Reactivity of Heavier Vinyl Anions [(CH3)2E E ... - ACS Publications

Article pubs.acs.org/Organometallics

Reactivity of Heavier Vinyl Anions [(CH3)2EE′(CH3)]− (E, E′ = C, Si, Ge) toward Carbon Monoxide: A Computational Study Cem B. Yildiz*,† and David Scheschkewitz*,‡ †

Department of Medicinal and Aromatic Plants, University of Aksaray, 68100, Aksaray, Turkey Krupp-Chair of General and Inorganic Chemistry, Saarland University, 66125, Saarbrücken, Germany

‡

S Supporting Information *

ABSTRACT: The structures of heavier vinyl anions [(CH3)2EE′(CH3)]− (E, E′ = C, Si, Ge) and their abilities to activate carbon monoxide were investigated by DFT. Particularly, heteronuclear species exhibit a strong influence of the position of the heavier of the two group 14 elements (E or E′) with strongly differing singlet− triplet gaps as a measure of tetrylene character. The reactions of CSi and CGe (E′ = Si, Ge) with CO proceed in a concerted manner via [1 + 2] or [2 + 2] cycloadditions to a variety of potential products, whereas those of positional isomers as well as digerma and sila-germa analogues occur in a stepwise fashion. The threemembered rings derived from tetrylene-like vinyl anions (E′ = Si, Ge and E = C) are dominated by keto resonance structures, while an enol structure is observed for the product obtained from SiC. Allene-like isomers could only be optimized in case of E = Si, Ge.

■

Scheme 1. Reductive Coupling of CO by Disilenide 1a

INTRODUCTION The activation and transformation of carbon monoxide (CO) has been a subject of continuous interest for more than a century.1 As the prime example of historic and current industrial relevance, the Fischer−Tropsch process converts synthesis gas (CO + H2) into hydrocarbons under heterogeneous transition-metal catalysis at elevated temperatures.2 Furthermore, a vast variety of homogeneous transition-metal catalysts promotes the reductive cleavage of the C−O triple bond of carbon monoxide under considerably milder conditions.3 The use of low-valent main-group species as catalysts in general is being actively pursued with renewed vigor due to the potentially reduced cost and environmental impact. Apart from the long-established reactivity of organolithium species such as phenyllithium,4 a few main-group systems have been shown to react with CO stoichiometrically.5 The exposure of Braunschweig’s NHC-stabilized BB triple-bond system to CO furnishes a bis(boralactone) (NHC = N-heterocyclic carbene).5a The formation of functionalized cyclic silenes from direct carbonylation of cyclotrisilenes was developed in a collaboration of the Scheschkewitz and Sekiguchi groups.5b In a very recent study, Kimjo and co-workers reported that a boryllithium species reacts with CO to afford a 1,2diborylalkene.5c We recently accomplished the first complete reductive cleavage of the C−O triple bond of carbon monoxide in the absence of a transition-metal catalyst by treatment of lithium disilenide 1 with CO to uniformly afford product 2. Despite the rather complicated structure of 2, its formation can plausibly be explained by invoking inter alia a bis(alkyne) intermediate (Scheme 1).5d In the present computational study, we distill a general message from the behavior of heavier vinyl anions © XXXX American Chemical Society

a

Abbreviations: dme, 1,2-dimethoxyethane; R, 2,4,6-iPr3C6H3.

([(CH3)2EE′(CH3)]−, denoted as EE′ with E, E′ = C, Si, Ge) toward carbon monoxide (CO). We set out to calculate the energy profiles for the oxidative addition reactions of CO to EE′ on the basis of the previously suggested mechanism.5d As we will show, the reactions proceed via different pathways depending on the nature of the group 14 element in the anionic position E′. Three key intermediates are identified on the potential energy surfaces: three-membered rings of type 3, open-chained allenic structures of type 4, and four-membered rings of type 5. The relative preference for each intermediate is shown to depend heavily on the nature of E and E′.

■

RESULTS AND DISCUSSION Structural Properties of Heavier Vinyl Anions. Sekiguchi and Scheschkewitz independently disclosed the synthesis of the first isolable disilenides in 2004.6a,b Since then, several other disilicon analogues of vinyl anions have become available, so that the reactivity of disilenides is probably the best understood of all heavier vinyl anions,7 despite the fact Received: April 28, 2017

A

DOI: 10.1021/acs.organomet.7b00327 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics

than those of positional isomers by ΔG = 30.1 and 41.1 kcal mol−1, respectively. The increased stability goes along with a longer Si−C bond distance of CSi (1.787 Å) in comparison to SiC (1.706 Å). While the SiC bond of SiC is at the short end of previously observed Si−C double bonds in silenes (1.702− 1.778 Å),12 that of CSi, with the anionic charge localized at the silicon atom, is more comparable to species with reduced double-bond character such as the inversely polarized silenes reported by Ottosson and ourselves despite the absence of pyramidalization at the tricoordinate terminus of the double bond.13 Apparently, if the heavier atom is located in the dicoordinate position, some of the unfavorable double-bond character as in resonance form A is avoided (Chart 1). Consequently, the contribution of a silylene resonance form B should be more prominent in that case. This is consistent with the calculations on singlet−triplet splitting of CSi and SiC (Table 1). It turns out that CSi has a singlet ground state by a significantly larger margin (ΔGS−T = −29.7 kcal mol−1) in comparison to that of the positional isomer SiC (ΔGS−T = −8.7 kcal mol−1). The heavier element in the anionic position thus increases the dominance of resonance form B for CSi: the more electronegative carbon atom at the tricoordinate position effectively reduces the accumulation of negative charge at the formally anionic terminus of the double bond, as is evident in the partial (Mulliken) charges of Si (−0.301) and C (0.046). In contrast, the isomeric SiC exhibits charges of 0.194 and −0.641 for Si and C atoms, respectively. The s character of the lone pair on the E′ = Si atom of CSi (sp0.46) is, as expected, larger than that of the positional isomer SiC (sp2.11). The resulting increase in p character of C−Si bonding orbitals in CSi (sp1.42 for C and sp3.96 for Si) leads to elongation of C−Si bond length in comparison to SiC (sp1.28 for Si and sp1.90 for C). The E− E′−C1 bond angle is significantly more acute in CSi (103.9°) than in the positional isomer SiC (129.3°), which also reflects the increased p character of the C−Si bonding orbitals of CSi and can be regarded as another manifestation of the characteristic reluctance of heavier elements to hybridize. A similar dominance of the silylene resonance B is also observed in the germanium−carbon case: the Ge−C bond lengths again depend on the position of the heavier atom (GeC, 1.789 Å; CGe, 1.844 Å). Although these values are within the typical range of germenes,14 the difference is nonnegligible and indeed the Ge−C−C1 bond angle of GeC (121.5°) is considerably wider than the C−Ge−C1 bond angle of CGe (104.5°). This is in line with the NBO calculations, which predict that the s character of Ge atom on CGe (sp0.38) is determined to be greater than that of the GeC case (sp1.32). On the basis of the much smaller difference in electronegativity between silicon and germanium, it was anticipated that the position of the heavier atom in the double bond should be less important. Indeed, the calculations reveal that the singlets of SiGe and GeSi are favored over triplets only by very small margins, as manifested in very similar ΔGS−T gaps (Table 1). The calculated Si−Ge bond lengths of GeSi and SiGe

that the first heavier vinyllithium species had been a digermenide reported by Masamune et al. as early as 1989.8 The only examples of a mixed heavier vinyl anion with silicon and carbon so far were reported by the Apeloig group and bear a negative charge at the silicon atom of the SiC bond.9 Despite considerable potential in this regard, reports on the activation of unreactive small molecules by heavier vinyl anions are still rare. The activation of carbon monoxide by heavier vinyl anions is therefore suitable to provide further impetus to the field and prompt additional experimental studies beyond the already reported reaction of CO with disilenide.5d Neutral disilenes, digermenes, and silagermenes adopt a trans-bent conformation in the presence of electronegative and π-donating substituents.10 In contrast, the anionic forms with electropositive substituents (for example, Li and K) are planar according to theoretical and experimental studies,6 in accordance with the so-called CGMT model.11 We started by fully optimizing the structures of the different permutations of heavier vinyl anions (Chart 1; [Me2E Chart 1. Calculated Heavier Vinyl Anions EE′ and Contributing Resonance Structures A and B

E′Me]−, denoted as EE′ with E, E′ = C, Si, Ge) at the B3LYP/ 6-31+G(d,p) level of theory. The optimized stationary points were characterized as minima by vibrational frequency calculations. As anticipated on the basis of preceding studies,7−9 all heavier vinyl anions considered here optimize to perfectly planar structures. The E and E′ atoms of [(CH3)2E E′(CH3)]− (E, E′ = C, Si, Ge) are located in one plane with the carbon atoms of the methyl substituents (sums of dihedral angles C1−E′−E−C2 and C1−E′−E−C3 = 0 and 180.0°, respectively). The dimethyl-substituted E atoms (E = C, Si Ge) adopt perfectly planar geometries, as is evident from the sum of the bond angles around the E atoms of 360.0°. The structures SiC, GeC, and GeSi are the positional isomers of CSi, CGe, and SiGe, respectively. The isomers with the negative charge at the heavier group 14 element are energetically favored in all cases (Table 1). For example, CSi and CGe are more stable

Table 1. Relative Free Energies and Singlet−Triplet Energy Splitting of Heavier Vinyl Anions [(CH3)2EE′(CH3)]− (E, E′ = C, Si, Ge)a ΔGrel ΔGS−T a

SiC

CSi

GeC

CGe

SiGe

GeSi

GeGe

SiSi

0 −8.7

−30.1 −29.7

0 −4.4

−41.1 −28.7

−4.7 −15.3

0 −15.2

− −12.3

− −15.2

ΔG energies are given in kcal mol−1. B


Article

Organometallics

Table 2. Calculated Free Energy Barriers for TS1EE′−TS6EE′ and Overall Energies of Proposed Products 3−5 at the B3LYP/631+G(d,p) and WB97XD/6-31+G(d,p) (in Italics) Levels of Theorya SiC (a) TS1 TS2 TS3 TS4 TS5 TS6 3 4 5 a

CSi (b)

