Distinct electronic structures and bonding interactions in inverse-sandwich samarium and ytterbium biphenyl complexes

Abstract

Inverse-sandwich samarium and ytterbium biphenyl complexes were synthesized by the reduction of their trivalent halide precursors with potassium graphite in the presence of biphenyl. While the samarium complex had a similar structure as previously reported rare earth metal biphenyl complexes, with the two samarium ions bound to the same phenyl ring, the ytterbium counterpart adopted a different structure, with the two ytterbium ions bound to different phenyl rings. Upon the addition of crown ether to encapsulate the potassium ions, the inverse-sandwich samarium biphenyl structure remained intact; however, the ytterbium biphenyl structure fell apart with the concomitant formation of a divalent ytterbium crown ether complex and potassium biphenylide. Spectroscopic and computational studies were performed to gain insight into the electronic structures and bonding interactions of these samarium and ytterbium biphenyl complexes. While the ytterbium ions were found to be divalent with a 4f¹⁴ electron configuration and form a primarily ionic bonding interaction with biphenyl dianion, the samarium ions were in the trivalent state with a 4f⁵ electron configuration and mainly utilized the 5d orbitals to form a δ-type bonding interaction with the π* orbitals of the biphenyl tetraanion, showing covalent character.

---

Introduction

The chemistry of rare earth elements (Sc, Y, and lanthanides) is still dominated by their trivalent ions, with the few exceptions of Ce(IV), Eu(II), Yb(II), and Sm(II), while the field of unusual redox chemistry is expanding.1–3 The overwhelming stability of the trivalent state results from the high electropositivity and the core-like 4f orbitals of the rare earth ions.4 Unlike the d orbitals of transition metals, the 4f orbitals of lanthanides are strongly shielded by the 5s2 and 5p6 electrons so that they barely interact with the ligand-based orbitals.4–9 As a result, the bonding interactions are primarily ionic with a negligible covalent character.10 This is in contrast to actinides, especially light actinides, which have been shown to exhibit distinctive redox chemistry and significant covalent interactions.10–14 However, a breakthrough in both high and low valent rare earth chemistry has emerged in recent years. For example, Mazzanti and La Pierre independently reported the synthesis and characterization of Tb(IV)15–17 and Pr(IV) complexes,18 which were the first examples of tetravalent molecular rare earth metal complexes beyond Ce(IV). In addition, Li and Zhou showed that pentavalent praseodymium oxides or nitrides could be trapped and characterized in a noble gas matrix.19,20 On the other hand, low valent rare earth chemistry had gradually progressed at the turn of the century from Sm(II)21 to Tm(II),22 Dy(II)23 and Nd(II).24 Further advance to afford divalent molecular complexes for other rare earth metals was once considered impractical due to the extreme negative M3+/2+ reduction potentials.25–28 This is in accordance with the solid state rare earth metal diiodides: Eu, Yb, Sm, Tm, Dy, and Nd form a genuine M(II)I2 salt, while other rare earth metal diiodides exist as metallic or semi-conducting M(III)I2(e−) with extra electrons delocalized in the band composed of the metal d orbitals (Chart 1).29–31 The seminal example of [Cp′′3La]− (Cp′′ = 1,3-C5H3(SiMe3)2) reported by Lappert et al. demonstrated that divalent molecular rare earth metal complexes beyond Nd(II) were accessible.32 Later, Evans et al. expanded this chemistry and completed the series of divalent complexes for all rare earth metals, except the radioactive Pm.33–37 Recently, the syntheses of the neutral sandwich complexes Ln(II) (CpiPr52) (Ln = Tb and Dy, CpiPr5 = C5(iPr)5) were reported by Long et al.38

Chart 1 Scale of M^3+/2+ (M = rare earth metal) reduction potentials²⁸ and electron configurations of divalent rare earth metal complexes in different coordination environment: pink, M(II)I₂;²⁹ blue, [K(crypt)](Cp′₃M);³⁹ green, [K(crypt)][((^Ad,MeArO)₃mes)M];⁴² red, [(NN^TBS)M]₂(μ-biphenyl)[K(solv)]₂ (this work).

Notably, in these compounds, rare earth(II) ions adopt a non-traditional 4fⁿ5d¹ configuration with the additional electron occupying on the 5d_z² orbital.^28,39–41 Different from the core-like 4f orbitals, the 5d orbitals are diffused and greatly influenced by the ligand field.¹⁰ Systematic studies of the [K(crypt)][Cp′₃M] (Cp′ = C₅H₄(SiMe₃), M = Y and lanthanides)^39,40 and [K(crypt)][((^Ad,MeArO)₃mes)M] (M = Y and lanthanides) ((^Ad,MeArO)₃mes = tris(aryloxide)arene)^42,43 series reveal that the electron configuration of rare earth(II) ions depend on the coordination environment: in the former series, the switch from a 4fⁿ⁺¹ to 4fⁿ5d¹ ground state takes place at Dy(II) (−2.5 V for Dy^3+/2+vs. standard hydrogen electrode, SHE),^39,40 while in the latter series, this switch does not happen even for Nd(II) (−2.6 V for Nd^3+/2+vs. SHE),⁴² and a ligand-centered instead of a metal-centered reduction occurred for rare earth ions that are more difficult to reduce than Nd (Chart 1).⁴³ This dependence of ground state electronic configuration for divalent rare earth ions suggests that it is possible to facilitate the transition from 4fⁿ⁺¹ to 4fⁿ5d¹ ground states through ligand manipulation.^31,44,45 Therefore, we were wondering how far this approach can push the limit for 4f → 5d transitions, especially for traditional divalent rare earth metals with a 4fⁿ⁺¹ electron configuration, such as Sm(II) and Yb(II).

Previously, we reported the synthesis and characterization of a series of inverse-sandwich rare earth metal biphenyl complexes [(NN^TBS)M]₂(μ-biphenyl)[K(solv)]₂ (M₂-biph-K₂, NN^TBS = fc(NSi^tBuMe₂)₂, fc = 1,1′-ferrocenediyl; M = Sc, Y, La, Lu, Gd, Dy, and Er) with an unprecedented tetranionic phenyl ring bridging two rare earth(III) ions.^46,47 Structural, spectroscopic, and computational studies of the representative yttrium complex supported the assignment that the tetranionic phenyl ring is a 10π-electron aromatic system stabilized by a δ-type bonding interaction between the 4d orbitals of Y(III) ions and the antibonding π* orbitals of the phenyl ring. This δ bonding interaction has a moderate covalent character since the contribution of yttrium orbitals is over 20% in HOMO and HOMO−1 (HOMO = highest occupied molecular orbital).⁴⁶ In the case of Dy₂-biph-K₂, instead of a reduction to 4f¹⁰ Dy(II), structural and magnetic data suggested the presence of 4f⁹ Dy(III) ions.⁴⁷ We have then become interested in the cases of other traditional divalent rare earth metals, in particular, Sm and Yb, because their M^3+/2+ reduction potentials are different from those of other rare earth metals and also apart from each other (Chart 1).

Herein, we report the synthesis of inverse-sandwich samarium and ytterbium biphenyl complexes, which are characterized by X-ray crystallography, NMR spectroscopy, elemental analysis, and UV/Vis/NIR spectroscopy. The structural and spectroscopic characterization supports the formulation of Sm(III) ions in the samarium biphenyl complexes and Yb(II) ions in the ytterbium biphenyl complexes. Upon encapsulation of the potassium ions, the coordination between rare earth ions with biphenyl remained intact for Sm but collapsed for Yb. Density functional theory (DFT) calculations show that the samarium and ytterbium biphenyl complexes have distinct electronic structures and bonding interactions: in the samarium biphenyl complex, Sm(III) ions have a 4f⁵ electron configuration and the 5d orbitals form a δ-type bonding interaction with the π* orbitals of the biphenyl tetraanion, showing a covalent character, while in the ytterbium biphenyl complex, Yb(II) ions have a 4f¹⁴ electron configuration and are bound to the biphenyl dianion via a primary ionic interaction.

Results and discussion

It has been shown that the ferrocene diamide ligand NN^TBS is well suited to stabilize rare earth metal complexes with a variety of reduced arenes, including biphenyl,^46,47 naphthalene,^48–50 and (E)-stilbene.⁵¹ Therefore, we prepared (NN^TBS)SmI(THF)₂ (Sm-I) and (NN^TBS)YbCl(THF)₂ (Yb-Cl) according to literature procedures,⁵² and employed them as trivalent metal precursors. Analogous to the synthesis of yttrium biphenyl complexes,⁴⁶ we attempted the reduction of Sm-I or Yb-Cl with potassium graphite (KC₈) in the presence of biphenyl.

