On the Reactivity of Mes*P(PMe₃) towards Aluminum(I) Compounds – Evidence for the Intermediate Formation of Phosphaalumenes

Graphical Abstract

Phospha-Wittig reagent Mes*P(PMe₃) reacts with different Al(I)-species to give C−H bond activated products, namely 1,2-P,Al-tetrahydronaphtalene derivatives, via an intermediate phosphaalumene. Depending on the Al(I) source the putative phosphaalumene was also found to dimerize or to add an additional Al(I)-fragment.

Abstract

Phosphaalumenes are the heavier isoelectronic analogs of alkynes and have eluded facile synthesis until recently. We have reported that the combination of a phosphinidene transfer agent, ^ArTerP(PMe₃) (^ArTer=2,6-Ar₂-C₆H₃), with (Cp*Al)₄ (Cp*=C₅(CH₃)₅) afforded the phosphaalumenes ^ArTerPAlCp* as isolable, violet, thermally stable compounds. In here we describe attempts to utilize Mes*P(PMe₃) (Mes*=2,4,6-tBu₃-C₆H₂) as a phosphinidene source in combination with different Al(I) precursors, namely ^DipNacnacAl (^DipNacnac=HC[C(Me)NDip]₂, Dip=2,6-iPr₂-C₆H₃), (Cp*Al)₄ and Cp^3tAl (Cp^3t=1,2,4-tBu₃-C₅H₂). In all cases the formation of phosphaalumenes was not observed, however, their intermediate formation is indicated by formation of the dimer [Cp*Al(μ-PMes*)]₂ (2) and C−H-bond activation products along the putative P=Al bond, giving unusual 1,2-P,Al-tetrahydronaphtalene derivatives 1 and 4, clearly underlining the role the sterically demanding group on phosphorus plays in these transformations. The reactivity studies are supported by theoretical studies, demonstrating a thermodynamic preference for the C−H activation products. Additionally, we show that there are potential pitfalls in the synthesis of Cp*₂AlH, the precursor to make (Cp*Al)₄ and give recommendations how to circumvent these.

Introduction

Multiple bonding between trielenes (group 13 elements, E¹³) and pnictogens (E¹⁵) is of general interest due to the isovalence electronic relationship with C−C multiple bonds. Nevertheless, the synthesis of such compounds is challenging, which can be attributed to the intrinsic weakness of the π-bond between E¹³−E¹⁵ elements. Moreover, adjacent Lewis acidic group 13 and Lewis basic group 15 elements result in a pronounced oligomerization tendency of pnictatrielenes.¹ Especially the heavier base-free multiple bonds between P and B, Al or Ga have eluded facile synthesis until recently. Only last year Liu and co-workers described the synthesis of the first free boraphosphene [(CH₂)(NDip)]₂B−P=B−NC[(NDip)(CH)]₂ (A, Figure 1, Dip=2,6-iPr₂-C₆H₃), enabled through push-pull substitution at the B and P atoms, respectively.² Base-stabilized acyclic boraphosphenes have been studied in detail,³ and NHC-stabilized phosphaborenes with small silyl-substituents (B, Figure 1) have been described, that can engage in metathesis reactions through Me₃Si-halide elimination.^3d Our group has used phosphanylidenephosphoranes of the type ArP(PMe₃) (Ar=sterically demanding aryl group),⁴ so-called phospha-Wittig reagents,⁵ as phosphinidene transfer reagents, by substitution of PMe₃ with stronger Lewis bases.⁶ The reaction of ^ArTerP(PMe₃) (Ar=Dip (C), Tip (D) 2,4,6-iPr₃-C₆H₂; ^ArTer=2,6-Ar₂-C₆H₃, Figure 1) with (Cp*Al)₄ (Cp*=C₅(CH₃)₅) afforded the first examples of phosphaalumenes ^ArTerP=AlCp*, as violet compounds with a highly polarized P−Al multiple bond.⁷ The size of the aryl group is important for the stabilization of phosphaalumenes, and using the minimally less demanding ^MesTerP(PMe₃) (^MesTer=2,6-(2,4,6-Me₃-C₆H₂)-C₆H₃), the three-membered species ^MesTerP(AlCp^x)₂ (E, Cp^x=Cp*, Cp^3t 1,2,4-tBu₃-C₅H₂; Figure 1) was obtained,^7b which according to theoretical aromaticity descriptors is a 2π-aromatic system.⁸ Attempts to use triphosphiranes Ar₃P₃ (Ar=Mes, Dip, Tip)⁹ in conjunction with Cp^3tAl or (Cp*Al)₄ afforded base-free diphosphadialanes, the formal dimers of the free phosphaalumenes.¹⁰ Depending on the aryl-group and the Cp-substituent on Al either the 1,3-isomers (F, Figure 1), with alternating P and Al atoms in the four-membered ring, or the 1,2 isomers (F’, Figure 1) with P−P and Al−Al bonds, were formed. Phosphagallenes were synthesised by Goicoechea and Schulz using phosphanyl- or gallaphosphaketenes, respectively. Both groups utilized the readily available Ga(I) precursor (^DipNacnac)Ga (^DipNacnac=HC[C(Me)NDip]₂) in conjunction with the phosphaketenes [(S)P]−PCO ([(S)P]=(H_nCNDip)₂P; n=1, 2) or (^DipNacnac)Ga(Cl)−PCO facilitating CO cleavage and formation of [(S)P]−P=Ga(^DipNacnac) (G, Figure 1),¹¹ or (^DipNacnac)Ga=P−Ga(Cl)(^DipNacnac) (H, Figure 1),¹² respectively. Early attempts to generate monomeric phosphagallenes, revealed the formation of the trimeric species [(2,4,6-Ph₃-C₆H₂)GaPcHex]₃ when (2,4,6-Ph₃-C₆H₂)GaCl₂ was treated with Li₂PcHex (cHex=cyclo-C₆H₁₁) in a 1 : 1 molar ratio.¹³ Similarly, when von Hänisch and co-workers utilized NHC-stabilized (NHC=N-heterocyclic carbene) metal silylphosphanides of the general formula (IMes)R₂MP(H)SitBuPh₂ (M=Al−In; R=Et, iPr; IMes={HCN(2,4,6-Me₃C₆H₂)}₂C:), oligomerization occurred and heterocubane structures, formally the tetramers of the phosphatrielenes, were obtained upon thermal intramolecular alkane (RH) elimination, rather than the corresponding NHC-stabilized phosphatrielenes (IMes)RM=PSitBuPh₂.¹⁴ The formation of dimers,^{3c, 15} trimers,¹⁶ and heterocubanes¹⁷ as oligomerization products of putative [RE¹³=E¹⁵R’] is well documented in the literature and underlines the importance of kinetic protection through sterically demanding groups and electronic effects invoked by the substituents.

Figure 1

Overview of isolable phosphatrielenes (A, C, D, G, H), along with dimers of phosphaalumenes (F, F’) and three-membered PAl₂-species (E).

