Coordination and Insertion: Competitive Channels for Borylene

Oct 26, 2017 - An insertion reaction of the form FH3M + BR → FH2MBHR competes against dative bonding, however, and the reaction is barrierless in se...
0 downloads 11 Views 4MB Size
Article Cite This: J. Phys. Chem. A XXXX, XXX, XXX-XXX

pubs.acs.org/JPCA

Coordination and Insertion: Competitive Channels for Borylene Reactions Kelling J. Donald,* Ezana Befekadu, and Supreeth Prasad Department of Chemistry, Gottwald Center for the Sciences, University of Richmond, Richmond, Virginia 23173, United States S Supporting Information *

ABSTRACT: Monovalent boron, free borylene species of the form B−R are notoriously unstable. Consequently, there are substantial gaps in the literature concerning the potential utility of those species in organic and inorganic synthesis either as ligands or as critical intermediates in reactions. We show that the relative stability of borylene complexes varies widely, depending on the electron donating ability of the R group. We find that borylenes can form, in the gas phase, weak sigma hole type interactions to saturated carbon centers and stronger dative bonds to tetravalent silicon and germanium. An insertion reaction of the form FH3M + BR → FH2MBHR competes against dative bonding, however, and the reaction is barrierless in several cases when M = Si and in a few cases when M = Ge. For M = C, the barriers are high enough to stabilize monovalent boron complexes. In each case, the barrier heights to M−H bond activation and BR insertion are very sensitive to the nucleophilicity of BR. We confirm, at the MP2(full) and CCSD(T) levels, a substantial preference in borylenes for the singlet over the triplet state. An account is provided at the B3LYP-D3 and MP2(full) levels for the facile insertion reaction on the singlet surface when M = Si and for the stability of FH3M·BR type complexes and the higher barriers to insertion when M = C and Ge.



INTRODUCTION Nitrogen and oxygen come readily to mind when we think of period 2 elements that are donor sites in Lewis bases. :NR3 and :Ö R2 compounds such as ammonia and water are viable bases that can form simple complexes or react in other ways with Lewis acids. In computational and experimental investigations across organic and inorganic chemistry, ammonia and water are common reagents. In addition to their electron donating tendencies, they are inexpensive, computationally less demanding than their heavier (e.g., P and S) analogues, and quite stable under a wide range of chemical and thermodynamic conditions. Bases of other period 2 elements, however, namely carbenes (:CR2) and borylenes (:BR), do not share the latter characteristic, and the stabilization and utility of such compounds remain very active areas of research. Yet carbene and borylene chemistry are at very different places in their development: carbenes have been bottled since the 1990s,1 but no persistent free borylene (B−R) is known. Boron monofluoride is isoelectronic with CO and N2, but BF and other f ree BR species that have been generated and detected experimentally2−8 are typically very reactive and are impractical as ligands or acceptors in reactions at ambient conditions.3,4,9 To be sure, simple borylene units are difficult to isolate, but borylenes have been identified in matrix isolation studies8,10,11 and have been inferred from so-called trapping reactions,12−15 even if one instance (ref 14) in which a mechanism invoking © XXXX American Chemical Society

the formation and trapping of a BR intermediate has been questioned.15 So, although B−R species are usually generated at high temperatures and tend to be unstable, the literature on monovalent boron continues to grow.9,16 Efforts have been made in recent years to locate free borylenes (B−R),15 terminal borylenes coordinated to other atomic centers (Q←B−R),17 and metal−borylene complexes (MB−R)examples of which are already known.9,16 A borylene dicarbonyl complex was reported recently,18 reaffirming the ability of borylenes to serve as Lewis acids (Q → B−R) and as “metallomimics”,16 forming coordination complexes with ligands that donate into one or both of the empty p-orbitals of boron in B−R. Borylenes stabilized by donors, including heterocyclic carbenes, have also been identified experimentally in the past few years.15,19 Additionally, borylene insertion into C−H and C−C single bonds and cycloaddition to higher order bonds have been identified experimentally4,10,12,20,21 and investigated computationally.22,23 In those insertion and addition reactions, the boron center in the BR molecule plays dual roles: as acceptor (electrophile), by its empty p-orbitals, and as σ-donor (nucleophile) by its lone pair. In previous contributions, we investigated the nature of different types of sigma hole (σ-hole) interactions,24−26 where a Received: September 28, 2017 Revised: October 25, 2017 Published: October 26, 2017 A

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

dispersion correction36 and at the CCSD(T)37 level as well. In each case, the correlation-consistent triple-ζ (cc-pVTZ) basis sets were employed for elements above iodine in the periodic table.38 A small (28-electron) core multielectron Dirac−Fock (MDF) relativistic effective core potential (without the spin− orbit part) and the corresponding cc-pVTZ basis set39 for valence electrons were employed for iodine. All of the molecular representations and images included in this article have been generated using the Gaussview graphical user interface40 and the Chemcraft program.41 The Synchronous Transit-guided Quasi-Newton−Raphson method as implemented in the Gaussian 09 software (i.e., the QST2 and QST3 options, the latter requiring a guess transition state (TS) structure) was employed to elucidate the nature of the potential energy surface (PES). In particular, the QSTn calculations aided us in our search for credible candidate transition state structures that link weakly bound or dative FH3M←BR pairs and insertion products. The scan option was employed to elucidate the nature of the potential energy surface (PES) between local minima as well. Structures obtained from the QSTn or scan data were reoptimized (as transition states) and confirmed by vibrational frequency calculations to be firstorder saddle points. A refined picture of each insertion reaction was achieved finally by calculating IRC paths using the confirmed TS structures. The number of IRC data points was effectively unrestricted; high “maxpoints” values were used so that the IRC terminated before any limit on the number of points was reached. The counterpoise correction42,43 as implemented in the G09 suite was used to correct for basis set superposition errors (BSSEs) in computed binding energies for the FH3M ← BR pair interactions. The Wiberg bond indices and other population analysis data that we report herein have been obtained from natural bond orbital (NBO) analyses on optimized geometries.

sigma hole is defined as a localized region of positive electrostatic potential induced on an atom, M, by a strongly electron withdrawing substituent, E, around the bond axis outside the E−M overlap region (e.g., by F on C in CF4).26,27 In sigma hole type interactions, an electron-rich site on a base aligns with the sigma hole on a polarized atomic center, M, to form a weakly bound pair: (i.e., E−M---Base). Now, even though “sigma hole bonds” tend to be weak electrostatic interactions, charge transfer to available (energetically and symmetrically suitable) orbitals on M can accompany such interactions, such that the overall outcome is a coordinate covalent (dative) E−M←Base interaction. Halogen bonding (e.g., Cl−I---NH3, where E = Cl and M = I) and the dative axial F−Si←NH3 bonding in SiF4·NH3 (where E = Faxial and M = Si) both involve sigma hole interactions. With or without attendant charge transfer, the location of the sigma hole on M causes sigma hole interactions to favor linear E−M−Base bond angles.28 We have shown recently that the availability of the lone pair on the base (hence the strength of the M−Base bond) is very sensitive to the identity of the substituents on the base.26,29 In the case of F4M←NR3 pairs for M = Si and Ge, for instance, the identity of R is decisive for the binding energies and the lengths of the M---N contacts. At the MP2(full) level of theory with a triple-ζ quality (cc-pVTZ) basis set, the Si−N bond distance in F4Si·N(CH3)3 is 2.078 Å, but it is 3.179 Å when R = Fover 1 Å longer!26 In the cases where the base is a free borylene species, B−R, we find herein that the identity of the substituent, R, is decisive for the nucleophilic character of the base. The electron donating power of R controls the stability of the base and determines whether B−R forms a stable acid←base pair as a local minimum or activates the M−H bond instead and forms an insertion product with a trivalent boron center. The substituent dependence of the reactivity of free borylenes, in the gas phase or otherwise, has not been systematically examined experimentally. That is undoubtedly because of the instability of borylenes, but the evidence that is available for CR2 and NR3 species suggests that strategically selected substituents can confer a substantial degree of stability upon the lone pairs of simple bases. Hence, our decision to assess the sensitivity of borylenes in that regard for substituents, R, with very different abilities as electron donating and withdrawing groups. The borylenes that have been studied to date are more stable as singlets,30 and general strategies for stabilizing or engaging them as ligands (σ-donors) or σ-acceptors are being developed.9,16,31 The implications of our results for progress in boron and group 14 chemistry is discussed. Our work is situated within the context of efforts in our group to understand the influence of sigma hole interactions on coordinate covalent bonding to group 14 compounds. Especially for M = Si, we find that the progress from dative FH3M←BR interactions to BR insertion to form FH2M−BHR may be fully suppressed or promoted depending on the identity of the substituents, R, on the monovalent boron center.



