Synthesis and Reactivity of a Cerium(iii) Scorpionate Complex Containing a Redox Non-innocent 2,2′-bipyridine Ligand

The Ce(III) hydrotris(3,5-dimethylpyrazolyl)borate complex [Ce(Tp Me2)2(κ 2-dmpz)] (1) (Tp Me2 = {HB(dmpz)3} − ; dmpz = 3,5-dimethylpyrazolide) was isolated in fair yield from the reaction of [Ce(I)3(THF)4] with two equivalents of [K(Tp Me2)] via the facile decomposition of Tp Me2. [Ce(Tp Me2)2(bipy)] (2) was synthesized in poor yield by the " one-pot " reaction of [Ce(I)3(THF)4], bipy (bipy = 2,2′-bipyridine), KC8 and two equivalents of [K(Tp Me2)] in tetrahydrofuran (THF). The reaction of 2 with N-methylmorpholine-N-oxide produced the known decomposition product [Ce(Tp Me2)(μ-BOp Me2)]2 (3) (BOp Me2 = {HBO(dmpz)2} 2−) in poor yield, presumably by N–O and B–N bond cleavage of a reactive intermediate. The reaction of 2 with trimethylsilylazide gave [Ce(Tp Me2)2(N3)] (4) in poor yield; the fate of bipy and the trimethylsilyl group is unknown. Complexes 1–4 were characterized by single crystal XRD, NMR and FTIR spectroscopy and elemental analysis. Complex 2 was additionally probed by UV/Vis/NIR and Electron Paramagnetic Resonance (EPR) spectroscopies, Cyclic Voltammetry (CV) and magnetometry, which together indicate a formal 4f 1 Ce(III) center coordinated by a bipy· − radical anion in this system.


Introduction
Complexes that exhibit terminal unsupported multiple bonds between a transition metal and a p-block element are legion, and interest in these species has surged in tandem with their increasing applicability in synthetic processes [1].In contrast, the corresponding lanthanide (Ln) chemistry is underdeveloped, as the large orbital energy mis-match between Ln valence orbitals and the 2pz orbital of C, N and O makes these bonds highly polarized, making them prone to decomposition and oligomerisation pathways [2,3].As such the isolation of Ln=CR2, Ln=NR and Ln=O bonds is a major synthetic challenge, with the first structurally characterized supported Ce(IV)=CR2 [4], terminal unsupported Ln=NR (Ln = Y, Lu) [5] and terminal supported Ce(IV)=O [6] complexes reported very recently, whilst the corresponding terminal unsupported Sc=NR chemistry has flourished since the first example was reported in 2010 by Chen [7].
Synthetic routes to terminal unsupported actinide (An) An=CR2, An=NR and An=O multiple bonds are far more developed than pathways to the corresponding Ln species [8,9].Andersen [10], Bart [11][12][13] and Zi and Walter [14] have shown that U(IV) and Th(IV) oxo and imido complexes can be prepared by the respective addition of N-oxides or organoazides to An(II) synthons, which mimic the reactivity of An(II) complexes although the metal is formally in the +4 oxidation state.We set out to find if this methodology could be extended to cerium, which is unique amongst the Lns as it has a readily accessible +4 oxidation state [E Ө Ce(IV) → Ce(III) = 1.74 V] [15].
We envisaged that an analogous Ce(II) synthon, [Ce(Tp Me2 )2(bipy)], could react with N-oxides or organoazides to give complexes that contain terminal unsupported Ce(IV)=O or Ce(IV)=NR multiple bonds, provided the Ce(III) → Ce(IV) oxidation potential in these complexes could be overcome by these reagents.[Ce(Tp Me2 )2(bipy)] should contain a formal Ce(III) 4f 1 center and the coupling of this unpaired spin with a ligand radical could lead to unusual physicochemical properties.It is noteworthy that interesting multiconfigurational behavior has previously been observed in [Yb(Cp*)2(bipy)] [19,20], where the electronic ground state contains contributions from Yb(III)/bipy• − (4f 13 ) and Yb(II)/bipy (4f 14 ) configurations, but the donor properties of Cp* ligands are not comparable to those of Tp Me2 ligands.The synthesis and characterization of [Ce(Tp Me2 )2(bipy)] and its reactivity towards N-oxides and organoazides is described herein.
The elemental analysis data for 1 correlated well with the predicted values, as with the other complexes (2-4, see below) reported in this paper.The 1 H NMR spectrum of 1 exhibited resonances over a wide chemical shift range (δH: −7.99 to +18.26 ppm) due to the 4f 1 formulation of Ce(III), and this precluded any interpretation of 13 C{ 1 H} NMR spectroscopic data.However, a single resonance was observed in the 11 B{ 1 H} NMR spectrum of 1 (δB: −0.66 ppm), which is downfield of that reported for [Ce(Tp Me2 )(COT)] (δB: −21.0 ppm) [27].The solution magnetic susceptibility of 1 was measured by the Evans method (μeff = 2.33 µB) [28], and this correlates well with the free ion approximation for a Ce(III) 4f 1 , 2 F5/2 ground term system (μeff: 2.54 µB [29]) considering weighing errors.The FTIR spectrum exhibited an absorption at 2547 cm −1 that is characteristic of the B-H stretching mode in tripodal Tp systems, and this was also seen in 2-4 (see below) [30,31].
The solid state structure of 1•C7H8 was determined by a single crystal XRD study and is depicted in Figure 1, with selected bond lengths and angles compiled in Table 1  Marques and co-workers previously reported that the one-pot reaction of [La(Cl)3(THF)1.5]with two equivalents of [K(Tp Me2 )] and one equivalent of bipy over a Na/Hg amalgam in THF gave [La(Tp Me2 )2(bipy)] in 90% yield [17].We modified this procedure to prepare [Ce(Tp Me2 )2(bipy)] ( 2) using [Ce(I)3(THF)4] [21], two equivalents of [K(Tp Me2 )] [22], one equivalent of bipy and KC8 [33] in THF (Scheme 2), in an effort to avoid the use of mercury.A poor crystalline yield of 2 (22%) was reproducibly obtained using our method, likely a result of the difficulties encountered during the synthesis of 1 (see above), though no other products could be identified in the reaction mixtures.The intense dark red color of 2 indicated that the complex contained a bipy• − radical monoanion; therefore we collected a wide range of characterization data to probe this system.

