Diverging Reactivity in Anilidophosphine Supported Group III Complexes

Abstract

We report the synthesis of scandium and yttrium halide complexes with bidentate, monoanionic anilidophosphine (PN) ligands and the general formula (PN)₂MX (M = Sc (1‐X), Y (2‐X); X = Cl (1‐Cl/2‐Cl), X = I (1‐I/2‐I). Attempts to functionalize these complexes by salt metathesis reaction revealed that the chlorido complexes are quite unreactive precursors, whereas the iodo complexes readily engage in a broad variety of reactions. We report the azide complexes (1‐N₃ and 2‐N₃ ) as well as the heavy cyanate complexes with the general formula (PN)₂M(OCPn) (M = Sc; Pn = P (1‐OCP), Pn = As (1‐OCAs) and M = Y; Pn = P (2‐OCP), Pn = As (2‐OCAs)). Furthermore, the amido and phosphanido complexes of the general formula (PN)₂M(PnHMesityl) with Pn = N (1‐NHMes, 2‐NHMes) and Pn = P (1‐PHMes, 2‐PHMes) are reported. Attempts to synthesize benzyl complexes of the general type (PN)₂M(Benzyl) with M = Sc, Y, La revealed drastic differences in the reactivity of the group III metal ions. While the scandium and yttrium complexes displayed elimination of KPN without the formation of defined metal complexes, for lanthanum, defined C‐H activation chemistry has been observed, yielding ‐ate complex 3‐CH.

The coordination chemistry and reactivity of anilidophosphine‐supported scandium and yttrium complexes is reported and compared to their lanthanum analogues previously reported by us. The manuscript shows striking differences in the stability and affinity of the anilidophosphine ligand toward these lighter group III metals, resulting in divergent reactivities between Sc, Y, and La.

1. Introduction

Amido‐ and anilidophosphine ligands of the PN(P)‐type, are well established in organometallic and coordination chemistry.^{[ 1 ]} Not only being prominently used in catalysis,^{[1b, 2]} they have also gained extensive interest to stabilize highly reactive compounds,^{[ 3 ]} paired with their (structural) characterization across the whole periodic table, ranging from p‐block elements^{[ 4 ]} to transition metals^{[ 1} , ³ , ^{5 ]} as well as the f‐elements.^{[ 6 ]} Especially, their ability to stabilize extremely reactive metal ligand multiple bonds such as early transition metal alkylidenes,^{[ 7 ]} phosphinidenes,^{[ 8 ]} alkylidynes,^{[ 9 ]} imidos,^{[ 10 ]} or highly reactive late transition nitrido/nitrenes is remarkable.^{[ 11 ]} In addition, these ligands were able to stabilize [P₂]^0/‐1/‐2 bridged complexes of platinum^{[ 12 ]} or nickel,^{[ 13 ]} a Pd(I) carbonyl complex^{[ 14 ]} or exotic magnetic properties, e.g. in iron(III) halide complexes.^{[ 1a ]}

