Coordination and electronic characteristics of a nitrogen heterocycle pincer ligand

Coordination and electronic characteristics of a nitrogen heterocycle pincer ligand

Inorganica Chimica Acta 451 (2016) 82–91 Contents lists available at ScienceDirect Inorganica Chimica Acta journal homepage: www.elsevier.com/locate...

2MB Sizes 0 Downloads 53 Views

Inorganica Chimica Acta 451 (2016) 82–91

Contents lists available at ScienceDirect

Inorganica Chimica Acta journal homepage: www.elsevier.com/locate/ica

Research paper

Coordination and electronic characteristics of a nitrogen heterocycle pincer ligand Brian J. Cook a, Chun-Hsing Chen a, Maren Pink a, Richard L. Lord b,⇑, Kenneth G. Caulton a,⇑ a b

Indiana University, Department of Chemistry, 800 E. Kirkwood Ave., Bloomington, IN 47405, USA Grand Valley State University, Department of Chemistry, Allendale, MI 49401, USA

a r t i c l e

i n f o

Article history: Received 8 June 2016 Received in revised form 1 July 2016 Accepted 4 July 2016 Available online 5 July 2016 Keywords: Pincer Iron Redox active ligand Pyrazole Noninnocent ligand

a b s t r a c t The Fe(II) coordination chemistry of bis(pyrazole-3-yl)pyridine ligands with both proton or methyl substituents on pyrazole nitrogen are investigated, including the willingness of the ligand to undergo redox change. Protons on the pyrazole nitrogen promote intermolecular hydrogen bonding and lead to redox irreversibility; N methylation of those nitrogens eliminates those intermolecular interactions and leads to reversible outer-sphere reducibility. The resulting anion radical of the N-methylated ligand has more spin in the pyridine moiety than in the pyrazolyl pincer ligand arms, but also detectably delocalized into the electron withdrawing pyrazolyl pincer ligand arms; EPR and density functional calculations assist in characterizing the ligand radical anion, as its potassium complex. Ó 2016 Published by Elsevier B.V.

1. Introduction

2. Results

We are attracted to ligands which connect two identical redox active heterocycles to a central linker to create a redox active pincer ligand. Perhaps the best known pincer with these characteristics is that which links two ortho phenylenediamine rings via a central amide bridge, and such chemistry has been actively studied [1] (A, Scheme 1). We are interested in a pincer with rings which are more difficult to reduce, for their resulting higher reducing power, and thus considered pyrazoles. Pyrazoles are heterocycles which combine a monoazabutadiene fragment with an electron donating amine functionality (B and C, Scheme 1). These have been explored as pincer arms simply for mer tridentate character [2–6], but never with an eye towards their redox activity. The connectivity between pyrazole and pyridine can have large impact on inter-ring communication, and connection via pyrazole nitrogen (pyrazol-1-yl, B) has been much more investigated than the carbon-connected pyrazol-3-yl isomer (C, Scheme 1) which we study here [7,8]. We describe here the characteristics of several iron complexes of this pyrazol-3-yl type C, both with and without reactive proton functionality on the pyrazole ring (substituent R), as a first step towards probing this ligand class’ redox activity [9,10].

2.1. Ligand electronic character

⇑ Corresponding authors. E-mail address: [email protected] (K.G. Caulton). http://dx.doi.org/10.1016/j.ica.2016.07.011 0020-1693/Ó 2016 Published by Elsevier B.V.

LR (Scheme 2; superscript indicates substituent on each pyrazole amine nitrogen) is comprised of two heterocycles: pyridine and pyrazole. Is pyrazole an electron donating or an electron withdrawing substituent on pyridine? It is valuable to dissect nitrogen donor ligands into two functionalities according to whether they are high oxidation state (imine R2C@NR0 ) or low oxidation state (amine R2NH). Pyrazole has, as ring elements, an electron-withdrawing imine, an electron-donating amine, and a vinyl component. The net outcome of these competing oxidized and reduced nitrogen contributors, conjugated via the vinyl component, leaves uncertain whether a pyrazole acts as an electron rich- or electron poor-substituent. To better understand which part of LR should accept the electron upon reduction, we performed a computational study of the electron affinities of some nitrogen heterocycles. The CBS-QB3 model [11,12] yields reliable electron affinities (to within 1 kcal/mol), the energy change upon removing an electron from the anionic species. For pyridine (Scheme 3) excellent agreement with experiment is seen (0.61 eV calculated vs. 0.62 eV experimental) [13]. Based on these agreements and the average electron affinity error of 1 kcal mol1 benchmarked by Petersson and co-workers [11,12], we felt confident interpreting these results. As seen in Scheme 3, the electron affinity of all of the heterocycles is negative,

83

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91 t

Bu

t

Bu

N H pyrrole -1.84 eV

NR

N

N

tBu

N

N1

3

N

NR

N pyridine -0.61 eV

N N N N H Me pyrazole 1-methylpyrazole -1.50 eV -1.32 eV

Scheme 3. Electron affinities of five- and six-membered nitrogen-containing heterocycles calculated at the CBS-QB3 level of theory.

N N R

t Bu

A

C

B Scheme 1. Redox active pincer ligands.

which indicates that in the gas phase, electron addition is thermodynamically unfavored. Both pyrazole and N-methyl pyrazole are more negative (less favorable) than that of pyridine, so this would anticipate that reduction of the LR pincer ligand would take place more in the pyridine portion. This conclusion is reinforced by an independent evaluation: the LUMO and LUMO + 1 of a truncated LMe pincer (Fig. 1) are localized on pyridine while the LUMO + 2 and LUMO + 3, which are higher in energy by 1 eV (consistent with the calculated electron affinity difference), are localized on the pyrazole rings. The truncated computational model for LMe is simplified with a methyl group instead of a t-butyl groups at the 5-position of the pyrazole. LUMO + 1 has greater C(pyridyl)–C (pyrazolyl) bonding character, and reduction thus strengthens this bond. This will have the effect of increasing diazabutadiene delocalization as these orbitals become populated upon reduction.

