Siloxide tripodal ligands as a scaffold for stabilizing lanthanides in the +4 oxidation state

Synthetic strategies to isolate molecular complexes of lanthanides, other than cerium, in the +4 oxidation state remain elusive, with only four complexes of Tb(iv) isolated so far. Herein, we present a new approach for the stabilization of Tb(iv) using a siloxide tripodal trianionic ligand, which allows the control of unwanted ligand rearrangements, while tuning the Ln(iii)/Ln(iv) redox-couple. The Ln(iii) complexes, [LnIII((OSiPh2Ar)3-arene)(THF)3] (1-LnPh) and [K(toluene){LnIII((OSiPh2Ar)3-arene)(OSiPh3)}] (2-LnPh) (Ln = Ce, Tb, Pr), of the (HOSiPh2Ar)3-arene ligand were prepared. The redox properties of these complexes were compared to those of the Ln(iii) analogue complexes, [LnIII((OSi(OtBu)2Ar)3-arene)(THF)] (1-LnOtBu) and [K(THF)6][LnIII((OSi(OtBu)2Ar)3-arene)(OSiPh3)] (2-LnOtBu) (Ln = Ce, Tb), of the less electron-donating siloxide trianionic ligand, (HOSi(OtBu)2Ar)3-arene. The cyclic voltammetry studies showed a cathodic shift in the oxidation potential for the cerium and terbium complexes of the more electron-donating phenyl substituted scaffold (1-LnPh) compared to those of the tert-butoxy (1-LnOtBu) ligand. Furthermore, the addition of the –OSiPh3 ligand further shifts the potential cathodically, making the Ln(iv) ion even more accessible. Notably, the Ce(iv) complexes, [CeIV((OSi(OtBu)2Ar)3-arene)(OSiPh3)] (3-CeOtBu) and [CeIV((OSiPh2Ar)3-arene)(OSiPh3)(THF)2] (3-CePh), were prepared by chemical oxidation of the Ce(iii) analogues. Chemical oxidation of the Tb(iii) and Pr(iii) complexes (2-LnPh) was also possible, in which the Tb(iv) complex, [TbIV((OSiPh2Ar)3-arene)(OSiPh3)(MeCN)2] (3-TbPh), was isolated and crystallographically characterized, yielding the first example of a Tb(iv) supported by a polydentate ligand. The versatility and robustness of these siloxide arene-anchored platforms will allow further development in the isolation of more oxidizing Ln(iv) ions, widening the breadth of high-valent Ln chemistry.

These species were thoroughly characterized in solution for the tris(amidyl)imidophosphorane system, 38 but could not be isolated in the solid state.Ln(III) complexes of the triphenylsiloxide (-OSiPh 3 ) ligand had been reported 30 years ago, 39,40 but only in 2020, our group demonstrated that such monodentate siloxide ligands can stabilize both Tb and Pr ions in the +4 oxidation state. 37,41The isostructural complexes, [Ln IV (OSiPh 3 ) 4 (MeCN) 2 ] 37,41 were isolated by the productive chemical oxidation of the analogous Ln(III) complexes and showed reasonable solution stability, which could be improved by the addition of the neutral triphenylphosphine oxide ligand (Ph 3 PO) 42 or more recently by the addition of the neutral bidentate ligands, 2,2 0 -bipyridine (bpy), 2,2 0 -bipyrimidine (bpym), and 1,10-phenanthroline (phen). 43The higher stability of the -OSiPh 3 complexes compared to the -OSi(O t Bu) 3 complexes was reected in their lower oxidation potentials, which was attributed to the stronger electron-donating ability of the -OSiPh 3 ligands.However, preliminary attempts to prepare Ln(IV) complexes with more electron-donating, monoanionic supporting ligands have so far been unsuccessful.Surprisingly, since 2020, no other anionic supporting ligands capable of stabilizing Tb(IV) or Pr(IV) have been identi-ed, demonstrating the difficulty in identifying ligands and conditions that are capable of stabilizing lanthanides, other than Ce, in the +4 oxidation state.
5][46][47][48] Moreover, Schelter and co-workers showed that tripodal ligands are effective for gaining better control of the Ln(III)/Ln(IV) redoxcouple by providing a single coordination site. 45Schelter and coworkers also demonstrated that tripodal ligands allow the implementation of simple solubility-based separation of early and late f-elements. 49,50n contrast, despite their potential relevance for the implementation of redox based separation of f-elements, synthetic routes for incorporating Ln(IV), other than cerium, in tripodalbased ligands remain unidentied.