Scheme 2. Synthesis of 2.
The 1 H NMR spectrum of 2 exhibits a similar range of resonances to 1 [δH: −9.30 to 19.84 ppm], with a wide spectral window of +200 to −200 ppm evaluated as bipy resonances were observed at large negative chemical shifts for [Ce(Cp*)2(bipy)] (Cp* = C5Me5) [34].The bipy proton resonances in 2 could only be tentatively assigned, as neutral and monoanionic bipy• − resonances may be difficult to distinguish in a paramagnetic spectrum.In Marques' report of the 1 H NMR spectrum of the lanthanum analogue [La(Tp Me2 )2(bipy)] these signals were not observed [17].One signal was seen in the 11 B{ 1 H} NMR spectrum of 2 (δB: 10.48 ppm); however the 13 C{ 1 H} NMR spectrum could not be interpreted.The FTIR spectrum of 2 exhibited characteristic absorptions for a bipy radical anion (v = 1541, 1494 and 941 cm −1 ) with the lowest frequency peak correlating with that assigned for [La(Tp Me2 )2(bipy)] (v = 940 cm −1 ) [17].The solution susceptibility of 2 (μeff = 1.59 µB) was reproducibly much lower than the value obtained for 1, which could be attributed to either partial decomposition of 2 or antiferromagnetic coupling between Ce(III) and bipy• − .To probe this result a variable temperature magnetic analysis was performed together with an EPR spectroscopy study (see below).
The solid state configuration of 2•(C4H8O)2 was confirmed by single crystal XRD (depicted in Figure 2, with selected bond lengths and angles compiled in Table 1).The metrical parameters of structurally analogous [La(Tp Me2 )2(bipy)]•(C4H8O)2 have been discussed previously [17], and as these are nearly identical to those in 2•(C4H8O)2 considering the small difference in ionic radii between lanthanum and cerium [32] these will not be commented on here for brevity.The salient point of the discussion of the structure of [La(Tp Me2 )2(bipy)] is that the distances throughout the bipy scaffold were used as evidence of its radical monoanionic formulation (the NMR and FTIR spectroscopy and magnetic data of [La(Tp Me2 )2(bipy)] concurred with this assignment) [17], and statistically identical intra-ligand bond distances were observed in the coordinated bipy framework in 2. The electronic structure of 2 was probed by additional techniques in an effort to unequivocally establish a formal Ce(III)/bipy• − configuration.The UV/Vis/NIR spectrum of 2 exhibits a strong absorption at 387 nm (εmax ≈ 4100 M −1 •cm −1 ) and a weaker, broad absorption at 812-972 (εmax ≈ 500 M −1 •cm −1 ) nm (Figure 3).These absorptions are comparable to those observed for [Yb(Cp * )2(bipy)], which were assigned as π→π* and π*→π* transitions that arise from the unpaired electron in the bipy radical π-system [19].
Bipy has neutral, monoanionic and dianionic forms that can be easily interchanged and these processes are readily observed by electrochemistry [35]; therefore the cyclic voltammogram of 2 was obtained (Figure 4).Only one Nernstian quasi-reversible process was observed within the range limited by solvent and electrolyte experimental conditions (E = −2.34V vs. Fc/Fc + , ΔE = 95 mV), which was assigned to a [Ce(Tp Me2 )2(bipy)] + / [Ce(Tp Me2 )2(bipy• − )] process based on its similarity to the first reduction potential of bipy that we measured using identical conditions (E = −2.72V vs. Fc/Fc + ).The second reduction potential of bipy and the oxidation of Ce(III) to Ce(IV) could not be observed under the conditions employed, which indicates that strong oxidizing agents are required to form Ce(IV) complexes supported by Tp Me2 ligands.This is unsurprising given both the considerable precedent of Tp ligands stabilizing metals in low oxidation states [18] and the dominance of electronrich donor ligands in Ce(IV) chemistry [36].The powder X-band EPR spectrum of 2 was collected in an effort to observe resonances arising from coupling of the Ce(III) and bipy• − radical unpaired electrons.No signal was seen at room temperature but a highly anisotropic complex spectrum was obtained at 5 K.We performed a variable temperature study to track the thermal evolution of the spectrum (Figure 5).The bulk features at 5 K are almost fully maintained at 10 K, but at 15 K the spectrum is less distinct and at higher temperatures the fine structure cannot be discerned from the baseline.The 5 K spectrum could not be modeled and therefore we were not able to extract any parameters but this highlights the complex nature of the coupling between the bipy• − radical electron and the 2 F5/2 ground state doublet deriving from Ce(III).Magnetic measurements were performed on a solid sample of 2 suspended in eicosane and the magnetic susceptibility product (χMT) was measured from 300 to 2 K in a 10,000 Oe applied dc field (Figure 6).At 300 K, χMT = 0.44 cm 3 •Kmol −1 (1.21 μB), which is comparable to the moment measured in solution at 298 K (0.58 cm 3 •Kmol −1 ), taking into account weighing errors.These values are much lower than is predicted for a system containing one organic radical and one Ce(III) center (1.1 cm 3 •Kmol −1 ) and are closer to a Ce(III)-only system.The discrepancy between predicted and observed χMT values at 300 K can be attributed to the presence of diamagnetic impurities, or more complex magnetic behavior in 2. The χMT value decreases to 0.15 cm 3 •Kmol −1 (0.41 μB) at 2 K.We analyzed 2 further by measuring the magnetization against variable applied dc field at 5 K from 0-70,000 Oe, finding a near-linear correlation of M vs. H, with M = 0.39 μB at 70,000 Oe (Figure 7).In contrast at 2 K the data forms a curve, with a linear dependence of M vs. H from 0 to around 20,000 Oe (M = 0.27 μB), but at higher fields the effect of field on M is reduced.At both temperatures saturation was not reached at the maximum obtainable field, and this was observed reproducibility on separate samples, consistent with a system containing two unpaired electrons.There is also an essentially linear relationship of M and H from 35,000-70,000 Oe (M = 0.50 μB at 70,000 Oe) with ∂M/∂H for this range of fields lower than for the corresponding data at 5 K.