LUMO +3

LUMO +2

LUMO +1

2.2. Ligand synthesis and characterization Ligand synthesis used conventional pyrazole methodology, which involves cyclization of a 1,3-diketone with hydrazine (Scheme 2) [2,6]. The 1H NMR spectrum of LH in CDCl3 shows all expected CH signals, and indicates twofold symmetry, consistent with equivalent pyrazole arms of the pincer form. The NH protons are not observed, apparently due to a dynamic process: the proton equilibrating between the two inequivalent pyrazole nitrogens in a given ring. Given the evidence (see below) for hydrogen bonding, it was thought that the NH exchange process might be suppressed in

LUMO

Fig. 1. Lowest unoccupied orbitals in LMe at the geometry of a complex (see below).

O t

MeO

OMe N O

O

BuCCH3

NaH THF 30 min, reflux

tBu

tBu

N

EtOH 2 h, reflux N2H4

O

t

O

O

O

N

Bu

LH N

NH

25oC NaH 4 h MeI

t Bu

HN

N

Yield: 52%

t

Bu

LMe

3

N1

N2

tBu

N

Yield: 35% (total)

Scheme 2. Route to ligand synthesis and alkylation.

N

N

84

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

THF solvent, as a hydrogen bonding partner. Indeed, the 1H NMR spectrum of LH dissolved in d8-THF shows not only all ring and t Bu protons in a twofold symmetric molecule, but also shows a broader resonance, intensity 2, at 12.4 ppm, assigned to the NH protons. We attribute the appearance of the NH signal in THF solvent to that solvent eliminating site exchange between NH tautomers at either inequivalent pyrazole nitrogen seen in CDCl3. This molecule tends to retain CH2Cl2 upon isolation even after drying in vacuo for several hours, and the 1H NMR of such a solid sample dissolved in CDCl3, CD2Cl2, or (CD3)2C@O shows a peak corresponding to that of CH2Cl2 in addition to ligand peaks. The integration of the CH2Cl2 peak (with respect to pyrazolyl CH) measures 0.34, 0.31, and 0.40 for these NMR solvents, respectively. This indicates that CH2Cl2 occupies the crystal lattice of the isolated product. Hydrogen bonding between Cl of CH2Cl2 and the acidic NH protons of two LH may be an explanation for the crystalline sample to comprise an idealized stoichiometry 2(LH)CH2Cl2. The generality for LH to interact with solvents in this way is best illustrated by the following single crystal X-ray structure determination of a sample grown from THF. Crystals of LH(THF) grown from THF by slow evaporation show (Fig. 2) the molecule to be fully planar with the NH protons directed inward, which creates two hydrogen bonds to a single THF oxygen which is a ‘‘guest” inside the pincer core. The N/O hydrogen bond distances are very symmetrical, 2.964(3) and 2.955(3) Å, and suggest involvement of the two ether oxygen lone pairs. Bond lengths within the pyrazole ring (See Table 1) show good consistency (<4 esd variation) of corresponding distances between the two rings. While distances within a given ring show some of the expected variations between formal single and double bonds, those differences are <0.03 Å (<9 esd’s).

2.3. Complexation to FeCl2 Reaction of LH with an equimolar slurry of anhydrous FeCl2(THF)1.5 in THF forms a single product as a red-orange solution in less than 30 min at 25 °C. The 1H NMR spectrum of the resulting product in THF shows five signals of appropriate intensity for a twofold symmetric pincer ligand. The range of chemical shifts, +66 to 29 ppm, indicates paramagnetism, but all signals, including the NH protons, are clearly evident; the signal at +66 ppm is broader than the others, which suggests proximity to the metal and is therefore assigned to the NH site. The pincer ligand is functioning as a tridentate tris-imine donor; in the absence of added base during the synthesis, the ligand LH is not deprotonated. An

Fig. 2. ORTEP view (50%) probability) of LH(THF), showing selected atom labeling; unlabeled atoms are carbon. Selected structural parameters: N3. . .O1S, 2.964(3) Å; N5. . .O1S, 2.955(3); N3-H3 N. . .O1S, 170(3)°; N5AH5N. . .O1S, 177(3).

Table 1 Selected intra-pyrazole bond lengths for LH. Bond distances (Å) in pyrazole rings of LH NAN

CAN

C@N

CAC

C@C

1.355(3) 1.356(3)

1.348(3) 1.349(3)

1.333(3) 1.334(3)

1.398(3) 1.405(3)

1.386(3) 1.373(3)

Evans method magnetic susceptibility determination, in THF with hexamethylbenzene reference, yields a value of 4.98 lB, which is consistent with four unpaired electrons (high spin FeII). Stripping a red-orange THF solution of this product to dryness in vacuum, causes a color change to pale yellow-orange. The 1H NMR spectrum of a golden yellow CD2Cl2 solution shows the absence of THF signals, indicating that THF has only weak binding to the complex, and symmetry equivalence of the ligand arms. The (still paramagnetically shifted) 1H NMR chemical shifts differ by up to 8 ppm from those in THF, but overall we attribute the modest color change and the chemical shift changes to the presence vs. absence of hydrogen bonding to solvent, and its impact on the ligand field strength of LH. The influence of hydrogen bonding on a ‘‘proton responsive ligand” like LH is widely studied [3,4,6]. The ESI mass spectrum in THF is informative of the reactivity of Fe(LH)Cl2. The positive ion spectrum shows not only protonated free ligand (i.e., [HLH]+), but also Fe(LH)+2 that indicates facile ligand redistribution under ion generating conditions. The negative ion mass spectrum shows deprotonated ligand (LH –H), the neutral ligand ion-paired with chloride LHCl, and deprotonated complex Fe(LH–H)Cl 2 . The latter are seen with the appropriate chlorine isotopic multiplet; no THF containing species are detected. Structure determination by single crystal X-ray diffraction of orange crystals grown by slow evaporation from THF shows (Fig. 3) the formula of the solid to be Fe(LH)Cl23THF, but none of the THF is coordinated to iron; the shortest Fe/O distance is 4.6 Å. Each NH proton is hydrogen bonded to one THF oxygen (O/N = 2.787(18) Å with angle OHN = 174.6°); the third THF merely occupies space in the unit cell. The structure determination shows that complexation of LH to the metal has caused the acidic hydrogens to migrate from the a to the b nitrogen, in comparison to the