RESULTS AND DISCUSSION The stability of borylene complexes of the form FH3M←:BR has been examined for M = C, Si, and Ge and R = H, F, Cl, Br, I, CN, CH3, CH3−CCH2, CH3−n(CH3)n (for n = 1, 2, and 3), and C(CH3)3−n(C2H5)n (for n = 1, 2, and 3). In each case, we started the geometrical optimization with the basic arrangement shown in Figure 1, with M---B separations that were close to the sum of the van der Waals radii44−46 of boron and the relevant M atom (for M = C, Si, and Ge). Singlet or Triplet? In contrast to CH2, which prefers the triplet state,47 the singlet has been found to be more stable for BH.22,30,48 To confirm that result at levels of theory used in this work, and to assess the state preferences for a few other cases



COMPUTATIONAL METHODS The geometrical, harmonic vibrational frequency, and internal reaction coordinate (IRC) data reported in this work have been obtained at the MP2(full) level of theory32 using the Gaussian 09 (G09) suite of programs,33 with some additional calculations carried out at the B3LYP34,35 level in tandem with the D3

Figure 1. Representation of the starting arrangement used for the optimization of FH3M←borylene complexes considered in this work. The arrow points to where the σ-hole is induced by F on the M center and where a dative M−F bond would be formed as BR gets closer. B

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

with monodentate BR bases and (ii) the sigma hole is always located immediately opposite the F−M bond (Figure 2),26 a

with either a strong donor or a strong acceptor, R, substituent, we calculated the difference in the zero point energy (ZPE) corrected energies, ΔE(S−T) = Esinglet − Etriplet (Table 1), for R = H, F, Cl, Br, I, CN, CH3, and C(CH3)3. By the definition that we used, ΔE(S−T) is negative in Table 1 if the singlet is more stable than the triplet. Table 1. Differences in the MP2(full) and CCSD(T) Optimized Singlet and Triplet Zero Point Energy (ZPE) Corrected Energies, ΔE(S−T) = Esinglet − Etriplet, for Selected Borylene Speciesa ΔE(S−T)/eV species

MP2(full)

CCSD(T)

:BF :BCl :BBr :BI

−3.36 −2.27 −2.06 −1.79

−3.62 −2.54 −2.32 −2.04

ΔE(S−T)/eV species b

:BH :BCN :BCH3 :BC(CH3)3

MP2(full)

CCSD(T)

−1.00 −1.31 −1.50 −1.48

−1.33 −1.29 −1.76 −c

ΔE(S−T) is negative if the singlet state is preferred. All values are in electronvolt (eV) units. bAn experimental value of 10 410 cm−1 (≡ 29.76 kcal mol−1 or 1.291 eV) was reported for BH in ref 48. cFor BC(CH3)3, the triplet calculation at the CCSD(T) level failed repeatedly to converge. a

Figure 2. Representations of the electrostatic potentials (ESPs) on the 0.001 au isodensity surface of (top) sample BR bases (for R = H and F) showing the location of the boron lone pairs (red) and (bottom) MH3F molecules showing the F induced σ-holes on M (blue region). The potential in the region of the lone pair on boron (top) is more negative when R = H vs R = F. The sigma hole on M in MH3F is smallest and weakest for C (bottom). The ESP range is ±3.102 × 10−2 au.

For borylenes, the preference for the singlet state is quite clearBH has a substantial ΔE(S−T) value of −1.33 eV at the CCSD(T) level in Table 1, and the haloborylenes are even more stable as singlets. ΔE(S−T) ≡ −3.62 eV for BF, which is some 2.7 times larger than ΔE(S−T) for BH. For their carbene (CR2) analogues, the sign of ΔE(S−T) is quite sensitive to R (CH2 is a triplet, for instance, but CF2 is a singlet),47,49 but the singlet state is rather strongly preferred for the borylenes in Table 1 regardless of the electron donating or withdrawing character of R. A preference for the singlet state has been discussed for other borylenes as well in ref 22. The MP2(full) method underestimates ΔE(S−T) by 0.26− 0.33 eV in six of seven cases in Table 1, betraying a known tendency for triplet spin unrestricted MP2 energies to lead to unreliable singlet−triplet gaps.50 Indeed, the near consensus between the MP2(full) and CCSD(T) methods on ΔE(S−T) for BCN in Table 1 may be due to a balancing of errors at the MP2(full) level. Nonetheless, the qualitative trend in the ΔE(S−T) values is consistent for both methods, and the CCSD(T) and experimental values48 for BH (see Table 1) are in close agreement. In the rest of this work, therefore we will consider only the singlet surfaces for borylenes. The chemistry of borylenes, including base-stabilized borylenes, has expanded in the past decade to include careful assessments of their potential viability as ligands.16,22,23,51 We consider in the next sections the tendencies of borylene species to engage in σ-hole interactions and to form complexes with Lewis acids, specifically monofluoromethane and its Si and Ge analogues. Weak and Not-So-Weak, Acid−Base Interactions. The 42 systems that we consider in this work (three MH3F electron acceptors paired with 14 BR bases) were all optimized to minimaeach confirmed by harmonic vibrational frequency analysesat the MP2(full) level. The arrow in Figure 1 points to the center of the sigma hole on M opposite the F−M bond in MH3F. The MH3F molecules were selected as our Lewis acids (instead of MH4 or MF4) because it is known that each fluorine substituent induces a sigma hole on the M centers, and since (i) we are considering only pairwise M---BR interactions

single F substituent is sufficient for our investigation. Why, though, is fluorine our polarizing substituent of choice? Fluorine-induced sigma holes are known to be stronger (i.e., more positive) than sigma holes induced by H, Cl, Br, or I.24,26 Fluorine carries, too, the extra advantages of being smaller, less complicated electronically, and less demanding computationally than those other credible alternatives (except H). In Figure 2, the positions of the lone pair on B in BR (red region, top) and the F-induced sigma hole on M (blue region, bottom) are identified by arrows. The computationally optimized (MP2(full)) M---B separations obtained for the FH3M---BR acid−base pairs considered in this work are summarized in Table 2. Some of the systems that we studied converged to an entirely different bonding motif from the arrangement shown in Figure 1, however, and those cases are blank in Table 2. We say more about them shortly. Among the cases that yielded optimized complexes in the general pattern indicated in Figure 1, two distinct bonding regimes emerged, depending on the identity of M: (i) For M = C, the long C---B contacts all exceed 3.440 Å, which is slightly shorter than the sum of the van der Waals radii of C and B,44−46 and well beyond typical covalent C−B bond distances. (ii) For M = Si and Ge, the M−B bonds are all shorter than 3.150 Å, in line with reasonable expectations for dative Si−B and Ge−B bonds. These observations on the M−B separations in FH3M←BR complexes are consistent with patterns observed previously for M−N contacts in F4M←NH3.26 In the latter systems, only very weak electrostatic sigma hole type interactions are formed when M = C, but stronger and shorter dative bonds are achieved when M = Si and Ge. Sigma holes are obviously present on Si and Ge, as shown in Figure 2, but the sigma hole interactions only reinforce the bonding between Si C