Furthermore, the use of PN(P) chelating ligands has been found to be highly beneficial for the stabilization of highly reactive group III metal complexes. In 1996, Fryzuk and co‐workers, started to investigate the chemistry of scandium amido‐phosphine complexes of the general formula ScCl₂[N(SiMe₂CH₂P ⁱ Pr₂)₂],^{[ 15 ]} which opened a large research area pf phosphorus‐derived ligands in group III chemistry. For example, Hou and co‐workers examined the use of PNP ligands related to A, in the synthesis of cationic or polyhydride complexes of yttrium and scandium^{[ 6} , ^{16 ]} and to stabilize rare polymetallic yttrium‐ruthenium complexes.^{[ 17 ]} Similarly, a bidentate phosphinoamide ligand was used by Cui and co‐workers to synthesize bimetallic Sc/Y‐Pd complexes.^{[ 18 ]} Furthermore, in catalytic applications, carbazole, acridane, and iminodibenzyl frameworks have efficiently stabilized a large variety of alkyl complexes for the living polymerization of 1,3‐conjugated dienes and ε‐caprolactam.^{[ 6e,f,h ]} Focusing on highly reactive intermediates, the Mindiola group has facilitated the isolation of Tebbe‐like alkylidene/alkylidyne complexes of scandium^{[ 19 ]} and successfully implemented PNP ligand A in the synthesis of scandium imido^{[ 10} , ^{20 ]} and phosphinidene^{[ 8a ]} complexes. Notably, these species can also be stabilized using other tridentate ligands, as has been shown by Chen and co‐workers, who synthesized similar imido complexes^{[ 21 ]} as well as phosphino‐^{[ 22 ]} and boronyl‐phosphinidenes^{[ 23 ]} of scandium stabilized by a BDI‐derived ligand framework. While most of these examples rely on the use of tridentate frameworks, bidentate ligands have been used more sparsely, but not with less success. As such, ligand framework D has been prominently used in the early transition metal regime, stabilizing a variety of highly reactive titanium and zirconium nitrido^{[ 24 ]} and phosphido^{[ 25 ]} complexes (E, G, H, Figure 1), the first terminal titanium arsenide complex (F),^{[ 26 ]} rare Zr^III and Hf^III complexes (I, J)^{[ 25} , ^{27 ]} as well as terminal titanium methylidenes (K).^{[ 28 ]} Additionally, it was used in uranium chemistry to stabilize a rare uranium(IV) parent imide complex L, formed via a transient uranium(IV) nitride complex.^{[ 29 ]} We have recently started to investigate the potential of bidentate anilidophosphine ligands in lanthanide chemistry.^{[ 30 ]} Thereby, we found that depending on the steric bulk of the N‐substituents, the formation of monomeric and halide‐bridged complexes can be easily controlled.^{[ 31 ]} Furthermore, the ligand led to the isolation of a unique η³ coordination of the phopshaethynthiolate anion [SCP]^– toward lanthanum (N).^{[ 32 ]} Targeting terminal metal ligand multiple bonds (namely lanthanum phosphinidenes) we found that bulky N‐substituents such as a terphenyl favors the formation of a bridging phosphanido‐phosphinidene complex O ^{[ 33 ]} (formed because of the severe steric bulk of the terphenyl residues) and that the mesityl substituted ligand gives rise to transient terminal phosphinidenes,^{[ 34 ]} degrading via a 1,2‐CH activation mechanism, resulting in the formation of CH activated complex M. Given the fact that in group IV chemistry, similar CH‐activation reactions occur for hafnium,^{[ 35 ]} but not for zirconium^{[ 24a ]} and titanium nitrido^{[ 24d ]} complexes, we envisioned that the size of the central metal ion has a major influence whether mesityl‐CH activation is happening or not.

Figure 1.

Overview of popular PN(P) ligands (top) and selected examples of the utility of monoanionic bidentate PN ligands in the chemistry of group IV, the lanthanide and the actinide metals.

Thus, we here present the synthesis of PN complexes of the remaining group III metals, namely yttrium and scandium of the general formula (PN)₂MX (M = Sc, Y and X = Cl, I). Their functionalization chemistry is explored, and structural and chemical differences to previously reported lanthanum complexes^{[ 30} , ³² , ^{34 ]} are discussed when possible and if necessary. The results show that despite being in the same group, significant differences in reactivity exist between lanthanum and its lighter congeners yttrium and scandium and that the chemistry of the group III complexes (especially the stability of the bis‐PN framework) drastically differs from group IV chemistry.