Recently, we reported a new tripodal ligand, (HOSi(O t Bu) 2 -Ar) 3 -arene, and its corresponding Ce complexes, demonstrating that four different redox states were accessible with this ligand scaffold. 48rein, we show that by combining a more electrondonating tripodal ligand, (HOSiPh 2 Ar) 3 -arene, 1,2 with the monodentate -OSiPh 3 ligand, the redox potential can be tuned to access Ln(IV), enabling the isolation of a new Tb(IV) complex containing a trianionic polydentate ligand.
Instead, we postulated that direct Ce(III) to Ce(IV) oxidation may become possible by coordination of an additional -OSiPh 3 ligand to 1-Ce OtBu , and/or by incorporation of phenyl substituents on the tripodal backbone, which would shi the potential cathodically. 3,4This approach could then be utilized to access Tb(IV) and Pr(IV) species.
Therefore, we next investigated the synthesis of the phenylsubstituted Ln(III) (Ln = Ce, Tb, Pr) tripodal complexes.
Comparison of the oxidation potentials for complexes 1-Ce OtBu and 1-Ce Ph (vide infra) displays a signicant cathodic shi (DE pa = 0.87 V), suggesting that 1-Ce Ph is signicantly more electron-rich than 1-Ce OtBu .For complexes 1-Tb Ph and 1-Pr Ph , no oxidation events were observed in the window permitted by THF and the [NBu 4 ][B(C 6 F 5 ) 4 ] electrolyte.Based on these electrochemical results, we next postulated that coordination of an additional monoanionic siloxide ligand could further shi the oxidation potential cathodically, allowing access to the desired high-valent Ln(IV) species.
As anticipated, the addition of the -OSiPh 3 ligand led to a signicant shi in the oxidation potential (DE pa = 1.57V) of the Ce(III) complexes 1-Ce OtBu and 2-Ce OtBu suggesting that a further shi could be expected by the addition of KOSiPh 3 to the 1-Ln Ph (Ln = Ce, Tb, Pr) complexes.
This was conrmed by DFT calculations (B3PW91 functional) including dispersion corrections (see the ESI † for computational details).The optimized geometry of 1-Ce Ph compares well within the experimental data (Table S4 †) with an average Ce-O siloxide bond distance of 2.2187 (7).The SOMO is a pure 4f orbital in line with a Ce(III) complex.The Ce-C centroid is 4.01 Å and is in good agreement with the experimental data.The Ce-C centroid in 1-Ce Ph is longer than that in 1-Ce OtBu (2.72 Å) and is consistent with a lack of a Ce-arene d bonding interaction.Indeed, computational studies show that the LUMO is a fully delocalized p* without any Ce contribution.
Complexes 1-Tb OtBu , 1-Tb Ph , and 2-Tb OtBu (Fig. S48, S49 † and 1) were found to be isostructural to the cerium analogs, 1-Ce OtBu , 1-Ce Ph , and 2-Ce OtBu , respectively.The average Tb-O siloxide distances in 1-Tb OtBu  Calculations were carried out on 2-Tb Ph , as a precursor of 3-Tb Ph , using the same computational method.The optimized geometry is in agreement with the experimental data (Table S10 †), and the Tb-O siloxide distances are reproduced with a maximum deviation of 0.05 Å.The Tb-O siloxide WBI (Table S12 †) are in the 0.40 range indicative of primarily ionic interactions, as expected for lanthanide complexes.The unpaired spin density is fully located at the Tb center in line with a Tb(III) complex.

Synthesis of the lanthanide(IV) complexes
With the tetrakis Ln(III) complexes in hand, we next investigated accessing high-valent Ln(IV) species by use of chemical oxidants.