Reactivity studies on phosphatrielenes have revealed a diverse reactivity towards small molecules. The in situ generated phosphaborene [Mes*PB(Tmp)] (Mes*=2,4,6-tBu₃-C₆H₂; Tmp=2,2,6,6-tetramethylpiperidyl) was shown to react with phenyl acetylene in [2+2] cycloadditions,^3c and to enable a bora-phospha-Wittig type chemistry, giving straightforward access to a variety of phosphaalkenes.¹⁸ A similar reactivity was reported for A with p-methyl-benzaldehyde, while p-fluoro-acetophenone reacted in C−H bond activation of the Me-group along the P=B fragment.² Phosphaalumenes C and D were found to react with alkynes, alkenes and butadiene, to give P,Al-heterocycles and P,Al-barrelenes, after initial [2+2] cycloaddition reactions at the P−Al double bond. Phosphanylphosphagallenes (G), mainly show 1,3-dipolar reactivity, akin to frustrated Lewis pairs (FLP), towards a range of protic substrates, with the anionic group being added to Ga, while the phosphanyl P atom is protonated.^{11, 19} H mainly reacts at the P−Ga double bond in 1,2-additions,²⁰ and the reversible activation of carbodiimides and CO₂ was reported.²¹ When attempting to generate the iminoalane ^DipNacnacAl=N^ArTer (Ar=Mes, Dip), Power, Roesky and co-workers observed CH-activation of one of the iPr-groups of the Dip-substituents in the ^DipNacnac-moiety, with the N atom being protonated and the carbanion binding to Al.²² This serves as another reminder that the correct choice of substituents on the group 13 and 15 elements is crucial to access pnictatrielenes.

In this manuscript we outline our attempts to synthesize phosphaalumenes using the phospha-Wittig reagent Mes*P(PMe₃) and different Al(I) sources. These studies clearly show that the Mes*-substituent is not ideal for the stabilization of highly reactive Al−P multiple bonds, as it is prone to be C−H activated at one of the o-tBu-groups, which has been documented on various occasions in the literature. Nevertheless, these studies show that unusual 6-membered heterocycles can be formed in selective fashion.

The intermediate formation of phosphaalumenes was authenticated by isolation of its dimer in one case and theoretical studies show a preference for tBu-activation along the putative Al=P bond. Moreover, we outline some practical recommendations in the synthesis of (Cp*Al)₄, which is a versatile and recently re-emerging Al(I) precursor.²³

Results and Discussion

We started our exploration of Mes*P(PMe₃) as Ar−P transfer reagent towards Al(I) precursors by combining it with ^DipNacnacAl. Treatment of ^DipNacnacAl with Mes*P(PMe₃) in C₆D₆ gave no discernable reaction at room temperature according to NMR spectroscopy. The mixture was thus heated to 60 °C for 2 h, which gave clean conversion into a new species, P,Al-tetrahydronaphtalene derivative 1, with a ³¹P{¹H} NMR signal at −158.7 ppm, which split into a doublet upon proton coupling (¹J_PH=209 Hz), with a corresponding doublet signal in the ¹H NMR spectrum at 3.48 ppm, indicating a PH-containing species. Evaporation of the solvent and recrystallization of the crude product from n-hexane afforded crystals suitable for SC-XRD, which revealed the formation of compound 1 (Scheme 1), which can be described as the deactivation product of the desired phosphaalumene ^DipNacnacAl=PMes* (1’), through activation of one of the o-tBu-groups along the putative Al=P bond. C−H bond activation of the o-tBu-group has been observed in many cases when the free phosphinidene Mes*−P has been generated intermediately, giving cyclic phosphaindane species.²⁴ A rather rigid framework in 1 is indicated by four doublet resonances for the iPr methyl groups, as well as two septets for the corresponding methine protons for the ^DipNacnac substituent. The C−H bond activation of the Mes*-substituent is indicated by multiple signals for the tBu-groups as well as by two signals for the m-C_ArH of the central aryl ring. The PH-functionality is further corroborated by the asymmetric P−H stretching mode in the IR spectrum at 2367 cm⁻¹, which agrees with the theoretically predicted value (calc. 2363 cm⁻¹).

Scheme 1

Synthesis of P,Al-tetrahydronaphtalene derivative 1.

1 crystallizes in the monoclinic spacegroup P2₁/c with four molecules in the unit cell (Figure 2). The P−Al bond (2.3275(6) Å) is rather short for a single bond (Σr_cov(P−Al)=2.37 Å,²⁵ (cf. {(Me₃Si)₂CH}₂Al−PPh₂ 2.3367(16) Å).²⁶ The central AlPC₄ six-membered ring adopts a butterfly conformation, with a fold angle along the P1−C44 axis of 127.93(8)° and the Al atom is located above the ^DipNacnac plane (18.2(2)°). The Al1−C47 distance of 1.9580(15) Å indicates a single bond, while the P1−C30 distance of 1.8587(14) Å is in the typical range of P−C_Ar distances (cf. Mes*₂P₂ 1.862(2) Å).²⁷ The P−H was identified in the difference fourier map, thus allowing to discuss the angles at P, which shows a distorted trigonal pyramidal coordination environment (Σ(∠P)=298.72(7)°), whereas according to a τ₄-value of 0.88, Al is in a minimally distorted tetrahedral coordination mode.²⁸

Figure 2

Molecular structure of 1. Ellipsoids drawn at 50 % with C-H-atoms (except those on C47 and P1) omitted, Dip- and tBu-groups rendered as wire-frame for clarity. Selected bond lengths [Å] and angles [°], 4: Al1−P1 2.3275(6), Al1−N1 1.9021(12), Al1−N2 1.9297(13), P1−C30 1.8587(14), Al1−C47 1.9580(15); C30−P1−Al1 100.72(5), C30−P1−H1 99.2(9), Al1−P1−H1 98.8(8), Al1−C16 1.973(2); C16−Al1−C3 118.26(9), C3−Al1−P2 113.89(7), C25−P2−Al1 96.83(7), N1−Al1−C47 118.84(6), N1−Al1−P1 117.00(4), Al1−C41 2.1603(18), C1−P1−Al1 112.00(5); Al1−P1−C30−C35−43.93(11).

Theoretical studies at the DLPNO-CCSD(T)/def2-TZVP²⁹ using the PBE0-D3/def2-SVP³⁰ optimized geometries for obtaining the corrections to the free enthalpy, revealed that the formation of 1 is strongly exergonic (Scheme 2, iii).³¹ The intermediate formation of phosphaalumene 1’ was considered, which is also exergonic, however, only by −70.9 kJ ⋅ mol⁻¹ (Scheme 2, i). The C−H bond activation along the Al=P double bond was found to be strongly exergonic (Δ_RG°=−119.4 kJ ⋅ mol⁻¹; Scheme 2, ii), thus clearly explaining the observed formation of 1 being the sole product of this transformation as the thermodynamic product. The formation of 1 is noteworthy, as six-membered rings containing Al and P in 1,2-position are rare. Hydroalumination of bis-alkynylphosphines with Et₂AlH afforded Al₂C₂P₂ rings, which can be regarded as masked FLPs,³² similar to reactivity observed for the six-membered species (R₂PCH₂AlMe₂)₂ (R=Me, Ph).³³ Considering the isoelectronic replacement of C by Al and P, compound 1 can be regarded as a 1,2-phosphaalumina-1,2,3,4-tetrahydronaphtalene derivative. By analogy to our recently described synthesis of the phosphaalumenes ^ArTerPAlCp* (Ar=Dip, Tip) we next focused on the reaction of Mes*P(PMe₃) with (Cp*Al)₄ under thermal conditions.