2. Results and Discussion

Synthesis of the chloride complex 1‐Cl and 2‐Cl was achieved following a similar strategy previously reported by us for the synthesis of lanthanide PN complexes.^{[ 30a ]} Mixing LiPN and the corresponding metal precursors MCl₃(THF)₃ (M = Sc or Y) in toluene, heating them overnight followed by filtration, evaporation, and washing of the residue with cold hexane affords the chloride complexes 1‐Cl and 2‐Cl in almost quantitative yields of 85% and 94% respectively (Scheme 1). The process seems to be scalable and can be performed on multigram scales (up to 5 g). It is important to activate the trichlorides before the reaction in THF as it first improves their solubility in toluene, but also for Sc, reduces the Lewis‐acidity, suppressing undesired side‐reactions. Successful formation of the chloride complexes was evident by several features in the NMR spectroscopic signatures. For example, in the ¹H NMR, the characteristic ortho‐arene proton to the phosphorus atom is shifted from 5.62 ppm in LiPN to 5.70 ppm in 1‐Cl and 5.77 ppm in 2‐Cl. Similarly, the ³¹P{¹H} NMR resonance of the phosphorus atom shifts from −14.9 ppm in LiPN to −6.8 ppm in 1‐Cl and −3.6 ppm in 2‐Cl (¹ J _P‐Y = 89.5 Hz). X‐ray diffraction experiments from crystals grown from a concentrated diethyl ether solution unambiguously proofed the synthesis of the desired chlorido complexes (Figure 2). The complexes crystallize in the monoclinic space groups P2₁/c and P2₁/n without any solvent lattice molecules being present. The central metal ions are coordinated in a distorted trigonal bipyramidal fashion, displaying τ ₅ values of 0.80 (for 1‐Cl) and 0.69 (for 2‐Cl). The metal‐nitrogen distances are 2.1294(13)/2.1097(13) Å in 1‐Cl and 2.275(3)/2.256(3) Å in 2‐Cl, and are slightly shorter than the ones found in the lanthanum analogue 3‐Cl or the lutetium analogue previously reported by us.^{[ 30a ]} Similarly, the metal phosphorus bonds decrease from 2.9681(10)/2.9235(11) Å in 2‐Cl to 2.7983(5)/2.7842(5) Å in 1‐Cl and are comparable to previously reported Sc and Y phosphine distances. This is all in line with the decrease of the electronic radii of the corresponding metal ions when going up group III elements. Similarly, the M‐Cl distances decrease from 2.5216(11) Å in 2‐Cl to 2.3683(6) Å in 1‐Cl.

Scheme 1.

Synthesis of PN‐supported scandium, yttrium, and lanthanum chlorido complexes 1‐Cl, 2‐Cl, 3‐Cl and subsequent halide exchange to the corresponding iodido complexes 1‐I and 2‐I serving as focal starting points in the chemistry discussed here. Please note that the synthesis of 3‐Cl has been described previously.

Figure 2.

Molecular structures of the chlorido complexes 1‐Cl and 2‐Cl (top) and the iodido complexes 1‐I and 2‐I (bottom). Hydrogen atoms have been omitted for clarity; ellipsoids are shown at a probability level of 50%.