Fig. 3. ORTEP view (50% probability) of the nonhydrogen atoms of Fe(LH)Cl2, showing selected atom labeling; unlabeled atoms are carbon. Lattice THF molecules are not shown. A crystallographic C2 axis is defined by Fe and N1. Selected structural parameters: Fe1AN1, 2.130(2) Å; Fe1AN3, 2.2587(13); Fe1ACl1, 2.3097 (5); N1AFe1AN3, 73.50(4)°; N3⁄AFe1AN3, 146.99(8); N1AFe1ACl1, 121.779(14); Cl1AFe1ACl1⁄, 116.44(3).

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

structure of LH in Fig. 2. The dihedral angle N1AC3AC4AN3 is 1.0°, hence the ligand rings are accurately eclipsed. The molecule has crystallographic C2 symmetry, so the chloride ligands are symmetry related; the N(pyridine)/Fe/Cl angle is 121.779(14). This is in full agreement with conclusions from the NMR spectra, but shows that the paramagnetism arises from unsaturated (five coordinate) iron, without coordination of THF to Fe. The Fe/N distances to pyridine are 0.13 Å shorter than to the pyrazole nitrogens. The three rings of the pincer ligand do not deviate significantly from coplanarity, and the 5-coordinate structure has a s parameter of 0.51 [14] where a trigonal bipyramidal structure has s = 1.0 and a square pyramid 0.0. Comparison to other 5 coordinate Fe structures shows distances here to be consistent with high spin Fe(II). The structure of solvent free LHFeCl2 has been reported [15], but without any comment about possible hydrogen bonding in that solid phase; also reported there is the structure of LHFeCl3. An examination of the five distances within the pyrazole ring here shows no significant (>0.02 Å) systematic bond length alternation, indicating aromatic delocalization of the six p electrons. This same statement applies to the corresponding distances in the pyrazolyl ring of complexes featuring HB(pz) 3 ligands in the Cambridge Crystallographic Database. 2.4. Cyclic voltammetry of Fe(LH)Cl2 Cyclic voltammetry was studied at a platinum working electrode in THF with 0.1 M [NBu4]PF6 at scan rates from 25 to 600 mV s1. The acid form of the ligand LH shows (See Supporting information) no oxidative wave and one irreversible reductive wave at about 2.2 V vs Fc/Fc+. This quite negative potential is consistent with pyrazole being relatively electron rich as concluded from our DFT analysis. Fe(LH)Cl2 shows (Fig. 4) one reversible (judging by ipa vs. ipc) oxidation at E1/2 = 0.10 V (vs. Fc/Fc+) which we assign to oxidation of Fe(II) to Fe(III). A reductive process is seen (Fig. 5) with an Epc of 2.3 V; this is less negative than that of free LH due to (ligand-centered) redox product stabilization by the attached electrophilic metal. Once this reduction has been effected, there are two peaks upon sweeping in the oxidative direction, one with an Epa of 2.05 V and a second weaker one at 1.85 V. This reductive process therefore proceeds through an EC type mechanism and rapidly gives rise to a new species with different redox chemistry detected in the oxidative direction. We suggest that the 2.3 V reduction involves an EC mechanism, due to chloride

85

loss but recognize that pyrazole proton loss as hydrogen atoms is also possible. 2.5. A pincer without acidic protons, and its FeCl2 complex For reductive applications, acidic protons are a vulnerable functionality on the ligand. The acidic protons of LH were replaced by methyl groups by double deprotonation of LH with NaH, then slow addition of 2 equivalents of [Me3O]BF4 in ethyl acetate. This reaction was regioselective for the desired isomer, as shown by 1H NMR, as well as by connectivity in LMe established by X-ray diffraction (see Supporting information). Note that the structure determination establishes that the methylation does not occur at the site of the acidic protons in LH, but rather at the ‘‘outer” nitrogens, which then moves the two imine functionalities to the inwardly directed positions of the pincer ligand. We attribute this methylation regioselectivity to sodium cation binding to the inner nitrogens in the Na2L synthetic intermediate in ethyl acetate [16]. Addition of LMe to anhydrous FeCl2 slurried in THF occurs rapidly at 25 °C, with color change to red orange. The product was isolated by removal of volatiles after 30 min stirring, and its 1H NMR spectrum in d8 THF shows signals over a chemical shift range which indicates paramagnetism, and shows the absence of coordinated THF. The compound is soluble in THF, poorly soluble in benzene, toluene, CH2Cl2 and Et2O, and is insoluble in pentane. The mother liquor of the synthesis shows only the signals of LMeFeCl2, so the synthesis is selective and high yielding. The chemical shifts of four of the five observed signals are very similar to those of (LH)FeCl2, but the +66 ppm signal of (LH)FeCl2 is gone in LMeFeCl2, and a new one of intensity six appears at +16 ppm and is assigned to the N-methyl groups. Evans method magnetic susceptibility studies carried out at 25 °C in d8-THF with hexamethylbenzene as standard gave a value of 4.7 Bohr magnetons, consistent with 4 unpaired electrons. The monocation (LMeFeCl2)+ is seen in both the ESI(+) and APCI(+) mass spectra in THF, indicating the molecule is readily oxidized, hence electron rich; this characteristic originates from both FeII and the LMe ligand. The corresponding negative ion spectra are absent, also consistent with electron rich character of the neutral. Crystals grown from THF/pentane were shown (Fig. 6) by X-ray diffraction to have formula LMeFeCl22THF, with the THF molecules merely occupying voids in the unit cell; these crystals are not isomorphous with those of (LH)FeCl23THF. The fact that these ethers do not coordinate in the solid state indicates that this iron complex

Fig. 4. CV (0.4 to 0.5 V) of FeLHCl2 in THF/0.1 M [NBu4]PF6 at 25–800 mV s1 [37].