Samarium biphenyl reduction

Upon the addition of 2.5 equivalents of KC₈ into a pre-cooled THF solution of an equivalent of Sm-I and 0.5 equivalents of biphenyl at −78 °C, the solution color turned immediately to dark (Scheme 1a). The reaction mixture was warmed up to room temperature and stirred for 10 min. After quickly passing it through a Celite column to remove insoluble solids, the filtrate was dried under a reduced pressure to yield a crude product. The NMR spectrum of the crude product showed only paramagnetic signals with the absence of starting materials. X-ray crystallography confirmed the product was [(NN^TBS)Sm]₂(μ-biphenyl)[K(toluene)]₂ (Sm₂-biph-K₂). An analytically pure material was obtained in a moderate yield of 55% after washing the crude product with Et₂O.

Scheme 1 Synthesis of Sm₂-biph-K₂ (a) and Sm₂-biph-[K(18-crown-6)]₂ (b).

The structure of Sm₂-biph-K₂ is similar to that of [(NN^TBS)Y]₂(μ-biphenyl)[K(toluene)]₂ (Y₂-biph-K₂), with two Sm ions coordinated to the same phenyl ring and two potassium ions coordinated to the other phenyl ring (Fig. 1a). The phenyl ring bridging the two Sm ions adopted a μ-η⁶:η⁶-coordination mode with Sm–C distances ranging from 2.549(4) to 2.639(4) Å (average Sm–C distance of 2.60 Å) and a Sm–Sm distance of 4.336(1) Å, comparable to the corresponding values in Gd₂-biph-K₂, when the difference between their ionic radii (average Gd–C distance of 2.58 Å, Gd–Gd distance of 4.27 Å, R(Sm) − R(Gd) = 0.03 Å)⁵³ is taken into account.⁴⁷ The C–C distances of the Sm-bound phenyl ring range from 1.421(5) to 1.476(1) Å, with an average value of 1.45 Å and a C_ipso–C_ipso distance of 1.413(4) Å, while the average C–C distance of the other phenyl ring is 1.41 Å. These values are close to those of Y₂-biph-K₂ (the average C–C distance of the Y-bound phenyl ring: 1.46 Å; C_ipso–C_ipso: 1.414(4) Å; the average C–C distance of the other phenyl ring: 1.41 Å).⁴⁶ The Sm1–N1 distance of 2.341(6) Å is 0.08 Å longer than the Sm–N distance of 2.263(3) Å in Sm-I.⁵² However, this elongation is likely due to the weakening of the Sm–N bond by the strong bonding between Sm and the phenyl ring, since a similar elongation of 0.07 Å was observed in the case of Y₂-biph-K₂ and (NN^TBS)YI(THF)₂ (2.292(2) Å vs. 2.222(3) Å).^46,49 The Sm–C_centroid distance is 2.196(7) Å, much shorter than the Sm–C_centroid distances found in Sm(III) neutral arene complexes 2.521(5) Å in (Sm(C₆Me₆) (AlCl₄)₃ (ref. 54) and 2.638(4) Å in (trans-calix[2]benzene[2]pyrrole)SmCl).⁵⁵ The angle Sm1–C_centroid–Sm2 is 164.3(9)°, indicating a nearly linear arrangement of the three atoms. The Sm-bound phenyl ring is slightly distorted from planarity with a torsion angle of 6.3(9)°, while the two phenyl rings have a dihedral angle of 5.9(2)°. Overall, the structural parameters of Sm₂-biph-K₂ are consistent with the assignment of two Sm(III) ions and a charge-localized biphenyl tetraanion, as in the analogous Y₂-biph-K₂.

Fig. 1 Thermal-ellipsoid (50% probability) representations of Sm₂-biph-K₂ (a) and Sm₂-biph-[K(18-crown-6)]₂ (b). Hydrogen and solvent atoms, disordered counterparts, and the counter cations for Sm₂-biph-[K(18-crown-6)]₂ were omitted for clarity. Selected distances [Å] and angles [°]: for Sm₂-biph-K₂: Sm1–N1 2.341(6), Sm1–N2 2.452(7), Sm1–C1 2.614(2), Sm1–C2 2.639(3), Sm1–C3 2.569(4), Sm1–C4 2.623(1), Sm1–C5 2.639(4), Sm1–C6 2.574(3), Sm1–C_centroid 2.196(7), Sm1–Fe1 3.225(2), C1–C2 1.421(5), C2–C3 1.464(6), C3–C4 1.476(1), C4–C5 1.476(1), C5–C6 1.464(6), C6–C1 1.421(5), C4–C4A 1.412(7), C4A–C3A 1.442(2), C3A–C2A 1.389(7), C2A–C1A 1.389(1), C1A–C6A 1.389(1), C6A–C5A 1.389(7), C5A–C4A 1.442(2), K1–N2 2.817(7), K1–C_centroid 2.813(6), Sm1–Sm2 4.336(3); N1–Sm1–N2 103.5(7), Sm1–C_centroid–Sm2 164.3(9), torsion angle defined by (C2–C3) vs. (C5–C6) 6.3(9), dihedral angle between the phenyl rings 5.9(2). For Sm₂-biph-[K(18-crown-6)]₂: Sm1–N1 2.380(5), Sm1–N2 2.409(2), Sm1–C1 2.603(2), Sm1–C2 2.652(1), Sm1–C3 2.512(3), Sm1–C4 2.615(6), Sm1–C5 2.625(6), Sm1–C6 2.549(4), Sm1–C_centroid 2.146(8), Sm1–Fe1 3.352(1), C1–C2 1.447(3), C2–C3 1.458(2), C3–C4 1.479(1), C4–C5 1.477(4), C5–C6 1.451(3), C6–C1 1.432(2), C4–C4A 1.439(8), C4A–C3A 1.407(9), C3A–C2A 1.392(1), C2A–C1A 1.374(3), C1A–C6A 1.379(9), C6A–C5A 1.384(2), C5A–C4A 1.428(4), Sm1–Sm2 4.300(8); N1–Sm1–N2 98.1(3), Sm1–C_centroid–Sm2 178.8(5), torsion angles defined by (C2–C3) and (C5–C6) 10.7(9), dihedral angle between the two phenyl rings 9.0(7).

It was previously found that the K ions are not required to maintain the inverse-sandwich structure in Y₂-biph-K₂.⁴⁶ Therefore, we probed whether this is also the case for Sm₂-biph-K₂. The addition of two equivalents of 18-crown-6 to a THF solution of Sm₂-biph-K₂ at −78 °C (Scheme 1b) led to no obvious change. Crystallization of the reaction mixture resulted in the isolation of an ion pair complex, Sm₂-biph-[K(18-crown-6)]₂, in a 66% yield. X-ray crystallography established the molecular structure of Sm₂-biph-[K(18-crown-6)]₂ (Fig. 1b), which is analogous to Y₂-biph-[K(18-crown-6)]₂.⁴⁶ The charge localization is more prominent in Sm₂-biph-[K(18-crown-6)]₂ than in Sm₂-biph-K₂ after the encapsulation of the potassium ions: average C–C distance of the Sm-bound phenyl ring increased to 1.46 Å and C_ipso–C_ipso distance increased to 1.440(6) Å, while the average C–C distance of the other phenyl ring decreased to 1.39 Å. The bonding between Sm and phenyl ring also strengthened as evidenced by the slightly shorter average Sm–C distance of 2.59 Å and shorter Sm–Sm distance of 4.301(1) Å compared to 2.60 Å and 4.336(1) Å in Sm₂-biph-K₂, respectively. The Sm–C_centroid distance also decreased from 2.196(7) Å to 2.146(8) Å with an essentially linear arrangement of Sm1–C_centroid–Sm2 (178.8(5)°). The torsion angle of the Sm-bound phenyl ring also increased slightly to 10.7(9)°, as did the dihedral angle between the two phenyl rings, to 9.0(7)°. These structural changes observed upon the removal of the potassium ions was also observed in the case of Y₂-biph-K₂ and Y₂-biph-[K(18-crown-6)]₂, interpreted as the strengthening of the bonding interaction between yttrium and the bound phenyl ring.⁴⁶