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Binding Energies and Competing Channels for Bonding. The BSSE corrected binding energies, Ebind, that we show in Figure 3 are defined as follows:

Table 2. M−B Distances in Weak FH3C←:BR and Dative FH3M←:BR Complexesa M---B separation/Å substituents

C

Si

Ge

−F −Cl −Br −I −CN −H −C(CH3)(CH2) −CH3 −CH2CH3 −CH(CH3)2 −C(CH3)3 −C(CH3)2C2H5 −C(CH3)(C2H5)2 −C(C2H5)3

3.480 3.482 3.441 3.478 3.517 3.560 3.492 3.525 3.522 3.519 3.512 3.511 3.508 3.508

3.135 2.995 2.912 2.887 2.921

3.136 3.048 3.007 3.008 3.032 2.950 2.756 2.828 2.794 2.768 2.742

E bind = E(FH3M·BR) − {E(MH3F) + E(BR)}

They confirm for us that the Si and Ge complexes are much more strongly bound than the longer and far less covalent C systems. In fact, the magnitudes of the binding energies for the Si or Ge systems (Figure 3) are between 2 and 3 times larger than they are when M = C. Other fundamental differences in the bonding of the C species vs the Si and Ge forms are apparent in the distance and binding energy data shown in Table 2 and Figure 3. The shortest and weakest C---B contacts are obtained when R is a halogen atom (or CN). For the other M centers, the longest and weakest Si−B and Ge−B bond distances are obtained when R is a halogen atom or CN. The latter trend is in line with general expectations of covalent chemical bondsstronger bonds between two specific atomic centers tend to be (though there are exceptions) shorter. For FH3Si·BR and FH3Ge·BR, the migration of the electron density in toward the B−R bonding region when R is electron withdrawing weakens the Si←B and Ge←B donation that is largely responsible for the dative bond. This effect has been observed already for M←N contacts26 and explains why the Si←B and Ge←B bonds are weaker and longer for the haloborylene complexes. In the alternative regime of subtle electrostatic and dispersive interactions, however, the long C---B contacts may be explained by the contraction of the boron lone pair due to the electron withdrawing power of the halides. The contraction of the electron density on B allows the haloborylenes to get a bit closer to the C center, still with no significant C←B charge transfer (Table 3). The exceptional sensitivity of these weak C---B contacts to changes in the immediate chemical environment is evident, in vacillations in the C---B separations as the halides get larger in Table 2. We do not explore this local trends in Table 2 but notice thatin a window of only 0.04 Åthe C---B distance is longest among the halides when R = Cl and shortest when R = Br. And, as we show in the Supporting Information, this anomaly is also observed at the B3LYP-D3 level. Guidance from Orbital Contributions. An important consideration in assessing the bonding of monohaloborylenes as bases is the extent of the π-donation from filled halide porbitals to the formally empty 2px and 2py orbitals on B. This interaction has several secondary effects on borylene bonding since it stabilizes the singlet state for the borylene22 and may compete with (E→B) back-donation in “E·BR” fragments if E has suitable occupied valence p- or d-orbitals. In Table 4, we show the Wiberg bond indices for bonds between B and the specific atom, A, to which it is bonded in BR. The net charge on B and the extent of the involvement of the boron σ(s−p) hybrid and p-orbitals in the B−A bonding are included. The data were obtained from NBO analyses; the orbital involvement is quantified as a percentage derived from the square of the polarization coefficient, cB, for B orbital(s) in the B−A bonds; for each B−A bond, cB2 + cA2 = 1. The NBO analysis uncovered π contribution to the bonding with B only for the monohalides. The cases for R = CN or CH3−CCH2 in which the B is bonded to an unsaturated C center showed no significant π involvement. The data summarized in Table 4 will help us again shortly.

a

The data were obtained by optimizing the complexes at the MP2(full) level.

or Ge and B; those bonds include significant amounts of M←B charge transfer. SiH3F and GeH3F have lower energy antibonding orbitals into which the lone pair electrons on B can be donated. The separations between the highest occupied and lowest unoccupied molecular orbitals (the HOMO−LUMO gaps, for which computed data are provided in the Supporting Information) decrease somewhat as well from 18.3 eV for M = C to 16.7 and 16.1 eV for M = Si and Ge. As we show in Table 3, some evidence for the noncovalent character of the Table 3. Wiberg Bond Indices for M---B Contacts in FH3M←:BR Complexes Obtained from a Natural Bond Orbital (NBO) Analysis on Geometries Optimized at the MP2(full) Level M---B bond indices substituents

C

Si

Ge

−F −Cl −Br −I −CN −H −C(CH3)(CH2) −CH3 −CH2CH3 −CH(CH3)2 −C(CH3)3 −C(CH3)2C2H5 −C(CH3)(C2H5)2 −C(C2H5)3

0.010 0.008 0.009 0.008 0.008 0.010 0.011 0.010 0.010 0.010 0.011 0.011 0.011 0.011

0.09 0.12 0.15 0.16 0.15

0.09 0.11 0.12 0.12 0.12 0.17 0.27 0.23 0.25 0.27 0.30

(1)

C←B bonds, compared to the Si←B and Ge←B, is provided by Wiberg bond indices for those M−B contacts in the optimized dative type structures. The indices are very low (to two decimal places, only 0.01 in each case) when M = C, but they jump by an order of magnitude when M = Si or Ge (Table 3); they are 0.09 when R = F and increase to 0.25−0.30 with the more electron donating alkyl substituents. D

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 3. Basis set superposition error (BSSE) corrected binding energies, ΔEbind = E(FH3M←:BR)BSSE − {E(MH3F) + E(BR)}, in kcal mol−1, for the optimized FH3M←:BR complexes relative to the isolated acid and base units. “−C” abbreviates the CH2Ċ −CH3 substituent.

directions in which the reactions proceeded for the acid−base pairs considered in this work. Figure 3 shows clearly that only the bases that are weakly bound to their CH3F and GeH3F counterparts in Figure 3, namely BF, BCl, BBr, BI, and BCN, actually persist in coordinate covalent bonding to SiH3F. BH and all of the alkylborylenes form weak electrostatic or dative complexes when M = C, and in most cases for Ge, but, when M = Si, they all collapse directly to the insertion product FH2Si−BHR, which has a trivalent B center and a simple covalent Si−B bond. Computed M−B distances for covalent FH2M−BHR compounds are included in the Supporting Information. For completeness, we directly optimized FH2M−BHR molecules for M = C, Si, and Ge for all 14 R groups. For the cases where the complexes did not collapse directly to that covalent molecule, we started the optimizations with guess structures in the geometry of Figure 4c with estimated covalent bond distances. Decline and Fall of Barriers. As we pointed out above, the borylene insertion into M−H bonds was not observed for M = C and occurred only with the largest alkylborylenes for Ge. But M−H activation and borylene insertion was the rule rather than the exception for SiH3F + BR. We find that the borylenes that form the strongest C---B and Ge−B contacts, which we expected to form strong dative bonds to Si as well, achieved direct Si−H activation, bypassing dative bonding completely. In more quantitative terms, there seems to exist a certain binding energy cutoff for coordination complexes beyond which M−H activation is stabilized relative to simple coordination such that the latter is not a favored arrangement at all and ceases to be a minimum on the potential energy surface, even at very low temperatures. For M = Si, the relevant |Ebind| cutoff is remarkably low; assuming that the Si curve would continue in Figure 3 to track closely with the Ge curve, the cutoff for Si is within the vicinity of 4−5 kcal mol−1 where the |Ebind| falls for BH for Ge. For Ge the cutoff appears much later on in Figure 3. The tendency toward BR insertion increases as R becomes more electron donating (on the right beyond R = CN in Figure 3) and as the lone pair on B, as a consequence, becomes more available. In language adopted from carbene chemistry,54−56 the monohaloborylenes are more nucleophilic in terms of interactions to its formally “vacant” boron p-orbitals due to π contributions to those p-orbitals from halide substituents, but