First, the oxidation of complexes 2-Ce OtBu and 2-Ce Ph was investigated with the oxidant AgBPh 4 (0.41 V vs. Fc 0 /Fc + in THF). 56ddition of 1.Since a quasi-reversible redox event was observed for 2-Tb Ph , with an oxidation potential of E pa = 0.53 V vs. Fc 0 /Fc + (vide infra), the oxidation of complex 2-Tb Ph was investigated with the commercially available oxidant [N(C 6 H 4 Br) 3 ][SbCl 6 ], commonly known as "magic blue" (0.70 V vs. Fc 0 /Fc + in MeCN). 56he addition of 1. Dark-orange X-ray quality crystals of 3-Tb Ph were obtained from a dilute reaction mixture (7 mM) aer standing overnight in MeCN at −40 °C.Bulk isolation of 3-Tb Ph can also be carried out in the absence of 2.2.2-cryptand, but in a slightly lower yield (43% yield).The 1 H NMR spectrum of the isolated complex 3-Tb Ph in toluene is silent, consistent with that of a Tb(IV) 4f 7 ion (Fig. S27 †).
Complex 3-Tb Ph is stable in solution for 2 days at room temperature in toluene, but decomposes rapidly in THF at room temperature, as shown by 1 H NMR studies, leading to the formation of 1-Tb Ph and unknown species immediately aer dissolution (Fig. S28 †).The decomposition of 3-Tb Ph into 1-Tb Ph in THF, as measured by 1 H NMR spectroscopy (using CH 2 Cl 2 as the internal standard), leads to the formation of 1-Tb Ph with 83% yield (Fig. S29 †).
Variable-temperature magnetic and EPR data were measured on the isolated complex 3-Tb Ph in order to conrm the presence of the Tb(IV) ion (Fig. 3).The measured c M T value for the 4f 7 complex 3-Tb Ph at 300 K (8.06 emu K mol −1 ) is consistent with the values found in the previously reported Tb(IV) complexes [Tb IV (OSi(O t Bu) 3 ) 4 ] (7.77 emu K mol −1 ), 35 [Tb IV (OSiPh 3 ) 4 (-MeCN) 2 ] (7.82 emu K mol −1 ), 41 and [Tb IV (NP(1,2-bis-t Budiamidoethane) 4 (NEt 2 ))] (8.55 emu K mol −1 ), 36 and are in agreement with the value of c M T predicted for a 4f 7 complex using the LS coupling for a 4f 7 ion (L = 0, S = 7/2). 19The X-band EPR spectrum of 3-Tb Ph measured at 6 K in the solid-state or toluene, displays strong features with g-values of [7.90, 5.00, 3.35] and [7.45, 4.55, 3.75], respectively, consistent with previously reported Tb(IV) complexes. 35,36,41he UV/vis spectrum of 3-Tb Ph measured immediately aer dissolution in toluene (1 mM) shows two absorption bands with l max at 285 and 355 nm, in which the absorption at 355 nm is most consistent with that of the previously reported siloxidesupported Tb(IV) complexes [Tb IV (OSi(O t Bu) 3 ) 4 ] (l max = 371 nm, toluene) 35 and [Tb IV (OSiPh 3 ) 4 (MeCN) 2 ] (l max = 386 nm, THF). 41onitoring the UV/vis spectra of 3-Tb Ph in toluene (Fig. S61 †) over time showed a high solution stability, where 82% of the complex remained aer 48 hours.This high stability is consistent with the 1 H NMR studies in toluene, and compares well to the stability of the previously reported Tb(IV) complex, [Tb IV (OSiPh 3 ) 4 (Ph 3 PO)(MeCN)], where 80% of the complex is present aer 96 hours at room temperature. 42Complex 3-Tb Ph is the rst example of an isolated molecular complex of Tb(IV) supported by a trianionic tripodal ligand.
Although the event observed in the cyclic voltammogram for 2-Pr Ph was found to be irreversible at slow scan rates (50-400 mV s −1 ) (vide infra), chemical oxidation with "magic blue" was investigated.The addition of 1.1 equiv. of [N(C 6 H 4 Br) 3 ][SbCl 6 ] to complex 2-Pr Ph with or without 2.2.2-cryptand in MeCN at −40 °C led to a dark brown-orange solution, which immediately faded to pale-yellow with the precipitation of a white solid.Colour-less crystals were isolated from the reaction mixture and characterized by X-ray diffraction studies as the MeCN adduct of complex 1-Pr Ph (Fig. S53 † We reasoned that the facile oxidation of 2-Tb Ph , leading to an insoluble Tb(IV) complex (3-Tb Ph ), could provide a pathway for the separation of the Tb ion from a neighboring lanthanide (for example Dy), which presents a more positive oxidation potential (5.0 V calculated).Preliminary experiments were conducted on a MeCN reaction mixture containing 2-Tb Ph and its Dy(III) analogue, 2-Dy Ph , prepared in situ, in a 1 : 1 ratio.
