Synthesis of Genuine Germenyl Lithiums and the First Persistent Germenyl Radicals.

The first isolated genuine germenyl lithiums (R3 Si)(1-Ad)C=Ge(SiMetBu2 )(Li⋅2 L) (R3 Si=tBu2 MeSi, L=THF (1 a), or L=12-crown-4 (1 b) and R3 Si=tBuMe2 Si, L=THF (2 a), or L=12-crown-4 (2 b)), were synthesized by reaction of the corresponding acyl germanes 3 and 4, respectively, with tBu2 MeSiLi in THF at 70 °C. The novel 1 a and 2 b were characterized by NMR and UV/Vis spectroscopy, and also by X-ray crystallography (r(C=Ge)=1.865 Å for 1 a and 1.877 Å for 2 b). Nucleophilic addition reaction of 1 a with MeI and a C-H insertion reactions to the C=Ge bond of 1 a, 2 a and 2 b, are reported. Oxidation of 1 a and 2 b (toluene, 230 K) produces the first persistent germenyl radicals (R3 Si)(1-Ad)C=Ge⋅-(SiMetBu2 ) (R3 Si=tBu2 MeSi (13 a), R3 Si=tBuMe2 Si (13 b)), which were characterized by EPR spectroscopy (t1/2 ≈30 min at 230 K, g=2.029, aav (73 Ge) is 55.0G for 13 a and 60.2G for 13 b). The experimental EPR parameters and DFT calculations indicate that 13 a and 13 b have a strongly bent structure at Ge (calc. ∡(C=Ge-Si)=136.7° (13 a), 135.9° (13 b)), and that the unpaired electron has a substantial s-character.

The heavier analogs of unsaturated hydrocarbons, featuring multiple bonds between one or two heavier Group 14 elements (E), continue to attract growing attention in main‐group chemistry. Of special interest are metal (M) substituted heavy alkenes, (R2E′=ER−M), the analogues of alkenyl anions.[ , , ] Such compounds have great synthetic potential, including transformations which incorporate E=E′ and C=E bonds into organic molecules. Several disilenyl anions (disilenides) (R2Si=SiR′−M, M=Li, Na, K) were synthesized,[ , , , , , ] and their use as precursors for more extended conjugated systems,[ , ] functionalized heavier alkenes, heterocycles and E‐clusters, was explored. In contrast, only one example of a digermenide, (R2Ge=GeR′−M), and several digermenide dianions were isolated. Synthesis of hetero‐nuclear heavier alkenyl anions, consisting of two different Group 14 elements, remains challenging. Recently, the first potassium silagermenide (R2Si=GeR′‐K) was isolated by Scheschkewitz et al. and was used for the synthesis of functional silagermenes. Metallo‐silenides I and II, were synthesized and isolated in our group (Figure 1). Tokitoh et al. reported the synthesis and isolation of metallo‐germenyl (germabenzenyl) IIIa and metallo‐stannenyl (stannabenzenyl) IIIb anions, heavier analogs of phenyl anion (Figure 1), and also of a trimer of germaanthracenyl anion IV. Silanylidene anion Va and germanylidene anions Vb and VI (Figure 1), were synthesized by the groups of Roesky and Tuononen, respectively. These low‐valent Ge anions, stabilized by cyclic alkyl(amino) carbene (cAAC), have significant C=E double bond character, but the authors preferred to classify them as germanylidene anions (i.e. germylene anions). Stable germenyl anions with a genuine C=Ge double bond, having substituents with mild electronic effects, have not been yet reported.

Figure 1

Isolated heavy alkenyl anions, R2C=ER−M.