86

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

Fig. 5. CV (1.1 to 2.8 V) of FeLHCl2 in THF/0.1 M [NBu4]PF6 at 25–800 mV s1 [37].

Fig. 6. ORTEP view (50% probability) of the nonhydrogen atoms of Fe(LMe)Cl2, showing selected atom labeling; unlabeled atoms are carbon. Selected structural parameters: Fe1ACl1, 2.2832(9) Å; Fe1ACl2, 2.3289(9); Fe1AN2, 2.222(3); Fe1AN3, 2.141(2); Fe1AN5, 2.213(3); Cl1AFe1ACl2, 118.58(3)°; Cl1AFe1AN3, 118.89(8); Cl2AFe1AN3, 122.42(8); N2AFe1AN3, 73.61(10); N2AFe1AN5, 147.01(10); N3AFe1AN5, 73.51(10).

is not significantly Lewis acidic. Five coordinate iron has idealized C2v symmetry, with N(pyridyl)AFeACl angles of 118.89(8) and 122.42(8)°. Distances and angles involving Fe and LMe are within 0.04 Å and 3° of those for (LH)FeCl2. However, the Fe/N(pyrazolate) distances are shorter (by 0.04 Å) in the LMe case, consistent with better ring donor power.

2.6. Cyclic Voltammetry of Fe(LMe)Cl2 Cyclic voltammetry (CV) was studied on LMe and FeLMeCl2 at a Pt working electrode, in THF containing 0.1 M TBAPF6 as supporting electrolyte, at scan rates from 25 to 200 mV s1. Open circuit potentials are 0.90 V and 0.99 V vs. Fc/Fc+ for the free ligand and Fe complex respectively, both indicative of electron rich character. Cathodic scans of LMe show (See Supporting information) negligible current flow from 0.2 to 2.8 V, but show oxidation with Epa  0.1 V, while LH shows no oxidative wave, consistent with

Fig. 7. CV (1.5 to 3.3 V) of FeLMeCl2 in THF/0.1 M [NBu4]PF6 at 25–800 mV s1 [38].

easier oxidation due to pyrazole nitrogen methyl substituent (vs. H in LH). Cathodic scans of FeLMeCl2 in the range 1.5 to 3.3 V (Fig. 7) show one quasi-reversible wave with E1/2 = 2.60 V, which is fully consistent with an electron rich ligand. The peak to peak separation, 0.14 V for the scan rate of 25 mV/s, is smaller than that measured for ferrocene in this medium and peak current is linear in (scan rate)1/2, both of which are consistent with reversibility (see Supporting information). This potential is more negative than that of Fe(LH)Cl2 (E1/2 = 2.20 V), consistent with electron donor methyl groups resisting reduction of the complex. This wave also has more similar ipa and ipc (i.e., is more reversible) than that of Fe(LH)Cl2. Fig. 8 shows anodic scans for FeLMeCl2. This shows an irreversible wave, with Epa = 0.66 V and Epc = 1.14 V. The open circuit potential for this compound is 0.99 V, so the scans were started at 1.5 V, where FeLMeCl2 is the dominant species. Oxidation of FeLMeCl2 is about 300 mV more negative than that of Fe (LH)Cl2, also consistent with methyl (vs. H) making the complex more electron rich. The irreversibility is attributed to coordination of THF to the more electron deficient cationic product; the proposed species FeLMeCl2(THF)+ would be harder to reduce, as observed with Epc = 1.14 V.

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

87

Fig. 8. CV (0.3 to 1.5 V) of FeLMeCl2 in THF/0.1 M [NBu4]PF6 at 25–600 mV s1. (⁄ is assigned as ligand centered oxidation) [39].

2.7. Chemical reduction of LMe To assess its potential for accepting an electron, and in spite of it showing no cathodic process out to 2.8 V, LMe was reacted with equimolar KC8 in THF (using 2–3 equivalents of KC8 forms the same product). Reaction occurs immediately at 25 °C as judged by color change to deep red. The gross features of the EPR spectrum, following removal of graphite, show major coupling, but a large amount of additional fine structure is reproducibly resolved with 0.025 G digital resolution and 0.1 G modulation amplitude. Since these features carry information which reveal the presence or absence of symmetry in the monoanion, simulation is important for understanding both geometric and electronic structure. Given the number of spin active nuclei in the species, simulation benefits from initial awareness of magnitudes of these coupling constants. The EPR parameters of KLMe were therefore calculated using density functional theory, to establish simulation starting values for these parameters. Full details are given in Supporting information. This gave the important information that coupling to the pyridyl para hydrogen was larger than that to the pyridyl nitrogen, a feature which has been seen before for pyridinyl radical anions with conventional saturated substituents [17]. This gives a good start to simulation of the gross spectral features, and allows implementing smaller coupling to all pyrazole nitrogens and two other hydrogens. At this point, small imperfections were eliminated by inclusion of coupling to 94% abundant I = 3/2 39K, to give the simulation shown in Fig. 9. This shows that (LMe) is an intimate ion pair with its counter cation. Further details of this simulation, as well as the suitability of slightly different parameter sets, are discussed in the Supporting information. Ion pairing of K+ with (LMe) is reasonable given that the mass spectrum shows ions where even neutral LMe bind the electrophile Na+. In summary, the EPR spectrum of KLMe is consistent with twofold symmetry relating the two halves of the entire species, and major spin density at pyridine. EPR spectra recorded 12 h after preparation retained the detailed fine structure, showing the radical to be persistent at least over this time frame at 25 °C. The persistence of this radical anion vs. pyridine radical anion itself indicates that the pyrazole delocalization increases the lifetime of this species. In order to further test our evidence that intimate ion pairing of the radical anion with K+ was present, we recorded the EPR spectrum again in the presence of added 18-crown-6 (10% molar excess over LMe). This showed (Fig. 10) both higher resolution and simpli-