Table 4. MP2(full) NBO Data for Borylenes: (i) Charge on Boron (q), (ii) Wiberg Bond Index (WBI) for Bond between B and A,a and (iii) Contributions of the B Valence (s−p Hybrid, px, and py) Orbitals to σ and Any π Bonding to Ab R

q/e

WBI

σ(s−p)/%

π(px)/%

π(py)/%

F Cl Br I CN H CH3−CCH2 CH3 CH2CH3 CH(CH3)2 C(CH3)3 C(CH3)2C2H5 C(CH3)(C2H5)2 C(C2H5)3

0.54 0.26 0.20 0.11 0.46 0.36 0.42 0.48 0.49 0.52 0.56 0.56 0.57 0.59

0.76 1.13 1.21 1.29 0.75 0.84 0.75 0.75 0.72 0.68 0.62 0.60 0.59 0.57

11 22 24 26 23 32 23 23 23 22 20 20 19 18

5 7 8 8

5 7 8 8

a

Here A is the atom in R to which B is bonded directly; A is shown in italics if R is polyatomic. bThe squares of the B valence orbital coefficients are in percentages.

Insertion vs Coordination. Despite the stark differences, there are some features that unite the data for M = C and for M = Si and Ge that we outlined above. Boron monofluoride has the lowest binding energy in each case (among the lowest, strictly speaking, for M = C, since the CN value is comparable (Figure 3)). For each M, the binding energies obtained for the alkyl R groups are consistently more negative than those for the halides. For M = Ge, the differences are quite substantial in Figure 3, with |ΔEbind| > 7 kcal mol−1 in some caseseven if the actual binding energies are rather low compared to other coordinate covalent bonds, which can be tens of kcal mol−1.52,53 As we mentioned above, however, none of the BR bases with organic R substituents converged to a dative complex for Si, and neither did the Ge systems (Figure 3) with the largest alkyl substituents: (C(CH3)3−n(C2H5)n). For those Si and Ge systems, the optimizations of starting geometries in the arrangement depicted in Figure 1 led directly to M−H bond activation and the insertion product FH2M−BHR with a trivalent boron center. Figure 4a−c shows stick diagrams of the E

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 4. Products obtained by optimizing FH3M + BR acid−base pairs, starting with M---B separations ≈ sum of the van der Waals radii. Outcomes: (a) weak van der Waals type complex for M = C, (b) dative covalent complex formed for M = Si and Ge with more electron withdrawing R groups, and (c) the B−R insertion product formed by all other Si and Ge systems.

they are worse σ-donors (nucleophiles) for dative interactions or, ultimately, insertion. The reverse is generally true for the alkyl substituents. Even if, as discussed in ref 22, hyperconjugation by alkyl substituents can also influence p-orbital occupation on the B center, the alkyl substituents are good σdonors. Consequently, the alkylborylenes coordinate more strongly to M as σ-donors in dative bonds. Tetravalent C will always be a poor bonding partner for bases, but as the M---B interaction becomes stronger (for M = Si and Ge), and the M− B contacts in the minimum-energy structures shrink, the prospect for M−H activation is greatly enhanced. The short M−B contact facilitates the H transfer from M to B by the donation of the B lone pair into a low-energy orbital on MH3F and the incipient donation from the M−H bond into an empty p-orbital on B. The resulting insertion product has the double thermodynamic advantage of oxidizing the monovalent boron to the preferred trivalent form and sacrificing one covalent (M−H) bond for two new ones: an M−B and a B−H bond. Reaction Paths. To understand better the tendency toward borylene insertion, we carried out IRC path calculations at the MP2(full) level for all five of the cases (R = F, Cl, Br, I, and CN) for which FH3M←BR type complexes were obtained for M = C, Si, and Ge. The IRC calculations were preceded by relaxed potential energy surface scans that gave us a general picture of the energy changes leading to insertions and allowed us to obtain reasonable guess structures for transitions states. The scans were carried out starting with the MH3F and BR molecules

separated by a distance equal to 2.5 Å plus the previously computed M−B distances in the insertion products (FH2M− BHR). We reduced that long distance incrementally during the scan down to the covalent M−B distance. The guess transition state structures obtained from those scans or from intuition were employed in QSTn calculations, and the transition state structure candidates obtained in that way were reoptimized to confirm that they were first-order saddle points. IRC path calculations were conducted using those confirmed structures. The transition state structures that were found to link the coordinated CF3H and BR systems (Figure 4a) to the insertion product (Figure 4c) are shown in Figure 5. They are quite similar to structures identified in refs 22 and 23 for borylene insertion into the CH bonds of methane. A competing transition state structure that we found for some of the haloborylenes, BX, is comparable as well to an arrangement identified in ref 22 for CH4 + BR (see Figure 6). In our case, however, we found that the IRC paths for that TS structure led to the insertion−elimination product (H2CB−R) shown in Figure 6. The optimized transition state and product structures of that form and the relevant IRC graphs are shown in the Supporting Information for X = Cl and Br. That dehydrohalogenation process depicted in Figure 6 is to be examined further in a separate discussion, but it is noteworthy that a transition state structure like that shown in Figure 6 leads to CH3−BHR22 for CH4 + BR. Substituting for F and the associated polarization of the C center is evidently F

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

quite decisive for the outcomes of BR insertion into C−H bonds and possible alkylideneborane formation. For M = Si and Ge, the transition states that we found are close geometrically to the latter form for M = C, and examples are shown in Figure 7 for the fluoride, iodide, and cyano bases. The actual Cartesian coordinates for the transition state structures and pictures for the TS structures for the BCl and BBr bases as well are provided in the Supporting Information. For the five cases (R = F, Cl, Br, I, and CN) that converged to dative type minimum for each M, we present in Figure 8 the computed (MP2(full)) IRC path that we obtained using the strategy described above in order to link the weak or datively bound pairs to the insertion products. The transition structures are unique for M = C for the BR insertion (cf. Figures 5 and 7), and the associated energy barriers are high compared to the Si and Ge cases (Figure 7). The differences in the transition states suggest a fundamental difference in the mechanism of the reactions for M = C vs M = Si and Ge, and that difference is summarized in Figure 9. The insertion occurs for C, as shown in Figure 9a, by a donation of the B lone pair to the terminal H, and a donation of the electrons of the activated C−H bond to an empty B porbital. For the polarized and more electropositive Si and Ge centers, however, the B lone pair is donated to the M center, leading to M−H activation and H migration to the B center (Figure 9b). Quantifying the Barriers. The IRC paths do not necessarily terminate, on either side of Figure 8, to fully optimized geometries, so we list in Table 5 the differences in the free energies of the optimized FH2M−BHR and the optimized weakly bound C or dative Si and Ge structures, ΔG, for the five R groups. The free energy differences between those minima and the relevant transition state structures, ΔG‡, have been computed as well. The actual energy and free energy data are given in the Supporting Information. In each case, the energy barrier to the TS going from the weak or dative structures on the right in Figure 8 (i.e., leading

Figure 5. Transition state structures linking the sigma hole or van der Waals and covalent minima on the potential energy surface of MH3F + BR, for M = C and R = F, Cl, Br, I, and CN.

Figure 6. Alternative insertion pathway observed for R = Cl and Br in IRC calculations for M = C.