Addition of [N(C 6 H 4 Br) 3 ][SbCl 6 ] to a 1 : 1 reaction mixture of 2-Tb Ph : 2-Dy Ph in the presence of 2.2.2-cryptand in MeCN at −40 °C led to the immediate precipitation of an orange solid.Extraction of the orange solid into toluene-d 8 resulted in a dark orange solution, which displayed featureless 1 H NMR spectra, suggesting that the Tb(IV) complex, 3-Tb Ph was the major species in solution.The 1 H NMR spectrum of the residue obtained aer removal of toluene was measured in THF-d 8 aer 24 hours, allowing the Tb(IV) to decompose, showing the presence of 1-Tb Ph as the major species and of a small amount of 1-Dy Ph (6%).In contrast, the 1 HNMR spectrum of the THF-d 8 solution obtained upon dissolution of the residual solid fraction, aer toluene extraction, shows a Tb : Dy ratio of 0.44 : 1.
The molar ratios of the two metals in the two fractions were determined both by 1 H NMR and ICP-MS analyses.From these data, the separation factor, S Tb/Dy , was calculated from the enrichment factor (S = D residual solid $D extracted solid ), D, which were determined from the molar ratios.
Overall, a separation factor of 10.1 was determined by 1 H NMR spectroscopy and a similar value was obtained by ICP-MS measurements (8.56).These values compare well with the separation factor reported for neighboring lanthanides ranging from 1 to 4 using size sensitive supramolecular encapsulation and precipitation techniques. 50,57These preliminary separation trials suggest that the Tb(IV)/Tb(III) redox couple may be used to implement effective separation of Tb from other lanthanides.
The molecular structure of 3-Ce Ph reveals the presence of a mononuclear Ce(IV) complex with a metal center bound by three -OSiPh 2 groups of the tripodal ligand, one -OSiPh 3 , and two THF molecules, adopting a distorted octahedral geometry.
The range of Ce-O siloxide distances (2.121(4)-2.170(4)Å) is consistent with the Ce-O siloxide length found in the previously reported Ce(IV) complexes [Ce IV (OSiPh 3 ) 4 (THF) 2 ] (2.109(3)-2.154(3)Å) 59 and [Ce IV (OSiPh 3 ) 4 (DME)] (2.10(1)-2.13(1)Å). 60 The Ce-O siloxide and Ce-O THF bond distances in 3-Ce Ph are shorter compared to those found in the Ce(III) complex 1-Ce Ph (0.078 Å and 0.100 Å, respectively), and are in agreement with the difference in ionic radii between Ce(III) and Ce(IV).Furthermore, the Ce-C centroid distance has a signicant increase in 3-Ce Ph (2.028(5)-2.087(5)Å) 41 and [Tb IV (OSi(O t Bu) 3 ) 4 ] (2.023(3)-2.093(3)Å). 35 The shorter Tb-O siloxide bond distances in 3-Tb Ph , compared to those of the Tb(III) complex 1-Tb Ph (2.139(3) Å) and 2-Tb Ph (2.16(1) Å), are in agreement with the +4 oxidation state of the terbium metal center.Furthermore, the Tb-C centroid distance has a signicant increase in 3-Tb Ph  Calculations were carried out on complex 3-Tb Ph .In order to ensure the formation of a Tb(IV) complex, the geometry optimization was carried out using f-in-core RECP for Tb which is adapted to the +4 oxidation state of Tb.This computational methodology has been shown to provide excellent geometric parameters. 61The optimized geometry of 3-Tb Ph is in excellent agreement with the experimental data (Table S15 †).The Tb-O siloxide distances are reproduced with an accuracy of 0.02 Å.Therefore, one can safely conclude that complex 3-Tb Ph implies the presence of a Tb(IV) center.As found experimentally and as expected with a stronger Lewis acid, the Tb-O siloxide distances in 3-Tb Ph are 0.15-0.18Å shorter than in 2-Tb Ph , further corroborating the presence of a Tb(IV) center.Finally, the Tb-O siloxide WBI (Table S17 †) are in the 0.45-0.50range indicating mainly ionic bonds, but a slightly greater covalency in Tb(IV) than in Tb(III), as expected with the shorter distances that allow better orbital overlap.Computational studies carried out on an analogous complex, replacing the monodentate triphenylsiloxide with a tert-butoxide ligand (Table S19 †), were also in agreement with the presence of a Tb(IV) center, suggesting that it may be possible to nd the appropriate experimental conditions for its isolation in the solid state.