Herein, we report the synthesis of the first stable genuine germenyl anions: (R3Si)(1‐Ad)C=Ge(SiMetBu2)(Li⋅2 L) (1: R3Si=tBu2MeSi, L=THF (1  a), or L=12‐crown‐4 (1  b), and 2: R3Si=tBuMe2Si, L=THF (2  a), or L=12‐crown‐4 (2  b)), obtained via metalation of the corresponding acylgermanes. Note that both 1 and 2 have an α‐tBu2MeSi substituent, but the trans‐β‐silyl group in 1 is (tBu2MeSi), much larger than in 2 (tBuMe2Si). Germenyl lithiums 1 and 2 were isolated and characterized by NMR and UV/Vis spectroscopy, and 1  a and 2  b also by X‐ray diffraction analysis. Several reactions of 1  a, 2  a and 2  b are reported. Oxidation of 1  a or 2  b generates the corresponding first persistent (t 1/2≈30 min at 230 K) germenyl radicals (R3Si)(1‐Ad)C=Ge⋅‐(SiMetBu2), characterized by electron paramagnetic resonance (EPR) spectroscopy and density functional theory (DFT) calculations.

Reaction of colorless acyl germanes (R3Si)2GeHC(O)(1‐Ad) (3 or 4) with 2.5 equivalents of tBu2MeSiLi in THF for 2 hours at 70 °C, produces a red colored solution [Eq. (1), step a]. Replacing THF by hexane leads to a color change to orange. Germenyl lithium 1  a crystallizes from hexane at room temperature as orange crystals in 71 % yield. The structure of 1  a was determined by X‐ray diffraction analysis, identifying it as a germenyl lithium contact ion pair (CIP) (Figure 2a). Crystallization of 2  a from hexane was not successful; therefore, 2 equivalents of 12‐crown‐4 were added to the reaction mixture, and subsequent crystallization from toluene yields orange crystals of 2  b in 74 % yield [Eq. (1), step b]. X‐ray diffraction analysis shows that 2  b is a free germenyl anion, associated by charge attraction to a lithium cation solvated by two crown ether molecules (Figure 2b).

Figure 2

X‐ray structure (Olex drawing) of: a) CIP germenyl lithium 1  a, and b) free germenyl anion 2  b. Hydrogen atoms are omitted for clarity. Some structural details are given in Table 1 and full details are given in the Supporting Information.

Table 1

Selected geometrical parameters of germenyl lithiums 1  a and 2  b.

Entry	Parameter	1  a	2  b
1	r(Ge=C1), [Å]	1.865	1.877
2	r(Ge−Li), [Å]	2.611	8.240
3	r(Ge−Si1), [Å]	2.469	2.502
4	r(C1−Si2), [Å]	1.901	1.825
5	r(C1−C2), [Å]	1.549	1.576
6	Σ∡(Ge)	359.5	360.0
7	Σ∡(C1)	360.0	359.9
8	∡(C1−Ge−Si1)	121.7	119.5
9	∡(Si2−C1−Ge)	110.2	109.2
10	∡(Si2−C1−C2)	125.3	125.7
11	∡ (Si1−Ge−C1−Si2)	10.1	13.8