Figure 7. Transition state structures linking the dative and covalent minima on the potential energy surface of MH3F + BR for M = Si and Ge, and R = F, I, and CN. The cases shown here are qualitatively identical to those for R = Cl and Br (see Figure S1 in the Supporting Information). G

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

nucleophilic (i.e., as the lone pair becomes more available for σdonation) going from R = F to I, the barriers fallslowly for C, rapidly for Ge, and even more precipitously for Si (Table 5). In general, the insertion product (left in Figure 8) becomes increasingly stable relative to the TS (ΔG‡(TS−Cov) in Table 5) as the dative structures on the right in Figure 8 become less stable (ΔG‡(TS−Dat) in Table 5) for M = Si and Ge. This result is in line with the Bell−Evans−Polanyi principle,57 though it does not hold up as well for M = C, where ΔG‡(TS− Cov) for the halides in Table 5 changes very slowly going from F to I, but in the same direction as ΔG‡(TS−Dat). Considering ΔG(Dat−Cov) and ΔG‡(TS−Dat) only in Table 5 for all three M centers, the results show clearly that as the R groups become more electron donating going from R = F to I, the insertion reaction becomes increasingly favored both thermodynamically and kinetically. And the situation is even more favorable for BCN insertion; the reactions are almost barrierless for Si and Ge! Indeed, as we will confirm shortly, the barriers disappear eventually as the R substituent becomes even more electron donating, which accounts in fact for our failure to find dative minima in most cases for Si and in some cases for Gethe missing data in Figure 3. Barrierless Convergence. To study the formation of FH2Si−BHR for all 14 bases, some of which are rather larger alkyl substituents, we carried out a series of IRC path calculations following the protocol mentioned above, using in this case, the B3LYP-D3 method. The outcomes are summarized in Figure 10, again using a common scale on the y-axis for easy comparison. The B3LYP-D3 data are very close qualitatively to those obtained at the MP2(full) level. The exceptions: BCN + SiFH3 goes directly to the insertion product, and for M = Ge, only B− C(CH3)(C2H5)2 failed to give a dative type minimum at the B3LYP-D3 level. As at the MP2(full) level, however, most of the Si systems (all except the halides) collapsed to FH2M− BHR. For the C, Si, and Ge systems with dative minima, we started the IRC path calculations from B3LYP-D3 optimized transition state structures. For the Si and Ge systems that had no dative minimum we carried out “downhill” IRC path calculations starting from long M---B separations equal to ∼3.5 Å. In a few cases where the potential energy surface was very flat, we accepted slightly shorter separations for the start of the IRC calculations. The most impressive observation from the IRCs for the latter systems is the barrierlessness of the insertions (Figure 10). That observation helps us to understand why no dative species was located in those cases. If the substituent is sufficiently electron donating and M is sufficiently electrophilic, the barrier to insertion shrinks relative to the energy of the dative complex andas we observe in most cases for M = Si the barrier disappears completely. Interpretations and Implications. The computed energy and free energy differences for the optimized weak and dative type pairs and insertion products relative to the transition state structures for the cases in Figure 10 are shown in Table 6. The actual coordinates for optimized structures, including the transition state structures (at the B3LYP-D3 level) where they exist, are provided in the Supporting Information. Unlike the Si cases for R = CN and H, barriers are actually observed (though they are very low; see the ΔG‡(TS−Dat) values in Table 6) when M = Ge. The same basic pattern (low barriers for Ge and none for Si) is observed in Table 6 for the alkyl groups. For R = C(CH3)(C2H5)2 we located no dative

Figure 8. IRC paths obtained at the MP2(full) level of theory linking the insertion product FH2M−BHR (left) to the more weakly bound FH3M---BR pairs (right) for M = C, Si, and Ge and R = F, Cl, Br, I, and CN. For comparison, the three graphs are presented on the same vertical scale. For M = C, the IRCs terminated in several cases at weak dipole−dipole or van der Waals type complexes as illustrated on the right in the C graph.

up to C−H activation) is highest for M = C and lowest for Si, even if the Ge barriers are only slightly higher than those for Si. For R = CN, the journey to the covalent structure (right to left in Figure 8) is nearly barrierless for both Si (ΔG = 0.9 kcal mol−1) and Ge (1.4 kcal mol−1) but quite high for C (15.4 kcal mol−1) (see Table 5). The largest barriers for each M is obtained when R = F, with ΔG = 57.1, 16.4, and 21.7 kcal mol−1 respectively for M = C, Si, and Ge in Table 5. As R becomes less electron withdrawing and the base becomes more H

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 9. Mechanistic differences between the BR insertion for M = C and for M = Si and Ge. For clarity we exclude one of empty p-orbitals on B.

Table 5. MP2(full) Free Energy Changes Going from Weak or Dative (Dat) Type Complexes to FH2MBHR (ΔGDat−Cov) and from the Covalent (Cov) and Dative Type Structures to the Transition Structure (TS)a ΔG‡(TS−Cov)

ΔG(Dat−Cov)

a

ΔG‡(TS−Dat)

R

C

Si

Ge

C

Si

Ge

C

Si

Ge

F Cl Br I CN

46.9 62.3 65.6 67.8 83.9

37.1 54.3 58.1 61.9 76.0

40.0 56.8 60.5 64.1 78.5

104.0 101.0 100.3 99.0 99.4

53.5 59.4 61.3 63.7 76.9

61.7 66.3 67.7 69.1 79.9

57.1 38.6 34.7 31.2 15.4

16.4 5.1 3.2 1.8 0.9

21.7 9.5 7.2 5.0 1.4

Sample equation: ΔG‡(TS−Cov) = GTS − GCov. All values are in kcal mol−1 units.

minimum for M = Ge. It is odd that BC(CH3)(C2H5)2 would form no dative minimum for M = Ge here since we found minima for the cases with one and three ethyl substituents. As we point out in the footnote to Table 6, however, the latter two cases have dative minima in which the F−Ge−B bond angles are slightly bent, so, possibly due to steric bulk, the simple linear F−Ge−B arrangement is not a minimum for Ge with any of the ethyl-substituted bases. We compare in Figure 11 the IRC graphs shown in Figure 10 for two sample cases (R = CH3, where there is no barrier for Si and a very small barrier for Ge, and R = F with the largest barriers for each M). We wanted you to see, for a direct comparison, a graphical representation of the effects of changing M on the nature of the potential energy surfaces. The insertion barrier is very small for GeFH3 + BCH3, so we have added an inset showing the curves in the region of the origin. The starting point in the downward IRC path for SiFH3 + BCH3 at the origin Figure 11 is arbitrary (a 3.5 Å Si---B separation in the pattern of Figure 1) since there is no TS in that case. Eventually, the PES becomes exceedingly flat, too, for M = Ge (when R is a large alkyl group; see ΔG‡(TS−Dat) in Table 6), as the base gets more electron donating and the boron lone pair becomes more diffuse. But why is the base insertion so facile for Si? We mentioned above the dependence of the barriers on the σ-donating and π-accepting ability of the bases (the former being enhanced substantially by alkyl, while both are diminished by halide R groups; see Table 6 and Figure 10). The electron accepting ability of B is inhibited somewhat by πdonations to B p-orbitals by halogen atoms (Table 4), and that helps to increase insertion barriers for the haloborylenes (Figures 8 and 10). The HOMO−LUMO gaps provided in the Supporting Information were computed at the CCSD(T) and MP2(full) levels. The B3LYP data reported in ref 22 (and B3LYP-D3 in the Supporting Information as well) under-