Having these potential starting materials in hand, we were of course interested in their further functionalization. However, contrasting the reactivity of the lanthanide precursors 3‐Cl ^{[ 30a ]} the chloride complexes 1‐Cl and 2‐Cl were found to be quite unreactive for the synthesis of further functionalized complexes via salt metathesis strategies. Especially in the case of 1‐Cl, conversions were usually below 10%, and heating led to decomposition rather than driving the reaction to completion. Also, and similar to the lanthanum complex 3‐Cl, the complexes 1‐Cl and 2‐Cl do not tolerate the presence of THF, which induced rapid hydrolysis, making salt metathesis reactions quite challenging. Thus, we envisioned that the use of a better leaving group might enhance their reactivity and therefore we replaced the chlorido ligand with an iodido ligand by stirring 1‐Cl and 2‐Cl in toluene with an excess of trimethylsilyl iodide (Scheme 1). This resulted in the clean formation of the corresponding iodido complexes 1‐I and 2‐I. Successful halide transfer was indicated by ¹H NMR spectroscopy, showing a shift of the ortho‐arene protons from 5.70 and 5.77 ppm in 1‐Cl and 2‐Cl to 5.66 and 5.75 ppm in 1‐I and 2‐I. At the same time, the phosphorus resonance of the phosphine donor shifts from −6.8 to −4.3 ppm and from −3.6 to −2.1 ppm (¹ J _P‐Y = 86.5 Hz) in the ³¹P{¹H} NMR spectra. Full halide exchange was furthermore confirmed by X‐ray diffraction studies performed on single crystals grown by slow evaporation of a concentrated diethyl ether solution at room temperature for both 1‐I and 2‐I, respectively (Figure 2). The complexes crystallize in the monoclinic space group P2₁ (1‐I) and P2₁2₁2₁ (2‐I) without any solvent lattice molecules. Similar to the chlorido complexes, the iodido complexes adopt a distorted trigonal pyramidal environment around the group III metal centers, displaying τ ₅ values of 0.74 and 0.72 for 1‐I and 2‐I. Of course, the metal halide distances increase to 2.761(2) and 2.9408(10) Å, which is in line with the larger radius of iodido versus chlorido ligands. Asides, no big changes in the bond metrics have been observed, and further information can be obtained from Tables S1–S3 in the Supporting Information.

To our satisfaction, the iodido complexes now proved to be very suitable precursors for salt metathesis reaction (Scheme 2), allowing the synthesis of an array of functionalized metal complexes. To be able to compare their reactivity and chemistry with the lanthanum complexes previously reported by us, we initially aimed at the synthesis of azide and phospha‐/arsaethynolate complexes of these complexes. While the lanthanide azide complex was only accessible with minor (unknown) impurities, the lanthanum phosphaethynolate complexes could not be accessed at that time. To our delight, the complexes 1‐I and 2‐I react cleanly with NaN₃ in toluene with a few drops of DME added. The DME is added to enhance the solubility of the sodium azide. Please note, that addition of THF induced decomposition / hydrolysis of the complexes and is therefore not tolerated. Similarly, the phospha‐/ and arsaethynolate complexes are cleanly accessible by the addition of NaOCP / NaOCAs respectively. Successful formation