Table 1 compares the most important bond lengths and bond angles of 1  a and 2  b. In both, 1 and 2, the α‐(tBu2MeSi) and β‐silyl substituents (tBu2MeSi in 1  a and tBuMe2Si in 2  b), are trans. There are only minor structural differences in the germenyl (C=Ge) unit between the CIP 1  a and the free germenyl anion 2  b, although the β‐silyl substituents have different size. The C=Ge double bond of 1  a (1.865 Å) and 2  b (1.877 Å) are similar to that in germanylidene anions Vb and VI (Figure 1) (1.872 Å and 1.879 Å, respectively), but are somewhat longer than previously reported C=Ge double bonds (1.772–1.859 Å)[ , ] (e.g., 1.808 Å in 2‐Ad=Ge(SiMe2tBu)2, 1.859 Å in a norbornene endocyclic germene ). Elongation of double bonds of heavy analogues of alkenyl anions relative to their neutral analogues, by 0.03–0.07 Å, was previously reported for Tip2E=E(Tip)‐Li (E=Si, Ge; Tip=2,4,6‐iPr3C6H2)[ , ] and (tBu2MeSi)(R)C=Si(SiR′3)‐Li (R=tBuMe2Si, 1‐Ad; R′3Si=tBu2MeSi, tBuMe2Si), and is also supported by calculations (by 0.046 Å for 1  a, in comparison to the corresponding H−Ge=C germene ). The calculated C=Ge Wiberg bond index (WBI) in 1  a of 1.456 is consistent with a C=Ge double bond (1.472 in (tBu2MeSi)(1‐Ad)C=Ge(SiMetBu2)(Me), and 1.553 in Me2C=GeMe2). The calculated WBI in 2  b of 1.375 is somewhat smaller, in line with its longer C=Ge bond. Interestingly, germanylidene anions Vb and VI (Figure 1) have similar WBI of 1.396 and 1.442, respectively, although they were classified as germanylidene anions.[ , ] The geometry around the C=Ge bond in both 1  a and 2  b is essentially planar (sum of bond angles around either the Ge or C atoms is nearly 360°). Twisting around the C=Ge double bonds is small (the ∡Si1−Ge=C1−Si2 dihedral angles are 10.1° (1  a) and 13.8° (2  b)). 1  a and 2  b exhibit a strongly bent geometry at Ge; ∡C1=Ge−Si1 is 121.7° in 1  a and 119.5° in 2  b, consistent with a sp2‐type anion. The significantly longer C1−Si2 bond in 1  a (1.901 Å) vs. 2  b (1.825 Å), by 0.076 Å, results from steric repulsions in 1  a, between the bulky tBu2MeSi substituent at C1 and Li⋅2 THF, repulsions which are absent in 2  b. The C1–C2 bond in 2  b is longer than in 1  a by 0.027 A, due to σ‐Ge lone pair‐σ*(C1–C2) hyperconjugation. Germenyl lithium 1  a is the heavier analogue of the previously reported silenyl lithium IIa (Figure 1). The structures of 1  a and Ila are very similar, except for the much shorter r(C=Si) of 1.773 Å (in IIa) vs. r(C=Ge) of 1.865 Å (in 1  a), due to the smaller atomic covalent radius of Si vs. Ge (r cov(Si)=1.169 Å, r cov(Ge) 1.223 Å).

In the 13C‐NMR spectrum (in THF), the doubly‐bonded carbon atoms of 1 and 2 resonate at ≈195 ppm, shielded relative to the alkenyl carbon of 2‐Ad=Ge(SiMe2tBu)2 (208.6 ppm), and of germanylidene anions Vb and VI (212 ppm and 226 ppm, respectively).[ , ] The alkenyl carbon of the analogous silenyl anion IIa resonates at 175 ppm. Similarly, for the isostructural 2‐Ad=E(SiMe2tBu)2 (E=Si, Ge), substitution of Si by Ge causes a downfield shift of the alkenyl carbon (198 ppm for E=Si, 208.6 ppm for E=Ge). In the 29Si‐NMR spectrum, two peaks, at −6.0 and 13.2 ppm for 1, and at −6.8 and 10.2 ppm for 2, were observed, corresponding to the silyl substituents bonded to the C=Ge atoms, respectively (based on DFT calculations, see Supporting Information).

The visible spectrum of 1  a in THF (most likely a free anion) shows a wide absorption with a peak at 495 nm (ϵ=85.7 M−1 cm−1), red‐shifted by 38 nm relative to that in benzene (λ=457 nm, ϵ=403.5 M−1 cm−1). A similar shift of the absorption between THF and benzene (585 and 526 nm, respectively) was reported for silenyl lithium IIc, corresponding to free‐anion⇌CIP transformation, controlled by THF⇌benzene solvent change. The visible absorption spectra of free anion 1  a (in THF) and CIP 1  a (in benzene), are reasonably well reproduced by TD‐DFT calculations (at M062X/def2svpp//M062X/6‐311+G(d,p) (for H, C, Si, O, Li), SDD (Ge)), λ(THF)=495 nm, ϵ=125 M−1 cm−1; λ (benzene)=448 nm, ϵ=333 M−1 cm−1). The observed absorptions are attributed to a forbidden HOMO‐LUMO transition, with calculated energy gaps of 5.62 eV for CIP 1  a, and 5.41 eV for free anion 1  a, explaining the observed red shift for free anion 1  a (observed in THF) relative to CIP 1  a (observed in benzene). In both species the HOMO is the in‐plane Ge σ‐anionic lone pair orbital (s72.5% p27.5% hybridization) and the LUMO is the π*(C=Ge) orbital (Figure 3). Upon dissociation of CIP 1  a both the HOMO and the LUMO shift to higher energy, but the change is larger for the HOMO (see Supporting Information), resulting in a smaller HOMO‐LUMO gap in the free anion (Figure 3). The UV spectra of 1  a in THF and in benzene also show a wide peak at 355 nm (ϵ(THF)=6496 M−1 cm−1, ϵ(benzene)=18 636 M−1 cm−1), attributed to a forbidden π (C=Ge) (HOMO‐1)–LUMO transition (see Supporting Information).