6.7 22.5 38.1 43.7

−44.5, −47.7 −34.5, −35.6 −37.8, −42.6

GeC (c)

CGe (d)

9.2 38.7 41.7 17.2, 15.6 62.3, 66.7 −0.7, −5.1 51.0, 46.5

20.1, 18.1 62.5, 67.2 −2.1, −5.8 −31.8, −35.2 −28.6, −30.6

49.9, 46.2

SiGe (e)

GeSi (f)

5.2, 6.0 11.9, 9.1 37.2, 35.2 22.7, 21.6

6.0, 13.5, 42.0, 25.7,

2.4, 0.2 −7.2, −8.2 22.3, 19.6

1.3, 0.2 −4.5, −6.1 24.7, 21.5

7.7 10.7 41.3 25.6

GeGe (g)

SiSi (h)

3.9 12.2 40.4 22.3

7.2, 9.4 10.2, 7.5 35.2, 36.6 25.5, 24.8

2.5, 1.4 −9.0, −9.8 24.6, 21.0

0.8, −2.7 −3.8, -5.7 20.9, 17.7

3.0, 14.6, 40.9, 21.7,

−1

ΔG energies are given in kcal mol .

(2.257 and 2.253 Å) are consistent with the reported value for a silagermene (2.276 Å).15 In the case of GeGe, the calculated Ge−Ge double-bond length of 2.317 Å is slightly longer than that of the neutral form of tetrakis(2,6-diethylphenyl)digermene (2.213 Å).16 Such a lengthening is typically observed for heavier vinyl anions.6 The E−E′−C1 bond angles of GeGe, GeSi, and SiGe are very similar to each other with values in the range of 99.8−101.9°. In the following, several conceptually different model pathways to explain the activation of CO by heavier vinyl anions are proposed. In all investigated cases, the activation is assumed to start with a nucleophilic attack of the anionic moieties of the heavier vinyl anions to the electrophilic carbon atom of CO. Nucleophilic Addition of Heavier Vinyl Anions to CO. The reaction mechanisms for both monosilicon analogues of vinyl anions SiC and CSi with carbon monoxide were explored at the B3LYP/6-31+G(d,p) level of theory. In order to evaluate the accuracy of the energetics of the proposed pathways, we also employed the WB97XD functional for comparison. Despite differences in the mechanisms in the cases of SiC and GeC, we found identical mechanistic scenarios for the positional isomers CSi and CGe as well as for the homonuclear systems, irrespective of the functional used (Figures S10−S17 in the Supporting Information). In none of the investigated species does the functional have a significant effect on the relative energies of 3−5 (Table 2). For this reason, only the B3LYP results are discussed in the following. In the case of E′ = C, the initial step is the nucleophilic attack of SiC to the carbon atom of CO via TS1SiC to furnish the ketene 6SiC as a key intermediate in a highly exergonic fashion (ΔG = −50.2 kcal mol−1). In TS1SiC, the nonbonding orbital HOMO of SiC and the LUMO (π*) of CO interact in the typical fashion of nucleophilic addition to carbonyl moieties to form 6SiC upon relaxation (Figure 1). Subsequently, three competitive pathways were considered for the intramolecular rearrangement of 6SiC to the heavier cyclopropene 3a, allene 4a, and cyclobutene 5a analogues (Figure 2). The relevant reaction channels are shown in Figure 3. The ring closure of ketene 6SiC to generate 3a requires an energy barrier of ΔG⧧ = +22.5 kcal mol−1 via TS2SiC, so that the overall pathway for 3a is decidedly exergonic by ΔG = −44.5 kcal mol−1 (Figure 3). The related energy of TS2SiC is determined to be lower than that of the previously proposed ring-closing step of the adduct from a disilenide reaction with CO.5d In the case of the disilicon analogue of vinyl anion, however, this step is considerably more endothermic in nature. The endocyclic C−E′ (E′ = C) bond length of 3a is calculated

Figure 1. Calculated frontier molecular orbitals of SiC, CSi, CO, TS1SiC, TS5CSi, and TS6CSi at the B3LYP/6-31+G(d,p) level of theory.

Figure 2. Proposed products from the reactions of heavier vinyl anions with CO.

to be 1.385 Å and is thus a typical CC double bond. The E′ atom is perfectly planar with ∑E′ = 360°, indicating sp2 hybridization. The formation of 1-silaallene 4a from the intermediate 6SiC necessarily requires a 1,2-methyl migration step, which is calculated to have a very high energy barrier of ΔG⧧ = +43.7 kcal mol−1 associated with TS4SiC (Table 2). Nonetheless, the overall pathway is strongly exergonic by ΔG = −34.5 kcal mol−1. The structural properties of 4a bear a reasonable resemblance to those of a reported stable 1-silaallene.17 The E− C


Article

Organometallics

3b via TS5SiC is within reach for a room-temperature process and is slightly exergonic by ΔG = −0.7 kcal mol−1 (Table 2). The nature of the E′ atom (the formally anionic position) exerts a substantial influence on the electronic structures of 3, 4, and 5 (enolate-like structures). According to Ottosson et al., an experimentally accessible 2-silenolate is best described by the keto resonance form rather than by the corresponding enol form.13b This had been attributed to the reversed polarity of the Si−C bond and the resulting marked pyramidalization of the silicon atom. In the case of 3a, the E′ = C atom is perfectly planar with the sum of angles around the E′ atom equal to 360°. The NBO analysis shows that the E′ atom of 3a is sp2.2 hybridized, inducing the presence of a classically planar CE′ bond (1.385 Å). Consequently, in the case of 3a the enolate resonance form 3 is strongly dominating (Figure 4). In analogy

Figure 3. Calculated reaction path for the proposed reactions of SiC and CSi with CO in the gas phase at the B3LYP/6-31+G(d,p) level of theory (ΔG energies in kcal mol−1).

