Volume 27, Issue 14 e202400069
Research Article
Open Access

Rational Design of Metallylenes for Hydrogenation Reactions

Eveline H. Tiekink

Eveline H. Tiekink

Department of Chemistry and Pharmaceutical Sciences, AIMMS, Vrije Universiteit Amsterdam, De Boelelaan 1108, 1081 HZ Amsterdam, The, Netherlands.

Search for more papers by this author
Siebe Lekanne Deprez

Siebe Lekanne Deprez

Department of Chemistry and Pharmaceutical Sciences, AIMMS, Vrije Universiteit Amsterdam, De Boelelaan 1108, 1081 HZ Amsterdam, The, Netherlands.

Search for more papers by this author
Dr. Pascal Vermeeren

Dr. Pascal Vermeeren

Department of Chemistry and Pharmaceutical Sciences, AIMMS, Vrije Universiteit Amsterdam, De Boelelaan 1108, 1081 HZ Amsterdam, The, Netherlands.

Search for more papers by this author
Dr. Trevor A. Hamlin

Corresponding Author

Dr. Trevor A. Hamlin

Department of Chemistry and Pharmaceutical Sciences, AIMMS, Vrije Universiteit Amsterdam, De Boelelaan 1108, 1081 HZ Amsterdam, The, Netherlands.

Search for more papers by this author
First published: 12 February 2024

Graphical Abstract

Theorydriven design! Using state-of-the-art quantum chemical analyses, we screened the hydrogenation capability of a range of diverse metallylene catalysts with varying Group 14 atoms and ligands. Our design features identify that stannylenes featuring a phosphorus–based ligand have the potential to activate dihydrogen and facilitate the hydrogenation of a spectrum of unsaturated bonds.

Abstract

The hydrogenation of alkenes, alkynes, carbonyls, and imines by H2-activated metallylenes (H3C−EH2−L, CELH2; E=Si, Ge, Sn and L=NMe2, PMe2) was studied using relativistic density functional theory at ZORA-BP86/TZ2P. By means of activation strain and Kohn–Sham molecular orbital analyses, the physical factors underlying the trends in reactivity were identified and quantified. Firstly, the hydrogenation reactivity increases on descending Group 14, that is from silylenes (E=Si) to stannylenes (E=Sn). This reactivity trend originates primarily from a reduced energy required to break the bonds between the Group 14 atom and the hydrogen atoms because the strength of these bonds decreases from Si−H to Ge−H to Sn−H. Secondly, the reactivity decreases as the Group 15 ligand changes from nitrogen to phosphorus (L=NMe2, PMe2) which stems from the trigonal pyramidal PMe2 geometry compared to the trigonal planar NMe2 geometry. The latter can, therefore, effectively engage in stronger orbital interactions with the substrate than the former, due to a more efficient HOMO–LUMO orbital overlap. By combining our insights, we rationally designed an optimally tuned catalyst (Group 14 atom and ligand) considering the overall catalytic cycle involving H2 activation of the metallylene and subsequent hydrogenation.

Introduction

Heavier main–group elements have emerged in the past two decades as suitable alternatives to transition metals.1 Main-group chemistry was originally considered to be limited by the large energy gaps in the valence s and p orbitals of the elements – either fully occupied or empty – and therefore having weak interactions with small molecules such as CO, C2H4, or H2. On the other hand, transition metal complexes exhibit a rich assortment of bonding capabilities due to the diversity in electronic structure as the complexes are often paramagnetic, have small energy gaps, and partially filled d orbitals. In recent years, however, a number of important discoveries have shown that compounds containing heavier main-group elements can, in fact, mimic transition metal catalysts.1, 2 One specific class that has gained considerable interest, is the heavier analogue of carbenes known as metallylenes with larger Group 14 elements such as Si, Ge, Sn. The key difference between carbenes and metallylenes is that the latter predominantly exists as a singlet state, because of a large singlet–triplet energy gap. Consequently, metallylenes are more stable than carbenes while the singlet–triplet energy gap remains small enough to facilitate the richness of transition metal–like chemistry. In addition, the larger atomic radii of heavy Group 14 elements allow accessibility of a large range of coordination numbers for ligands. Metallylenes may therefore mimic transition metals and offer an alternative to transition metals catalysts.