Synthesis and Structural Characterization of [Ce(Tp
We were interested in finding out if 2 could react with 2e − oxidizing agents as a Ce(II) synthon, mimicking the reactivity profile of a Ce(II) complex, even though the measurements in Section 2.1.3indicate that Ce is in the +3 oxidation state in 2. The reactions of 2 with N-oxides were investigated as Bart previously showed that the treatment of [U(Tp Me2 )(bipy)] with pyridine-N-oxide gave the U(IV) terminal oxo complex [U(Tp Me2 )(O)] [11].In contrast the reaction of 2 with N-methylmorpholine-N-oxide, pyridine-N-oxide or TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl) all gave the Ce(III) complex [Ce(Tp Me2 )(μ-BOp Me2 )]2 (3) as the only isolable product in ca.20% crystalline yield in each case (Scheme 3).Complex 3 was previously reported by Sella and co-workers to form in trace amounts via the partial hydrolysis of a range of heteroleptic complexes, [Ce(Tp Me2 )2(X)] (X = anionic ligand), and a mechanism for this decomposition was postulated whereby dmpzH was generated [37].As no other products could be identified during the formation of 3 it is not certain if the oxygen atom derives from initial coordination of the N-oxide or adventitious O2/H2O; however the isolation of 3 was reproducible, performed with the strict exclusion of oxygen and water, and was formed in higher yields than previously reported.