E′ and E′C (E = Si, E′ = C) bond lengths in 4a are calculated to be 1.662 and 1.382 Å, respectively. The Si−C double bond in 4a is slightly shorter than that of the reported 1-silaallene (1.704 Å), whereas the C−C bond length in 4a is elongated. Additionally, the Si−C−C bond angle (179.0°) of 4a is slightly wider in comparison to that of the 1-silaallene (173.5°). The formation of 5a is another plausible outcome of the reaction of SiC with CO (Figure 3). After formation of the common intermediate 6SiC, a subsequent cyclization step is necessary in order to give 5a via TS3SiC. The DFT calculations indicate that the ring closure of the intermediate 6SiC leads to 5a with an energy barrier of ΔG⧧ = +38.1 kcal mol−1 (Table 2). The activation energy for this step is determined to be higher than that of the ring-closing step to 3a by ΔG = 15.6 kcal mol−1. Overall, the reaction is strongly exergonic by ΔG = −38.7 kcal mol−1. As in the three-membered 3a, the fourmembered 5a has a planar structure. The calculated C−E′ (E′ = C) bond length in 5a (1.382 Å) is almost identical with that in 3a (1.385 Å). Additionally, the WBO value of 1.729 reflects a double-bond character for C−E′ in 5a, whereas that of C−O is 0.834, indicating a single bond. As shown in Figure 3, the exchange of positions of the carbon and silicon atoms (E′ = Si) on CSi has a considerable effect on the mechanistic scenario: instead of stepwise mechanisms, now the concerted [1 + 2] and [2 + 2] cycloadditions of carbon monoxide lead to 3b and 5b, respectively. Clearly in this case, the π → π* interaction (HOMOCSi and LUMOCO) enables the reduction of CO via TS5CSi and TS6CSi (Figure 3). Unlike the formation of 5a, however, the [2 + 2] addition for 5b is endergonic by ΔG = 51.0 kcal mol−1 via TS6CSi, which is largely responsible for the appreciable barrier of ΔG⧧ = +62.3 kcal mol−1. In contrast, the energy barrier of ΔG⧧ = +17.2 kcal mol−1 for the formation of

Figure 4. Proposed resonance structures of 3, 4, and 5.

to 2-silenolate, however, the structure of 3b with E′ = Si is dominated by the resonance structure of 3′. The strong pyramidalization of the E′ atom in 3b is reflected in the sum of angles about the silicon atom, ∑E′ = 243.4°. The C−O and C−E′ bond lengths of 1.240 and 1.920 Å in 3b are only negligibly shorter than those of Ottossons’s 2-silenolate. A suitable reaction channel leading to the structure 4b (E′ = Si) could not be determined at the level of theories used herein. As for 4a, the structure is dominated by the resonance structure 4 due to the reasons described above. Expectedly, the same trend in the electronic structures of 3a,b depends on the E′ atom, as is apparent for structures 5a,b. The observed C−E′ and C−O bond lengths of 1.961 and 1.347 Å and WBO values of 1.048 and 1.212 in 5b respectively indicate major contributions of the resonance structure 5′, whereas those of positional isomer 5a appear to be best described by resonance structure 5. To conclude, the relative position of the different group 14 atoms has a drastic influence on the electronic structures of the involved species and thus dictates the formation of intermediates 3−5. The reaction of GeC with CO proceeds in a fashion similar to that of the monosilicon analogue of vinyl anion SiC (Figure 5). The formation of the key intermediate 6GeC requires an energy barrier of ΔG⧧ = +9.2 kcal mol−1 via TS1GeC but is highly exergonic by ΔG = −56.3 kcal mol−1. The subsequent rearrangements of 6GeC to 4c and 5c need relatively high energy barriers of ΔG⧧ = +41.7 and +38.7 kcal mol−1, respectively, but are also exergonic by ΔG = −31.8 and −28.6 kcal mol−1. Again, as in the cases of 4a and 5a, the favored resonance forms are 4 D


Article

Organometallics

Figure 5. Calculated reaction path for the proposed reactions of GeC and CGe with CO in the gas phase at the B3LYP/6-31+G(d,p) level of theory (ΔG energies in kcal mol−1).