estimate the gaps significantly relative to the CCSD(T) and MP2(full) values, but they agree qualitatively. The trend in the HOMO−LUMO gaps, which vary in terms of R as F > Cl > Br > I, CN, and alkyl groups (see the Supporting Information), helps to account as well for the significant barriers for the haloborylenes. As we show in Figure 9, the insertion process requires BR to function as both an electron (σ) donor and as an acceptor using its empty p-orbitals (Figure 9), and the latter is facilitated by the availability of low energy LUMOs. Regardless of the electron donating or accepting abilities of the bases, however, Si has in each case (for R = F, Cl, Br, I, and CN) the lowest barriers to insertion, as we illustrate specifically for BF in Figure 11. The exceptional behavior for silicon reflects the generally anomalous character of Si in group 14, in particular its more electropositive (and electrophilic) character compared to C and arguably Ge, the smaller HOMO−LUMO gap mentioned above, and larger atomic size compared to C. On the Pauling scale, the electronegativity, χP, of Si is lower than those of both C and Ge, with χPSi (1.90) slightly smaller than χPGe (2.01). That order is reversed on the basic Mulliken (absolute) electronegativity scale (with χMSi > χMGe for the isolated atoms), but the more relevant tetrahedral valence state Mulliken electronegativities58 are in full agreement with the Pauling scale that χC > χGe > χSi. On the corresponding valence state chemical hardness scale,58 the Si atom is also the softest (least hard) of the three atoms. And given that softness varies directly (though not linearly) with polarizability,59 it is not surprising that the size and strength of the sigma hole on M increase substantially going from C to Si and are comparable for Si and Ge (Figure 2). Indeed, considering the whole atom beyond the localized sigma hole, the MP2(full) NBO charge is slightly negative for C (−0.071|e|) and more positive for Si (+1.37|e|) than it is for Ge (+1.20|e|). I

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 10. B3LYP-D3 IRC paths for the interactions of BR bases and MH3F for M = C, Si, and Ge. For convenience, the curves are shown in three separate sets; the relevant R groups are indicated on each graph, and the same scale is used on the y-axes. *For SiH3F + BCN (top, center) the IRC path terminated initially at a dative second-order saddle point. We continued the optimization with a new input obtained by following one of two degenerate bending vibrations with imaginary frequencies. **For SiH3F + BH (middle), the usual product (FH2Si−BH2) transformed in the last few IRC steps to a product with one of the H atoms bonded to Si moved into a bridging position above the Si−B bond.

The availability of low-energy orbitals on Si and Ge (relative to C) makes it possible for stable dative interactions (supported by strong sigma holes induced by F on M) to be formed with

BR. But, as the lone pair on B becomes more diffuse and available for bonding, and as M becomes more electropositive, the interactions between the boron center and M (Figure 9) J

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Table 6. B3LYP-D3 Free Energy Differences (ΔG) for the Covalent (Cov), Dative (Dat) and Transition State Structures (TS)a ΔG‡(TS−Cov)

ΔG(Dat−Cov)

ΔG‡(TS−Dat)

R

C

Si

Ge

C

Si

Ge

C

Si

Ge

F Cl Br I CN H CH3−CCH2 CH3 CH2CH3 CH(CH3)2 C(CH3)3 C(CH3)2C2H5 C(CH3)(C2H5)2 C(C2H5)3

44.8 58.7 60.7 62.4 80.5 81.1 70.4 67.2 67.7 67.8 67.5 66.5 66.6 67.0

36.0 51.6 54.3 57.0

38.7 53.7 56.2 58.7 75.4 74.6 65.2 61.7 63.0 61.7 61.6 62.0b

97.1 95.9 95.7 94.8 93.7 94.4 94.6 93.4 92.7 93.5 95.5 95.1 95.8 96.5

52.4 57.5 58.9 60.2

61.6 65.1 65.7 65.9 77.0 75.5 69.0 65.4 66.3 65.1 64.6 64.0b

52.4 37.1 35.0 32.5 13.2 13.3 24.1 26.2 25.0 25.8 28.0 28.5 29.1 29.4

16.4 5.9 4.6 3.2

22.9 11.4 9.5 7.2 1.6 0.9 3.8 3.6 3.4 3.4 3.0 2.0b

61.8b



63.2b −1

1.5b

Sample equation: ΔG (TS−Cov) = GTS − GCov. All values are in kcal mol units. The complexes formed by these two systems with shorter dative bonds for M = Ge had F−M−B bond angles that were tilted noticeably from the 180° alignment that is typical for σ-hole supported interactions. They were 171.6° and 166.7° for C(CH3)2C2H5 and C(C2H5)3, respectively. We located no minimum for C(CH3)(C2H5)2. a

b

Figure 11. IRC paths for the reaction of borylenes (BR, for R = CH3 and F) with MH3F (for M = C, Si, and Ge) (right in graphs) to form FH2MBHR (left in graphs). FH3Si + BCH3 is barrierless.

for FH2Si−H may be comparable to or even larger than that for FH2Ge−H. Yet GeH3F is somewhat more resistant to insertion than SiH3F, so the ease of the insertion in the latter case is probably not accounted for by any particular weakness of the Si−H bond, but by a somewhat more favorable Si---B interaction. Barrierless insertion is promoted by (i) increasing the electrophilic character of the M center and (ii) increasing the σ-donating character of the base, which depends on the electron donating ability of the R group. Delineating clear-cut general guidelines for predicting new barrierless insertions for other M centers is difficult, however, since that will depend partly on the particular low-energy symmetry appropriate orbitals that are available on M. And, as we saw above (Figures 6 and for FH2Si−BHR in Figure 10), there are competing reaction directions as well. For the various Ge systems where the barriers are present but very low, the likelihood of Ge−H activation and BR insertion is expected to be very sensitive to the actual thermodynamic conditions. We have not been able to find any experimental or previous computational investigation of borylene insertion reactions with Si or Ge halohydrides. C−H bond activation is hardest to achieve, but examples of bond activation for organic species that we mentioned above are encouraging for the future

and one of the M−H bonds, too, if M is small enough, become stronger. For Si, which is a bit smaller than Ge (especially as a formally 4+ cation)45,46,60 but is larger and softer than C, and more electropositive, the M−H interaction (i) is more stabilizing and (ii) happens early (Table 6) as R becomes more electron donating. Consequently, BR insertion is practical, and even facile, when M = Si, except in cases where BR is a poor σ-donor and a compromised Lewis acid due to B←halide π-donation. Ge is similar in many respects to Si, but the collapse in the barrier to insertion is postponed for Ge to the most nucleophilic cases where the barriers (ΔG‡(TS−Dat) in Table 6) are miniscule compared to those for carbon. Overall, the structural and electronic evidence suggests that the lower barriers to BR insertion into Si−H bonds relative to C−H and Ge−H bonds are achieved by a conspiracy of subtle physicochemical properties and not by any single aspect of bonding in SiH3F. We found M−H bond dissociation energies for FH2M−H only for M = C (423.8 ± 4.2 kJ mol−1). If the H3M−H bond dissociation energies are reliable indicators, however (with D°298(H3M−H) = 383.7 ± 2.1 and 348.9 ± 8.4 kJ mol−1 for M = Si and Ge, respectively, and 439.3 ± 0.4 kJ mol−1 for C),61 D° K

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A Notes

of borylene mediated chemistry in organic synthesis under appropriate conditions.

The authors declare no competing financial interest.