Scheme 2

Thermodynamic considerations for the formation of compound 1 at the DLPNO-CCSD(T)/def2-TZVP//PBE0-D3/def2-SVP level of theory.

The combination of Mes*P(PMe₃) and (Cp*Al)₄ in a 4 : 1 ratio in C₆H₆ afforded after heating to 80 °C without stirring for 16 h a clear yellow solution, from which a yellow crystalline solid precipitated after cooling to 8 °C (Scheme 3, top). ³¹P NMR spectroscopy of the precipitate revealed the formation of the four-membered ring [Cp*Al(μ-PMes*)]₂ (2), a 1,3-diphospha-2,4-dialane with a characteristic ³¹P NMR signal at −183.8 ppm (δ_calc.(³¹P)=−195.2 ppm). After work-up 2 was recovered in a yield of 9 %, making this only a side-product. X-ray quality crystals of 2 were obtained directly from the cooled reaction mixture and confirmed the formation of 2 (Figure 3), as a centrosymmetric dimer of the phosphaalumene Mes*PAlCp* (2’), with a deltoid central four-membered ring (P−Al 2.3166(10), 2.3473(10) Å) with the P-atoms in a significantly flattened trigonal pyramidal coordination environment (Σ(∠P)=344.53°) and nearly rectangular angles at Al (87.05(4)°) and P (92.95(4)°).The planarization at P is more pronounced than in the related [Cp*Al(μ-PTip)]₂ (cf. Σ(∠P)=332.816°),¹⁰ whereas the P−Al distances in 2 are similar (cf. 2.3099(6), 2.3395(6) Å) and also compare well with the six-membered species (Mes*PAlPh)₃ (cf. d_avg. (P−Al)=2.328(3) Å).^16b

Scheme 3

Reaction of Mes*P(PMe₃) with (Cp*Al)₄ (top) and thermodynamic considerations to understand the preference for compound 4.

Figure 3

Molecular structure of 2. Ellipsoids drawn at 50 % with C-H-atoms omitted, tBu- and Cp*-Me-groups rendered as wire-frame for clarity. Selected bond lengths [Å] and angles [°]: Al1−P1 2.3166(10), Al1−P1’ 2.3473(10), P1−C1 1.865(3), C1−P1−Al1 115.29(9), C1−P1−Al1’ 136.29(9), Al1−P1−Al1’ 92.95(4), P1−Al1−P1’ 87.05(4); Al1−P1−C1−C6 110.1(2), Al1−P1−Al1’−P1’ 0.000(35).

With 2 identified we turned to the supernatant solution, which revealed a broad singlet signal at −107.6 ppm and two additional major singlet signals at −48.6 and −163.6 ppm in the ³¹P NMR spectrum before further work-up. Removal of C₆H₆ from the supernatant solution and addition of n-pentane, again gave a yellow precipitate, which was isolated by decantation of the mother liquor. This precipitate was identified as the 2π-aromatic three-membered species Mes*P(AlCp*)₂ by its characteristic ³¹P NMR signal at −107.6 ppm (δ_calc.(³¹P)=−132.1 ppm; cf. ^MesTerP(AlCp*)₂ −116.4 ppm)^7b and a Mes* to Cp* ratio of 1 : 2 in the ¹H NMR spectrum. The formation of 2 and 3 can be understood when inspecting the thermodynamics of these transformations (Scheme 3, bottom). First the intermediate formation of the phosphaalumene 2’ was considered (Mes*P(PMe₃)+Cp*Al→Mes*PAlCp*+PMe₃; Scheme 3, i), which is exergonic by −14.3 kJ ⋅ mol⁻¹ (note that the monomerization of (Cp*Al)₄ is endergonic, but it was shown that the monomer is present in solution at elevated temperatures).³⁴ Subsequent dimerization of 2’ to give half an equivalent of 2 is exergonic by −75.0 kJ ⋅ mol⁻¹ (Scheme 3, ii). While the reaction of 2’ with a second equivalent of Cp*Al was found to be exergonic by −66.1 kJ ⋅ mol⁻¹ (Scheme 3, iii). One more parameter to differentiate between four-membered species 2 and three-membered 3 are the different ring deformation stretching modes in the IR spectrum. For 2 there are two strong modes at 513 and 493 cm⁻¹ (calc. 504 and 480 cm⁻¹), which according to DFT calculations correspond to the asymmetric deformation modes. In contrast 3 only shows a strong vibration at 470 cm⁻¹ (calc. 461 cm⁻¹), clearly assigned as a ring deformation mode, in addition to a band at 544 cm⁻¹ (calc. 545 cm⁻¹), which corresponds well with the theoretically predicted value for the Al−Al symmetric stretching vibration.

With 2 and 3 identified we were able to recover the main product in pure form by placing the n-pentane supernatant solution in the freezer at −30 °C for 24 h. This yielded pale yellow crystals, suitable for SC-XRD analysis and showed the formation of the C−H bond activation product 4 akin to 1, in which one of the o-tBu-groups was activated along the putative Al=P bond, while the Al-atom is coordinated by PMe₃. The ³¹P NMR spectrum of isolated 4 showed two singlet resonances at −48.6 and −163.2 ppm, with no coupling between the phosphorus nuclei. Upon ¹H coupling the signal at −163.2 ppm split into a doublet, with a corresponding doublet in the ¹H NMR spectrum at 3.42 ppm (¹J_PH=197.7 Hz), signifying the PH-moiety. The CH₂-group attached to Al is shielded significantly and gives a broad singlet at 0.47 ppm in the ¹H NMR spectrum with the C(CH₃)₂ part of the activated tBu-group at 1.88 ppm. In the solid state the Cp*-substituent on Al is η¹-coordinated, whereas a singlet in the ¹H NMR spectrum indicates a fast sigmatropic shift on the NMR time scale. The PMe₃ coordinated to Al shows a deshielded signal in the ³¹NMR spectrum, while the doublet in the ¹H NMR spectrum at 0.22 ppm is shielded compared to free PMe₃. The presence of the P−H group was further confirmed by IR spectroscopy, showing the P−H stretching mode at 2335 cm⁻¹ (calc. 2346 cm⁻¹).

In the solid state 4 (Figure 4) adopts a similar butterfly structure as found in 1 (fold angle along the C17⋅⋅⋅P2 axis 126.84(16)°), with a P(H)−Al atom distance of 2.3440(8) Å, while the Al−P_PMe3 distance is 2.4381(8), in line with a coordinative bond (cf. Cl₃Al(PMe₃)₂ 2.452(2), 2.459(2) Å).³⁵ Also the sum of angles at P (Σ(∠P)=293.79°) indicate a trigonal pyramidal coordination environment, whereas the coordination at Al is nearly tetrahedral (τ₄=0.91).²⁸ The formation of 4 again points to the intermediate formation of the phosphaalumene 2’, and to verify our assumption of a thermodynamically driven process we again turned to calculations on the DLPNO-CCSD(T)/def2-TZVP//PBE0-D3/def2-SVP level of theory. Starting from 2’ the C−H bond activation to give the PMe₃-free activation product 4’, was found to be exergonic by −80.6 kJ ⋅ mol⁻¹ (Scheme 3, iv), while the addition of PMe₃ to give 4 provided an additional energy gain of −14.5 kJ ⋅ mol⁻¹ (Scheme 3, v). Considering that the reaction takes place at 80 °C the formation of 4 as the main product is clearly thermodynamically driven.