of complexes is evident by ¹H NMR spectroscopy showing a shift of the ortho‐arene protons from 5.66 and 5.75 ppm in 1‐I and 2‐I to 5.67 and 5.73 ppm in 1‐N₃ and 2‐N₃ , to 5.62 and 5.68 ppm in 1‐OCP and 2‐OCP, and to 5.62 and 5.67 ppm in 1‐OCAs and 2‐OCAs. Similarly, the ³¹P NMR resonances shift to −5.9/−2.8 in 1‐N₃ /2‐N₃ (¹ J _YP = 90 Hz), −5.5/−2.6 in 1‐OCP/2‐OCP (¹ J _YP = 90 Hz) and to −5.6/−2.5 in 1‐OCAs/2‐OCAs (¹ J _YP = 95 Hz). In addition, the resonances at −334.3 and −341.7 ppm in the ³¹P NMR spectra of the complexes 1‐OCP and 2‐OCP show the presence of the [OCP]⁻ anion being coordinated to the Sc/Y center. Finally, the IR resonances at 2097 cm⁻¹ (1‐N₃ /2‐N₃ , Figures S112 and S119), 1691 (1‐OCP/2‐OCP, Figures S113 and S120) and 1631 cm⁻¹ (1‐OCAs/2‐OCAs, Figures S114 and S121) confirm the presence of the azide, phospha‐, and arsaethynolate anions, respectively. Unambiguous formation of the pseudohalogenido complexes was given by X‐ray diffraction analysis of crystals grown from vapor diffusions of hexane into toluene solutions of the complexes (Figure 3). All complexes except for 1‐OCP yielded X‐ray satisfactory crystals using this method. Unfortunately, for 1‐OCP no crystals could be obtained under any conditions tried, but given the NMR (Figures S22–S28) and IR (Figure S113) spectroscopic signatures we are confident to also assign a κ¹‐O coordination mode of the phosphaethynolate anion in 1‐OCP. Complexes 1‐N₃ , 2‐N₃ , 2‐OCP, and 1‐OCAs, 2‐OCAs crystalize isostructural in the monoclinic space group P2₁/c with one molecule in the asymmetric unit. The scandium and yttrium metals are in a slightly distorted trigonal bipyramidal coordination environment showing τ ₅ values between 0.78 and 0.81. The M1‐N10 distances in 1‐N₃ and 2‐N₃ are 2.0454(14) and 2.217(2) Å, respectively. Here, the Sc1‐N10 distance is remarkable shortened in comparison to the only other terminal Sc‐azido complex [K(2.2.2‐Crypt)][Cp*(Amid)Sc(N₃)₂] known (2.14–2.17 Å).^{[ 36 ]} Notably, for yttrium no terminal but only few 1,3‐bridged azido complexes have been structurally characterized.^{[ 37 ]} The phospha‐ and arsaethynolate complexes 2‐OCP, 1‐OCAs, and 2‐OCAs contain κ¹‐O coordinated phospha‐ and arsaethynolate anions, with M1‐O1 distances of 2.166(3) Å (2‐OCP) (similar to previous reported Y‐OCP complexes^{[ 38 ]}), 2.012(3) Å (1‐OCAs) and 2.176(2) Å (2‐OCAs). Structurally related κ¹‐O end‐on bound phosphaethynolate complexes of praseodymium and terbium have also been reported by Yu and co‐workers.^{[ 39 ]} All complexes are monomeric, which might also explain their relatively simple isolation compared to their lanthanum analogues, which were found to always contain minor impurities. ^{[ 32 ]} We would like to further note, that the stoichiometry in this reaction is highly important and that using a higher excess than 1.1 equiv of (for example) sodium azide produces several side products, of which one could be identified as the 1,3‐bridged ‐ate complex [Na(THF)₄][(PN)₂Y(N₃)₂] 2‐N₃’ forming an indefinite chain in the solid state (Figure S125 and Tables S1–S3). Finally, the complexes 1‐OCAs and 2‐OCAs state the first isolated and structurally characterized examples of arsaethynolate complexes of the rare‐earth metals.