Ytterbium biphenyl reduction

Initially, we attempted the synthesis of Yb biphenyl complexes through the one-pot reduction of Yb-Cl and biphenyl by KC₈, similarly to the synthesis of Sm₂-biph-K₂ (Scheme 2a). The reaction proceeded with the formation of multiple diamagnetic products corresponding to various Yb(II) products. Crystallization of the crude products from Et₂O yielded some dark single crystals that were determined by X-ray crystallography to be an ytterbium biphenyl complex [(NN^TBS)Yb]₂(μ-biphenyl)[K(Et₂O)]₂ (Yb₂-biph-K₂). However, the corresponding ¹H NMR spectrum showed a persistent contamination from other Yb(II) side products, even after a number of purification attempts by multiple fractional crystallization. In addition, we found that the relative amount of Yb(II) side products was sensitive to the reaction conditions. Since the Yb^3+/2+ reduction potential is −1.15 V vs. SHE,²⁶ which is far less negative than the reduction potential of −2.45 V vs. SHE for (biphenyl)^0/1−,⁵⁶ we rationalized that a simultaneous reduction of Yb-Cl and biphenyl resulted in the formation of multiple products. Therefore, we tried pre-mixing an excess biphenyl and KC₈ in order to generate a reduced biphenyl species (to simplify, we use [K₂(biphenyl)] for representation) and then added Yb-Cl (Scheme 2b). Through this protocol, we were able to obtain Yb₂-biph-K₂ reproducibly in high purity and a moderate yield of 52%.

Scheme 2 (a) One-pot reduction of Yb-Cl and biphenyl by KC₈; (b) optimized synthesis of Yb₂-biph-K₂ by pre-mixing an excess of biphenyl and KC₈; (c) direct reduction of Yb-Cl and transformations of Yb(II) products; (d) reaction of Yb₂-biph-K₂ with an excess of 18-crown-6.

The ¹H NMR spectrum of Yb₂-biph-K₂ in THF-d₈ at room temperature showed broad peaks indicating the structure of compound is fluxional in solution. Signals of SiCH₃ and SiC(CH₃)₃ of the NN^TBS ligands are in the typical diamagnetic region, consistent with the presence of 4f¹⁴ Yb(II) ions. A broad peak at 5.41 ppm likely belongs to the meta proton of the biphenyl anion, based on a comparison with a previously reported rare earth metal stabilized biphenyl dianion.⁵⁷ Four peaks at 127.9, 102.2, 101.8, and 87.2 ppm in the ¹³C NMR spectrum are assigned to the reduced biphenyl ligand, indicating that the two phenyl rings are equivalent. These ¹³C chemical shifts are distinct from those of Y₂-biph-K₂ (four peaks for the Y-bound phenyl ring at 86.8, 79.0, 73.1, and 58.6 ppm; four peaks for the K-bound phenyl ring at 138.5, 128.5, 114.8, and 103.5 ppm)⁴⁶ but are similar to those of the alkali metal biphenyl dianion (four peaks at 128.5, 104.6, 102.3 and 74.2 ppm).⁵⁸ Overall, the ¹H and ¹³C NMR spectra support the assignment of two 4f¹⁴ Yb(II) ions and a charge-delocalized biphenyl dianion.

Despite of the similar molecular formula, the molecular structure of Yb₂-biph-K₂ is distinct from Sm₂-biph-K₂ and other M₂-biph-K₂ compounds (Fig. 2a). The two Yb and K ions coordinate to different phenyl rings. The Yb1–N2 distance of 2.312(2) Å is about 0.13 Å longer than the average Yb–N distance of 2.182(5) Å in Yb-Cl. This elongation is likely due to the change in ionic radii upon going from Yb(III) to Yb(II) (R(Yb(II)) − R(Yb(III)) = 0.15 Å).⁵³ The two phenyl rings are equivalent, consistent with the symmetric solution structure determined by ¹H and ¹³C NMR spectra. The alternating C–C distances (1.41, 1.38, and 1.47 Å) within the phenyl ring and the shortening of the C_ipso–C_ipso distance to 1.396(4) Å imply that the negative charges are equally delocalized over the two phenyl rings and the biphenyl ligand is best described as a dianion.⁵⁷ The average Yb–C distance of 2.80 Å and the Yb1–C_centroid distance of 2.415(6) Å are much longer than those of 2.60 Å and 2.196(7) Å in Sm₂-biph-K₂, respectively, even after adjusting for the difference in ionic radii (R(Sm(III)) = 1.08 Å, R(Yb(II)) = 1.14 Å)⁵³, implying a much weaker interaction between the Yb ions and the phenyl rings. In addition, the two phenyl rings are not twisted but essentially coplanar, differing from other M₂-biph-K₂ but similar to the yttrium biphenyl dianion complex previously reported by Fryzuk et al.⁵⁷

Fig. 2 Thermal-ellipsoid (50% probability) representations of Yb₂-biph-K₂ (a) and [(NN^TBS)Yb(THF)]₂ (b). Hydrogen (except H involved in agostic interaction), disordered counterparts, and solvent atoms were omitted for clarity. Selected distances [Å] and angles [°]: For Yb₂-biph-K₂: Yb1–N1 2.395 (9), Yb1–N2 2.312(2), Yb1–C1 2.773(3), Yb1–C2 2.775(9), Yb1–C3 2.797(5), Yb1–C4 2.860(3), Yb1–C5 2.812(5), Yb1–C6 2.791(7), Yb1–C_centroid 2.415(6), Yb1–Fe1 3.097(7), C1–C2 1.412(1), C2–C3 1.378(8), C3–C4 1.471(9), C4–C5 1.469(8), C5–C6 1.376(7), C6–C1 1.413(7), C4–C4A 1.396(0), K1–N1 2.881(4), K1–C_centroid 2.829(8); N1–Yb1–N2 104.2(1), torsion angle defined by (C2–C3) vs. (C5–C6) 0.0(8), dihedral angle between the phenyl rings 0.0(1). For [(NN^TBS)Yb(THF)]₂: Yb1–N2 2.360(5), Yb1–N1 2.558(7), Yb1–O1 2.443(1), Yb1–Fe1 3.251(4), Yb1–C1 3.019(6), Yb1–H1 2.747(8); N1–Yb1–N2 133.8(5), Yb1–N1–Yb1A 92.7(1).

Since our initial attempt to synthesize a ytterbium biphenyl complex by the one-pot reduction procedure resulted in the concomitant formation of unknown divalent ytterbium side product(s) together with Yb₂-biph-K₂, we were interested in finding out their identity. The corresponding ¹H NMR spectra showed multiple peaks other than those belonging to Yb₂-biph-K₂ in the SiCH₃ region of the NN^TBS ligand. This implies that there are either multiple Yb(II) side products or one Yb(II) species of low symmetry. The direct reduction of Yb-Cl by 1.5 equivalents of KC₈ in THF yielded a single Yb(II) product with high symmetry as determined by ¹H NMR spectroscopy (Scheme 2c). However, upon removing volatiles and washing with hexanes, a poorly soluble brownish solid was obtained. The solid is insoluble in aliphatic solvents and barely soluble in aromatic solvents. The corresponding ¹H NMR spectrum in C₆D₆ showed weak signals but matched those of the side product from the one-pot reduction.