Scheme 3. Synthesis of 3.
Whilst the solid state structure of 3 was obtained previously by Sella, no other characterization data were reported [37]; therefore we acquired elemental analysis and collected spectroscopic data for 3. Unusually in the original report 3 was reported to exhibit a dark red color [37], whilst the samples we obtained were pale yellow.The 1 H NMR and FTIR spectra of 3 were unremarkable and as with 1 and 2 the 13 C{ 1 H} NMR spectrum could not be interpreted, though the 11 B{ 1 H} NMR spectrum of 3 contained one resonance (δB: 0.60 ppm).The Evans method susceptibility was also determined (μeff: 2.71 µB).Although the solid state structure of 3 has been reported previously [37], we include this data here for completeness as we obtained a polymorph (depicted in Figure 8, with selected bond lengths and angles compiled in Table 1).

Synthesis and Structural Characterization of [Ce(Tp Me2 )2(N3)] (4)
Andersen [10], Bart [11] and Walter and Zi [13] have all shown that azides react with Th(II) synthons to yield Th(IV) imido complexes.Therefore we added trimethylsilyl azide to 2, and when no reaction was observed at room temperature over three days the reaction mixture was refluxed for 16 h and the Ce(III) complex [Ce(Tp Me2 )2(N3)] (4) was obtained as the only isolable product (Scheme 4).Whilst the mechanism for the formation of 4 and the fate of the silyl group is unclear, N-Si bond cleavage is relatively facile and Bart recently reported a range of reductive heterocouplings of the coordinated reduced bipy ligand in [U(Cp*)2(bipy)] with ketones [38], though the electronics of the spectator ligands in this system are vastly different to 2. Germane to this, no reaction was observed between 2 and mesityl azide, even after prolonged heating, showing that the preparation of Ce(IV) imido complexes by a formal two electron oxidation of 2 by silyl and aryl azides is not favored.Gratifyingly, the 1 H and 13 C{ 1 H} NMR spectra of 4 could be fully assigned, with the 13 C signals all observed between 0 and 200 ppm.The Evans method susceptibility determined for 4 (μeff = 1.94 µB) was higher than that obtained for 2. Only one signal was observed in the 11 B{ 1 H} NMR spectrum of 4 (δB: −25.29 ppm), albeit at a higher field than 1-3, which we attribute to variable paramagnetic shift effects in the four complexes.The FTIR spectrum of 4 exhibited the diagnostic B-H stretching mode at 2555 cm −1 , together with characteristic absorptions for an azide group (v = 2083 and 2065 cm −1 ) that are comparable to the absorption observed for [Ce(TRENDSAL [39].The composition of 4 was determined by single crystal XRD (depicted in Figure 9, with selected bond lengths and angles compiled in Table 1