Scheme 2.

Synthesis of functionalized azido, phospha‐ and arsaethynolato, as well as amido and phosphanido complexes.

Figure 3.

Molecular structures of the (pseudohalogenide) azido (1‐N₃ and 2‐N₃ , top), phosphaethynolate (2‐OCP, middle), and arsaethynolate (1‐OCAs and 2‐OCAs, bottom) complexes. Hydrogen atoms have been omitted for clarity; ellipsoids are shown at a probability level of 50%.

After the synthesis of the pseudohalogenide complexes, we further aimed toward the synthesis of amido and phosphanido complexes using KNHMes and KPHMes as the corresponding precursors.^{[ 30a ]} In all four cases, the reaction proceeded smoothly by mixing the starting materials in toluene, followed by stirring the reactions overnight and subsequent work‐up. This yielded the amido complexes 1‐NHMes and 2‐NHMes as pale‐yellow solids in good yields of 89% and 93%. The phosphanido complexes 1‐PHMes and 2‐PHMes show an intense orange color and could be obtained in satisfactory yields of 87% and 83%. The successful formation of the complexes is indicated by a variety of spectroscopic features in the NMR signatures. For example, aside from the appearance of the typical mesityl ligand resonances in the ¹H and ¹³C NMR spectra, the presence of a NH resonance at 6.57 and 5.62 ppm for 1‐NHMes and 2‐NHMes is unambiguous proof of the desired amido complexes. For the phosphanido complexes 1‐PHMes and 2‐PHMes, the presence of a doublet at 3.99 (¹ J _PH = 217 Hz) and 3.59 (¹ J _PH = 209 Hz, ² J _YH = 1.9 Hz) ppm in their ¹H NMR spectra is further indicative of the presence of the phosphanido ligand. Likewise, the ³¹P{¹H} NMR spectra of the complexes display in addition to the expected PN ligand resonances (−7.1 and −2.9 ppm [¹ J _YP = 84.2 and ² J _PP = 19.8 Hz]), a new resonance at −38.7 ppm and −65.2 ppm (¹ J _YP = 96 Hz and ² J _PP = 19.8 Hz) for 1‐PHMes and 2‐PHMes, further confirming the presence of the phosphanido ligands. Especially the observed yttrium coupling resulting in a doublet of triplets in the ³¹P NMR spectrum of 2‐PHMes is an unambiguous proof for the phosphanido ligand being bound to yttrium (Figure S92). Crystals suitable for X‐ray diffraction experiments of all complexes 1‐EHMes and 2‐EHMEs (E = N or P) could be grown from concentrated diethyl ether solutions at room temperature (Figure 4). While the scandium and yttrium centers in the amido complexes 1‐NHMes and 2‐NHMes are in a nearly trigonal bipyramidal coordination environment (τ ₅ = 0.94 and 0.97), the phosphanido complexes 1‐PHMes and 2‐PHMes are strongly distorted between trigonal bipyramidal and square pyramidal (τ₅ = 0.6 for both complexes). Compared to previously reported PN lanthanum complexes,^{[ 30a ]} the metal‐mesityl amide/phosphanide distances are shortened to 2.0421(18) and 2.208(5) Å in 1‐NHMes and 2‐NHMes and to 2.6841(10) and 2.7933(16) Å in 1‐PHMes and 2‐PHMes. Besides these values, the bond distances/angles are comparable to the other structures reported in this study, wherefore a detailed description is omitted here. For more structural information, please see the Supporting Information, Tables S1–S3. Unfortunately, all attempts to either reduce the azide complexes 1‐N₃ or 2‐N₃ or to deprotonate the amido and phosphanido complexes 1‐PnHMes or 2‐PnHMes (Pn = N, P) have led to intractable mixtures, from which we could only identify KPN or LiPN as potential side products so far.

Figure 4.

Molecular structures of the amido (1‐NHMes and 2‐NHMes) and the phosphanido complexes (1‐PHMes and 2‐PHMes) of scandium and yttrium stabilized by monoanionic, bidentate anilidophosphine ligands. Hydrogen atoms have been omitted for clarity; ellipsoids are shown at a probability level of 50%.