Figure 6. Calculated reaction path for the proposed reactions of SiGe and GeSi with CO in the gas phase at the B3LYP/6-31+G(d,p) level of theory (ΔG energies in kcal mol−1).

and 5 for 4c and 5c, as is evident from the observed E′−E (1.760 and 1.911 Å) and E′−C (1.374 and 1.380 Å) bond lengths, respectively. However, all attempts to locate the related transition state structure for 3c at the levels of theories used herein lead directly to 6GeC. In stark contrast to the proposed mechanisms for the reactions of SiC and GeC, the allenic alcoholate 4d does not play any role in the reaction of CO with CGe (E′ = Ge). Instead, the cyclic products 3d and 5d are much more relevant. The proposed pathways of the reaction of CGe with CO are presented in Figure 5. The heavier vinyl anions CSi and CGe behave similarly toward CO due to a electron distribution comparable to that in resonance structures 3′ and 5′ ascribed to the Si or Ge atoms in the anionic position E′. The CO reduction by CGe proceeds in a concerted manner similar to that for CSi for both 3d and 5d. The intermolecular rearrangement of CGe + CO to 5d requires a very high energy barrier of ΔG⧧ = +62.5 kcal mol−1, so that the overall pathway is also decidedly endergonic by ΔG = +49.9 kcal mol−1. In contrast, the formation of 3d requires a relatively lower energy barrier (ΔG⧧ = +20.1 kcal mol−1) with an exergonic nature shown by ΔG = −2.1 kcal mol−1. The reaction mechanisms of GeGe, SiGe, and GeSi with CO have also been separately investigated and predicted to result in similar products 3e−g, 4e−g, and 5e−g (Table 2 and Figures 6 and 7). Furthermore, the reaction channels of EE′ (E, E′ = Si, Ge) are found to be in agreement with those previously obtained for the disilicon analogue SiSi. For the sake of comparison, the generation of the intermediate 5h from the reaction of SiSi ([Me2SiSiMe]−) with CO is included in the following discussion, taking into account the formation mechanism of 3h and 4h (Figure S9 in the Supporting Information), even though these results were already reported in our previous experimental study.5d The initial barriers for

Figure 7. Calculated reaction path for the proposed reactions of GeGe with CO in the gas phase at the B3LYP/6-31+G(d,p) level of theory (ΔG energies in kcal mol−1).

EE′ + CO (E, E′ = Si, Ge) are very similar irrespective of the nature of E and E′ (Table 2). The products are generated in concerted manners after forming key intermediates 6EE′ (E, E′ = Si, Ge). In accordance with SiSi, the calculations reveal that E


Article

Organometallics the lowest energy barriers after the formation of 6EE′ are determined for the keto forms of 3e−g, respectively. However, those processes are calculated to be slightly endergonic. In a similar fashion, the overall pathways for 5e−g are endergonic and the highest energy barrier of ΔG⧧ = 42.0 kcal mol−1 is associated with TS3GeSi, whereas the lowest energy barrier is related to TS3SiSi (Table 2). The endothermicities of 5e−g are relatively higher than that of 5h. Another interesting feature of the cyclic compounds 3e−h and 5e−h is that they are dominated by resonance forms 3′ and 5′ despite the presence of two heavier group 14 atoms (E, E′ = Si, Ge) (Figure 4). The compounds 4 are potential products in terms of energetics of the reactions (Table 2). The electronic structures of 4e−h are rather different from those of 4a,c. In marked contrast, their dominating resonance form is 4′ on the basis of the observed C−E′ (E′ = Si, Ge) bond lengths (2.022, 1.969, 2.007, and 1.967 Å for 4e−h), which are close to the corresponding single bonds. In addition, the strongly bent structures around the E− E′−C bond angle (100.2, 98.0, 96.7, and 98.8° for 4e−h) and the negative charges situated on the central E′ atoms (−0.246, − 0.073, − 0.218, − 0.045 for 4e−h) indicate that the resonance form 4′ resembles the electronic structure of homonuclear heavier group 14 allenes and Bertrand’s bent allenes.18 In conclusion, we predict that homonuclear systems with comparable substitution patterns will behave similarly toward CO irrespective of the nature of the group 14 elements E and E′. The mechanistic effect of the position of the anionic heavier atom (E′ = C, Si, Ge) can mainly be attributed to the structural peculiarities of nevertheless related intermediates and products. The energy barriers to arrive at TS1EE′−TS6EE′ and overall energies for 3a−g, 4a−g, and 5a−g are graphically displayed in Figure 8. Although all of the proposed products 3a−g except

be generated spontaneously. On the other hand, the lower energy barrier of TS5CSi for 3b makes this route more facile than TS6CSi in the case of CSi. It is also noteworthy that the 1,2-Me migration pathway to afford the proposed product of 1silaallene 4a has the strongest exergonic nature and the highest energy barrier for TS4SiC among the suggested routes toward 4c,e,f. A comparison of the free energy profiles of 5a−g through the proposed mechanisms provides an important result. In the case of mono heavier anionic atom analogues of heavier vinyl anions (SiC and GeC, E = Si, Ge), the optimized structures 5a,c can be produced as possible products with considerably exergonic natures, whereas those of other forms (5b,d−g) cannot be considered with the observed energetics of the reactions. Second Molecule of CO Reduction by 5a,b. The subsequent reactions of 5a,b with a second equivalent of CO were also investigated in order to better understand the effect of the E′ atom on the steps required to complete the experimental reaction of Tip2SiSiTipLi with CO reported earlier.5d The calculations indicate that CO can feasibly bind to the anionic carbon atom in 5a, initiating the reduction of the second molecule of CO, as the required energy barrier is only ΔG⧧ = +5.8 kcal mol−1 (Figure S1 in the Supporting Information). This activation energy is much lower than previously determined values for disilicon analogues.5d For this reason, we propose that the second molecule of CO reduction by a heteronuclear system with E′ = C will occur more readily than in case of the homonuclear analogue (E, E′ = Si). The subsequent rearrangements yield the key intermediate 7, which lead to the diverse products 8 and 9 (Scheme 2). In contrast to Scheme 2. Proposed Products from the Reaction of 5a with CO