General Procedures
All manipulations were carried out using standard Schlenk and glove box techniques under an atmosphere of dry argon.Solvents were dried by refluxing over potassium and degassed before use.All solvents were stored over potassium mirrors (with the exception of THF which was stored over activated 4 Å molecular sieves).Deuterated solvents were distilled from potassium, degassed by three freeze-pump-thaw cycles and stored under argon.[Ce(I)3(THF)4] [21], [K(Tp Me2 )] [22] and KC8 [33] were prepared according to published procedures.Most solid reagents were dried under vacuum for 4 h and most liquid reagents were dried over 4 Å molecular sieves and distilled before use. 1 H, 13 C{ 1 H} and 11 B{ 1 H} NMR spectra were recorded on a spectrometer operating at 400.2, 100.6 and 128.3 MHz, respectively; chemical shifts are quoted in ppm and are relative to TMS ( 1 H, 13 C) or external BF3•Et2O (1.0 M in CDCl3) ( 11 B).Magnetic susceptibility was determined by the Evans method [27].FTIR spectra were recorded as Nujol mulls in KBr discs.Elemental microanalyses were carried out by Stephen Boyer at the Microanalysis Service, London Metropolitan University, UK.UV/Vis spectra were recorded in sealed 10 mm pathlength cuvettes.Redox potentials are referenced to the [Fe(Cp)2] + /[Fe(Cp)2] couple, which was used as an internal standard.Cyclic voltammetry was carried out using a sealed cell and a three-electrode arrangement, with a Pt wire working electrode, Pt flag secondary electrode and an AgCl/Ag wire pseudo-reference electrode.The susceptibility and magnetization curves were obtained using a SQUID magnetometer (Quantum Design, San Diego, CA, USA).The powdered sample was suspended in eicosane and sealed in a borosilicate glass NMR tube under vacuum, with diamagnetic contributions subtracted from the data.X-band EPR spectra were recorded using a MD5 dielectric resonator (Bruker, Billerica, MA, USA).The spectrometer is equipped with a CF935 cryostat connected to an Intelligent Temperature Controller (Oxford Instruments, Abingdon, UK).

X-ray Crystallography
The crystal data for complexes 1-4 are compiled in Tables 2 and 3. Crystals were examined on CCD area detector diffractometers using graphite-or mirror-monochromated Mo Kα (λ = 0.71073 Å) or Cu Kα (λ = 1.54184Å) radiation.Intensities were integrated from data recorded on 1° frames by ω (1-3) or φ and ω rotation (4).Cell parameters were refined from the observed positions of all strong reflections in each data set.A Gaussian grid face-indexed (1-3) or multi-scan (4) absorption correction with a beam profile correction was applied.The structures were solved variously by direct and heavy atom methods and were refined by full-matrix least-squares on all unique F 2 values, with anisotropic displacement parameters for all non-hydrogen atoms, and with constrained riding hydrogen geometries; Uiso(H) was set at 1.2 (1.5 for methyl groups) times Ueq of the parent atom.The largest features in final difference syntheses were close to heavy atoms and were of no chemical significance.CrysAlisPro [40] was used for control and integration, SHELXTL [41] and OLEX2 [42] were employed for structure solution and refinement and POVRAY [43] was used for molecular graphics.Highly disordered lattice solvent in 3 could not be modelled and Platon SQUEEZE was used to resolve this component [44].CCDC 1423263-1423266 (1-4) contain the supplementary crystal data for this article.These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.  params)] 1/2 for all data.

Conclusions
We have investigated novel synthetic routes to prepare heteroleptic cerium tris(pyrazolyl)borate complexes, including a Ce(II) synthon, 2, that formally contains a monoreduced bipy• − ligand and a Ce(III) center.We used a wide range of techniques to probe the electronic structure of 2 and conclude that the coordinated bipy exhibits features of a monoreduced radical.However, the EPR spectrum and magnetic measurements indicate that the coupling of this radical with the formal 4f 1 electron is complex, and future analysis of a more extensive library of Ce(II) synthons containing radical ligands would facilitate a deeper understanding of this coupling.A selected reactivity study of 2 with potential 2e − oxidants was performed and whilst bipy was eliminated in each isolated product cerium was not found in the +4 oxidation state.The formation of 3 and 4 suggests that the reaction of N-oxides and azides with Ce(II) synthons supported by bipy• − and Tp Me2 is not a favored route to Ce(IV) complexes exhibiting Ce=O or Ce=NR multiple bonds under the conditions we employed.

Figure 1 .
Figure 1.Molecular structure of 1•C7H8 with selective atom labeling.Displacement ellipsoids set at 30% probability level and hydrogen atoms (except B-H) and lattice solvent omitted for clarity.

Figure 9 .
Figure 9. Molecular structure of 4•C7H8 with selective atom labeling.Displacement ellipsoids set at 30% probability level and hydrogen atoms (except B-H) and lattice solvent omitted for clarity.
a Conventional R =