The cyclic voltammogram of 1-Ce OtBu shows an oxidation feature at E pa = 1.36 V vs. Fc 0 /Fc + , which is very close to the solvent window of THF, and a related reduction feature at E pc = −0.98V vs. Fc 0 /Fc + , with a peak separation of DE pa = 2.33 V.
Interestingly, addition of the -OSiPh 3 ligand, forming 2-Ce OtBu , signicantly shis the oxidation potential cathodically (E pa = −0.21V vs. Fc 0 /Fc + ) compared to that of 1-Ce OtBu (E pa = 1.36 V vs. Fc 0 /Fc + ).The complexes of the -OSiPh 2 substituted tripodal ligand, 1-Ce Ph and 2-Ce Ph , exhibit a similar trend, in which the oxidation potential for 2-Ce Ph (E pa = −0.60V vs. Fc 0 /Fc + ) shied 1.09 V more negative compared to that of 1-Ce Ph (E pa = 0.49 V vs. Fc 0 /Fc + ).The cyclic voltammograms of the isolated Ce(IV) species, 3-Ce OtBu and 3-Ce Ph , exhibit comparable redox events (3-Ce OtBu : E pa = −0.30V and E pc = −1.22V; 3-Ce Ph : E pa = −0.53V and E pc = −1.54V) to their Ce(III) precursors, 2-Ce OtBu and 2-Ce Ph .In the case of complexes 1-Tb Ph and 1-Pr Ph , no redox processes were observed in the window permitted by THF and the [NBu 4 ][B(C 6 F 5 ) 4 ] electrolyte.This is in agreement with the oxidation potentials previously reported for Tb(III) and Pr(III) siloxide complexes, 35,37 which where ∼1.1 V higher than those observed for the analogous Ce(III) derivatives.Therefore, electrochemical oxidation of the complexes 1-Ln Ph (Ln = Tb, Pr) is inaccessible in this potential range.However, when the -OSiPh 3 ligand was added to form complexes 2-Ln Ph (Ln = Tb, Pr), the cyclic voltammograms display redox events that are quasireversible for 2-Tb Ph , but appear to be irreversible at slow scan rates for 2-Pr Ph .The Tb(III) derivatives follow the same trend as the analogous Ce(III) complexes, where the oxidation potential shis cathodically in 2-Tb Ph (E pa = 0.53 V vs. Fc 0 /Fc + ) compared to 2-Tb OtBu (1.36 V vs. Fc 0 /Fc + ).
Interestingly, the oxidation potential of complex 2-Tb Ph is 1.13 V higher than the one found for 2-Ce Ph , and is only 0.12 V higher than that of the previously reported Tb(III) complex, [KTb III (OSiPh 3 ) 4 ] (0.41 V vs. Fc 0 /Fc + ). 41A reduction wave is also observed in 2-Tb Ph (E pc = −0.08V vs., Fc 0 /Fc + ), suggesting that the Tb(IV) species is stable in solution at least within the electrochemical timeframe.Interestingly, changing the nature of the additional siloxide ligand by reacting 1.0 equiv. of KOSi(O t Bu) 3 or KOSiMe 3 with 1-Tb Ph , aer recording their cyclic voltammograms, allowed signicant shis of the oxidation potentials (0.14 V higher and 0.19 V lower, respectively) (Fig. 5a).We also investigated the effect of adding different alkoxide (NaO t Bu and 2-KOAd (AdOH = 2-adamantanol), and amide (KN(SiMe 3 ) 2 ) ligands on the redox properties of the terbium complex (see the ESI †).The addition of NaO t Bu, 2-KOAd, and  The reduction wave was found to depend on the scan rate and occurs at 0.17 V vs. Fc 0 /Fc + at 800 mV s −1 (Fig. S79).KN(SiMe 3 ) 2 to 1-Tb Ph resulted in the in situ formation of the respective ate-complexes, as indicated by 1 H NMR studies.The -O t Bu, -OAd, and -N(SiMe 3 ) 2 complexes showed E pa values of 0.11 V, 0.23 V, and 0.35 V, respectively, at a scan rate of 100 mV s −1 (Fig. S82-S84 †).