Experimental work has shown that metallylenes can be used to activate small molecules such as H2.3, 4, 5 Recently, we have performed a computational study to unravel the mechanism of, and the underlying factors behind, reactivity trends in metallylene activation of H2.6 Model metallylene systems CEL were studied with E being a Group 14 element C, Si, Ge or Sn, and L a Group 15 ligand NMe2 or PMe2 (Scheme 1a). Upon going down Group 14, the reaction barrier increases from Si to Sn due to less back-donation from the lone–pair of the metallylene to the σ*-orbital of hydrogen. When changing the ligand from NMe2 to PMe2, the reaction barrier decreases because of a reduced steric (Pauli) repulsion between the metallylene and the incoming substrate, which is a result of a deformation of the PMe2 ligand in CEP from trigonal planar to trigonal pyramidal, while the NMe2 ligand in CEN remains trigonal planar. A promising subsequent reaction for activated metallylenes is the hydrogenation of unsaturated bonds. Conventional catalysts for hydrogenation reactions include transition metal catalysts such as palladium, ruthenium, nickel, and rhodium complexes.7 However, metallylenes might offer a safer and more sustainable alternative.8 As a result, the question can be raised whether metallylenes can be used in hydrogenation reactions without the need for transition metals.

Details are in the caption following the image

a) The activation of H2 by a metallylene CEL that forms the activated metallylene CELH2 and b) hydrogenation using the activated metallylenes CELH2.6

This work aims to investigate the ability of H2-activated metallylenes to facilitate hydrogenation reactions by means of a systematic computational study. Our model activated metallylenes H3C−EH2−L (CELH2, with E=Si, Ge, Sn and L=NM2, PMe2) are considered to hydrogenate different types of bonds: carbonyls (C=O), alkenes (C=C), alkynes (C≡C) and imines (C=N) (Scheme 1b). For analyzing the reactions, relativistic density functional theory (DFT) is employed at a ZORA-BP86/TZ2P level using the Amsterdam Density Functional (ADF) program. Within the framework of Kohn-Sham molecular orbital (KS-MO) theory,9 the activation strain model (ASM)10 and the energy decomposition analysis (EDA)11 are utilized to determine and rationalize the hydrogenation reactivity trends, when varying i) the central atom E, and ii) the ligand L. In addition, we have evaluated the energetics of the overall catalytic cycle of metallylene H2 activation and hydrogenation to identify the most viable metallylene catalysts to serve as alternatives to transition metal catalysts. With these considerations, a rational design of metallylene hydrogenation is established to provide a foundation for further developments regarding realistic metallylene catalysis.

Results and Discussion

Reaction profiles of the hydrogenation

We start with studying the reaction profiles for the hydrogenation of substrates by activated metallylenes CELH2 with varying Group 14 elements E=Si, Ge, Sn, and varying Group 15 ligands L=NM2, PMe2. Table 1 shows the (Gibbs free) reaction barriers (ΔE, ΔG) and (Gibbs free) reaction energies (ΔErxn, ΔGrxn) for the hydrogenation of acetylene. The complementary results for other substrates, that is, ethylene, formaldehyde, methanimine, acetone, and propan-2-imine, are provided in the SI in Table S1. We have selected these substrates to compare the hydrogenation of the C≡C, C=C, C=N, and C=O bonds (see the section Expanding the substrate scope). These reactions follow a concerted asynchronous reaction path, where one bond forms ahead of the other bond. From Table 1, two trends can be observed. Firstly, the reaction barrier decreases, and the reaction becomes more exergonic, when varying E by going down Group 14 from Si to Ge to Sn while keeping the ligand L fixed. Secondly, the reaction barrier increases, and the reaction becomes less exergonic, when varying L from NMe2 to PMe2 while keeping E consistent. Both trends hold for the other substrates as well (Table S1). These two trends in varying E and L in the hydrogenation are the opposite of those in the H2 activation.6 Whereas the reaction barrier increases from Si to Ge to Sn with a fixed ligand for H2 activation, the barrier decreases for hydrogenation. Similarly, the reaction barrier decreases from NMe2 to PMe2 regarding H2 activation while the barrier increases in the same order for hydrogenation. Further attention to these opposing trends is given in a subsequent section wherein both H2 activation and hydrogenation are examined. Next, we will examine the hydrogenation trends when varying the central atom E from Si to Ge to Sn and when varying the ligand L from NMe2 to PMe2. The trends in electronic energies are the same as for the Gibbs free energies and we will, therefore, focus solely on the electronic energies.