Figure 4

Molecular structure of 4. Ellipsoids drawn at 50 % with C-H-atoms (except those on C16 and P2) omitted, Tip-groups rendered as wire-frame for clarity. Selected bond lengths [Å] and angles [°], 4: Al1−P1 2.4381(8), Al1−P2 2.3440(8), P2−C25 1.867(2), Al1−C3 2.032(2), Al1−C16 1.973(2); C16−Al1−C3 118.26(9), C3−Al1−P2 113.89(7), C25−P2−Al1 96.83(7), Al1−C41 2.1603(18), C1−P1−Al1 112.00(5); Al1−P2−C25−C20 −43.16(16).

In a last series of experiments we investigated the reaction of Mes*P(PMe₃) with monomeric Cp^3tAl³⁶ in C₆D₆ at room temperature in a 1 : 1 ratio at first, giving one product with a ³¹P{¹H} NMR signal at −104.5 ppm and unreacted Mes*P(PMe₃) in a 1 : 1 ratio, which is in line with the formation of Mes*P(AlCp^3t)₂ (5) and full conversion was achieved when starting from a 1 : 2 mixture of Mes*P(PMe₃) and Cp^3tAl after 16 h (Scheme 4). After slowly evaporating a n-pentane solution of 5, pale yellow crystals suitable for SC-XRD experiments were obtained.

Scheme 4

Synthesis of compound 5.

In the ¹H NMR spectrum the characteristic signals for the Cp^3t- and Mes*-substituent show the expected 2 : 1 ratio, with two singlet signals for the Cp^3t tBu-groups at 1.45 and 1.43 ppm and two singlets for the o-tBu and p-tBu-groups of the Mes*-substituent at 2.00 and 1.37 ppm, respectively. The aromatic protons of the Mes*-group at 7.39 ppm and of the Cp^3t-group at 6.29 ppm show the expected 1 : 2 ratio. The ³¹P{¹H} NMR shift of −104.5 ppm (δ_calc.(³¹P)=−123.3 ppm) compares well with that of 3 and is in the range of ^MesTerP(AlCp^3t)₂ (cf. −79.7 ppm). In case of 5 we were also able to determine the ²⁷Al NMR shift at −43 ppm as a broad singlet,³¹ which is in the range of Cp^3tAlBr₂ (δ(²⁷Al)=−46 ppm).³⁶ The Al₂P-ring in 5 is not isosceles (Figure 5) in contrast to previously reported ^MesTerP(AlCp^3t)₂ (cf. d(Al−P)=2.3543(8), 2.3495(7) Å). The Al2−P1 (2.2965(4) Å) is considerably shorter than the Al1−P1 distance (2.3664(4) Å), while the Al1−Al2 distance (2.5464(5) Å) is only minimally longer than that in ^MesTerP(AlCp^3t)₂ (2.5265(9) Å) and compares well with Al−Al distance in strained cyclic dialanes.³⁷ The P atom in 5 is in a trigonal planar coordination environment (Σ(∠P)=359.97°), which again indicates a 2π-aromatic system as was shown for ^MesTerP(AlCp^3t)₂ and the isovalence electronic species Na₂[^MesTer₃Al₃].³⁸ In the IR spectrum two characteristic modes for the PAl₂-ring were identified, with the mode at 528 cm⁻¹ (calc. 519 cm⁻¹) showing the ring breathing, whereas the mode at 459 cm⁻¹ (calc. 450 cm⁻¹) corresponds to ring deformation.

Figure 5

Molecular structure of 5. Ellipsoids drawn at 50 % with C-H-atoms omitted, tBu-groups rendered as wire-frame for clarity. Selected bond lengths [Å] and angles [°], 5: Al1−P1 2.3664(4), Al2−P1 2.2965(4), P1−C1 1.8862(11), Al1−Al2 2.5464(5); C1−P1−Al2 134.62(4), C1−P1−Al1 159.17(4), Al2−P1−Al1 66.177(14), P1−Al1−Al2 55.594(12), P1−Al2−Al1 58.229(13).

For the synthesis of ^DipNacnacAl we have used a modified synthesis that was recently described by Kretschmer and co-workers.³⁹ This route is based on the reaction of ^DipNacnacNa with (Cp*Al)₄ and gave ^DipNacnacAl (under formation of NaCp*) in reasonable, reproduceable yields when strictly following the literature procedure. However, we have encountered some challenges when preparing (Cp*Al)₄, which we want to disclose here. Two pathways for (Cp*Al)₄ are described in the latest edition of Inorganic Syntheses,⁴⁰ and we use the reductive elimination approach originally outlined by Fischer et al., using the thermal elimination of Cp*H from Cp*₂AlH, which gives (Cp*Al)₄ in pure form in a high yielding process.⁴¹ During the synthesis of Cp*₂AlH we have observed two potential pitfalls, which we want to report here. Firstly, the preparation of Cp*K is based on the reaction of Cp*H with KH in THF,⁴² using an excess of Cp*H, which can be removed afterwards by filtration. While stating that Cp*K is obtained in pure form after decantation of the supernatant THF-solution, washing with THF and careful drying, we have found that this can result in some remaining THF in the thus-prepared Cp*K. THF was found to interfere with the following reaction of Cl₂AlH and by fractional crystallization [Cp*AlHCl(thf)] (6) was identified as an undesired by-product (Scheme 5, top), by re-crystallization from Et₂O. In the ¹H NMR spectrum in C₆D₆ 6 was identified by three signals, with a singlet signal for the Cp*-substituent at 2.06 ppm, as well as two broad signals at 1.01 and 3.50 ppm for the coordinated THF molecule. To the best of our knowledge 6 is the first mixed chloride/hydride AlCp* derivative, in which the SC-XRD analysis revealed the Cp* substituent to be on the continuum of being η¹ and η³-coordinated, with one short Al−C3 (2.1032(18) Å) and two longer Al−C2 (2.3763(17) Å) and Al−C4 (2.3151(17) Å) (Figure 6, left). In contrast, the Cp*-substituent in the related [Cp*AlCl₂(thf)] complex is η¹-coordinated (Al−C 2.009(2) Å) and the Al−O_thf in 6 is considerably longer compared to [Cp*AlCl₂(thf)] (1.8332(15) Å),⁴³ in line with a less Lewis acidic Al-center in the mixed hydride/chloride 6. When extracting the reaction mixture of Cp*₂AlH with n-hexane and concentrating the filtrate 6 crystallizes first. After removal of the supernatant solution from 6 the desired Cp*₂AlH crystallized in analytically pure form from this solution. To circumvent the formation of 6 during the preparation of Cp*₂AlH we recommend washing the Cp*K with n-pentane rather than using only THF. Since adapting the n-pentane washing step we have not encountered the formation of 6 anymore.