Figure 3

Calculated frontier molecular orbitals of CIP‐1  a and free anion of 1  a.

Reaction of acyl germane 3 with only one equivalent of tBu2MeSiLi in THF at room temperature also produces 1  a, but only in 30 % yield, after 18 h. In contrast, reaction of the less bulky acyl germane 4 with one equivalent of tBu2MeSiLi (THF, room temperature), yields in a few minutes quantitatively, acyl germyl lithium/germenolate 5  a (Scheme 1, path a). The formation of 5  a was confirmed by NMR spectroscopy, and by reactions with HCl and Me2SiHCl that produce the expected substitution products (Scheme 1, path b) (for details see Supporting Information). Heating 5  a with 1.5 equivalents of tBu2MeSiLi in THF for an hour at 70 °C yields germenyl lithium 2  a (Scheme 1, path c).

Scheme 1

Synthesis of acyl germyl lithium 5  a and its reactions with electrophiles (path b) and with tBu2MeSiLi (path c).

The proposed mechanism for the formation of germenyl lithiums 1 and 2 (equation 1) is presented in Scheme 2. It is analogous to the previously described mechanism for formation of the analogous silenyl lithums II, and it is supported by DFT calculations (see Supporting Information). Acyl germyl lithium 5  a (observed) or 5  b (not observed) undergo interconversion to the isomeric germenolate 6  a or 6  b, respectively (Scheme 2, step a). Elimination of R3SiOLi from 6  a or 6  b yields germyne 7  a or 7  b, respectively (Scheme 2, step b), which isomerizes by silyl group migration to the more stable germylidene 8  a or 8  b, respectively (Scheme 2, step c). Insertion of germylidene 8  a or 8  b into the Si−Li bond of tBu2MeSiLi yields germenyl lithiums 1 or 2, depending on the identity of the β‐R3Si group (Scheme 2, step d). The fact that reaction of acylgermane 4 (or 5  a) with tBu2MeSiLi, produces 2  b, in which the tBuMe2Si group is bonded to C and the tBu2MeSi group to Ge, supports the mechanism in Scheme 2. Only the E‐isomers of 1 and 2 are produced in Equation (1), probably due to smaller steric repulsions in the insertion transition state of 8 (a,b) with tBu2MeSiLi, leading to the E‐isomer (Scheme 2, step d). Similarly, only the E‐isomer is produced in the analogous reaction producing silenyl lithiums II. Elimination of R3SiOLi from 6  a occurs at 70 °C, while for 6  b with the bulkier β‐tBu2MeSi substituent, elimination occurs at room temperature, apparently because of a more favorable release of steric strain (see Supporting Information). This is consistent with the fact that reaction of tBu2MeSiLi with 4 stops at 5  a (Scheme 1, path a), while reaction with the bulkier 3 proceeds directly to 1  a [Eq. (1), step a], and 5  b was not observed.

Scheme 2

Proposed mechanism for the formation of 1 and 2.

Several reactions of germenyl lithiums 1  a and 2  a were studied. Addition of iodomethane to 1  a in THF at room temperature yields germene 9 [Eq. 2] in 80 % yield. The 13C‐NMR spectrum of 9 shows a characteristic peak of the alkenyl carbon atom at 178.6 ppm (calculated 180 ppm ), upfield shifted in comparison with that of 1  a (195 ppm) and of 2‐Ad=Ge(SiMe2tBu)2 (208.6 ppm ). Germene 9 was also characterized by high‐resolution mass spectrometry.