X-ray crystallography determined the molecular structure of this Yb(II) product to be a dimer, [(NN^TBS)Yb(THF)]₂ (Fig. 2b). Each Yb(II) ion is coordinated to one terminal amide, two bridging amides and a THF molecule. The Yb–N distance of 2.558(7) Å for the bridging amide is much longer than the terminal one of 2.360(5) Å, which is close to 2.312(2) Å in Yb₂-biph-K₂. An agostic interaction of the Yb ion and a C–H bond from the ferrocene backbone (Yb1–H1: 2.747(8) Å and Yb1–C1: 3.019(6) Å) may play a role in stabilizing this dimeric structure. This rigid structure may be responsible for its poor solubility and very low symmetry in solution. When dissolved in THF-d₈, the ¹H NMR spectrum showed a highly symmetric pattern for the NN^TBS ligand. We anticipated that the dimeric structure could be intercepted by the coordination of additional THF molecules. Crystallization from THF/hexanes yielded single crystals determined by X-ray crystallography to be the monomeric (NN^TBS)Yb(THF)₃ (Fig.S14†). The structural parameters are consistent with an Yb(II) ion. However, attempts to obtain an analytically pure sample of (NN^TBS)Yb(THF)₃ were not successful due to the loss of the coordinating THF molecules upon drying. This observation contrasts what we know about trivalent rare earth metal complexes supported by the NN^TBS ligand and indicates a weak interaction between Yb(II) and THF. In an effort to obtain an Yb(II) complex that is stable in both the solid and solution state, a chelating ligand, 18-crown-6, was employed. The chelation effect allowed a complete replacement of the coordinating THF molecules (Scheme 2c). The ¹H NMR spectrum of the resulting product, (NN^TBS)Yb(18-crown-6), showed a single peak at 3.43 ppm for 18-crown-6 protons, indicating a fluxtional behavior in solution. In the solid state, we obtained two coordination isomers from different crystals of the same crystallization batch, (NN^TBS)Yb(κ³-18-crown-6) (Fig. S16†) and (NN^TBS)Yb(κ⁴-18-crown-6) (Fig. S17†), which differed by the number of coordinated O-donors from 18-crown-6.

In previously reported M₂-biph-K₂ (ref. 46 and 47) and Sm₂-biph-K₂, it was found that upon the addition of 18-crown-6, the inverse-sandwich structure remained intact and the bonding interaction between rare earth ions and biphenyl strengthened in the ion pair complexes M₂-biph-[K(18-crown-6)]₂. However, the reaction of 18-crown-6 with Yb₂-biph-K₂ led to the collapse of the whole structure (Scheme 2d). Upon the addition of excess 18-crown-6, the dark solution of Yb₂-biph-K₂ became clouded. After crystallization, the dark crystals that precipitated out were determined by X-ray crystallography to be the previously reported [K(18-crown-6)][biphenyl],⁵⁹ while the orange supernatant was identified to be (NN^TBS)Yb(18-crown-6) by ¹H NMR spectroscopy (Fig. S10†). [K(18-crown-6)][biphenyl] is likely to be the decomposition product of the initially generated bare biphenyl dianion, which was not stable under the reaction conditions. In addition, we also identified [K(18-crown-6)][(NN^TBS)₂Yb] by ¹H NMR spectroscopy. We rationalized that the decomposition of Yb₂-biph-K₂ upon the addition of 18-crown-6 is due to the much weaker interaction between the Yb(II) ions and the biphenyl dianion.

Spectroscopic and magnetic characterization

To probe the electronic structures of samarium and ytterbium biphenyl complexes further, we collected the UV/Vis/NIR spectra of Sm₂-biph-K₂, Sm₂-biph-[K(18-crown-6)]₂, and Yb₂-biph-K₂ in THF. The spectra of Sm₂-biph-K₂ and Sm₂-biph-[K(18-crown-6)]₂ were almost identical, with an intense band around 403 nm (ε > 10⁴ M⁻¹ cm⁻¹) and a broad intense band centered at 638 nm (ε > 10⁴ M⁻¹ cm⁻¹, Fig. 3). The broad intense band in the visible region is responsible for the extremely dark color of their crystals and solutions; it may be attributed to ligand to metal charge transfer (LMCT) or an excitation of the ligand-based orbitals. However, the absorption spectrum of Yb₂-biph-K₂ lacked this broad intense band in the visible region (Fig. S18†). The near-infrared (NIR) region of Sm₂-biph-K₂ and Sm₂-biph-[K(18-crown-6)]₂ has a higher intensity than that of Yb₂-biph-K₂, in agreement with the latter having a 4f¹⁴ electronic configuration, and, thus, no f–f transitions. Moreover, there was no Yb(III) characteristic f–f transition observed around 980 nm (ref. 60) for Yb₂-biph-K₂, further supporting the presence of Yb(II) instead of Yb(III).

Fig. 3 UV/Vis/NIR spectra of Yb₂-biph-K₂ (black), Sm₂-biph-[K(18-crown-6)]₂ (red), and Sm₂-K₂-biph (blue) in THF at room temperature. The NIR region from 1000–1600 nm is enlarged in the upper right corner.

Magnetic susceptibility has been used to probe the ground state electronic configuration of divalent rare earth ions.⁶¹ In particular, Sm(II) and Sm(III) have distinct magnetic susceptibilities, with the normal range for Sm(II) being 3.4–3.8 μ_B, while the normal range for Sm(III) is 1.3–1.9 μ_B.^62–67 The room temperature magnetic susceptibilities of Sm-I, Sm₂-biph-K₂, and Sm₂-biph-[K(18-crown-6)]₂ were determined by the Evans method to be 1.53, 2.47, and 2.39 μ_B, respectively (per formula unit). These results are consistent with the assignment of Sm(III) having a 4f⁵ electron configuration in Sm₂-biph-K₂ and Sm₂-biph-[K(18-crown-6)]₂.

Preliminary X-ray absorption near edge structure (XANES) spectroscopy studies were conducted on Sm₂-biph-K₂ and Yb₂-biph-K₂ (see ESI for details†). Although the collected data agreed with the findings discussed above, the quality of the data was not high enough to allow decisive conclusions about the oxidation state of samarium and ytterbium in these biphenyl complexes.

Quantum chemical calculations

Theoretical calculations were carried out at the level of scalar relativistic density functional theory (DFT) to probe the electronic structures and bonding interactions of (Sm₂-biph)²⁻ (the structure is based on Sm₂-biph-[K(18-crown-6)]₂ with the potassium counter cations omitted for simplification) and Yb₂-biph-K₂ as well as two hypothetical isomeric structures of (Sm₂-biph)_iso²⁻ and [Yb₂-biph-K₂]_iso for comparison (in (Sm₂-biph)_iso²⁻, the two Sm ions are bound to different phenyl rings; in [Yb₂-biph-K₂]_iso, the two Yb ions are bound to the same phenyl ring while the two potassium ions are bound to the other phenyl ring, see also Fig. S19†). Herein, the alkyl substituents on the silicon atoms were replaced by hydrogen atoms for simplification. In order to determine the ground state of the (Sm₂-biph)²⁻ complex, two electronic configurations ¹¹A and ¹³A with different multiplicities were performed. The calculated energy of (Sm₂-biph)²⁻ (¹³A) is 26.6 kcal mol⁻¹ higher than that of (Sm₂-biph)²⁻ (¹¹A). Furthermore, there is also a larger discrepancy in the optimized structure parameters for (Sm₂-biph)²⁻ (¹³A) and those of the experimental compound (Table S3†). For example, the calculated average Sm–C distance of 2.68 Å and Sm-C_centroid distance of 2.33 Å are significantly longer than the experimental ones of 2.59 Å and 2.15 Å. The calculated torsion angle of the Sm-bound phenyl ring is only 0.9°, while the experimental one is 10.7(9)°. In contrast, the optimized structure parameters for (Sm₂-biph)²⁻ (¹¹A) are in good agreement with those of the experimental structure (Table S3†). The calculated average Sm–C distance of 2.56 Å and Sm–C_centroid distance of 2.10 Å are comparable to the experimental ones of 2.59 Å and 2.15 Å, respectively. The torsion angle of the Sm-bound phenyl ring is also reproduced well (12.4° vs. 10.7(9)°). Moreover, the calculated average C–C distance of the Sm-bound phenyl ring is 1.47 Å, which is close to the experimental average value of 1.46 Å. For the unbound phenyl ring, the calculated average C–C distance is 1.40 Å, similar to the experimental one of 1.39 Å. Overall, according to the DFT results, the (Sm₂-biph)²⁻ complex has a ¹¹A ground state, in which the oxidation state of the samarium ions is 3+ with a 4f⁵ electronic configuration.