Scheme 5

Pitfalls in the synthesis of Cp*₂AlH, which afforded the side-products 6 and 7.

Figure 6

Molecular structure of 6 (left) and 7 (right). Ellipsoids drawn at 50 % with C-H-atoms (except H on Al) omitted. Selected bond lengths [Å] and angles [°], 6: Cl1−Al1 2.1935(8), Al1−O1 1.9209(13), Al1−C3 2.1032(18), Al1−C4 2.3151(17), Al1−C2 2.3763(17), Al1−H2 1.703(17); Cl1−Al1−H2 103.3(5), O1−Al1−Cl1 95.30(5), O1−Al1−H2 100.6(5). 7: Al1−Cl1 2.1743(8), Al1−H3 1.674(18), Al1−H1 1.657(18), Al2−H1 1.752(17), Al2−H2 1.725(17), Al3−H4 1.689(17), Al3−H3 1.722(18), Al3−H2 1.782(17).

Another potential pitfall was identified when skipping a step in the literature procedure. According to this, after the reaction of Cp*K with Cl₂AlH the reaction mixture in Et₂O is filtered, then the filtrate is dried and this filtrate is again extracted with n-hexane and filtered to give Cp*₂AlH after recrystallization at −30 °C (Scheme 5, bottom). We attempted to directly dry the reaction mixture in Et₂O and then extract with n-hexane and filter afterwards, thereby skipping one filtration. This however resulted in the isolation of another undesired by-product, which was identified as a trimeric mixed Cp*₃Al₃Cl_1.5H_4.5 (7) species by SC-XRD analysis (Figure 6, right), which is in line with ¹H NMR spectroscopy, showing a singlet at 1.94 ppm for the Cp*-substituent in C₆D₆ solution along with a broad resonance for the hydrides in a 45 : 4.3 ratio, which agrees well with the ratio obtained from SC-XRD analysis (45 : 4.42). We note that there is probably fast exchange between [Cp*₃Al₃Cl₂H₄] and [Cp*₃Al₃ClH₅] in solution, thus giving the observed singlet resonance for the Cp*-groups in the ¹H NMR spectrum. In the solid state three hydrides were found to be μ-bridging, whereas Cl and the additional hydrides occupy end-on positions on Al, with the Cp*-substituents being η⁵-coordinated on Al1 and Al2, and showing a rather unsymmetric coordination on Al3 in line with a η³-coordination mode. The Al−(μ-H)−Al distances range from 1.65–1.75 Å, which agrees with the Al−H distances in the dimeric (ArAlH₂)₂ (1.60–1.72 Å, Ar=2,6-(CH(SiMe₃)₂)₂-4-tBu-C₆H₂).⁴⁴ Again, after removal of the supernatant solution from 7, the desired Cp*₂AlH was isolated in good yields by crystallization. We have not encountered the formation of 7 when strictly following the literature procedure for Cp*₂AlH.

Conclusion

Using the phospha-Wittig reagent Mes*P(PMe₃) the synthesis of phosphaalumenes was attempted. With the Al(I) precursor ^DipNacnacAl the 1,2-phosphaalumina-1,2,3,4-tetrahydronaphtalene derivative 1 was obtained, which signifies the intermediate formation ^DipNacnacAl=PMes*. However, the C−H-activation of an o-tBu-group of the Mes*-substituent along the putative Al=P bond occurred in a thermodynamically driven process. A similar outcome was observed for the reaction of Mes*P(PMe₃) with (Cp*Al)₄, which gave C−H-activation product 4 as the main species. However, in this case the dimer of a putative phosphaalumene, the four-membered species 2, as well as the three-membered ring system 3 were identified as by-products. The formation of 4 as the main product is again thermodynamically driven. Lastly, the selective formation the 2π-aromatic system 5 was achieved when Mes*P(PMe₃) was treated with the monomeric Cp^3tAl at room temperature. All this evidence points to the effect the sterically demanding group on P plays, and in particular the Mes*-group is shown to be not an ideal candidate as it is prone to C−H-activation of the o-tBu-groups. Moreover, the isolation of 2 clearly shows that the steric bulk of Mes* is not sufficient to stabilize a monomeric phosphaalumene. In addition, we have identified two potential side-products when preparing Cp*AlH, the direct precursor of (Cp*Al)₄, which is an essential starting material for the synthesis of low-valent aluminum species. Future work will focus on further fine tuning the sterically demanding groups on phosphorus and aluminium to access novel phosphaalumenes and to exploit their reactivity.

Experimental Section

If not stated otherwise, all manipulations were carried out under oxygen- and moisture-free conditions under an inert atmosphere of argon using standard Schlenk or Drybox techniques. All glassware was heated three times in vacuo using a heat gun and cooled under argon atmosphere. Solvents were transferred using syringes, which were purged three times with argon prior to use. Solvents and reactants were either obtained from commercial sources or synthesized as detailed in Table S1.³¹ NMR spectra were recorded on Bruker spectrometers (AVANCE 300, AVANCE 400 or Fourier 300) and were referenced internally to the deuterated solvent (C₆D₆ δ_ref=128.06 ppm), to protic impurities in the deuterated solvent (C₆HD₅ δ_ref=7.16 ppm) or externally (³¹P: 85 % H₃PO₄ δ_ref=0 ppm). The broad resonance at δ=70 ppm observed in the ²⁷Al NMR spectrum corresponds to a background from Al-nuclei in the probe head. All measurements were carried out at ambient temperature unless denoted otherwise. NMR signals were assigned using experimental data (e. g. chemical shifts, coupling constants, integrals where applicable) in conjunction with computed NMR data (GIAO method, cf. Computational details, p. S31 ff.).³¹ IR spectra of crystalline samples were recorded on a Bruker Alpha II FT-IR spectrometer equipped with an ATR unit at ambient temperature under argon atmosphere. Relative intensities are reported according to the following intervals: very weak (vw, 0–10 %), weak (w, 10–30 %), medium (m, 30–60 %), strong (s, 60–90 %), very strong (vs, 90–100 %). Elemental analyses were obtained using a Leco Tru Spec elemental analyzer. Important Note: Suitable CHN values for compounds 2–7 were not obtained, with deviations in between 3 to 15 % for both C and H values. As CHN analysis is unfortunately already prone to manipulations⁴⁵ we decided to not provide any values because in our eyes there is no evidence beyond a reasonable doubt that the material is pure at the point of measurement. However, the state-of-the-art characterization such as with NMR spectroscopy (see displayed spectra in the ESI) demonstrates the existence and analytical pureness of the herein published compounds if not stated otherwise. Crystallographic details are given in the ESI (p. S4 ff).³¹

Synthesis of compound 1: ^DipNacnacAl (0.045 mg, 0.1 mmol) and Mes*P(PMe₃) (0.036 g, 0.1 mmol) were combined in an NMR tube equipped with a J-Young valve and benzene-d₆ (0.6 mL) was added at room temperature. The NMR tube is sonicated until all solids have dissolved and the NMR tube is then placed in an oil bath set to 60 °C. The sample is heated to 60 °C over a period of 2 h resulting in a colour change from bright orange to yellow. After checking the reaction by ¹H and ³¹P{1H} NMR spectroscopy the reaction mixture is transferred into a Schlenk tube and the volatiles are removed in vacuo. The yellow solid residue is then extracted with n-heptane (2 mL) and filtered through a canula fitted with a glass filter paper to give a clear yellow solution. This solution is then placed in a freezer at −30 °C and compound 1 is afforded as yellow crystalline solid after 48 h. Yield: 45 mg (0.062 mmol, 62 %). Single crystals suitable for X-ray diffraction can be grown from saturated n-hexane solution at 5 °C.