Table 1. Electronic reaction barriers (ΔE) and reaction energies (ΔErxn) (in kcal mol−1) of the hydrogenation of acetylene by H3C−EH2−L (CELH2) metallylenes.[a]

image

E

L

ΔE≠[b]

ΔErxn[b]

Si

NMe2

32.2 (42.6)

−14.8 (−8.9)

Ge

NMe2

25.3 (34.7)

−37.0 (−30.8)

Sn

NMe2

20.1 (31.6)

−52.1 (−44.1)

Si

PMe2

39.9 (50.2)

−5.5 (−0.9)

Ge

PMe2

30.9 (41.0)

−23.7 (−21.3)

Sn

PMe2

24.3 (35.2)

−38.6 (−32.0)

  • [a] Computed at ZORA-BP86/TZ2P. [b] Gibbs free energies are given in parenthesis.

Varying the metallylene central atom (E) along Group 14

To understand the trend of decreasing barriers as E enlarges (from Si to Sn), we employ the activation strain model (ASM)10 for the CELH2 systems. The ASM is a fragment-based approach where the electronic energy (ΔE) is decomposed into the strain energy (ΔEstrain), which is the prerequisite in order to deform the reactants, and the interaction energy (ΔEint) between the deformed reactants. Different series are studied in which the central atom E of CELH2 is varied from Si to Ge to Sn for both ligands (NMe2 and PMe2) and a substrate (ethylene and acetylene). The activation strain analyses are shown in Figure 1 and Figures S5–S7.12 Since all series exhibit the same trends, we will only focus on the hydrogenation with activated metallylenes CEPH2 (E=Si, Ge, Sn) and acetylene as substrate. The interacting fragments in the activation strain analyses are CEPH2 and acetylene (C2H2) (Figure 1a). The trend in energy (ΔE) originates from a consistent decrease in strain energy (ΔEstrain) from Si to Sn. On the other hand, the interaction energies are nearly superimposed around the transition state region and hence are not responsible for setting the observed reactivity trend. To investigate the strain energy trend further, we decompose the total strain energy into the strain of the fragments CEPH2 and C2H2 (Figure 1b). The strain energies of both fragments decrease when going from Si to Sn and follow the trend in the total strain energy and thus the trend in the reaction barrier.

Details are in the caption following the image

a) Activation strain analyses and b) decomposition of the strain of the hydrogenation reaction of acetylene (C2H2) by metallylenes CEPH2 with variating Group 14 central atom (E=Si, Ge, Sn), where the transition states are indicated with a dot and the energy terms along the IRC are projected on the shortest of the newly forming C⋯H bonds. Computed at ZORA-BP86/TZ2P.

The strain of CEPH2 is larger in magnitude around the transition state region and thus, the strain energy is primarily dictated by the geometrical deformations of the activated metallylene. The decisive factor for the lowering of the strain energy from CSiPH2 to CGePH2 to CSnPH2 is the weaker E−H bond along this series. The homolytic E−H bond dissociation energy and thus the E−H bond strength decreases from 85.4 kcal mol−1 for Si−H to 69.3 kcal mol−1 for Sn−H (Table 2). The E−H bonds break during the hydrogenation reaction, and breaking a stronger bond requires a higher strain energy. One may then wonder why the bond weakens from Si−H to Sn−H. Our energy decomposition analyses (see SI Figure S9) identify that the weaker bond strength from Si−H to Sn−H stems from an increasing steric (Pauli) repulsion with the effectively larger atom size of Sn, thereby lengthening, and correspondingly weakening, the E−H bond along the series. See Ref. [13] for a comprehensive elucidation of the origin of HnX−YHn bond strengths.

Table 2. Homolytic bond dissociation enthalpies (in kcal mol−1) of the E−H bonds in CEPH2 systems and their corresponding E−H bond lengths (Å).[a]

CEPH2

ΔHBDE[b]

E−H (Å)

Si−H

85.4

1.502

1.497

Ge−H

77.6

1.551

1.544

Sn−H

69.3

1.738

1.730

  • [a] Computed at ZORA-(U)BP86/TZ2P at 298.15 K and 1 atm. [b] The ΔHBDE value is the same for both E−H bonds.