As depicted in Fig. 4a, the highest occupied molecular orbital (HOMO, 40b) and HOMO−1 (41a) confirm the δ bonding interaction between the samarium ions and the bound phenyl ring in (Sm₂-biph)²⁻. Moreover, the contribution from the C 2p orbitals of the bound phenyl ring is around 60% and the contribution from the 5d orbitals of the samarium ions is over 20% on average. Furthermore, the natural population analysis (NPA) charges for the bound ring (−1.56) and the unbound ring (−0.25) are significantly different (Table S5†). Indeed, the δ bonding molecular orbital (MO) is formed by the interaction between the 5d atomic orbitals of samarium and the π* MOs of the Sm-bound phenyl ring. Based on the MO component percentage of the δ bonding interaction and NPA results, the four δ bonding electrons are mainly localized on the bound phenyl ring, resulting in a 10π-electron tetraanionic system in (Sm₂-biph)²⁻, similar to the previously reported Y₂-biph-K₂.⁴⁶ Meanwhile, the proposed isomer (Sm₂-biph)_iso²⁻ can be formulated as two 8π-electron dianionic systems (Fig. S21†). Apart from that, the energy of the hypothetical structure (Sm₂-biph)_iso²⁻ is calculated to be 30.3 kcal mol⁻¹ higher than the experimental structure (Sm₂-biph)²⁻ (Fig. S24†). For Yb₂-biph-K₂, the biphenyl ring contributes around 70% to the HOMO with a small amount (ca. 10%) of 4f ytterbium atomic orbitals, indicating that there is a weak interaction between the ytterbium ions and the biphenyl ring (Fig. 4b). Furthermore, HOMO−1 to HOMO−14 are mostly composed of the 4f electrons of ytterbium with almost no biphenyl contribution. A similar bonding pattern is found for the Kohn–Sham frontier MOs in the proposed isomer [Yb₂-biph-K₂]_iso (Fig. S22†).

Fig. 4 Kohn–Sham representations of HOMO and HOMO−1 and main atomic orbital contributions for (Sm₂-biph)²⁻ (a) and Yb₂-biph-K₂ (b). The alkyl substituents on the silicon atoms were replaced by hydrogen atoms for simplification (iso = 0.03).

The δ bonding assignment in the simplified (Sm₂-biph)²⁻, represented as (Sm₂-biph)_Cl²⁻ hereafter, in which the NN^TBS ligands are replaced by chlorine atoms, is also supported by adaptive natural density partitioning (AdNDP)⁶⁸ analysis as shown in Fig. S20.† The semi-localized results indicate that the bonding interaction between the samarium ions and the bound phenyl ring involve two 8c–2e δ bonds, which resemble the Kohn–Sham frontier molecular orbitals. Fig. 4 illustrates the bonding interaction scheme of (Sm₂-biph)_Cl²⁻ between the fragments of (Sm₂Cl₄)²⁺, with two Sm(III) ions each in a 4f⁵ electron configuration, and (biphenyl)⁴⁻.

The 4f orbitals are known to be radially too contracted in lanthanides to participate in chemical bonding, and form a nonbonding f-band shown as a red shadow in Fig. 5. On the other hand, the 5d orbitals are more radially extended than the 4f orbitals. Thus, the two unoccupied δ-type 5d orbitals (30a, 30b) in the [Sm₂Cl₄]²⁺ fragment and the two occupied π* orbitals (15b, 18a) of the biphenyl ring form two δ bonding MOs (41a, 40b). Furthermore, the stability of Sm(III) with a 4f⁵ electronic configuration is evident by the energy gap of 2.19 eV between the HOMO−1 and the 4f band. The stability of the whole compound is also revealed by the large HOMO–LUMO gap of 3.14 eV, indicating that the ground state is dominated by a single configuration and the description of structure and bonding based on single-configurational DFT methods is sensible.

Fig. 5 Bonding scheme of (Sm₂-biph)_Cl²⁻ illustrating the MO interactions between [Sm₂Cl₄]²⁺ and the biphenyl tetraanion (iso = 0.03).

The nature of the above-mentioned pairwise orbital interactions can be further analyzed by the energy decomposition analysis with natural orbitals for chemical valence (EDA-NOCV).^69,70 Table S4† shows the numerical results for the interaction between fragments of [Sm₂Cl₄]²⁺ and (biphenyl)⁴⁻ in (Sm₂-biph)_Cl²⁻. It shows that 60% of the attractive interaction ΔE_int is from the electrostatic attraction ΔE_elstat, 39% from the covalent orbital interaction ΔE_orb and the rest 1% is from the dispersion interaction ΔE_disp. The breakdown of the ΔE_orb into individual orbital contributions reveals that there are two major terms ΔE_orb(δ1) and ΔE_orb(δ2) due to the δ-type bonding with the energies of −278.75 and −192.77 kcal mol⁻¹, respectively. The total amount of −471.52 kcal mol⁻¹ for these two terms ΔE_orb(δ1) and ΔE_orb(δ2) is about the same as the total ΔE_orb term for Sm(C₅Me₅)₂ and Sm(C₄Me₄P)₂, with values of −494.4 and −484.8 kcal mol⁻¹, respectively.⁷¹ The other two terms ΔE_orb(π1) and ΔE_orb(π2) also show a charge flow from the bound phenyl ring to Sm centers, accounting for about 10% in the ΔE_orb. The δ bonding interaction in (Sm₂-biph)²⁻ is further characterized using the principal interacting orbital (PIO) approach.⁷² The calculated PIOs and principal interacting molecular orbitals (PIMOs) shown in Fig. S23† clearly reveal the δ-type bonding interaction between samarium ions and the biphenyl fragment in (Sm₂-biph)²⁻, which further supports the above electronic structure assignment. The δ bonding interaction in these inverse sandwich samarium biphenyl complexes is reminiscent of recently identified inverse sandwich lanthanide-boron binary clusters^73,74 and actinide metallacycles.⁷⁵

Discussion

Sm₂-biph-K₂ and Yb₂-biph-K₂ extend the series of M₂-biph-K₂ (M = Sc, Y, La, Lu, Gd, Dy, and Er) to traditional divalent lanthanide ions and provide a unique case study of the electronic structures of low valent rare earth ions. In the previously reported (Cp′₃M)⁻ series, despite the fact that Nd(II) and Dy(II) adopted a non-traditional 4fⁿ5d¹ electronic configuration, rare earth metals with less negative M^3+/2+ reduction potentials, including Tm, Sm, and Yb, maintained a traditional 4fⁿ⁺¹ electronic configuration for their divalent ions.^39,40 However, in the series of [((^Ad,MeArO)₃mes)M]⁻, all traditional divalent rare earth ions kept a 4fⁿ⁺¹ electron configuration.^42,43 In the M₂-biph-K₂ series, not only was dysprosium found unambiguously in the 3+ oxidation state,⁴⁷ but also samarium was confirmed to be Sm(III) with a 4f⁵ electronic configuration. The latter is remarkable since the (biphenyl)^0/1− reduction potential of −2.45 V vs. SHE⁵⁶ is ca. 0.90 V more negative than the Sm^3+/2+ reduction potential of −1.55 V vs. SHE.²⁵ Instead of a simultaneous reduction to the 4f⁶ Sm(II) ions and a biphenyl dianion or even a neutral biphenyl, Sm₂-biph-K₂ contained two Sm(III) ions and a biphenyl tetraanion with most negative charges localized in the Sm-bound phenyl ring.

To the best of our knowledge, Sm₂-biph-K₂ and Sm₂-biph-[K(18-crown-6)]₂ are the first structurally characterized and spectroscopically confirmed Sm(III) complexes with anionic arene ligands that are stable in the presence of coordinating solvents like THF.⁷⁶ The syntheses of (C₁₀H₈)Sm(THF)₃,⁷⁷ [(CpV(C₁₀H₈)SmCp(THF)_n] and [CpV(C₁₀H₈)]₂Sm(THF) (DME),⁷⁸ and (CpSm)₂(C₁₀H₈) (THF)₄ (ref. 79) were reported but their structures were not determined by X-ray crystallography. Magnetic and spectroscopic data suggested the formation of Sm(II) and a naphthalene dianion in these compounds.⁷⁶ The only other structurally characterized Sm(III) complexes with anionic arene ligands are (Cp*₂Sm)₂(C₁₄H₁₀), (Cp*₂Sm)₂(C₁₆H₁₀), and (Cp*₂Sm)₂(C₁₈H₁₂), which were synthesized by the reduction of the corresponding arene by Cp*₂Sm.⁸⁰ The structures of these compounds were characterized by X-ray crystallography and showed an unusual η³-allylic type of bonding between Sm ions and the arene ligand.⁸⁰ The structural and spectroscopic data supported the formulation of two Sm(III) ions and dianionic arene ligands. However, upon the addition of THF to their toluene solutions, the regeneration of neutral arenes and Cp*₂Sm(THF)₂ immediately occurred, implying that these Sm(III) arene complexes are intrinsically unstable and readily decompose to Sm(II) and neutral arenes.⁸⁰ Two structurally relevant samarium inverse-sandwich toluene complexes (Sm₂L₃)₂(μ-η⁶:η⁶-C₇H₈) and (KSmL₃)₂(μ-η⁶:η⁶-C₇H₈) (L = OSi(O^tBu)₃) were recently reported by Mazzanti et al.⁸¹ The former tetranuclear complex was formulated with four Sm(II) ions and a toluene dianion bridging between two Sm(II) ions, while the latter was assigned to two Sm(II) ions bridged by a neutral toluene ligand. It is also worth noting that in the series of formal zero-valent rare earth metal bis(arene) complexes reported by Cloke et al., Sm(1,3,5-^tBu₃C₆H₃)₂ is the least stable one among all M(1,3,5-^tBu₃C₆H₃)₂ (except for Ce, Eu, Tm, and Yb, which did not form isolable products).^82,83 The difference in the thermal stability of rare earth metal bis(arene) complexes is rationalized by the difference in the promotion energy from fⁿs² to fⁿ⁻¹d¹s² for different rare earth metals since the bonding interaction is mainly the back-donation from the d and s orbitals of the rare earth metals to the π* orbitals of the arenes.^82–85 These literature precedents suggest that samarium should have a low tendency to form a strong bonding interaction with arenes due to the relative stability of the 4f⁶ electronic configuration of the Sm(II) ion.