To further validate our hypothesis that Mes‐CH₃ activation is hindered in smaller group III metals yttrium and scandium, but not in lanthanum we set out to synthesize benzyl complexes from all precursors 1‐I, 2‐I, and the lanthanum complex 3‐Cl (Scheme 3). This revealed a striking difference in the chemistry of group III PN complexes as well as undesired reactivities. Addition of benzyl potassium to 1‐I or 2‐I results in the clean formation of KPN and other (unidentifiable) products observed by ³¹P NMR spectroscopy. Given the fact that KPN might be more stable than the anticipated scandium or yttrium alkyl complexes, we tried to encapsulate the potassium atom using 18crown6. Unfortunately, this strategy did not help to resolve the problem, and we repeatedly find the formation of KPN salts in the reaction mixtures and could (so far) not identify any other scandium or yttrium containing products. Contrary, using the lanthanum complex 3‐Cl and benzyl potassium in the presence of 18crown6 resulted in the clean formation of a new species, showing two distinct PN ligands by ¹H and ³¹P NMR spectroscopy (Figures S96 and S106). The latter shows to distinct doublets at 4.6 and 3.1 ppm with a ² J _PP coupling of 40.7 Hz. The presence of a ² J _PP coupling across the lanthanum center, was already observed in a C‐H activated PN mesitolate complex,^{[ 34 ]} indicating deprotonation of the mesityl‐CH₃ instead of the synthesis of a desired alkyl complex. Indeed, crystals grown from cold DME solutions confirmed the presence of the mesityl C‐H activated complex 3‐CH (Figure 5). The lanthanum center is in a strongly distorted octahedral coordination environment, coordinated by the two PN ligands, one chloride atom, and the mesityl‐CH₂ alkyl group. The La‐C25 distance is 2.644(6) Å and in line with previous examples of CH activated PN lanthanum complexes.^{[ 34 ]} Furthermore, the CH activation induced a slight asymmetry in the lanthanum PN bond parameters and the La1‐N1 and La1‐N2 distances are 2.401(4) Å for the activated and 2.470(4) Å for the non‐activated PN ligand. Contrasting the cyclometallated phosphanido complex M,^{[ 34 ]} in which the potassium ion coordinats via mesityl‐arene interactions, in complex 3‐CH, the potassium ion is coordinated towards the terminal chloride ligand, displaying K1‐Cl1 distances of 3.2456(16) Å. The K1‐Cl1 interaction furthermore results in an elongation of the La1‐Cl1 distance from 2.691(1) Å^{[ 30a ]} in 3‐Cl to 2.7843(14) Å in 3‐CH. The potassium coordination toward the halide atom instead of the mesityl‐arene framework also influences the packing of the different complexes. While complex M forms a 1D chain bridged by K‐arene interactions,^{[ 34 ]} 3‐CH builds discrete molecular units.

Scheme 3.

Attempted synthesis of scandium, yttrium, and lanthanum alkyl complexes using benzyl potassium (KBn) as an alkyl source.

Figure 5.

Molecular structure of the cyclometallated lanthanum chlorido complex 3‐CH. Hydrogen atoms have been omitted for clarity; ellipsoids are shown at a probability level of 50%.

3. Conclusion

To conclude, we have presented the synthesis and first reactivities of scandium and yttrium bis‐PN complexes 1‐I and 2‐I and compared them with their lanthanum analogue 3‐Cl recently reported by us. While for the coordination chemistry with (hetero‐)cumulene anions, the scandium and yttrium complexes seem to give more reliable and cleaner transmetallation reactions, the chemistry with functionalized amido or phosphanido complexes 1‐PnHMes and 2‐PnHMes (Pn = N, P) is strongly limited and all attempts to deprotonate the amido or phosphanido ligands in these complexes resulted in either ligand elimination (LiPN or KPN) together with intractable reaction mixtures. Similarly, in the attempt of synthesizing alkyl complexes of scandium and yttrium, KPN elimination was found to be favored over alkyl coordination. However, when using the bigger lanthanum complex, mesityl‐CH activation was again observed, leading to the clean formation of an anionic chloride ‐ate complex 3‐CH. The latter could be a useful precursor to study further protonolysis reactivity and potential catalytic hydroelementation reactions. Overall, the study manifests that the monoanionic PN ligand seems to be a very good supporting ligand for lanthanides/lanthanum but has only limited suitability in the chemistry of lighter group III elements scandium and yttrium, which is surprising given its importance and versatility in titanium and zirconium chemistry.

Supporting Information

The authors have cited additional references within the Supporting Information.^{[ 40 ]} The supporting information contains all relevant synthetic procedures and corresponding NMR and IR spectra. Furthermore, crystallographic details are found in the Supporting Information, and the CCDC numbers are listed in Tables S1 and S2.^{[ 41 ]}

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Dedicated to Franc Meyer and Christian Limberg on the occasion of their 60th birthday

Hett F., Wittwer B., Bereiter S., Seidl M., Hohloch S., Chemistry - An Asian Journal. 2025, e202500364. 10.1002/asia.202500364

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.