Addition of tert‐butylacetylene to 1  a (in hexane or toluene) produces hydrido germene 10 in 97 % yield (Scheme 3, path a). 10 exhibits a characteristic alkenyl 13C‐NMR signal at 187 ppm (calculated 191 ppm ), slightly downfield shifted in comparison to germene 9. The alkenyl hydrogen of 10 absorbs at 6.66 ppm, in the range of known hydrido digermenes, L(H)Ge=Ge(H)L (L=C6H3‐2,6(C6H3‐2,6‐iPr2)2; N(Ar)[Si(iPr)3], Ar=2,6‐[C(H)Ph2]2‐4‐iPrC6H2 ), (5.87 and 8.21 ppm, respectively). For comparison, the resonances of the alkenyl carbon and hydrogen of the analogous hydrido silene (tBu2MeSi)(1‐Ad)C=Si(H)(SiMetBu2) (11) are at 165 ppm and 5.03 ppm, respectively, upfield shifted compared to 10. A characteristic GeH stretch at 2062 cm−1 is observed in the IR spectrum of 10, slightly higher than for GeII [ , ] and GeIV  hydrides (1900–2000 cm−1).

Scheme 3

Reactions of germenyl lithiums 1  a and 2  a with tert‐butylacetylene.

Germene 10 does not react with an additional equivalent of tert‐butylacetylene in toluene or hexane. However, changing the hydrocarbon solvent to THF promotes addition of the acetylenic C−H bond across the Ge=C bond of 10, producing germylacetylene 12  a, as a single (R*,S*)‐diastereoisomer in 98 % yield (according to NMR analysis, see Supporting Information) (Scheme 3, path b). The anti‐(R*,S*) stereochemistry of the silyl groups in 12  a was determined by X‐ray diffraction analysis (see Supporting Information). C−H insertion reactions of germene Mes2Ge=CH(CH2tBu) with several aryl‐, silyl‐ and alkyl‐substituted alkynes, were previously reported, and the reaction mechanism was elucidated. According to these studies, C−H insertion of tert‐butylacetylene to the Ge=C bond probably proceeds through a chain reaction, initiated by trace amounts of acetylenide, tBu−C≡C− (present as a byproduct from the synthesis of germene 10 (Scheme 3, path a)). Addition of acetylenide anion to Ge of the Ge=C bond of 10 yields a carbanion intermediate, which can then abstract a proton, probably from tert‐butylacetylene, producing 12  a (and the acetylenide). The observed high anti‐stereoselectivity in the addition of tert‐butylacetylene to the Ge=C bond in 10 is surprising, as calculations show that the barrier for rotation around the Ge−C bond in the carbanion intermediate is only ca. 1 kcal mol−1 (see Supporting Information). Additional studies are required to unravel the reasons for the observed stereospecific anti‐addition of tBuC≡CH to 10. Interestingly, the smaller 2  a reacts with one equivalent of tert‐butylacetylene (in toluene or hexane), but the reaction does not stop at the germene, producing directly a single diastereoisomer of germylacetylene 12  b (Scheme 3, path c), according to NMR spectroscopy (see Supporting Information), by analogy probably also with anti‐stereochemistry of the silyl groups.

Reaction of 1  a with Re(CO)5Br or fullerene (C60) in toluene at 230 K [Eq. 3] produces the EPR spectrum shown in Figure 4, exhibiting a superposition of two signals: Signal A, of the known (tBu2MeSi)3Ge⋅ radical (g=2.0229, a(73Ge)=20.0G, a(29Siβ)=8.0G), probably a byproduct obtained from (tBu2MeSi)3GeLi in the synthesis of 1  a [Eq. (1)]; Signal B, corresponds to the novel germenyl radical, 13  a. Radical 13  a is persistent at 230 K, having a half‐life of about 30 min. The g‐value of 13  a is 2.029, slightly higher than the reported range of Ge‐centered radicals, 1.9991–2.0107. The Ge signal is expected to show splitting to 10 satellite lines, due to coupling with the 73Ge nucleus (I=9/2), but due to overlap, only 7 signals (3+4) are observed, with a hyperfine coupling constant (hfcc) of a av(73Ge)=55.0G (Figure 4, Table S8 in Supporting Information). This a(73Ge) hfcc is somewhat smaller (55.0G vs 68.4–173G) in comparison to previously reported alkyl‐ and aryl‐ tri‐substituted germyl radicals having a pyramidal structure at Ge. However, a(73Ge) of 13  a is significantly larger than that of planar (tBu2MeSi)3Ge⋅ (a(73Ge)=20.0G) and planar cyclotrigermenyl radical, (GeC6H3mes2‐2,6)3⋅, (a(73Ge)=16.0G). Thus, a(73Ge) of 13  a indicates a significant s‐character in the singly occupied orbital and a bent structure at the Ge atom. Reaction of 2  b with fullerene (C60) in toluene at 230 K produces the analogous germenyl radical 13  b [Eq. (3)], characterized by EPR spectroscopy (a(73Ge)=60.2G, t 1/2≈30 min at 230 K) (Figure S3 in Supporting Information).