The above observations lead to the question why in the series of M₂-biph-K₂, most rare earth metals, including Sm, are in the trivalent state and utilize primarily the d_xy and d_x²−y² orbitals to form a strong δ bonding interaction with the π* orbitals of the bridging arene ligand. We propose two main reasons. First, the resulting arene tetraanion is a 10π-electron aromatic system, giving an extra aromatic stability to these inverse-sandwich compounds. In the series of (Cp*₂Sm)₂(arene), neither are the Sm ions bound to the π surface of the arene nor do the arenes fulfill the (4n + 2) Hückel rule.⁸⁰ In the case of (KSmL₃)₂(μ-η⁶:η⁶-C₇H₈), a formulation of two Sm(III) ions would result in a dianionic toluene ligand, which is anti-aromatic.⁸¹ Similarly, in the series of [((^Ad,MeArO)₃mes)M]^–, a single electron reduction could take place at either the metal center or the arene, but the latter would result in a mesitylene radical anion.⁴³ However, this aromatic stabilization could not explain why in other inverse-sandwich samarium arene complexes, such as (Sm₂L₃)₂(μ-η⁶:η⁶-C₇H₈), the toluene is a dianion and all Sm ions were divalent.⁸¹ It is also worth noting that a relevant trinuclear cerium toluene complex, [K(2.2.2-crypt)]₂[((KL₃Ce) (μ-η⁶:η⁶-C₇H₈))₂Ce], supported by the same siloxide ligand reported by Mazzanti et al. was formulated as three Ce(II) ions and two toluene dianions,⁸⁶ despite the much more negative Ce^3+/2+ reduction potential of −3.2 V vs. SHE, compared to Sm^3+/2+ reduction potential of −1.55 V vs. SHE.²⁸ Therefore, other than aromaticity, we considered that the ferrocene diamide ligand NN^TBS may also play a key role in stabilizing the unique electronic structure of Sm₂-biph-K₂. The δ bonding interaction in the inverse-sandwich complex Sm₂-biph-K₂ could be viewed by analogy to the sandwich compounds M(1,3,5-^tBu₃C₆H₃)₂. In the latter, the strength of the M–arene bond is correlated to the promotion energy from 4fⁿ6s² to 4fⁿ⁻¹5d¹6s²: the lower the promotion energy, the stronger the M–arene bond.⁸² While the energy of the 4f orbitals is barely changed by the ligand field, the 5d orbitals interact strongly with the ligand orbitals and their energy could be readily tuned by the ligand field. Compared to O-donors in (Sm₂L₃)₂(μ-η⁶:η⁶-C₇H₈), N-donors in Sm₂-biph-K₂ provide a stronger ligand field, which is also responsible for lowering the energy level of the 5d orbitals. Moreover, the flexibility of the NN^TBS ligand may play a role here, since it allows a much smaller bite angle (N–Sm–N) of 103.5(7)° and 98.1(3)° in Sm₂-biph-K₂ and Sm₂-biph-[K(18-crown-6)]₂, respectively, which are very close to the bite angle (Cl–Sm–Cl) of 101.9° in optimized (Sm₂Cl₄)⁴⁺ structure, but much smaller than that of 133.6(1)° in Sm-I. Combined, the strong ligand field of N-donors and the flexibility in coordination geometry of the NN^TBS ligand may explain why Sm₂-biph-K₂ has a unique electronic structure among low valent samarium arene complexes.

Unlike Sm₂-biph-K₂, Yb₂-biph-K₂ contains two Yb(II) ions and a biphenyl dianion with the negative charges equally delocalized over the two phenyl rings. The Yb(II) ions and the biphenyl dianion were only weakly bound through a primarily ionic interaction, which could be readily disrupted by the addition of 18-crown-6. The divalent 4f¹⁴ electron configuration of ytterbium in Yb₂-biph-K₂ was supported by structural parameters, reactivity studies, absorption spectra, and diamagnetic nature of the compound. It is worth noting that ytterbium was reported to have a multiconfigurational character when complexed with redox active ligands, such as 2,2′-bipyiridine.^87–90 However, we did not observe any evidence for such a multiconfigurational character in Yb₂-biph-K₂. This is likely due to the already highly reducing nature of the biphenyl dianion.

The distinct electronic structures of samarium and ytterbium biphenyl complexes can be rationalized by the difference in their M^3+/2+ reduction potentials,²⁵ as well as their fⁿ to fⁿ⁻¹d¹ promotion energies.⁸² The preference to stay in the completely filled 4f¹⁴ electronic configuration led to a switch in the electronic structure of Yb₂-biph-K₂ compared to other M₂-biph-K₂ with more negative M^3+/2+ reduction potentials (Chart 1).

Conclusions

We successfully synthesized the inverse-sandwich samarium and ytterbium biphenyl complexes Sm₂-biph-K₂ and Yb₂-biph-K₂ and characterized them by X-ray crystallography, elemental analysis, NMR and UV/Vis/NIR spectroscopy, and room temperature magnetic susceptibility measured by the Evans method. The structural and spectroscopic data are consistent with the assignment of 4f⁵ Sm(III) ions and a biphenyl tetraanion in Sm₂-biph-K₂ and 4f¹⁴ Yb(II) ions and a biphenyl dianion in Yb₂-biph-K₂. The reaction of Sm₂-biph-K₂ or Yb₂-biph-K₂ with 18-crown-6 proceeded with distinct results: while the Sm–arene interaction remained intact, the Yb–arene interaction was readily disrupted by the coordination of 18-crown-6. DFT calculations were carried out on model complexes for Sm₂-biph-K₂ and Yb₂-biph-K₂ to elucidate the electronic structures and bonding interactions. Sm₂-biph-K₂ was confirmed to bear two Sm(III) ions and a charge-localized biphenyl tetraanion. The bonding interaction between the Sm(III) ions and the biphenyl tetraanion involves the δ-type 5d orbitals of Sm and the π* orbitals of the bound phenyl ring and features two δ bonds with a covalent character accounting for a 39% attractive interaction by computational analysis. On the contrary, Yb₂-biph-K₂ was found to contain two Yb(II) ions and a charge-delocalized biphenyl dianion, which are only weakly bound through a primarily ionic interaction. Sm₂-biph-K₂ and Yb₂-biph-K₂ extended the series of M₂-biph-K₂. The electronic structure of rare earth metals in this series was compared with the previously reported (Cp′₃M)⁻ and [((^Ad,MeArO)₃mes)M]⁻ series as well as the solid state rare earth metal diiodides. It was found that the switch point for rare earth ions to adopt a divalent 4fⁿ electronic configuration changed from Nd in MI₂ and [((^Ad,MeArO)₃mes)M]⁻ to Tm in (Cp′₃M)⁻ and further to Yb in M₂-biph-K₂, showing that the appropriate choice of ligands could compensate the positive shifts in M^3+/2+ reduction potentials and the increase in the promotion energy for the fⁿ⁺¹ to fⁿd¹ transition. The stability of Sm₂-biph-K₂ and Sm₂-biph-[K(18-crown-6)]₂ is remarkable since they are the only structurally characterized and stable Sm(III) complexes with a highly reducing anionic arene ligand. In comparison with other samarium arene complexes, the unique electronic structure of Sm₂-biph-K₂ is attributed to the features of the ferrocene diamide NN^TBS ligand. Overall, this study extends the series of the inverse-sandwich complexes M₂-biph-K₂ to traditional divalent rare earth metals Sm and Yb, and sheds light on the relationship between the electronic structures of rare earth ions and their coordination environment, as well as the bonding interaction between rare earth ions and arene ligands in low valent metal chemistry. Future studies will focus on detailed XANES spectroscopic and magnetometry studies to gain further insight into the electronic structure of these highly reducing systems.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