CHN calc. (found) in %: C 78.29 (78.09), H 9.79 (9.71), N 3.89 (3.70). ³¹P{¹H} NMR (C₆D₆, 162.01 MHz): δ=−158.7. ³¹P NMR (C₆D₆, 162.01 MHz): δ=−158.7 (d, ¹J_PH=201.9 Hz). ¹H NMR (C₆D₆, 400.1 MHz): δ=0.47 (s, 2H, Al-CH₂(CMe₂)), 1.07 (d, ³J_HH=6.7 Hz, 6H, CH(CH₃)₂), 1.14 (d, ³J_HH=6.8 Hz, 6H, CH(CH₃)₂), 1.22 (d, ³J_HH=6.7 Hz, 6H, CH(CH₃)₂), 1.29-1.34 (m, 21 H, o-tBu, Al-CH₂(C(CH₃)₂), CH(CH₃)₂)*, 1.54 (s, 6 H, CH₃-^DipNacnac), 1.75 (s, 9 H, p-tBu), 3.25 (hept, ³J_HH=6.7 Hz, 2H, CH(CH₃)₂), 3.47 (hept, ³J_HH=6.7 Hz, 2H, CH(CH₃)₂), 3.48 (d, ¹J_PH=202.3 Hz, PH), 4.85 (s, 1H, CH-^DipNacnac), 7.01-7.13 (m, 6H, C_ArylH-Dip), 7.34 (m, 1 H, m-C_ArylH), 7.45 (m, 1 H, m-C_ArylH). ¹³C{¹H} NMR (C₆D₆, 100.6 MHz): δ=23.9 (s (br), AlCH₂), 24.2 (s, CH₃-^DipNacnac), 24.8, 24.9 (d, J_PC=7.4 Hz), 25.1, 25.3 (CH(CH₃)₂), (s, CH₃-^DipNacnac), 28.7 (s, CH(CH₃)₂), 29.3 (d, J_PH=6.1, CH(CH₃)₂), 31.7 (s, p-C(CH₃)₃), 32.6 (d, J_PC=6.2 Hz, o-C(CH₃)₃), 34.8 (s, p-C(CH₃)₃), 35.5 (d, J_PC=10.5 Hz, C(CH₃)₂), 38.5 (o-C(CH₃)₃), 40.6 (C(CH₃)₂), 98.6 (HC(C(CH₃)NDip)₂), 120.5, 121.6 (CH_aryl, Mes*), 124.4, 125.1 (m-CH_Dip), 127.6,* 128.6* (o-CH_Dip), 132.1 (¹J_PC=34.3 Hz, C_ipso), 141.0, 144.0, 145.0 (C_q,Dip), 145.4 (p-C_q,Mes*), 152.7 (d, ²J_PC=4.2 Hz, o-C_q,Mes*), 154.4 (d, ²J_PC=5.2 Hz, o-C_q,Mes*, activated tBu-group), 170.9 (C_q, HC(C(CH₃)NDip)₂). *overlap with the C₆D₆ signal, assigned via HSQC, and DEPT-135 spectra. IR (ATR, 32 scans, cm⁻¹): =3067 (vw), 3057 (vw), 2956 (s), 2925 (m), 2867 (m), 2366 (vw), 1663 (vw), 1622 (vw), 1591 (w), 1550 (m), 1521 (s), 1515 (s), 1461 (m), 1435 (s), 1385 (vs), 1358 (s), 1315 (s), 1296 (m), 1253 (m), 1214 (w), 1208 (m), 1193 (w), 1173 (m), 1132 (w), 1105 (m), 1084 (m), 1053 (m), 1041 (m), 1016 (m), 969 (w), 936 (m), 884 (m), 872 (m), 860 (m), 822 (w), 794 (s), 759 (s), 726 (w), 711 (m), 678 (m), 653 (m), 641 (m), 610 (m), 596 (w), 557 (m), 528 (m), 495 (m), 474 (m), 443 (s).

Reaction of Mes*P(PMe₃) with (Cp*Al)₄.

Isolation of compound 4: (Cp*Al)₄ (0.162 mg, 0.25 mmol) and Mes*P(PMe₃) (0.353 g, 1.0 mmol) were combined in a Schlenk tube and benzene (3 mL) was added at room temperature. The Schlenk tube was then heated to 80 °C (oil bath). The sample was heated at 80 °C over a period of 16 h without stirring resulting in a clear yellow solution. Upon cooling to 8 °C a yellow crystalline solid precipitated (compound 2, see below), from which the supernatant was removed via a canula fitted with a glass filter paper to give a clear yellow filtrate. From this filtrate the volatiles were removed in vacuo (1 ⋅ 10⁻³ mbar), giving a yellow solid. This solid was then extracted with n-pentane (2 mL), again giving a yellow precipitate (compound 3, see below), and the supernatant solution was filtered and the filtrate was then concentrated to incipient crystallization and placed in a freezer at −30 °C. Compound 4 was afforded as a pale yellow crystalline solid after standing at −30 °C for 24 h and removal of the supernatant solution and drying in vacuo (1 ⋅ 10⁻³ mbar). Yield: 0.155 mg (0.030 mmol, 30 %; based on Mes*P(PMe₃)). Single crystals suitable for X-ray diffraction can be grown from a saturated n-pentane solution at −30 °C.