The lowering of the strain energy and consequently the lowering of the reaction barrier of hydrogenation with CELH2 from E=Si to Ge to Sn is reinforced by the difference in degree of asynchronicity of the transformation. During the hydrogenation by CSnPH2, less geometrical deformations occur due to a large asynchronicity, that is a large difference in length between the newly forming C⋯H bonds.14 The transition state structures in Figure S8 show that the difference in length between the newly forming C⋯H bonds is larger for CSnPH2rC⋯H=0.92 Å) than for CSiPH2rC⋯H=0.43 Å). Because of the large asynchronicity for CSnPH2, only one E−H bond is elongated and hence only one HCC angle of acetylene is bent largely at its transition state (∠CCH1=135° and ∠CCH2=179°, Figure S8). In contrast, for CSiPH2, both HCC angles of acetylene are bent (∠CCH1=141° and ∠CCH2 153°), giving a larger geometrical deformation and hence cause a higher strain energy. So, the reaction barrier of hydrogenation with CELH2 decreases from E=Si to Ge to Sn due to a reducing activation strain caused by a weakening of the E−H bond strength in the same order and a larger degree of asynchronicity.

Varying the ligand (L) from NMe2 to PMe2

Next, we focus on the origin of the increase in reaction barrier as the ligand of the activated metallylenes CELH2 changes from NMe2 to PMe2. Solely the reactivity trends of CGeLH2 with L=NMe2, PMe2, and acetylene as substrate are discussed here since the other metallylenes (CSiLH2 and CSnLH2) and substrate (ethylene) show similar characteristics (Figure S10–S14). We examine the physical factors leading to the lower barrier for L=NMe2 by using the activation strain model. Figure 2a shows that the hydrogenation involving CGeNH2 reaches the transition state earlier and is associated with a lower reaction barrier than CGePH2. The interaction energy (ΔEint) is more stabilizing for NMe2 than for PMe2 and, therefore, determines the trend in total energy (ΔE). The trend in strain energy (ΔEstrain) is, on the other hand, opposite to the trend in total energy, due to their difference in asynchronicity (vide supra). To understand where the trend in interaction energy originates from, we applied our canonical energy decomposition analysis (EDA).11 Our EDA decomposes the interaction energy (ΔEint) into three physically meaningful energy terms: the classical electrostatic interaction (ΔVelstat), the destabilizing Pauli repulsion (ΔEPauli) that comprises the repulsion between closed–shell orbitals, and the orbital interactions (ΔEoi) that accounts for charge transfer (HOMO–LUMO interactions) and polarization. Decomposing the interaction energy reveals that the orbital interactions and to a lesser extent the electrostatic interactions determine the trend in interaction energy (Figure 2b). In the next part, we exclusively focus on elucidating the orbital interactions because they are the most dominant for setting the trend in interaction energy from PMe2 to NMe2.

Details are in the caption following the image

a) Activation strain analysis and b) energy decomposition analyses of the hydrogenation reaction of acetylene by metallylenes CGeLH2 when L=NMe2, PMe2. Transition states are indicated with a dot and the energy terms along the IRC are projected on the shortest of the newly forming C⋯H bonds. Computed at ZORA-BP86/TZ2P.

A Kohn–Sham molecular orbital (KS-MO) analysis is performed to further understand why the hydrogenation with CGeNH2 and acetylene proceeds with more stabilizing orbital interactions. We find that the stronger orbital interactions emerge from the symmetrically shaped occupied orbitals of CGeNH2 compared to the asymmetrically shaped counterparts of CGePH2. This difference in orbital shape can be traced back to the geometry of the ligand being trigonal planar for CGeNH2 versus trigonal pyramidal for CGePH2. We utilize consistent transition state-like geometries with a C⋯H bond length of 1.60 Å. By using consistent geometries, the resulting orbital analysis is not skewed by the different locations of the transition states.15 The primary orbital interaction is donation from the σ-bonding E−H orbital of the activated metallylene CGeLH2 (HOMOCGeLH2) into the π* orbital of acetylene (LUMOC2H2, Figure 3).16 The KS-MO analysis reveals that the HOMOCGeLH 2−LUMOC2H2 interaction is stronger for CGeNH2 compared to CgePH2 due to both a smaller HOMO–LUMO orbital energy gap, 2.9 eV compared to 4.1 eV, respectively, and a larger orbital overlap, 0.18 and 0.09, respectively.