YX, X-KZ, H-SH, and WH thank Peking University and Beijing National Laboratory for Molecular Sciences for funding. JTM was supported by the National Science Foundation under Cooperative Agreement no. EEC1647722. YX thanks Dr Jie Su for help with X-ray crystallography and Dr Hui Fu and Dr Xiu Zhang for help with NMR spectroscopy. PLD acknowledges support from NSF Grant CHE-1809116. The use of the Advance Photon Source was supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under contract no. DE-AC02-06CH11357. MRCAT operations were supported by the Department of Energy and MRCAT member institutions. The Tsinghua Xuetang Talents Program is acknowledged for providing computational resources. We are grateful to Ms Rong Sun and Prof. Bing-wu Wang at Peking University for helpful discussion about magnetic data.

References

1. N. Kaltsoyannis and P. Scott, The f Elements, Oxford University Press, New York, 1999.
2. S. Cotton, Lanthanide and Actinide Chemistry, Wiley, Hoboken, NJ, 2006.
3. D. A. Atwood, The Rare Earth Elements: Fundamentals and Applications, Wiley, Hoboken, NJ, 2012.
4. A. J. Freeman and R. E. Watson, Phys. Rev., 1962, 127, 2058–2075.
5. G. H. Dieke, Spectra and Energy Levels of Rare Earth Ions in Crystals, Wiley, New York, 1968.
6. K. A. Gschneidner, J. Less-Common Met., 1971, 25, 405–422.
7. H. J. Emeléus and A. G. Sharpe, Modern Aspects of Inorganic Chemistry, Wiley, New York, 4th edn, 1973.
8. J. C. G. Bünzli and G. R. Choppin, Lanthanide probes in life, chemical, and earth sciences: theory and practice, Elsevier, Amsterdam, 1989.
9. K. A. Gschneidner, in Handbook on the Physics and Chemistry of Rare Earths, ed. J.-C. G. Bünzli and V. K. Pecharsky, Elsevier, 2016, vol. 50, pp. 1–18.
10. M. L. Neidig, D. L. Clark and R. L. Martin, Coord. Chem. Rev., 2013, 257, 394–406.
11. G. R. Choppin, J. Alloys Compd., 2002, 344, 55–59.
12. N. Kaltsoyannis, Inorg. Chem., 2013, 52, 3407–3413.
13. S. M. Butorin, K. O. Kvashnina, J. R. Vegelius, D. Meyer and D. K. Shuh, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 8093–8097.
14. J. Su, E. R. Batista, K. S. Boland, S. E. Bone, J. A. Bradley, S. K. Cary, D. L. Clark, S. D. Conradson, A. S. Ditter, N. Kaltsoyannis, J. M. Keith, A. Kerridge, S. A. Kozimor, M. W. Löble, R. L. Martin, S. G. Minasian, V. Mocko, H. S. La Pierre, G. T. Seidler, D. K. Shuh, M. P. Wilkerson, L. E. Wolfsberg and P. Yang, J. Am. Chem. Soc., 2018, 140, 17977–17984.
15. C. T. Palumbo, I. Zivkovic, R. Scopelliti and M. Mazzanti, J. Am. Chem. Soc., 2019, 141, 9827–9831.
16. N. T. Rice, I. A. Popov, D. R. Russo, J. Bacsa, E. R. Batista, P. Yang, J. Telser and H. S. La Pierre, J. Am. Chem. Soc., 2019, 141, 13222–13233.
17. A. R. Willauer, C. T. Palumbo, R. Scopelliti, I. Zivkovic, I. Douair, L. Maron and M. Mazzanti, Angew. Chem., Int. Ed., 2020, 59, 3549–3553.
18. A. R. Willauer, C. T. Palumbo, F. Fadaei-Tirani, I. Zivkovic, I. Douair, L. Maron and M. Mazzanti, J. Am. Chem. Soc., 2020, 142, 5538–5542.
19. Q. Zhang, S.-X. Hu, H. Qu, J. Su, G. Wang, J.-B. Lu, M. Chen, M. Zhou and J. Li, Angew. Chem., Int. Ed., 2016, 55, 6896–6900.
20. S.-X. Hu, J. Jian, J. Su, X. Wu, J. Li and M. Zhou, Chem. Sci., 2017, 8, 4035–4043.
21. W. J. Evans, J. W. Grate, H. W. Choi, I. Bloom, W. E. Hunter and J. L. Atwood, J. Am. Chem. Soc., 1985, 107, 941–946.
22. M. N. Bochkarev, I. L. Fedushkin, A. A. Fagin, T. V. Petrovskaya, J. W. Ziller, R. N. R. Broomhall-Dillard and W. J. Evans, Angew. Chem., Int. Ed. Engl., 1997, 36, 133–135.
23. W. J. Evans, N. T. Allen and J. W. Ziller, J. Am. Chem. Soc., 2000, 122, 11749–11750.
24. M. N. Bochkarev, I. L. Fedushkin, S. Dechert, A. A. Fagin and H. Schumann, Angew. Chem., Int. Ed., 2001, 40, 3176–3178.
25. L. R. Morss, Chem. Rev., 1976, 76, 827–841.
26. L. J. Nugent, R. D. Baybarz, J. L. Burnett and J. L. Ryan, J. Phys. Chem., 1973, 77, 1528–1539.
27. N. B. Mikheev and A. N. Kamenskaya, Coord. Chem. Rev., 1991, 109, 1–59.
28. W. J. Evans, Organometallics, 2016, 35, 3088–3100.
29. G. Meyer, Chem. Rev., 1988, 88, 93–107.
30. G. Meyer, Z. Anorg. Allg. Chem., 2007, 633, 2537–2552.
31. G. Meyer, J. Solid State Chem., 2019, 270, 324–334.
32. P. B. Hitchcock, M. F. Lappert, L. Maron and A. V. Protchenko, Angew. Chem., Int. Ed., 2008, 47, 1488–1491.
33. M. R. MacDonald, J. W. Ziller and W. J. Evans, J. Am. Chem. Soc., 2011, 133, 15914–15917.
34. M. R. MacDonald, J. E. Bates, M. E. Fieser, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2012, 134, 8420–8423.
35. M. R. MacDonald, J. E. Bates, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2013, 135, 9857–9868.
36. D. H. Woen, G. P. Chen, J. W. Ziller, T. J. Boyle, F. Furche and W. J. Evans, Angew. Chem., Int. Ed., 2017, 56, 2050–2053.
37. A. J. Ryan, L. E. Darago, S. G. Balasubramani, G. P. Chen, J. W. Ziller, F. Furche, J. R. Long and W. J. Evans, Chem.–Eur. J., 2018, 24, 7702–7709.
38. C. A. Gould, K. R. McClain, J. M. Yu, T. J. Groshens, F. Furche, B. G. Harvey and J. R. Long, J. Am. Chem. Soc., 2019, 141, 12967–12973.
39. M. E. Fieser, M. R. MacDonald, B. T. Krull, J. E. Bates, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2015, 137, 369–382.
40. M. E. Fieser, M. G. Ferrier, J. Su, E. Batista, S. K. Cary, J. W. Engle, W. J. Evans, J. S. Lezama Pacheco, S. A. Kozimor, A. C. Olson, A. J. Ryan, B. W. Stein, G. L. Wagner, D. H. Woen, T. Vitova and P. Yang, Chem. Sci., 2017, 8, 6076–6091.
41. V. E. Fleischauer, G. Ganguly, D. H. Woen, N. J. Wolford, W. J. Evans, J. Autschbach and M. L. Neidig, Organometallics, 2019, 38, 3124–3131.
