Fe-Fe interactions. Reactivity studies with small molecules (H2, CO2 and CO) yielded insoluble or unstable powders with the exception of [(MeSNNSMe)Fe]2 and CO. Two CO molecules were added to each iron center to yield the fivecoordinate square pyramidal complex Fe(κ3-MeSNNSMe)(CO)2 with one of the SMe groups not bound to the metal. As [(MeNNNNMe)Fe]2 did not result in isolable complexes with small molecules, the oxidation chemistry of the dimer was investigated. Homolysis of diphenyldisulfide and [(MeNNNNMe)Fe] resulted in the isolation of a five-coordinate square pyramidal complex (MeNNNNMe)Fe(SPh) with an axial phenylthiolate ligand. To determine if the ligand was involved in the oxidation process, we reduced the complex with cobaltocene. Comparison of the isolated complex salt [(MeNNNNMe)Fe(SPh)]CoCp2 against (MeNNNNMe)Fe(SPh) suggested a metal based reduction from FeIII ® FeII. These complexes are awaiting study with CO2 and other substrates. During the investigation of the coordination chemistry of H2(MeSNNSMe) with Ru and BH3,we found that BH3 reacted with the o-phenylenediamine backbone by concomitant loss of two equiv. of H2 to yield 1,3-bis(2-(methylthio)phenyl)-2,3-dihydro-1H-benzo[d][1,3,2]diazaborole, H(MeSBSMe), a precursor to an SBS pincer ligand. Almost all boryl centered pincer ligands feature flanking phosphine donor groups, and SBS complexes with Rh and Ir were prepared so their structures and reactivity could be compared to those with PBP ligands. Metalation of H(MeSBSMe) was conducted by facile B-H oxidative addition to common Rh(I) and Ir(I) starting materials, resulting in isostructural chloride bridged dimers of [(MeSBSMe)MH(μ-Cl)]2. Reactivity studies of the Rh complex with CO resulted in decomposition. However, reactions with the Ir dimer resulted in a monomeric CO complex that displayed interconversion of the pincer ligand coordination mode from meridional in the dimer to facial coordination. Anion exchange conducted with Li[N(SiMe3)2], resulted in monomeric (MeSBSMe)Ir(H)[N(SiMe3)2] and dimeric [(MeSBSMe)Rh]2(μ-H)[N(SiMe3)2] featuring a hydrido and boryl supported RhI®RhIII dative metal-metal bond. These latter MeSBSMe complexes were tested for catalytic alkane dehydrogenation transfer activity in the benchmark reaction of tert-butylethylene and cyclooctane, but unlike similar PBP complexes, they showed no appreciable reactivity for this transformation. However, the complexes were amenable to alkene isomerization of 1-hexene at relatively low temperatures (< 60 ºC).

Design and synthesis of multifunctional noninnocent ligands: electronic characterization and redox analysis with transition metals

Files and links (1)

Abstract

Details

Metrics