Details are in the caption following the image

a) Molecular orbital diagram with energy gaps (in eV) and overlaps (S) for the HOMOCGeLH 2−LUMOC 2H 2 interaction; and b) the fragment molecular orbitals (isovalue=0.03 Bohr−3/2), computed at ZORA-BP86/TZ2P on consistent geometries with a C⋯H bond length of 1.60 Å.

The larger orbital overlap for CGeNH2 originates from a symmetrically shaped HOMOCgeNH 2 that leads to i) orbital overlap at both E−H bonds; and ii) a more synchronous pathway (Figure 3b). The HOMOCGeNH2 has a symmetric shape and has 1s atomic orbital character of both hydrogen atoms through which the HOMOCGeNH2 can overlap at both E−H bonds with the LUMOC2H2. The asymmetric HOMOCGePH2 has solely 1s atomic orbital character on one E−H bond and overlaps with the LUMOC2H2 at one side. Additionally, for CGeNH2, both C−H bonds are formed in a rather synchronous fashion: one C⋯H bond is 1.60 Å and the other C⋯H bond is 1.80 Å. For CGePH2, however, one C⋯H bond is 1.60 Å, while the other C⋯H bond is 2.14 Å. Hence, the lower orbital overlap for CGePH2 is further assisted by the difference in bond length between the newly forming C⋯H bonds, i. e., the asynchronicity.14 As explained in the SI Figures S17–S21, the more synchronous pathway for CGeNH2 arises from the symmetric shape of the occupied orbitals. Thus, the symmetric shape of the orbitals of CGeNH2 compared to the asymmetric shape of the orbitals of CGePH2 causes the difference in HOMO–LUMO orbital overlap.

Now, we will examine the origin of the smaller HOMOCGeLH2−LUMOC2H2 energy gap for NMe2 compared to PMe2 which mainly results from a higher HOMO energy of −4.9 eV for CGeNH2 versus −5.7 eV for CGePH2. A lower–lying LUMO of acetylene contributes only slightly to the smaller orbital energy gap.17, 18 The HOMO of CGePH2 is more stable (lower in energy) than the HOMO of CGeNH2 due to the σ-bonding between phosphorus and germanium: the 3p atomic orbital of phosphorus mixes in-phase with the 4p atomic orbital of germanium, whereas for nitrogen, almost no in-phase mixing occurs (Figure S22). The in-phase mixing is caused by a trigonal pyramidal structure of the PMe2 ligand, whereas the NMe2 ligand is trigonal planar which leads to phase cancellation when the 2p atomic orbital of nitrogen overlaps with the 4p atomic orbital of germanium.

Not only the difference in orbital energies but also the difference in shape of the occupied orbitals between CGeNH2 and CGePH2 are caused by the geometry of the ligands. The nitrogen ligand in the transition state of CGeNH2 is trigonal planar while the phosphorus ligand in CGePH2 is trigonal pyramidal (Figure 4). Both ligands are to different extents trigonal pyramidal in the equilibrium structures of CGeNH2 and CGePH2 (R) with pyramidalization sums of angles (SoA)19 around L of respectively 342.6° and 297.0°. During the hydrogenation reaction, the ligands planarize when the Ge−H bonds break. The ligands in the metallylene catalysts CGeL (P) are trigonal planar because of a hyperconjugation interaction between the empty 4p atomic orbital of germanium and the filled np atomic orbital of N or P (see Ref. [6]). The nitrogen ligand in CGeNH2 has a stronger preference for trigonal planarity than the phosphorus ligand in CGePH2 and hence already adopts a trigonal planar structure in the transition state. The energy and shape of the HOMOCGeNH2 and HOMOCGePH2 are strongly dependent on their structural configuration; the HOMO lowers in energy and gets an asymmetric shape when the ligand undergoes pyramidalization (Figure S23). So, the high reaction barrier for CGePH2 originates from the asymmetric orbitals in CGePH2, which arises from a trigonal pyramidal PMe2. The reaction barrier of CGeNH2, on the other hand, is lower, because the NMe2 ligand adopts a trigonal planar geometry making the occupied orbitals of CGeNH2 symmetrically shaped.