³¹P{¹H} NMR (C₆D₆, 121.5 MHz): δ=−162.0 (PH), −48.6 (PMe₃). ³¹P NMR (C₆D₆, 121.5 MHz): −162.0 (d, ¹J_PH=197.7 Hz, PH), −48.6 (decett, ²J_PH=7.7 Hz, PMe₃). ¹H NMR (C₆D₆, 300.1 MHz): δ=0.22 (d, ²J_PH=7.7 Hz, 9 H, P(CH₃)₃), 0.47 (s (br), AlCH₂), 1.34 (s, p-C(CH₃)₃), 1.81 (s, o-C(CH₃)₃), 1.88 (AlCH₂(C(CH₃)₂), 1.93 (s, C₅(CH₃)₅), 3.42 (d, ¹J_PH=197.7 Hz), 7.50 (m, 1 H, m-CH), 7.52 (m, 1 H, m-CH). ¹³C{¹H} NMR (C₆D₆, 75.5 MHz): δ=10.1 (d, ¹J_PC=20.5 Hz, P(CH₃)₃), 12.3 (d, J_PC=2.7 Hz, C₅(CH₃)₅), 24.5 (s (br), AlCH₂), 31.7 (s, p-C(CH₃)₃), 32.4 (d, J_PC=4.6 Hz, o-(C(CH₃)₃), 34.8 (s, p-C(CH₃)₃), 36.1 (s (br), AlCH₂C(CH₃)₂), 38.5 (s, o-C(CH₃)₃), 41.3 (s (br), AlCH₂C(CH₃)₂), 117.8 (s, C₅(CH₃)₅), 121.0 (d, J_PC=2.7 Hz, m-C, PAr), 121.8 (s, m-C, PAr), 133.6 (J_PC=37.9 Hz, ipso-C, PAr), 145.8 (p-C, Ar), 152.2 (o-C, PAr), 154.2 (J_PC=7.4 Hz, o-C (C(CH₃)₂CH₂Al), PAr). IR (ATR, 32 scans, cm⁻¹): =2956 (s), 2906 (m), 2865 (m), 2853 (m), 2335 (w), 1593 (w), 1536 (w), 1476 (w), 1457 (m), 1445 (m), 1428 (w), 1414 (m), 1408 (m), 1391 (m), 1379 (m), 1358 (m), 1307 (w), 1284 (w), 1253 (m), 1239 (m), 1210 (m), 1183 (m), 1169 (w), 1130 (w), 1086 (w), 1041 (m), 961 (m), 946 (m), 903 (w), 890 (w), 874 (m), 843 (s), 824 (m), 804 (m), 748 (m), 699 (m), 658 (s), 641 (m), 631 (m), 589 (m), 561 (m), 542 (m), 534 (m), 493 (s), 466 (vs), 443 (s), 416 (m).

Isolation of compound 2: After heating a mixture (Cp*Al)₄ (0.162 mg, 0.25 mmol) and Mes*P(PMe₃) (0.353 g, 1.0 mmol) to 80 °C for 16 h and cooling to 8 °C the four-membered ring-species [Cp*Al(μ-PMes*)]₂ (2) (0.040 g, 0.045 mmol, 9 % based on Mes*P(PMe₃)), was obtained as a yellow precipitate, which was isolated by decantation of the supernatant solution. X-Ray quality crystals were obtained directly from the reaction mixture.

³¹P{¹H} NMR (C₆D₆, 121.5 MHz): δ=−183.8. ¹H NMR (C₆D₆, 300.1 MHz): δ=1.37 (s, 9 H, p-t-Bu), 1.89 (s, 15 H, C₅(CH₃)₅), 2.00 (s, 18 H, o-t-Bu), 7.50 (s, 2 H, m-H). ¹³C{¹H} NMR (C₆D₆, 75.5 MHz): δ=11.7 (s, C₅(CH₃)₅), 31.6 (s, p-C(CH₃)₃), 34.9 (s, p-C(CH₃)₃), 35.4 (s, o-C(CH₃)₃, PMes*), 40.0 (s, o-C(CH₃)₃), 116.8 (s, C₅Me₅), 122.3 (s, m-C, PMes*), 147.2 (s, p-C, PMes*), 156.9 (s, o-C, PMes*), ipso-C at PMes* not observed. IR (ATR, 32 scans, cm⁻¹): =2956 (m), 2908 (m), 2867 (w), 1593 (w), 1542 (vw), 1534 (vw), 1521 (vw), 1476 (w), 1459 (w), 1408 (w), 1391 (w), 1381 (w), 1360 (m), 1309 (vw), 1278 (w), 1235 (m), 1210 (w), 1195 (w), 1165 (w), 1119 (w), 1107 (w), 1066 (w), 1020 (w), 998 (w), 983 (w), 921 (w), 901 (w), 876 (m), 831 (m), 814 (m), 802 (m), 752 (m), 730 (m), 709 (m), 680 (m), 672 (m), 662 (m), 645 (m), 625 (m), 612 (m), 596 (m), 513 (s), 493 (vs), 437 (s).

Isolation of compound 3: After heating a mixture (Cp*Al)₄ (0.162 mg, 0.25 mmol) and Mes*P(PMe₃) (0.353 g, 1.0 mmol) to 80 °C for 16 h and cooling to 8 °C the supernatant solution was removed from a yellow precipitate by decantation. From the supernatant solution the volatiles were removed in vacuo and n-pentane (2 mL) was added, resulting in the precipitation of a yellow solid. This yellow solid was isolated by decantation of the supernatant n-pentane solution and was identified as the three-membered ring species Mes*P(AlCp*)₂ (3, 0.030 g, 0.05 mmol, 5 % based on Mes*P(PMe₃)). X-Ray quality crystals of 3 could not be obtained.

³¹P{¹H} NMR (C₆D₆, 121.5 MHz): δ=−107.6. ¹H NMR (C₆D₆, 300.1 MHz): δ=1.30 (s, 9 H, p-t-Bu), 1.88 (s, 30 H, C₅(CH₃)₅), 2.14 (s, 18 H, o-t-Bu), 7.51 (s, 2 H, m-H). ¹³C{¹H} NMR (C₆D₆, 75.5 MHz): δ=10.4 (d, J_PC=2 Hz, C₅(CH₃)₅), 31.6 (s, p-C(CH₃)₃), 34.2 (d, J_PC=5.3 Hz, o-C(CH₃)₃, PMes*), 35.1 (s, p-C(CH₃)₃), 38.2 (s, o-C(CH₃)₃), 114.0 (s, C₅Me₅), 121.2 (s, m-C, PMes*), 147.5 (s, p-C, PMes*), 154.1 (s, o-C, PMes*), ipso-C at PMes* not observed. IR (ATR, 32 scans, cm⁻¹): =2954 (m), 2906 (m), 2865 (m), 1595 (w), 1546 (vw), 1529 (vw), 1459 (w), 1447 (w), 1410 (w), 1387 (w), 1377 (w), 1360 (m), 1280 (vw), 1237 (w), 1212 (w), 1189 (w), 1022 (w), 983 (w), 923 (w), 876 (m), 831 (m), 814 (m), 800 (m), 752 (m), 730 (m), 715 (m), 678 (m), 666 (m), 658 (m), 647 (m), 629 (m), 596 (m), 544 (s), 470 (vs), 437 (s), 420 (s).

Synthesis of compound 5: Cp^3tAl (0.030 g, 0.11 mmol) and Mes*PPMe₃ (20.0 mg, 0.056 mmol) were mixed in C₆H₆ (0.6 mL). While stirring for a day at room temperature, the solution turned orange. Benzene and the forming PMe₃ were removed under vacuum and the residue was extracted with n-pentane. Slow evaporation of the solvent at −30 °C yielded pale yellow crystals of Mes*P(AlCp^3t)₂ (6), suitable for single crystal X-ray diffraction. Yield: 20.8 mg, 0.026 mmol (46 %).