fication of the previous fine structure, clearly showing a change in species composition/structure attributable to removal of K+ from the radical anion nitrogens. Simulation using the same hyperfine parameters in Fig. 9 (see Supporting information) revealed satisfactory agreement with this simplified spectrum without any coupling to potassium, thus supporting the above assignment of some of the hyperfine structure as originating from quadrupolar potassium. Finally the fine structure of this spectrum in the presence of the crown ether was unchanged in a sealed EPR tube under inert atmosphere after 24 h at 25 °C, showing at least this lifetime for the radical under these conditions. In summary, the EPR study of KLMe confirms the SOMO composition as forecast by DFT calculations, with the additional result that intimate ion pairing occurs. Hyperfine coupling confirms that some spin is on the pyrazoles, but most is on the pyridyl; in the pyridyl, the largest hyperfine coupling is to hydrogen on the para carbon (see also Fig. 1), reflecting the strong participation of that carbon in the LUMO of unreduced LMe. The fact that (LMe) radical anion shows hyperfine coupling to potassium [18] in their ion pair emphasizes the high reducing power of (LMe), and illustrates its potential to reduce any redox active transition metal ion to which it might be coordinated. 2.8. Product of Reaction of KLMe with O2 Radical KLMe is very air sensitive in solution, and is decolorized by contact with air. This product was investigated by mass spectrometric and NMR analysis. The KLMe radical frozen in THF was reacted at low temperature by addition of molecular oxygen. Decolorization was immediate even with the frozen solution; ultimately 99% of the EPR signal disappears. The proton NMR spectrum of the resulting material in d8-THF clearly shows a diamagnetic product with two arms of the pincer now inequivalent, as judged by pyridyl ring and pyrazolyl ring hydrogens. Particularly diagnostic are the presence of two resonances in the region for alkyl protons attached to pyrazole amine nitrogen with an intensity ratio of 2–3. These we interpret as the N-methyl group of one pyrazole arm having undergone O atom insertion into its CH bond. The remainder of the proton NMR spectrum has signals appropriate for inequivalent arms on the pincer ligand, including those of inequivalent tBu groups. ESI+ mass spectrometry shows a peak at 16 mass units heavier than LMe, and APCI+ shows (M + O + H)+. This result is particularly important since it shows alkyl

88

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

Fig. 9. Above: X-band EPR spectrum of KLMe in THF at 25 °C. Microwave frequency, 9.859 MHz, modulation amplitude, 0.10 G, modulation frequency 100 kHz. Below: Simulated spectrum, using parameters shown.

CH hydroxylation without need for any transition metal. In addition it shows efficient scission of the OO bond to deliver single oxygen atoms from what is probably a hydroperoxy intermediate. 3. Discussion and conclusions Both pyridine and the neutral pyrazole ring contain imine functionality and are therefore in principle reducible. The electron donor NMe ring member is a substituent on the pyrazole imine nitrogen that diminishes its reducibility and directs the arriving electron into the pyridyl ring. A pincer ligand is characterized by substituents at the 2 and 6 positions of the central ring. This mutu-

ally meta relationship means that these arm functionalities cannot conjugate with each other, but each conjugates separately with the pyridine nitrogen. This last point shows that pyridine as a linker is much more effective for redox activity than phenyl, since phenyl is inferior at accepting negative charge at the central ring atom (i.e., C vs. N). The work with LH has defined the coordinating properties of this ligand and shown the significant influence of the two NH functionalities, which is being investigated by others as an example of ‘‘proton responsive ligands” [19–26]. For the goal of reductive applications of bis-pyrazolyl pyridines, we thought that these protons would be vulnerable to loss by H2 evolution, following

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

89

Fig. 10. X-band EPR spectrum of [K(18-crown-6)]LMe in THF at 25 °C. Microwave frequency, 9.838 MHz, modulation amplitude, 0.10 G, modulation frequency 100 kHz.

reductions of metal complexes with LH. Therefore, we moved to alkylated analogs, which have a more robust tertiary amine functionality in each ring. In the introduction, we compared the electron affinities (‘‘reducibility”) of a number of nitrogen heterocycles, as this is central to their redox activity, but interaction between pyridyl and arm heterocycles also impacts their electronic structure. Note the similarity (red substructures in Scheme 4) of one resonance form of the alkylated pincer to the reducible 1,4-diazabutadiene moiety in the 2,6 bis-imino pyridyl. Scheme 4 also compares deprotonated pyrazolate and recently employed pyrrolide[27] as pincer arms. Clearly the carbanionic character of the diazabutadiene resonance form of the pyrrolide shows that it is inferior to the pyrazolate arms in accepting electron density in this pincer class. Our electron affinity calculations described above show that pyrrole is the hardest to reduce. Indeed pyrrolide has been incorporated into macrocyclic, bidentate and tripodal variants, and has been shown to make the metal electron rich [28–34]. We recently reported a bis-pyrrolyl pyridine pincer and several of its metal complexes [27]. We have also reported a pincer with two easily reduced tetrazine arms [35]. The difference between C3 symmetric tripodal ligands and C2 symmetric pincers remains to be seen in various applications. However, it is clearly valuable to have pincers of different ease of reduction, since the most difficult to reduce is, once reduced, the strongest reducing agent. The

LMe: tBu

N R N

N R

Pyrrolide:

tBu

N N N

N N

Pyrazolate:

N

N

N N

N

N

N N

N N

Scheme 4.

bis-pyrrolide pyridyl has the advantage of being the strongest p donor even without reduction. The tunability of this palette of ligands is thus one of their advantages, and pyrazoles should play a central role in redox active ligand construction.