the heavily exothermic disilicon process by ΔE = −109.2 kcal mol−1,5d we find that the formation process of 8 from initial structures 5a + CO is endergonic by ΔG = +2.5 kcal mol−1. The complete C−O triple-bond cleavage of one out of two CO equivalents to generate 9 is also slightly endergonic in nature (ΔG = 0.2 kcal mol−1). Prior to the intermediate 7, the formation of the intermediate 10 is reasonable (Scheme 3). In accordance with our previous study, it can be suggested that the silylene liberation (:SiMe2) from the structure 10 is also possible in the presence of donor solvent. An anionic ketenylidene compound 11 would result,

Figure 8. Graphical visualization of the energy barriers for TS1EE′− TS6EE′ and overall energies for 3a−3g, 4a−g, and 5a−g (E, E′ = C, Si, Ge).

Scheme 3. Unfavored Route for 11 and 12a

for 3c are expected to be formed from the reaction of CO with the heavier vinyl anions, the strongest exergonic nature is determined for the formation of 3a (Figure 8). Furthermore, the favored reaction route for SiC is predicted for the process of 3a with the lowest energy barrier of TS2SiC in comparison to TS3SiC and TS4SiC. Taking account of the slightly endergonic nature of the processes, the three-membered rings 3e−g cannot

ΔG energies are given in kcal mol−1 at the B3LYP/6-31+G(d,p) level of theory.

a

F


Organometallics which should undergo dimerization to yield 12. However, the calculations show that the formation of 12 is strongly endergonic by ΔG = +119.6 kcal mol−1 and thus nowhere nearly as favorable as in the case of the disilicon analogue. This is expected on the grounds of the reduced oxophilicity of carbon so that monomeric products may be a viable alternative in that case. Furthermore, we also investigated the possible reaction of CO with 5b, the positional isomer of 5a, to yield the intermediate 13 (Scheme 4). Although the addition step is

Article

■

EXPERIMENTAL SECTION

■

ASSOCIATED CONTENT

Initially, all calcualtions were performed using the Gaussian 09 suite of programs.19 In the all systems, methyl groups were used instead of bulky substituents. The calculations have been performed in the absence of solvent and in the presence of tetrahydrofuran (THF) solvent considering anionic forms of the title structures. In order to optimize the anionic structures on their potential energy surfaces in the gas phase, Becke’s three-hybrid method and the exchange functional of Lee, Yang, and Parr (B3LYP) theory was employed with the 6-31+G(d,p) basis set.20,21 Further, the solvent effects were also taken into account by using the polarizable continuum model (PCM) at the B3LYP/6-311++G(d,p) level of theory.22 The gas-phase calculations were repeated with full geometry optimizations at the newer theory level of WB97XD/6-31+G(d,p), since it includes both long-range corrected hybrid and empirical dispersion functions.23 The stationary points were characterized as minima or transition structures by vibrational frequency calculations, and all relative energies reported here are Gibbs free energies in kcal mol−1. The intrinsic reaction coordinates (IRC) were also followed to verify the energy profiles connecting each transition state to the correct local minima, by using the second-order Gonzalez−Schlegel method.24 The computed structures were visualized by using the GaussView 5.0 program.25

Scheme 4. Proposed Products from the Reaction of 5b with CO

exergonic by ΔG = −36.0 kcal mol−1, the required energy barrier of ΔG⧧ = +44.9 kcal mol−1 is found to be higher than that of the disilicon analogue.5d The calculations suggest that the formation of 2-propanone and silaketenylidene 14 by fragmentation of 13 requires an energy barrier of ΔG⧧ = +27.7 kcal mol−1 (Figure S3 in the Supporting Information). The formation of 14 is nonetheless exergonic by ΔG = −28.0 kcal mol−1 (Scheme 4). The position of the E′ atom on 5a,b entails differences even in the reduction of the second molecule of CO. As can be seen from Figures S1 and S3 in the Supporting Information, the calculated initial barrier for the process of 13 is higher than in case of the structural isomer 7. Therefore, we focused on the mechanistic scenario for the fragmentation of 7 in the following. The proposed product 14 can be produced in a concerted fashion after formation of the initial intermediate 13 from 5b and CO. However, the stepwise evolution of the key intermediate 7 is necessary for the different products 8 and 9 in the case of 5a. This may be ascribed to the fact that the E′ = C, Si atoms in 5a,b can promote the dissimilar processes, respectively. In summary, we have shown that there is a substantial influence of the nature of the formally anionic atom on the dominant resonance forms A (E′ = C) and B (E′ = Si, Ge) for the heteronuclear vinyl anions. The proposed reaction mechanisms of [(CH3)2EE′(CH3)]− (E = C and E′ = Si, Ge) with CO proceed in concerted manners to afford the threeand four-membered rings dominated by the keto forms 3′ and 5′, respectively. Conversely, those positional isomers as well as homonuclear analogues indeed occur in stepwise manners via the intermediacy of ketene species 6EE′. In this case, the openchain allenic structure types 4 appear as additional minima on the potential energy surface. The preferred resonance structures of the generated products from the proposed reaction mechanisms of heteronuclear heavier vinyl anions (E′ = C) are the enol forms, whereas those of digerma and silagerma analogues prefer the keto type products 3′−5′. On the basis of thermochemical data, we predict that the vinyl anion species with carbon in the two-coordinate position are more reactive toward CO. The same trend is observed for the reaction of a second molecule of CO with the intermediates 5a,b. Experimental validation of our predictions is encouraged.