SUMMARY AND OUTLOOK The nature of dative and sigma hole type interactions by monovalent boron (B−R) as a base is investigated. Apart from those weak modes of bonding, however, which leave the original molecular units essentially intact, we find that the electron deficiency of boron in B−R opens up alternative channels for bonding. For Lewis acids with the general formula FH3M, where M = C, Si, and Ge, BR insertion into one of the M−H bonds to form FH2M−BHR is barrierless in most cases for M = Si and for Ge, too, if BR is sufficiently σ-donating. For M = C, the barriers are relatively high. For the monohaloborylenes, BR is weakened as both a Lewis acid (due to π contributions from the halogen atoms to the valence p-orbitals of B; Table 4) and a Lewis base (due to polarization by the halides). For those reasons, barriers exits to insertion even for M = Si in those cases. Nonetheless, as soon as the conditions become more favorablewhere the sigma hole on M is strong, the net charge on M is sufficiently positive, low-energy empty orbitals are available on M for charge transfer, and the base is a good σdonor and a good acceptorBR insertion is barrierless and dative bonding is bypassed completely. Different mechanisms are implied by the transition state structures that we observe (Figures 5 and 7). The lone pair on B is donated to H followed by C−B bond formation when M = C. For the more electrophilic Si and Ge centers, the B lone pair is donated to the M center with donation from the M−H bond into an empty B p-orbital. New routes for Si−H and Ge−H activation and Si−B and Ge−B bond formation are identified, which represent new departure points from which to build in the synthetic chemistry of those elements. C−H activation with BR insertion is not barrierless; it is relatively hard to achieve. The experimental evidence that is available to date, however, shows that barriers to insertion for organic and organometallic compounds are surmountable under reasonable reaction conditions. The barriers, where they exist at all, should be even more readily transgressed for Si and Ge.



ACKNOWLEDGMENTS Our work was supported by the National Science Foundation (NSF-CAREER Award CHE-1056430 and NSF-MRI Grants CHE-0958696 (University of Richmond (UR)) and CHE1229354 (the MERCURY consortium). E.B. is grateful to the Howard Hughes Medical Institute for support through the URISE and SMART programs at UR. K.J.D. acknowledges the support of the Henry Dreyfus Teacher-Scholar Awards Program.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.7b09656. Representations of transition state (TS) structures obtained in this work, the zero point corrected energies and free energies for reactions, binding energies for weak electrostatic and dative type complexes, coordinates for the dative complexes, TSs, and covalent insertion products optimized in this work, binding energies and free energies for minima and TSs, and orbital energies and HOMO−LUMO gaps for borylenes at different levels of theory and for MH3F molecules (PDF)



REFERENCES

(1) Arduengo, A. J.; Harlow, R. L.; Kline, M. A Stable Crystalline Carbene. J. Am. Chem. Soc. 1991, 113, 361−363. (2) Blauer, J.; Greenbaum, M. A.; Farber, M. The Heat and Entropy of Formation of Boron (I) Fluoride(g). J. Phys. Chem. 1964, 68, 2332− 2334. (3) Timms, P. L. Boron-Fluorine Chemistry. I. Boron Monofluoride and Some Derivatives. J. Am. Chem. Soc. 1967, 89, 1629−1632. (4) Timms, P. L. Boron-Fluorine Chemistry. II. Reaction of Boron Monofluoride with Acetylenes. J. Am. Chem. Soc. 1968, 90, 4585− 4589. (5) Hildenbrand, D. L.; Murad, E. Dissociation Energy of Boron Monofluoride from Mass-Spectrometric Studies. J. Chem. Phys. 1965, 43, 1400−1403. (6) Lovas, F. J.; Johnson, D. R. Microwave Spectrum of BF. J. Chem. Phys. 1971, 55, 41−44. (7) Pianalto, F. S.; O’Brien, L. C.; Keller, P. C.; Bernath, P. F. Vibration-Rotation Spectrum of BH X1Σ+ by Fourier Transform Emission Spectroscopy. J. Mol. Spectrosc. 1988, 129, 348−353. (8) Bettinger, H. F. Phenylborylene: Direct Spectroscopic Characterization in Inert Gas Matrices. J. Am. Chem. Soc. 2006, 128, 2534−2535. (9) Braunschweig, H.; Dewhurst, R. D.; Gessner, V. H. Transition Metal Borylene Complexes. Chem. Soc. Rev. 2013, 42, 3197−3208. (10) Andrews, L.; Hassanzadeh, P.; Martin, J. M. L.; Taylor, P. R. Pulsed Laser Evaporated Boron Atom Reactions with Acetylene. Infrared Spectra and Quantum Chemical Structure and Frequency Calculations for Several Novel BC2H2 and HBC2 Molecules. J. Phys. Chem. 1993, 97, 5839−5847. (11) Thompson, C. A.; Andrews, L.; Martin, J. M. L.; El-Yazal, J. Infrared Spectra of Boron-Ammonia Reaction Products in Solid Argon. J. Phys. Chem. 1995, 99, 13839−13849. (12) Pachaly, B.; West, R. Photochemical Generation of Triphenylsilylboranediyl (C6H5)3SiB: from Organosilylboranes. Angew. Chem., Int. Ed. Engl. 1984, 23, 454−455. (13) Ito, M.; Tokitoh, N.; Kawashima, T.; Okazaki, R. Formation of a Borylene by Photolysis of an Overcrowded Bis(methylseleno)borane. Tetrahedron Lett. 1999, 40, 5557−5560. (14) Bissinger, P.; Braunschweig, H.; Kraft, K.; Kupfer, T. Trapping the Elusive Parent Borylene. Angew. Chem., Int. Ed. 2011, 50, 4704− 4707. (15) Curran, D. P.; Boussonnière, A.; Geib, S. J.; Lacôte, E. The Parent Borylene: Betwixt and Between. Angew. Chem., Int. Ed. 2012, 51, 1602−1605. (16) Bertrand, G.; Soleilhavoup, M. Borylenes: an Emerging Class of Compounds. Angew. Chem., Int. Ed. 2017, 56, 10282−10292. (17) Cowley, A. H.; Lomelí, V.; Voigt, A. Synthesis and Characterization of a Terminal Borylene (Boranediyl) Complex. J. Am. Chem. Soc. 1998, 120, 6401−6402. (18) Braunschweig, H.; Dewhurst, R. D.; Hupp, F.; Nutz, M.; Radacki, K.; Tate, C. W.; Vargas, A.; Ye, Q. Multiple complexation of CO and Related Ligands to a Main-Group Element. Nature 2015, 522, 327−330.

AUTHOR INFORMATION

Corresponding Author

*Tel 804-484-1628; Fax 804-287-1897; e-mail kdonald@ richmond.edu (K.J.D.). ORCID