Figure 4

a) Experimental EPR spectrum of the reaction mixture [Eq. (3)] (230 K, toluene): A) signal of (tBu2MeSi)3Ge⋅; B) signal of germenyl radical 13  a; b) simulated EPR spectra of germenyl radical 13  a. (For details see Supporting Information.)

Additional insight is provided by quantum‐mechanical calculations. The calculated EPR hfcc for 13  a (a(73Geα)=54.3G) is in good agreement with the experimental hfcc(Table S8 in Supporting Information). 13  a is calculated to have a strongly bent structure at Ge with a ∡Si−Ge=C bond angle of 136.7° (Figure 5a, Table S8 in Supporting Information), by 15° larger than that in corresponding germenyl lithium 1  a (121.7°). The calculated C=Ge bond in 13  a is 1.835 Å, by 0.03 Å and 0.068 Å shorter than in 1  a (1.865 Å) and in free anion 1  a (calculated 1.903 Å), respectively. The C=Ge bond in 13  a is slightly twisted, by 12.4°, similarly to 1  a (10.1°). Natural bond orbital (NBO) analysis indicates that the SOMO of 13  a has s(54%) p(46%) hybridization with occupancy of 0.9 electrons, consistent with the large a(73Ge) value.

Figure 5

Calculated SOMOs of a) germenyl radical 13  a; b) silenyl radical 14. Hydrogen atoms were omitted for clarity.

Comparison of 13  a with the analogous silenyl radical (tBu2MeSi)(1‐Ad)C=Si⋅‐(SiMetBu2) (14) is of interest (Figure 5, Table S8 in Supporting Information). Silenyl radical 14 (t 1/2≈30 min at 300 K) is significantly more stable kinetically than germenyl radical 13  a (t 1/2≈ca. 30 min at 230 K). Replacing Si by Ge elongates the C=E bond by 0.094 Å, decreasing steric protection by the substituents at the radical center, explaining the lower kinetic stability of 13  a compared with 14. The ∡C=E−Si angles in 13  a and 14 are similar (136.7° and 137.1°, respectively) (Figure 5), indicating a similar s‐contribution to the orbital of the unpaired electron in these radicals (see Supporting Information, Table S8).

In conclusion, we report the synthesis, isolation, and spectroscopic characterization, including X‐ray crystallography of 1  a and 2  b, the first genuine stable germenyl lithiums. These germenyl lithiums have substituents with a mild electronic effect, in contrast to Vb and VI having a perturbing nitrogen‐substituent at the alkenyl carbon, compounds, which are better described as germanylidene anions.[ , ] The potential of 1  a and 2  a for the synthesis of new germanium compounds is demonstrated by their reactions with MeI and tert‐butylacetylene. Oxidation of 1  a or 2  b yields the first persistent germenyl radicals 13  a or 13  b, respectively (t 1/2≈30 min at 230 K), which however are kinetically significantly less stable than the analogous silenyl radical 14. EPR spectroscopy of radicals 13  a and 13  b points to significant s‐character of the Ge‐centered singly occupied orbital (a(73Ge) is 55.0G for 13  a and 60.2G for 13  b)), in agreement with their calculated strongly bent structures at Ge (∡Si−Ge=C=136.7° (13  a), 135.9° (13  b)). Reactions of germenyl lithiums 1 and 2 with other electrophiles and with small molecules such as CO and CO2, as well as the chemistry of germenes 9 and 10, are being currently explored in our group.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.