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.organomet.7b00327. Full reaction channels, energies, and imaginary vibrational frequencies of all the calculated species (PDF) All computed molecule Cartesian coordinates (XYZ)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail for C.B.Y.: [email protected]. *E-mail for D.S.: [email protected]. ORCID

Cem B. Yildiz: 0000-0002-0424-4673 David Scheschkewitz: 0000-0001-5600-8034 Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS Financial support by the Aksaray University is gratefully acknowledged. REFERENCES

(1) (a) Macho, V.; Kralik, M.; Komora, L. Pet. Coal 1997, 39, 6−12. (b) Tolman, W. B. Activation of Small Molecules: Organometallic and Bioinorganic Perspective; Wiley: Weinheim, Germany, 2006. (2) (a) Fischer, F.; Tropsch, H. Brennst.-Chem. 1926, 7, 97−104. (b) Schulz, H. Catal. Today 2014, 228, 113−122. (c) Maitlis, P. M. J. Organomet. Chem. 2004, 689, 4366−4374. (3) (a) Christian, G.; Stranger, R.; Petrie, S.; Yates, B. F.; Cummins, C. C. Chem. - Eur. J. 2007, 13, 4264−4273. (b) Whyman, R.; Wright, A. P.; Iggo, J. A.; Heaton, B. T. J. Chem. Soc., Dalton Trans. 2002, 5, 771−777. (c) West, N. M.; Miller, A. J. M.; Labinger, J. A.; Bercaw, J. E. Coord. Chem. Rev. 2011, 255, 881−889. (d) Batsanov, A. S.; Cabeza, J. A.; Crestani, M. G.; Fructos, M. R.; García-Á lvarez, P.; Gille, M.; Lin, Z.; Marder, T. B. Angew. Chem., Int. Ed. 2016, 55, 4707−4710. (4) (a) Wittig, G. Angew. Chem. 1940, 53, 241−264. (b) Jutzi, P.; Schröder, F. W. J. Organomet. Chem. 1970, 24, 1−5. (5) (a) Braunschweig, H.; Dellermann, T.; Dewhurst, R. D.; Ewing, W. C.; Hammond, K.; Jimenez-Halla, J. O. C.; Kramer, T.; Krummenacher, I.; Mies, J.; Phukan, A. K.; Vargas, A. Nat. Chem. G