4. Experimental 4.1. General All manipulations were carried out under an atmosphere of ultra high purity nitrogen using standard Schlenk techniques or in a glovebox under argon. Solvents were purchased from commercial sources, purified using Innovative Technology SPS-400 PureSolv solvent system or by distilling from conventional drying agents and degassed by the freeze-pump-thaw method twice prior to use. Glassware was oven-dried at 150 °C overnight and flame dried prior to use. THF, including d8-THF was stored over activated 4 Å molecular sieves or sodium metal pieces. The synthesis of ligand LH2 has been reported [36]. NMR spectra were recorded in various deuterated solvents at 25 °C on a Varian Inova-400 spectrometer (1H: 400.11 MHz). Proton chemical shifts are reported in ppm versus solvent protic impurity, but referenced finally to SiMe4. Mass spectrometry analyses were performed in an Agilent 6130 MSD (Agilent Technologies, Santa Clara, CA) quadrupole mass spectrometer equipped with a Multimode (ESI and APCI) source. All starting materials have been obtained from commercial sources and used as received without further purification. Cyclic voltammetry was done with Pt as the working electrode, Pt as the counter electrode, and Ag/AgCl wire as the reference electrode. 0.1 M TBAPF6 was employed as a supporting electrolyte, in THF solvent. In this medium, ferrocene has a peak-to-peak separation of 0.50 V. 13.5 and 15.5 mg of free ligand and Fe complex were used, respectively, in 10 mL solvent. All CVs are referenced to internal Fc/Fc+ as the standard, added at the end of a study of the experimental sample. Electrodes were polished when a new molecule was studied. Open circuit potentials are 0.90 V and 0.99 V vs. Fc/Fc+ for LMe and (LMe)FeCl2, respectively; we designate ‘‘open circuit potential” as the potential where there is negligible current flow for the species dissolved. EPR simulations used SimFonia; in the complicated spectra in Fig. 9, the effect of small coupling constants (e.g., to K) can be comparably compensated by simply increasing linewidth.

90

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

The color of KLMe in THF ranges from orange when dilute (EPR samples) to purple when very concentrated. 4.2. Fe(LH)Cl2 3.00 g (9.29 mmol) of LH was dissolved in 20 mL of THF and added dropwise to a stirring slurry of 1.17 mg (9.28 mmol) FeCl2(THF)1.5 in 30 mL THF. The solution turned to a red/orange color and became homogenous over the course of 1 h. The solution was allowed to stir at RT for another 12 h without any additional color change. Solvent was removed under reduced pressure to yield an orange solid. This solid was extracted with 2  50 mL Et2O until the final color was a pale apricot-orange color. This extraction is done to remove free ligand from the product solid, but also to remove THF from the solid. Single orange needle crystals suitable for X-ray diffraction can be obtained via slow evaporation from a concentrated THF solution into silicon grease in a sealed vessel. Yield: 3.89 g (93%). 1H NMR (THF-d8, 400MHz) d(ppm), line width (Hz): 29.1 (1H, Ar-p, 100), 7.09(18H, tBu, 30), 42.7 (2H, CH, 60), 49.2 (2H, Ar-m, 80), 66.9 (2H, NH, 400). In CD2Cl2, 21.9 (1H, Ar-p), 8.02 (18H, tBu), 42.6 (2H, CH), 57.9(2H, Ar-m), 72.4 (2H, NH). ESI-MS () in THF: [Fe(LH)Cl2] 448.08 (100), 450.08(65.3), 449.09(21.1). Evans Method magnetic susceptibility (THF-d8, hexamethylbenzene reference): 5.0 lB. 4.3. LMe NaH (60 wt.% in mineral oil, 155 mg, 3.87 mmol) was slurried with 20 mL EtOAc under an atmosphere of N2 and stirred until no clumps remained. LH2 (500 mg 1.55 mmol) in 10 mL EtOAc was added to this slurry dropwise over the course of 5 min. Once the addition was completed, this mixture was allowed to stir at RT until bubbling had ceased. Solid [Me3O][BF4] (500 mg 3.38 mmol) was then added in small portions to this mixture over the course of 30 min, and the reaction allowed to stir for 12 h. The mixture was then cooled back to RT and filtered through a medium porosity ground glass frit to remove a white solid. Solvent was removed from the filtrate under reduced pressure and the resulting white solid washed with 20 mL H2O. This slurry was then extracted with EtOAc (2  20 mL) and the organic layer collected. This organic was then dried over MgSO4, filtered to remove solids and solvent removed under reduced pressure. The resulting solid was washed with 3  30 mL of n-hexane to remove the inner/inner dimethyl isomer, and the undissolved solid was then taken to dryness. The purity of this solid was sufficient that column chromatography was unnecessary. Single crystals suitable for X-ray diffraction can be grown from slow diffusion of cyclohexane vapors into a saturated CH2Cl2 solution. Yield: 66.9% (365 mg, 1.04 mmol). 1 H NMR (400 MHz, CDCl3 298 K): d (ppm) 7.80 (d, 8 Hz, 2H Ar-m), 7.70 (d, 8 Hz, 1H Ar-p), 6.79 (2H, pyz CH), 4.04 (6H, pyz CH3), 1.44 (18H, pyz C(CH3)3). MS(ESI-Positive) in CHCl3: 352 [M+H]+, C21H30N5, 338 [M+HMe]+ C20H28N5. A peak at m/z = 725 is occasionally seen in the ESI positive ion spectrum and has been assigned to Na(LMe)+2, due to scavenging of Na+ in the injector. This was confirmed by preparing a solution of sodium acetate in ethanol, adding LMe, and observing this same 725 peak. X-ray diffraction data (see Supporting information) collected on a single crystal of LMe gave connectivity showing methylation of the ‘‘outer” pyrazole nitrogens. 4.4. Fe(LMe)Cl2 FeCl2(THF)1.5 (67 mg, 0.285 mmol) was slurried in 3 mL THF. To this, LMe (100 mg, 0.310 mmol) in 2 mL THF was added dropwise. The slurry changed color to red/orange over the course of the addition and the slurry became homogenous. This mixture was allowed