Details are in the caption following the image

Structures for the hydrogenation reaction of acetylene by a) CGeNH2 and b) CGePH2 metallylene catalysts with the sum of angles (SoA) around nitrogen or phosphorus, computed at ZORA-BP86/TZ2P.

The last question to address is why NMe2 has a stronger preference for a trigonal planar structure than PMe2. We find that pyramidalization costs more energy for NMe2 than for PMe2 because of shorter N−C bonds compared to P−C bonds.20 Pyramidalization of NMe2 results in a large steric repulsion between the methyl groups while the steric repulsion for PMe2 remains low due to the longer P−C bonds (Table S3). Therefore, NMe2 does not undergo significant pyramidalization upon activation by hydrogen, in contrast to PMe2. Similar reasons apply for why the CH3⋅ radical is trigonal planar whereas its heavier central atom analogs are trigonal pyramidal.21

Expanding the substrate scope

In the previous two sections, the trends in hydrogenation reactions of acetylene by CELH2 metallylenes were studied when changing E down Group 14 from Si to Ge to Sn, and when changing L down Group 15 from NMe2 to PMe2. In this section, we expand the substrate scope by comparing the hydrogenation of various substrates, such as acetylene (C2H2), ethylene (C2H4), methanimine (CH2NH), propan-2-imine (C3H6NH), formaldehyde (CH2O) and acetone (C3H6O). The reactivity trends when changing the central atom E and the ligand L hold for all substrates (Table S1, Scheme 2). Focusing on the hydrogenation of acetylene (C≡C) and ethylene (C=C) shows that C=C hydrogenation is in general less favorable than C≡C hydrogenation since the former has slightly lower reaction barriers and the products are thermodynamically more stable. Comparing the hydrogenation of C=C (C2H4) to C=N (CH2NH) to C=O (CH2O) reveals that the reaction barriers increase from C=O to C=N to C=C, but the reaction becomes more exergonic (or less endergonic) from C=O to C=N to C=C. Thus, whereas the C=C hydrogenation proceeds with the highest barriers for all CELH2 catalysts, the reaction energies are the most stabilizing. Lastly, the significance of larger substituents on the substrate has been studied by comparing methanimine (CH2NH) and formaldehyde (CH2O) with their methyl-substituted analogues (C3H6NH and C3H6O, respectively). When enlarging the substituents, the reaction barriers increase, and the reactions become more endergonic/less exergonic.

Details are in the caption following the image

Gibbs free energy reaction barriers (ΔG, beside arrow) and reaction energies (ΔGrxn, beside product) (in kcal mol−1) of the hydrogenation of various small molecules by CEPH2 metallylenes, where E=Si, Ge, Sn, computed at ZORA-BP86/TZ2P.

Closing the catalytic cycle

Thus far, solely the metallylene hydrogenation reaction has been considered wherein the CELH2 catalyst transfers two hydrogen atoms to the substrate, yielding H3C−E−L (CEL) and the hydrogenated product. However, CELH2 first needs to be formed through the activation of H2 by CEL. This section aims to combine the reaction profiles of metallylene H2 activation and hydrogenation, thereby completing the catalytic cycle of CEL which enables for identification of the best catalyst based on its reactivity. CEL first activates H2 (TS1) and subsequently transfers the hydrogen atoms to a substrate (TS2, Figure 5). Depending on the composition of CEL – the choice of the central atom E and the ligand L -- the reaction barriers can be raised or lowered as has been shown for H2 activation,6 and in the current work for hydrogenation. The trends in the reaction barriers when varying E and L in the hydrogenation are the opposite of those in the H2 activation. Figure 5 shows the reaction profile for CSiL and CSnL (L=NMe2, PMe2) with acetylene as substrate. All reaction profiles including CGeL and all substrates are shown in the SI Figures S24–S29. The opposing trends between H2 activation and hydrogenation can be observed by considering the barrier from R to TS1 and the barrier from INT to TS2, respectively. When varying the central atom E from Si to Sn with a fixed ligand, the barrier for H2 activation (TS1) increases and the CELH2 intermediate destabilizes (INT). Due to the destabilization of the CELH2 intermediate, the second barrier decreases for hydrogenation (TS2) from Si to Sn. On the contrary, varying the ligand from NMe2 to PMe2 lowers the first barrier while the second barrier increases.