Article

Organometallics 2013, 5, 1025−1028. (b) Cowley, M. J.; Ohmori, Y.; Huch, V.; Ichinohe, M.; Sekiguchi, A.; Scheschkewitz, D. Angew. Chem., Int. Ed. 2013, 52, 13247−13250. (c) Lu, W.; Hu, H.; Li, Y.; Ganguly, R.; Kinjo, R. J. Am. Chem. Soc. 2016, 138, 6650−6661. (d) Majumdar, M.; Omlor, I.; Yildiz, C. B.; Azizoglu, A.; Huch, V.; Scheschkewitz, D. Angew. Chem., Int. Ed. 2015, 54, 8746−8750. (e) Fukazawa, A.; Dutton, J. L.; Fan, C.; Mercier, L. G.; Houghton, A. Y.; Wu, Q.; Piers, W. E.; Parvez, M. Chem. Sci. 2012, 3, 1814−1818. (f) Sajid, M.; Elmer, L. M.; Rosorius, C.; Daniliuc, C. G.; Grimme, S.; Kehr, G.; Erker, G. Angew. Chem., Int. Ed. 2013, 52, 2243−2246. (g) Dobrovetsky, R.; Stephan, D. W. J. Am. Chem. Soc. 2013, 135, 4974−4977. (6) (a) Scheschkewitz, D. Angew. Chem., Int. Ed. 2004, 43, 2965− 2967. (b) Ichinohe, M.; Sanuki, K.; Inoue, S.; Sekiguchi, A. Organometallics 2004, 23, 3088−3090. (c) Pak, C.; Rienstra-Kiracofe, J. C.; Schaefer, H. F., III J. Phys. Chem. A 2000, 104, 11232−11242. (7) (a) Inoue, S.; Ichinohe, M.; Sekiguchi, A. Chem. Lett. 2005, 34, 1564−1565. (b) Iwamoto, T.; Kobayashi, M.; Uchiyama, K.; Sasaki, S.; Nagendran, S.; Isobe, H.; Kira, M. J. Am. Chem. Soc. 2009, 131, 3156− 3157. (c) Yamaguchi, T.; Ichinohe, M.; Sekiguchi, A. New J. Chem. 2010, 34, 1544−1546. (d) Cowley, M. J.; Abersfelder, K.; White, A. J. P.; Majumdar, M.; Scheschkewitz, D. Chem. Commun. 2012, 48, 6595−6597. (8) Park, J.; Batcheller, S. A.; Masamune, S. J. Organomet. Chem. 1989, 367, 39−45. (9) (a) Bravo-Zhivotovskii, D.; Dobrovetsky, R.; Nemirovsky, D.; Molev, V.; Bendikov, M.; Molev, G.; Botoshansky, M.; Apeloig, Y. Angew. Chem., Int. Ed. 2008, 47, 4343−4345. (b) Pinchuk, D.; Mathew, J.; Kaushansky, A.; Bravo-Zhivotovskii, D.; Apeloig, D. Angew. Chem., Int. Ed. 2016, 55, 10258−10262. (10) Lee, V.; Sekiguchi, A. Organometallic Compounds of LowCoordinate Si, Ge, Sn, and Pb; Wiley: Chichester, U.K., 2010. (11) (a) Driess, M.; Grutzmacher, H. Angew. Chem., Int. Ed. Engl. 1996, 35, 828−856. (b) Grutzmacher, H.; Fassler, T. F. Chem. - Eur. J. 2000, 6, 2317−2325. (12) (a) Ottosson, H.; Eklöf, A. M. Coord. Chem. Rev. 2008, 252, 1287−1314. (b) Majumdar, M.; Huch, V.; Bejan, I.; Meltzer, A.; Scheschkewitz, D. Angew. Chem., Int. Ed. 2013, 52, 3516−3520. (c) Bejan, I.; Inoue, S.; Ichinohe, M.; Sekiguchi, A.; Scheschkewitz, D. Chem. - Eur. J. 2008, 14, 7119−7122. (d) Zborovsky, L.; Dobrovetsky, R.; Botoshansky, M.; Bravo-Zhivotovskii, D.; Apeloig, Y. J. Am. Chem. Soc. 2012, 134, 18229−18232. (13) (a) El-Sayed, I.; Guliashvili, T.; Hazell, R.; Gogoll, A.; Ottosson, H. Org. Lett. 2002, 4, 1915−1918. (b) Guliashvili, T.; El-Sayed, I.; Fischer, A.; Ottosson, H. Angew. Chem., Int. Ed. 2003, 42, 1640−1642. (c) Bejan, I.; Güclü, D.; Inoue, S.; Ichinohe, M.; Sekiguchi, A.; Scheschkewitz, D. Angew. Chem., Int. Ed. 2007, 46, 3349−3352. (14) Baines, K. M.; Stibbs, W. G. Coord. Chem. Rev. 1995, 145, 157− 200. (15) Igarashi, M.; Ichinohe, M.; Sekiguchi, A. Heteroat. Chem. 2008, 19, 649−741. (16) Snow, J. T.; Murakami, S.; Masamune, S.; Williams, D. J. Tetrahedron Lett. 1984, 25, 4191−4194. (17) Miracle, G. E.; Ball, J. L.; Powell, D. R.; West, R. J. Am. Chem. Soc. 1993, 115, 11598−11599. (18) (a) Ishida, S.; Iwamoto, T.; Kabuto, C.; Kira, M. Nature 2003, 421, 725−727. (b) Iwamoto, T.; Masuda, H.; Kabuto, C.; Kira, M. Organometallics 2005, 24, 197−199. (c) Tanaka, H.; Inoue, S.; Ichinohe, M.; Driess, M.; Sekiguchi, A. Organometallics 2011, 30, 3475−3478. (d) Dyker, C. A.; Lavallo, V.; Donnadieu, B.; Bertrand, G. Angew. Chem., Int. Ed. 2008, 47, 3206−3209. (e) Fernandez, I.; Dyker, C. A.; DeHope, A.; Donnadieu, B.; Frenking, G.; Bertrand, G. J. Am. Chem. Soc. 2009, 131, 11875−11881. (f) Dahcheh, F.; Martin, D.; Stephan, D. W.; Bertrand, G. Angew. Chem., Int. Ed. 2014, 53, 13159− 13163. (g) Li, Z.; Chen, X.; Benko, Z.; Liu, L.; Ruiz, D. A.; Peltier, J. L.; Bertrand, G.; Su, C. Y.; Grützmacher, H. Angew. Chem., Int. Ed. 2016, 55, 6018−6022. (19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.

P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision E.01; Gaussian, Inc., Wallingford, CT, 2009. (20) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (21) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (22) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995− 2001. (b) Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 106, 5151− 5158. (23) Chai, J. D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10, 6615−6620. (24) (a) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1991, 95, 5853− 5860. (b) Wiberg, K. B. Tetrahedron 1968, 24, 1083−1096. (25) Dennington, R. I. I.; Keith, T.; Millam, J.; Eppinnett, K.; Hovell, W. L.; Gilliland, R. GaussView v.5.0.9 Visualizer and Builder, Gaussian 09; Gaussian, Inc., Wallingford, CT, 2009.

H


Reactivity of Heavier Vinyl Anions [(CH3)2E E ... - ACS Publications

Recommend Documents