³¹P{¹H} NMR (C₆D₆, 162MHz): δ=−104.5 ppm. ²⁷Al NMR (C₆D₆, 104 MHz): δ=-43 (s, (br)) ppm. ¹H NMR (C₆D₆, 400 MHz): δ=1.37 (s, 9H, p-C(CH₃)₃), 1.42 (s, 18H, C(CH₃)₃, Cp^3t), 1.45 (s,36H, C(CH₃)₃, Cp^3t), 2.00 (s,18H, o-C(CH₃)₃), 6.29 (s, CH_Cp, 4H), 7.39 (s, 2H, CH_Aryl). ¹³C{¹H} NMR (C₆D₆, 101 MHz): δ=31.7 (CH(CH₃)₃)), 31.9 (CH(CH₃)₃), 32.3 (CH(CH₃)₃), 33.9(CH(CH₃)₃), 34.1 (d, J_PC=3.4 Hz, CH(CH₃)₃, o-tBu), 34.9 (CH(CH₃)₃, p-tBu), 38.5 (CH(CH₃)₃, o-tBu), 107.2 (CH_Cp), 121.1 (CH_Aryl), 129.8 (C_q,_Aryl), 131.3 (C_q,Aryl), 145.0 (d, JPC=175.6 Hz, C_ipso), 146.9 (C_q,Aryl), 152.8 (C_q,Aryl). IR (ATR, 32 scans, cm⁻¹): =2954 (vs), 2903 (m), 2866 (m), 2281 (vw),* 1624 (vw), 1587 (w), 1479 (m), 1459 (m), 1389 (m), 1361 (s), 1240 (s), 1204 (m), 1167 (m), 1124 (m), 1024 (m), 1006 (m), 965 (w), 953 (w), 922 (w), 902 (w), 871 (m), 838 (vs), 814 (m), 800 (m), 745 (m), 669 (m), 590 (m), 563 (m), 528 (m), 496 (m), 459 (vs), 420 (m), 408 (m). *artefact or minimal amounts of a C−H-activation product.

Isolation of compound 6: To AlCl₃ (1.506 g, 11.3 mmol) in a 100 mL Schlenk flask, which was cooled to −78 °C (EtOH/CO₂), was added Et₂O (50 mL). To this LiAlH₄ (4 M in Et2O, 0.94 mL, 3.76 mmol) was added dropwise over a period of 5 min. The reaction mixture is allowed to warm to room temperature over a period of 1 h and stirred for an additional hour at that temperature. Afterwards the now cloudy suspension (LiCl) is filtered via a canula equipped with a glass filter paper and from the filtrate the volatiles are removed in vacuo. The Schlenk flask containing the sticky white residue of Cl₂AlH is brought inside the glovebox and Cp*K (4.94 g, 28.3 mmol) are added and the mixture is then suspended in Et₂O (100 mL) and stirred at room temperature for 4 h. Afterwards the volatiles were removed in vacuo and to the sticky yellow residue n-hexane (50 mL) and Et₂O (5 mL) was added. The resulting suspension was filtered using a Schlenk-frit, giving a pale yellow, clear filtrate. After concentration of the filtrate to incipient crystallization and placement in the fridge at 6 °C for 16 h a first crop of crystals was afforded, which were identified as [Cp*Al(H)(thf)] (6) (0.760 g, 2.81 mmol, 24.9 % based on AlCl₃).

¹H NMR (C₆D₆, 300.2 MHz): δ=1.01 (t, ³J_HH=6.8 Hz, OCH₂CH₂), 2.06 (s, C₅(CH₃)₅), 3.50 (t, ³J_HH=6.8 Hz, OCH₂CH₂). ¹³C{¹H} NMR (C₆D₆, 75.5 MHz): δ=11.7 (s, C₅(CH₃)₅), 24.9 (s, OCH₂CH₂), 71.3 (s, OCH₂CH₂), 115.3 (s, C₅(CH₃)₅).

The desired Cp*₂AlH then crystallized from the supernatant solution upon placement at −30 °C as a colourless crystalline solid (1.618 g, 5.42 mmol, 47.9 % based on AlCl₃). Cp*₂AlH: ¹H NMR (C₆D₆, 300 MHz): δ=1.92 (s, C₅(CH₃)₅).

We want to note that THF in the side product (most likely) stems from the synthesis of Cp*K, which was afforded from the reaction of Cp*H with KH in THF as a white solid,⁴² after drying in vacuo (10⁻³ mbar) for 4 h. Consequently, we recommend to not only decant the supernatant solution of the Cp*H/KH reaction mixture, but to also wash the crude Cp*K with n-pentane (2x).

Isolation of compound 7: To AlCl₃ (1.614 g, 12.0 mmol) in a 100 mL Schlenk flask, which was cooled to −78 °C (EtOH/CO₂), was added Et₂O (75 mL). To this LiAlH₄ (1 M in Et2O, 4 mL, 4 mmol) was added dropwise over a period of 5 min. The reaction mixture is allowed to warm to room temperature over a period of 1 h and stirred for an additional hour at that temperature. Afterwards the now cloudy suspension (LiCl) is filtered via a canula equipped with a glass filter paper and from the filtrate the volatiles are removed in vacuo. The Schlenk flask containing the sticky white residue of Cl₂AlH is brought inside the glovebox and Cp*K (5.23 g, 30 mmol) are added and the mixture is then suspended in Et₂O (100 mL) and stirred at room temperature for 4 h. Afterwards the volatiles were removed in vacuo and to the sticky yellow residue n-hexane (60 mL) was added. The resulting suspension was filtered using a Schlenk-frit, giving a pale yellow, clear filtrate. After concentration of the filtrate (ca. 30 mL) an amorphous white precipitate formed, which was isolated by decantation of the supernatant solution via a cannula. The amorphous residue was identified as a mixture of [Cp*₃Al₃H₂(μ-H)₃Cl] and [Cp*₃Al₃H(μ-H)₃Cl₂] (7, ca. 1 : 1) and X-ray quality crystals were obtained by recrystallization from Et₂O at −30 °C (0.215 mg, 0.39 mmol, 9.7 % based on AlCl₃). ¹H NMR (C₆D₆, 300.2 MHz): δ=1.94 (s, C₅(CH₃)₅), 3.36 (s (br), AlH). ¹³C{¹H} NMR (C₆D₆, 75.5 MHz): δ=10.7 (s, C₅(CH₃)₅), 114.4 (s, C₅(CH₃)₅).

The desired Cp*₂AlH then crystallized from the supernatant solution upon placement at −30 °C as a colourless crystalline solid (2.6 g, 8.71 mmol, 73 % based on AlCl₃). We therefore recommend to strictly follow the synthesis of Cp*₂AlH as outlined in Inorganic Syntheses,⁴⁰ which states that the reaction mixture of Cp*K and Cl₂AlH in Et₂O is first filtered, then dried and extracted with fresh n-hexane followed by a second filtration, instead of first removing the Et₂O directly from the reaction mixture before filtration.

Acknowledgments

He. Be. and C. H.-J. thank the Leibniz Association for funding within the scope of the Leibniz ScienceCampus Phosphorus Research Rostock (www.sciencecampus-rostock.de). We thank Till Brockmann for synthetic assistance, the technical and analytical staff at LIKAT for assistance, especially Dr. Anke Spannenberg for her support regarding X-ray analysis. We want to thank Prof. Dr. Torsten Beweries for helpful discussions. Ho. Br. wishes to acknowledge financial support by the Deutsche Forschungsgemeinschaft, DFG. Open Access funding enabled and organized by Projekt DEAL.

Conflict of interest

The authors declare no conflict of interest.

Open Research

Data Availability Statement

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