Kelling J. Donald: 0000-0001-9032-4225 L

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A (19) Kinjo, R.; Donnadieu, B.; Celik, M. A.; Frenking, G.; Bertrand, G. Synthesis and Characterization of a Neutral Tricoordinate Organoboron Isoelectronic with Amines. Science 2011, 333, 610−613. (20) Meller, A. The Borylation of Aromatic Compounds by Dehalogenation Products of Dialkylamino(dihalogeno)boranes. Pure Appl. Chem. 1991, 63, 395−398. (21) Grigsby, W. J.; Power, P. P. Isolation and Reduction of Sterically Encumbered Arylboron Dihalides: Novel Boranediyl Insertion into CC σ-Bonds. J. Am. Chem. Soc. 1996, 118, 7981−7988. (22) Krasowska, M.; Bettinger, H. F. Reactivity of Borylenes toward Ethyne, Ethene and Methane. J. Am. Chem. Soc. 2012, 134, 17094− 17103. (23) Krasowska, M.; Bettinger, H. F. Computational Study of van der Waals Complexes between Borylenes and Hydrocarbons. Chem. - Eur. J. 2014, 20, 12858−12863. (24) Donald, K. J.; Wittmaack, B. K.; Crigger, C. Tuning σ-Holes: Charge Redistribution in the Heavy (Group 14) Analogues of Simple and Mixed Halomethanes can Impose Strong Propensities for Halogen Bonding. J. Phys. Chem. A 2010, 114, 7213−7222. (25) Tawfik, M.; Donald, K. J. Halogen Bonding: Unifying Perspectives on Organic and Inorganic Cases. J. Phys. Chem. A 2014, 118, 10090−10100. (26) Donald, K. J.; Tawfik, M. The Weak Helps the Strong: SigmaHoles and the Stability of MF4·Base Complexes. J. Phys. Chem. A 2013, 117, 14176−14183. (27) Politzer, P.; Murray, J. S.; Lane, P.; Concha, M. C. Electrostatically Driven Complexes of SiF4 with Amines. Int. J. Quantum Chem. 2009, 109, 3773−3780. (28) Wilcken, R.; Zimmermann, M. O.; Lange, A.; Joerger, A. C.; Boeckler, F. M. Principles and Applications of Halogen Bonding in Medicinal Chemistry and Chemical Biology. J. Med. Chem. 2013, 56, 1363−1388. (29) Donald, K. J.; Tawfik, M.; Buncher, B. Weak Interactions as Diagnostic Tools for Inductive Effects. J. Phys. Chem. A 2015, 119, 3780−3788. (30) Krasowska, M.; Edelmann, M.; Bettinger, H. F. Electronically Excited States of Borylenes. J. Phys. Chem. A 2016, 120, 6332−6341. (31) Vidovic, D.; Aldridge, S. Coordination and Activation of the BF Molecule. Angew. Chem., Int. Ed. 2009, 48, 3669−3672. (32) Head-Gordon, M.; Head-Gordon, T. Analytic MP2 Frequencies Without Fifth Order Storage: Theory and Application to Bifurcated Hydrogen Bonds in the Water Hexamer. Chem. Phys. Lett. 1994, 220, 122−128 and references therein. (33) 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.; et al. Gaussian 09, Revision D.01; Gaussian, Inc.: Wallingford, CT, 2013. The full reference for this suite of programs is in the Supporting Information. (34) Becke, A. D. Density-functional thermochemistry. III. The Role of Exact Exchange. J. Chem. Phys. 1993, 98, 5648−5652. (35) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. Ab Initio Calculation of Vibrational Absorption and Circular Dichroism Spectra Using Density Functional Force Fields. J. Phys. Chem. 1994, 98, 11623−11627. (36) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010, 132, 154104. (37) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. A Fifth-Order Perturbation Comparison of Electron Correlation Theories. Chem. Phys. Lett. 1989, 157, 479−483. (38) Dunning, T. H. Gaussian Basis Sets for use in Correlated Molecular Calculations. The Atoms Boron through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007−1023. (39) Peterson, K.; Shepler, B.; Figgen, D.; Stoll, H. On the Spectroscopic and Thermochemical Properties of ClO, BrO, IO, and Their Anions. J. Phys. Chem. A 2006, 110, 13877−13883. (40) Dennington, R.; Keith, T.; Millam, J. GaussView, Version 5; Semichem Inc.: Shawnee Mission, KS, 2009.

(41) Chemcraft: Andrienko, A. G.; Senchenya, I. N.; Romanov, A.; http://www.chemcraftprog.com. (42) Boys, S. F.; Bernardi, F. Calculation of Small Molecular Interactions by Differences of Separate Total Energies − Some Procedures with Reduced Errors. Mol. Phys. 1970, 19, 553−566. (43) For a description of this procedure, see: Jensen, F. Introduction to Computational Chemistry; Wiley: New York, 1999; pp 172−173. (44) The van der Waals (vdW) radii of B, C, Si, and Ge are 1.92, 1.70, 2.10, and 2.11 Å, respectively, in ref 46. Different sources may list slightly different values for these radii, however. In ref 45, the vdW radii are 2.05, 1.85, 2.25, and 2.23 Å for B, C, Si, and Ge, respectively, with the vdW radius a bit larger for Si than it is for Ge. Even in the latter source, however, the covalent radius of Si is smaller than that for Ge (1.176 and 1.225 Å, respectively). (45) Batsanov, S. S. van der Waals Radii of Elements. Inorg. Mater. 2001, 37, 871−885. (46) Atomic Radii of the Elements. In CRC Handbook of Chemistry and Physics, 97th ed.; Haynes, W. M., Ed.; CRC Press: Boca Raton, FL, 2017 (Internet version). (47) Bauschlicher, C. W., Jr.; Schaefer, H. F., III; Bagus, P. S. Structure and Energetics Of Simple Carbenes Methylene, Fluoromethylene, Chloromethylene, Bromomethylene, Difluoromethylene, And Dichloromethylene. J. Am. Chem. Soc. 1977, 99, 7106−7110. (48) Brazier, C. R. Emission Spectroscopy of the Triplet System of the BH Radical. J. Mol. Spectrosc. 1996, 177, 90−105. (49) Schwartz, R. L.; Davico, G. E.; Ramond, T. M.; Lineberger, W. C. Singlet−Triplet Splittings in CX2 (X = F, Cl, Br, I) Dihalocarbenes via Negative Ion Photoelectron Spectroscopy. J. Phys. Chem. A 1999, 103, 8213−8221. (50) Ess, D. H.; Cook, T. C. Unrestricted Prescriptions for OpenShell Singlet Diradicals: Using Economical Ab Initio and Density Functional Theory to Calculate Singlet−Triplet Gaps and Bond Dissociation Curves. J. Phys. Chem. A 2012, 116, 4922−4929. (51) Shang, R.; Braunschweig, H. Reactivity of Transition-Metal Borylene Complexes: Recent Advances in B−C and B−B Bond Formation via Borylene Ligand Coupling. Inorg. Chem. 2015, 54, 3099−3106. (52) Fiorillo, A. A.; Galbraith, J. M. A Valence Bond Description of Coordinate Covalent Bonding. J. Phys. Chem. A 2004, 108, 5126− 5130. (53) Jonas, V.; Frenking, G.; Reetz, M. T. Comparative Theoretical Study of Lewis Acid-Base Complexes of BH3, BF3, BC13, AlC13, and SO2. J. Am. Chem. Soc. 1994, 116, 8141−8753. (54) Moss, R. A.; Mallon, C. B. The Characterization of Carbene Selectivity. Applications to Difluorocarbene. J. Am. Chem. Soc. 1975, 97, 344−347. (55) Moss, R. A. Carbenic Selectivity in Cyclopropanation Reactions. Acc. Chem. Res. 1980, 13, 58−64. (56) Moss, R. A Carbenic Philicity. In Carbene Chemistry: From Fleeting Intermediates to Powerful Reagents; Bertrand, G., Ed.; Fontis Media: Chapter 3, pp 57−101. (57) IUPAC Compendium of Chemical Terminology, 2nd ed. (the “Gold Book”); compiled by D. McNaught and A. Wilkinson; Blackwell Scientific Publications: Oxford, 1997. XML on-line corrected version: http://goldbook.iupc.org (2006) created by M. Nic, J. Jirat, and B. Kosata; updates compiled by A. Jenkins. This term: https://goldbook. iupc.org/html/B/B00628.html. (58) Bratsch, S. G. Revised Mulliken Electronegativities - I. Calculation and Conversion to Pauling Units. J. Chem. Educ. 1988, 65, 34−41. The valence state Mulliken electronegativities for tetravalent C, Si, and Ge are 8.15, 7.30, and 7.53 eV, respectively. (59) Fuentealba, P.; Reyes, O. Atomic Softness and the Electric Dipole Polarizability. J. Mol. Struct.: THEOCHEM 1993, 282, 65−70. (60) Pyykkö, P. Additive Covalent Radii for Single-, Double-, and Triple-Bonded Molecules and Tetrahedrally Bonded Crystals: A Summary. J. Phys. Chem. A 2015, 119, 2326−2337. (61) Bond Dissociation Energies. In CRC Handbook of Chemistry and Physics, 97th ed.; Haynes, W. M., Ed.; CRC Press: Boca Raton, FL, 2017 (Internet version). M

DOI: 10.1021/acs.jpca.7b09656 J. Phys. Chem. A XXXX, XXX, XXX−XXX