Details are in the caption following the image

Reaction profile of the H2 activation and subsequent hydrogenation of acetylene involving metallylenes CEL with L =NMe2, PMe2 and E=Si, Sn. The stationary points are further specified by their ΔG values, computed at ZORA-BP86/TZ2P.

The ‘best’ catalyst would be a moderate catalyst that has a low barrier for the H2 activation, but not too low as this results in an overly stable intermediate that leads to a high hydrogenation barrier. The catalyst with the highest reactivity should have the lowest barrier in the rate-determining step. For instance, the most suitable catalyst for acetylene hydrogenation would be CSnP in which the H2 activation (TS1) is rate-determining with an energy of 37.8 kcal mol−1 (Figure S24). Table S4 shows an overview of the most suitable catalysts for the different substrates. All contain the PMe2 ligand and either Ge or Sn (i. e., CSnP and CGeP). Two main trends are responsible for this preference in metallylene catalysts. First, the replacement of NMe2 with PMe2 causes the TS1 barrier to decrease significantly, while the TS2 barrier increases to a much lesser extent. Second, to undo the increase in TS2 barriers, a larger central atom is preferred that lowers the TS2 barriers while raising the TS1 barriers. Thus, the preference for a larger central atom and a larger ligand is due to the cancellation of opposing trends for the TS1 and TS2 barrier. So, experimentalists could try a large central atom, for instance, Ge or Sn, to decrease the hydrogenation barrier and could optimize the phosphorus ligand to decrease the H2 activation barrier.

Conclusions

Our computational study provides physical insights into hydrogenation reactions of alkenes, alkynes, carbonyls, and imines by various metallylenes H3C−EH2−L (CELH2). The reactivity of the hydrogenation reaction can be tuned by varying the central Group 14 atom and ligand of the metallylenes (CELH2 : E=Si, Ge, Sn, and L=NMe2, PMe2). When changing the central atom E down Group 14, the hydrogenation barrier lowers from Si to Ge to Sn. Upon varying the ligand from NMe2 to PMe2 (while keeping E constant), the hydrogenation barrier increases.

Our activation strain analyses reveal that the reaction barrier lowers from Si to Ge to Sn because of i) a larger degree of asynchronicity for CSnLH2 compared to CSiLH2; and ii) a weaker Sn−H bond compared to the stronger Si−H bond. The Sn−H bond is weaker due to an increasing Pauli repulsion between the hydrogen with the effectively larger atom size of Sn. The reaction barrier increases from CGeNH2 and CGePH2 due to their difference in ligand geometry: the nitrogen ligand in the transition state of CGeNH2 is trigonal planar whereas the phosphorus ligand in CGePH2 is trigonal pyramidal. During the hydrogenation reaction, the two Ge−H bonds of CGeLH2 break which induces planarization of the ligand. NMe2 has a stronger preference for a trigonal planar structure than PMe2. The former ligand experiences a larger steric repulsion between the methyl groups when pyramidalizing than the latter ligand, which arises from the shorter C−N bonds compared to the C−P bonds. The trigonal planar structure of the nitrogen ligand of CGeNH2 leads to symmetrically shaped occupied orbitals, which promote orbital interactions with the substrate, thereby, lowering the reaction barrier.

Lastly, we examined the step prior to the hydrogenation, that is, H2 activation by the metallylene catalyst, and compared this reaction step to the hydrogenation. The trends in the reaction barriers when varying E and L in the hydrogenation are the opposite to those in the H2 activation. When varying the central atom E from Si to Sn, the H2 activation barrier increases and the CELH2 intermediate destabilizes. Due to the destabilization of the CELH2 intermediate, the hydrogenation barrier decreases from Si to Sn. On the contrary, varying the ligand from NMe2 to PMe2 lowers the H2 activation barrier while the hydrogenation barrier increases. A good candidate for a metallylene catalyst could contain a PMe2 ligand that lowers the H2 activation barrier significantly and a large central atom E, such as Sn or Ge, that lowers the hydrogenation barrier.

Acknowledgments

This work was supported by the Netherlands Organization for Scientific Research (NWO). DFT calculations were carried out on the Dutch national e-infrastructure with the support of SURF Cooperative.

    Conflict of interests

    The authors declare no conflict of interest.

    Data Availability Statement

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