Only for the adducts of 1-Tb Ph with NaO t Bu and 2-KOAd, the associated reduction processes could be observed at higher scan rates (Fig. 5b and S82-S83 †).Preliminary attempts to isolate Tb(IV) products were unsuccessful due to their high reactivity in the solvents utilized, but these electrochemistry studies suggest that alkoxides may also provide suitable ligands for the isolation of Tb(IV) species.Although the Tb(IV) complex 3-Tb Ph was found to be unstable in THF by 1 H NMR studies, it was possible to record its cyclic voltammogram immediately aer dissolution at room temperature.Scanning reductively, an event was observed (E pc = −0.33V vs. Fc 0 /Fc + ), which is ∼0.3 V lower than that observed for the parent 2-Tb Ph complex (E pc = −0.08V vs., Fc 0 /Fc + ).This suggests an enhanced kinetic stability of the Tb(IV) complex in the absence of K + cations, similar to what was previously observed in the homoleptic Tb(IV) complex, [Tb IV (OSiPh 3 ) 4 (MeCN) 2 ]. 41 Compared to the complexes 2-Ln Ph (Ln = Ce, Tb), the Pr(III) derivative, 2-Pr Ph , displays an oxidative event (E pa = 0.60 V vs. Fc 0 / Fc + ) with no corresponding reduction wave at a slow scan rate (50-400 mV s −1 ).However, a reduction feature could be observed at faster scan rates (600-1200 mV s −1 ) (Fig. S79 †).Additionally, for 1-Pr Ph , additional ligands, KOSiMe 3 and 2-KOAd, were examined.They showed lower oxidation potentials (E pa = 0.27 and 0.48 V, 100 mV s −1 ) (Fig. S85-S86 †) compared to 2-Pr Ph .Addition of KOSiMe 3 also allowed the observation of a reduction process at different scan rates (50-1000 mV s −1 ) (Fig. S84 †).However, in the case of 2-KOAd, no reduction processes could be observed in the cyclic voltammograms at different scan rates (50-1000 mV s −1 ) (Fig. S86 †).Attempts to isolate a molecular Pr(IV) species using different solvents and varying temperatures were not successful.The CV results suggest that transient, highly reactive Pr(IV) species are generated from the electrochemical oxidation of 2-Pr Ph , and of the -OSiMe 3 adduct, but they are unstable under the solvent/electrolyte conditions employed, as observed for other Pr(III) systems. 38verall, the cyclic voltammograms of the Ce(III) and Tb(III) complexes show that replacing the tris(tert-butoxy) moieties for phenyl substituents shis the oxidation potentials to more negative values.This may be attributed to the combination of two inuences; (1) the presence of the more electron-donating -OSiPh 2 groups, compared to the tert-butoxy moieties on the tripodal ligand, makes the oxidation to Ln(IV) (Ln = Ce, Tb) more accessible; and (2) the presence of a metal-arene d bonding interaction in 1-Ln OtBu , which is absent in 1-Ln Ph as corroborated by structural and DFT studies, is likely to stabilize the Ln(III) metal center, rendering the oxidation in 1-Ln OtBu more difficult.Furthermore, the coordination of an additional ligand, -OSiPh 3 , -OSi(O t Bu) 3 , or -OSiMe 3 , further shis the potential cathodically, making the Ln(IV) even more accessible.This may be attributed to the increased electron donation from the additional siloxide ligand and to the formation of neutral Ln(IV) ions, which should be more favorable than the formation of a positively charged Ln(IV) ion in complexes 1-Ln OtBu/Ph (Ln = Ce, Tb).

Fig. 3
Fig. 3 (a) Plot of c M T versus temperature for isolated 3-Tb Ph under an applied field of 1 T and (b) the X-band (9.4 GHz) EPR spectrum of complex 3-Tb Ph in toluene (20 mM) at 6 K (bottom).

Table 1
Selected structural parameters of 2-Ln and 3-Ln