42. M. E. Fieser, C. T. Palumbo, H. S. La Pierre, D. P. Halter, V. K. Voora, J. W. Ziller, F. Furche, K. Meyer and W. J. Evans, Chem. Sci., 2017, 8, 7424–7433.
43. C. T. Palumbo, D. P. Halter, V. K. Voora, G. P. Chen, A. K. Chan, M. E. Fieser, J. W. Ziller, W. Hieringer, F. Furche, K. Meyer and W. J. Evans, Inorg. Chem., 2018, 57, 2823–2833.
44. K. E. Johnson and J. N. Sandoe, J. Chem. Soc., Abstr., 1969, 1694–1697.
45. S. A. Kulyukhin, Radiochemistry, 2018, 60, 451–469.
46. W. Huang, F. Dulong, T. Wu, S. I. Khan, J. T. Miller, T. Cantat and P. L. Diaconescu, Nat. Commun., 2013, 4, 1448.
47. W. Huang, J. J. Le Roy, S. I. Khan, L. Ungur, M. Murugesu and P. L. Diaconescu, Inorg. Chem., 2015, 54, 2374–2382.
48. W. Huang, S. I. Khan and P. L. Diaconescu, J. Am. Chem. Soc., 2011, 133, 10410–10413.
49. W. Huang and P. L. Diaconescu, Chem. Commun., 2012, 48, 2216–2218.
50. W. Huang and P. L. Diaconescu, Eur. J. Inorg. Chem., 2013, 2013, 4090–4096.
51. W. Huang, P. M. Abukhalil, S. I. Khan and P. L. Diaconescu, Chem. Commun., 2014, 50, 5221–5223.
52. W. Huang, J. L. Brosmer and P. L. Diaconescu, New J. Chem., 2015, 39, 7696–7702.
53. R. D. Shannon, Acta Crystallogr., 1976, 32, 751–767.
54. F. A. Cotton and W. Schwotzer, J. Am. Chem. Soc., 1986, 108, 4657–4658.
55. S. Ilango, B. Vidjayacoumar and S. Gambarotta, Dalton Trans., 2010, 39, 6853–6857.
56. N. G. Connelly and W. E. Geiger, Chem. Rev., 1996, 96, 877–910.
57. M. D. Fryzuk, J. B. Love and S. J. Rettig, J. Am. Chem. Soc., 1997, 119, 9071–9072.
58. W. Huber, A. M. Und and K. Müllen, Chem. Ber., 1981, 114, 1318–1336.
59. M. Castillo, A. J. Metta-Magaña and S. Fortier, New J. Chem., 2016, 40, 1923–1926.
60. J.-C. G. Bünzli and C. Piguet, Chem. Soc. Rev., 2005, 34, 1048–1077.
61. K. R. Meihaus, M. E. Fieser, J. F. Corbey, W. J. Evans and J. R. Long, J. Am. Chem. Soc., 2015, 137, 9855–9860.
62. W. J. Evans, J. W. Grate, I. Bloom, W. E. Hunter and J. L. Atwood, J. Am. Chem. Soc., 1985, 107, 405–409.
63. W. J. Evans, J. W. Grate and R. J. Doedens, J. Am. Chem. Soc., 1985, 107, 1671–1679.
64. W. J. Evans, D. K. Drummond, S. G. Bott and J. L. Atwood, Organometallics, 1986, 5, 2389–2391.
65. W. J. Evans and T. A. Ulibarri, J. Am. Chem. Soc., 1987, 109, 4292–4297.
66. J. Jubb and S. Gambarotta, J. Am. Chem. Soc., 1994, 116, 4477–4478.
67. A. N. Selikhov, A. V. Cherkasov, G. K. Fukin, A. A. Trifonov, I. del Rosal and L. Maron, Organometallics, 2015, 34, 555–562.
68. D. Y. Zubarev and A. I. Boldyrev, Phys. Chem. Chem. Phys., 2008, 10, 5207–5217.
69. A. Michalak, M. Mitoraj and T. Ziegler, J. Phys. Chem. A, 2008, 112, 1933–1939.
70. M. P. Mitoraj, A. Michalak and T. Ziegler, J. Chem. Theory Comput., 2009, 5, 962–975.
71. S. Labouille, C. Clavaguéra and F. Nief, Organometallics, 2013, 32, 1265–1271.
72. J.-X. Zhang, F. K. Sheong and Z. Lin, Chem.–Eur. J., 2018, 24, 9639–9650.
73. W.-L. Li, T.-T. Chen, D.-H. Xing, X. Chen, J. Li and L.-S. Wang, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, E6972–E6977.
74. T.-T. Chen, W.-L. Li, J. Li and L.-S. Wang, Chem. Sci., 2019, 10, 2534–2542.
75. M. P. Kelley, I. A. Popov, J. Jung, E. R. Batista and P. Yang, Nat. Commun., 2020, 11, 1558.
76. M. N. Bochkarev, Chem. Rev., 2002, 102, 2089–2118.
77. M. N. Bochkarev, A. A. Trifonov, E. A. Fedorova, N. S. Emelyanova, T. A. Basalgina, G. S. Kalinina and G. A. Razuvaev, J. Organomet. Chem., 1989, 372, 217–224.
78. I. L. Fedushkin, V. K. Nevodchikov, V. K. Cherkasov, M. N. Bochkarev, H. Schumann, F. Girgsdies, F. H. Görlitz, G. Kociok-Köhn and J. Pickardt, J. Organomet. Chem., 1996, 511, 157–162.
79. A. V. Protchenko, L. N. Zakharov, G. K. Fukin, Y. T. Struchkov and M. N. Bochkarev, Russ. Chem. Bull., 1996, 45, 950.
80. W. J. Evans, S. L. Gonzales and J. W. Ziller, J. Am. Chem. Soc., 1994, 116, 2600–2608.
81. R. P. Kelly, D. Toniolo, F. F. Tirani, L. Maron and M. Mazzanti, Chem. Commun., 2018, 54, 10268–10271.
82. D. M. Anderson, F. G. N. Cloke, P. A. Cox, N. Edelstein, J. C. Green, T. Pang, A. A. Sameh and G. Shalimoff, J. Chem. Soc., Chem. Commun., 1989, 53–55.
83. F. G. N. Cloke, Chem. Soc. Rev., 1993, 22, 17–24.
84. W. A. King, T. J. Marks, D. M. Anderson, D. J. Duncalf and F. G. N. Cloke, J. Am. Chem. Soc., 1992, 114, 9221–9223.
85. W. A. King, S. Di Bella, G. Lanza, K. Khan, D. J. Duncalf, F. G. N. Cloke, I. L. Fragala and T. J. Marks, J. Am. Chem. Soc., 1996, 118, 627–635.
86. R. P. Kelly, L. Maron, R. Scopelliti and M. Mazzanti, Angew. Chem., Int. Ed., 2017, 56, 15663–15666.
87. M. Schultz, J. M. Boncella, D. J. Berg, T. D. Tilley and R. A. Andersen, Organometallics, 2002, 21, 460–472.
88. R. E. Da Re, C. J. Kuehl, M. G. Brown, R. C. Rocha, E. D. Bauer, K. D. John, D. E. Morris, A. P. Shreve and J. L. Sarrao, Inorg. Chem., 2003, 42, 5551–5559.
89. C. H. Booth, M. D. Walter, D. Kazhdan, Y.-J. Hu, W. W. Lukens, E. D. Bauer, L. Maron, O. Eisenstein and R. A. Andersen, J. Am. Chem. Soc., 2009, 131, 6480–6491.
90. C. H. Booth, D. Kazhdan, E. L. Werkema, M. D. Walter, W. W. Lukens, E. D. Bauer, Y.-J. Hu, L. Maron, O. Eisenstein, M. Head-Gordon and R. A. Andersen, J. Am. Chem. Soc., 2010, 132, 17537–17549.

---

Footnotes

† Electronic supplementary information (ESI) available: Synthetic procedures, NMR spectra, X-ray crystallography, XANES and UV/Vis/NIR spectroscopic data, and DFT calculation details. CCDC 2012143–2012149. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d0sc03555f

‡ Y. X. and X.-K. Z. contributed equally to this work.

§ Present address: X-ray Science Division, Argonne National Laboratory, Argonne, Illinois 60439, USA.

¶ Present address: Davidson School of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907, USA.

---