to stir for 10 min at RT, during which precipitation was observed. The mixture was filtered through Celite to remove an orange precipitate and yield an orange filtrate. The orange precipitate was washed through the Celite with fresh THF. Independent NMR analysis of both the filtrate and orange precipitate revealed the two to be identical. Single crystals suitable for X-ray diffraction were grown from slow vapor diffusion of pentane vapors into a saturated THF solution. Yield: quantitative. 1H NMR (400 MHz, THFd8 298 K) d(ppm), line width (Hz): 51.0 (2H, 100), 45.7 (2H, 100), 16.5 (6H, 200), 6.91 (18H, 120), 28.1(1H, 100). MS(ESI Positive) in THF: (M = LMeFeCl2) MeLMe+, 366.3; M+, 477.1; KM+, 517.2; Me + + Fe2LMe 2 Cl3, 919.3; Fe2L2 Cl3O , 935.3. These last two strong show the Lewis acidity developed when one chloride has been removed. The APCI(+) spectrum also shows these same peaks, but also shows (HLMe)+ (352.3) and HM+ (478.1). Evans Method magnetic susceptibility (25 °C, THF-d8, hexamethylbenzene reference): 4.7 Bohr magnetons. Acknowledgments This work was supported in part by the IU Office of Vice President for Research. RLL acknowledges support from GVSU start-up funds and a GVSU-CSCE Faculty Research Grant-in-Aid. Computational resources were provided through NSF-MRI support to the Midwest Undergraduate Computational Chemistry Consortium (CHE-1039925). We also thank Prof. Josh Telser for advice on EPR simulations, Keith Searles for assistance in the cyclic voltammetry, and Adam D. Miller for crystallographic work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ica.2016.07.011. References [1] R.F. Munha, R.A. Zarkesh, A.F. Heyduk, Group transfer reactions of d0 transition metal complexes: redox-active ligands provide a mechanism for expanded reactivity, Dalton Trans. 42 (11) (2013) 3751–3766. [2] M.A. Halcrow, The synthesis and coordination chemistry of 2,6-bis(pyrazolyl) pyridines and related ligands – Versatile terpyridine analogues, Coord. Chem. Rev. 249 (24) (2005) 2880–2908. [3] G.L. Taghizadeh, S. Farsadpour, Y. Sun, W.R. Thiel, New N,N,N-donors resulting in highly active ruthenium catalysts for transfer hydrogenation at room temperature, Eur. J. Inorg. Chem. 2011 (23) (2011) 3431–3437. [4] A. Yoshinari, A. Tazawa, S. Kuwata, T. Ikariya, Synthesis, structures, and reactivities of pincer-type ruthenium complexes bearing two protonresponsive pyrazole arms, Chem. Asian J. 7 (6) (2012) 1417–1425. [5] M.A. Halcrow, Pyrazoles and pyrazolides-flexible synthons in self-assembly, Dalton Trans. 12 (2009) 2059–2073. [6] K. Umehara, S. Kuwata, T. Ikariya, N–N bond cleavage of hydrazines with a multiproton-responsive pincer-type iron complex, J. Am. Chem. Soc. 135 (18) (2013) 6754–6757. [7] G.A. Craig, O. Roubeau, G. Aromi, Spin state switching in 2,6-bis(pyrazol-3-yl) pyridine (3-bpp) based Fe(II) complexes, Coord. Chem. Rev. 269 (2014) 13–31. [8] K. Umehara, S. Kuwata, T. Ikariya, Synthesis, structures, and reactivities of iron, cobalt, and manganese complexes bearing a pincer ligand with two protic pyrazole arms, Inorg. Chim. Acta 413 (2014) 136–142. [9] O.R. Luca, R.H. Crabtree, Redox-active ligands in catalysis, Chem. Soc. Rev. 42 (4) (2013) 1440–1459. [10] P.J. Chirik, K. Wieghardt, Radical ligands confer nobility on base-metal catalysts, Science 327 (5967) (2010) 794–795. [11] J.A. Montgomery Jr., M.J. Frisch, J.W. Ochterski, G.A. Petersson, A complete basis set model chemistry. VI. Use of density functional geometries and frequencies, J. Chem. Phys. 110 (6) (1999) 2822–2827. [12] J.A. Montgomery Jr., M.J. Frisch, J.W. Ochterski, G.A. Petersson, A complete basis set model chemistry. VII. Use of the minimum population localization method, J. Chem. Phys. 112 (15) (2000) 6532–6542. [13] I. Nenner, G.J. Schulz, Temporary negative ions and electron affinities of benzene and N-heterocyclic molecules: pyridine, pyridazine, pyrimidine, pyrazine, and S-triazine, J. Chem. Phys. 62 (5) (1975) 1747–1758. [14] A.W. Addison, T.N. Rao, J. Reedijk, J. Van Rijn, G.C. Verschoor, Synthesis, structure, and spectroscopic properties of copper(II) compounds containing nitrogen-sulfur donor ligands: the crystal and molecular structure of aqua[1,7-

B.J. Cook et al. / Inorganica Chimica Acta 451 (2016) 82–91

[15]

[16]

[17]

[18]

[19]

[20]

[21] [22]

[23] [24]

bis(N-methylbenzimidazol-20 -yl)-2,6-dithiaheptane]copper(II) perchlorate, J. Chem. Soc., Dalton Trans. 7 (1984) 1349–1356. Y.-Y. Fang, W.-J. Gong, X.-J. Shang, H.-X. Li, J. Gao, J.-P. Lang, Synthesis and structure of a ferric complex of 2,6-di(1H-pyrazol-3-yl)pyridine and its excellent performance in the redox-controlled living ring-opening polymerization of ?-caprolactone, Dalton Trans. 43 (22) (2014) 8282–8289. D. Zabel, A. Schubert, G. Wolmershäuser, R.L. Jones, W.R. Thiel, Iron and cobalt complexes of tridentate N-donor ligands in ethylene polymerization: efficient shielding of the active sites by simple phenyl groups, Eur. J. Inorg. Chem. 2008 (23) (2008) 3648–3654. A.R. Buick, T.J. Kemp, G.T. Neal, T.J. Stone, Electron spin resonance studies of reduction by solvated electrons in liquid ammonia. II. Pyridines, J. Chem. Soc. A 11 (1969) 1609–1613. Much detailed information about ion pairing by radical anions has been derived from solvent and temperature dependent studies of such compounds but primarily by observation of changing hyperfine coupling constants to ring protons. Occasionally coupling to Li, Na and K cations has been observed, and interpreted in terms of spin densities at the metal as low as 0.14%. See Y. Mizuta, M. Kohno, K. Fujii, K. Kuwata, Chem. Lett. (1995) 573; M.P. Khakhar, B.S. Prabhananda, M.R. Das, J. Am. Chem. Soc. 89 (1967) 3100. Y.M. Badiei, W.-H. Wang, J.F. Hull, D.J. Szalda, J.T. Muckerman, Y. Himeda, E. Fujita, Cp*Co(III) catalysts with proton-responsive ligands for carbon dioxide hydrogenation in aqueous media, Inorg. Chem. 52 (21) (2013) 12576–12586. C.M. Conifer, D.J. Law, G.J. Sunley, A. Haynes, J.R. Wells, A.J.P. White, G.J.P. Britovsek, Dicarbonylrhodium(I) complexes of bipyridine ligands with proximate H-bonding substituents and their application in methyl acetate carbonylation, Eur. J. Inorg. Chem. 2011 (23) (2011) 3511–3522. T. Ikariya, Chemistry of concerto molecular catalysis based on the metal/NH bifunctionality, Bull. Chem. Soc. Jpn. 84 (1) (2011) 1–16. J.A. Kitchen, S. Brooker, Spin crossover in iron(II) complexes of 3,5-di(2pyridyl)-1,2,4-triazoles and 3,5-di(2-pyridyl)-1,2,4-triazolates, Coord. Chem. Rev. 252 (18–20) (2008) 2072–2092. S. Kuwata, T. Ikariya, Quest for metal/NH bifunctional bioinspired catalysis in a dinuclear platform, Dalton Trans. 39 (12) (2010) 2984–2992. D.C. Marelius, S. Bhagan, D.J. Charboneau, K.M. Schroeder, J.M. Kamdar, A.R. McGettigan, B.J. Freeman, C.E. Moore, A.L. Rheingold, A.L. Cooksy, D.K. Smith, J. J. Paul, E.T. Papish, D.B. Grotjahn, How do proximal hydroxy or methoxy groups on the bidentate ligand affect [(2,20 ;60 ,200 -terpyridine)Ru(N,N)X| wateroxidation catalysts? synthesis, characterization, and reactivity at acidic and near-neutral pH, Eur. J. Inorg. Chem. 2014 (4) (2014) 676–689.

91

[25] Z. Ni, M.P. Shores, Supramolecular effects on anion-dependent spin-state switching properties in heteroleptic iron(II) complexes, Inorg. Chem. 49 (22) (2010) 10727–10735. [26] D. Wang, S.V. Lindeman, A.T. Fiedler, Intramolecular hydrogen bonding in CuII complexes with 2,6-pyridinedicarboxamide ligands: synthesis, structural characterization, and physical properties, Eur. J. Inorg. Chem. 2013 (25) (2013) 4473–4484. [27] N. Komine, R.W. Buell, C.-H. Chen, A.K. Hui, M. Pink, K.G. Caulton, Probing the steric and electronic characteristics of a new bis-pyrrolide pincer ligand, Inorg. Chem. 53 (3) (2014) 1361–1369. [28] S. Ilango, B. Vidjayacoumar, S. Gambarotta, S.I. Gorelsky, Low-Valent vanadium complexes of a pyrrolide-based ligand. electronic structure of a dimeric V(I) complex with a short and weak metal–metal bond, Inorg. Chem. 47 (8) (2008) 3265–3273. [29] E.R. King, G.T. Sazama, T.A. Betley, Co(III) imidos exhibiting spin crossover and C–H bond activation, J. Am. Chem. Soc. 134 (43) (2012) 17858–17861. [30] I. Korobkov, S. Gambarotta, G.P.A. Yap, Highly reactive uranium(III) polypyrrolide complexes: intramolecular C–H bond activation, ligand isomerization, and solvent deoxygenation and fragmentation, Organometallics 20 (12) (2001) 2552–2559. [31] G.T. Sazama, T.A. Betley, Multiple, disparate redox pathways exhibited by a Tris(pyrrolido)ethane iron complex, Inorg. Chem. 53 (1) (2014) 269–281. [32] A.B. Scharf, T.A. Betley, Electronic perturbations of iron dipyrrinato complexes via ligand b-halogenation and meso-fluoroarylation, Inorg. Chem. 50 (14) (2011) 6837–6845. [33] M. Tayebani, S. Conoci, K. Feghali, S. Gambarotta, G.P.A. Yap, Tri- and tetravalent and mixed-valence niobium complexes supported by a tripodal tripyrrolylmethane trianion, Organometallics 19 (2000) 4568–4574. [34] J. Jubb, L. Scoles, H. Jenkins, S. Gambarotta, Formation of bridging nitride versus terminal oxovanadium promoted by a vanadium(II) macrocyclic complex, Chem. Eur. J. 2 (7) (1996) 767–771. [35] C.R. Benson, A.K. Hui, K. Parimal, B.J. Cook, C.-H. Chen, R.L. Lord, A.H. Flood, K.G. Caulton, Multiplying the electron storage capacity of a bis-tetrazine pincer ligand, Dalton Trans. 43 (17) (2014) 6513–6524. [36] A.K. Hui, B.J. Cook, D.J. Mindiola, K.G. Caulton, Transition metal chlorides are lewis acids toward terminal chloride attached to late transition metals, Inorg. Chem. 53 (7) (2014) 3307–3310. [37] CV scan rates are as follows: 25, 50, 75, 100, 200, 400, 600, and 800 mV s1. [38] CV scan rates are as follows: 25, 50, 100, 200, 400, 600, 800 mV s1. [39] CV scan rates are as follows: 25, 50, 100, 200, 400, 600 mV s1.