The activation strain model and molecular orbital theory.

The activation strain model is a powerful tool for understanding reactivity, or inertness, of molecular species. This is done by relating the relative energy of a molecular complex along the reaction energy profile to the structural rigidity of the reactants and the strength of their mutual interactions: ΔE(ζ) = ΔEstrain(ζ) + ΔEint(ζ). We provide a detailed discussion of the model, and elaborate on its strong connection with molecular orbital theory. Using these approaches, a causal relationship is revealed between the properties of the reactants and their reactivity, e.g., reaction barriers and plausible reaction mechanisms. This methodology may reveal intriguing parallels between completely different types of chemical transformations. Thus, the activation strain model constitutes a unifying framework that furthers the development of cross-disciplinary concepts throughout various fields of chemistry. We illustrate the activation strain model in action with selected examples from literature. These examples demonstrate how the methodology is applied to different research questions, how results are interpreted, and how insights into one chemical phenomenon can lead to an improved understanding of another, seemingly completely different chemical process. WIREs Comput Mol Sci 2015, 5:324-343. doi: 10.1002/wcms.1221.

INTRODUCTION

Chemistry is, roughly speaking, the branch of the natural sciences investigating the properties, composition, and transformation of matter. Within theoretical chemistry, this is done not by observation, but by a mathematical description of the physical system of interest. The constant improvement in the quality of mathematical descriptions, combined with the enormous advancement of computer technology in the past decades, has allowed the field of theoretical chemistry to advance as well. Nowadays, it is feasible to computationally study a large variety of molecular systems and chemical processes with, for many purposes, sufficient accuracy. Although some of the romance of doing practical experiments is lost, theoretical chemistry opens up a whole new world of research by eliminating many practical limitations. Theoreticians can, for example, study a synthetically useful and thus desired type of reaction which, however, does not proceed. Such a reaction can not be studied experimentally because it does not occur. Theory, on the other hand, can examine this process in silico and therefore reveal the reasons why it is not viable and what can be done to change this situation for the better.

In the following, we focus on understanding chemical reactivity, i.e., the reaction energy profile that accompanies the transformation of molecular species into new species. We will assume that the energy profile of a chemical reaction has been obtained to sufficient accuracy, and will discuss the application of the activation strain model of chemical reactivity,1–5 which has been developed to obtain more insight into the qualitative and quantitative features of the energy profile. This is done by splitting the relative energy of a molecular complex along the reaction coordinate into two separate terms, originating from the deformation of the reacting species and the interaction between them. Similar approaches have been explored before,6–8 but have not evolved into the generally applicable model that the activation strain model, as it is known in short, has become. Note also, that more recently the group of Houk has adopted the same approach, the distortion/interaction model, differing only in terminology.9

Conveniently, there is no special computer code required to perform activation strain analyses: all necessary quantities can be computed using any of the regular quantum-chemical software packages available. As a result, the activation strain model has been applied by various research groups, on a range of chemical processes, such as nucleophilic substitution,1,10–19 cycloaddition,20–33 oxidative addition,34–49 isomerization,50–55 and many other processes from organic56–69 and organometallic70–91 chemistry.

The activation strain model has been around for a number of years and has been reviewed before.3,5 In this advanced review, we will therefore place extra emphasis on how this model connects to the well-known, and even more mature, framework of quantitative (frontier) molecular orbital (MO) theory.92–97 We limit our discussion primarily to SN2 and oxidative addition reactions (Figure 1(a) and (b), respectively), but discuss related chemical processes when relevant. Also, we limit ourselves to analyses of gas-phase results, to avoid complications introduced by solvation effects. These complications arise from the fact that one has to take into account the partial desolvation that occurs at the interacting sites of approaching reactants. It should be noted that this is inherent to any fragment-based description, and that solutions have been proposed.41,98

FIGURE 1

Schematic representation of (a) oxidative addition and (b) SN2 reaction mechanisms.

THE ACTIVATION STRAIN MODEL

The activation strain model1–3,5 is a fragment-based approach, and the first necessary step is therefore to choose suitable reference fragments. There is freedom of choice, but for the bimolecular processes that we will be dealing with in the following sections, the obvious and straightforward choice is to use the initial reactants as reference fragments. For bimolecular reactions via a transition state (TS), the activation strain model can be used to obtain insight into, for example, the height of a reaction barrier. This is done by splitting its energy at this point, ΔE‡, into the strain energy term ΔE‡strain, and the interaction energy term ΔE‡int:

The strain energy ΔE‡strain is the energy required for the geometrical deformations of the fragments from a reference geometry (often, but again not necessarily, their equilibrium geometry) to the geometry they acquire at the transition state. It is therefore strongly related to the structural rigidity of the fragments. As the reference geometries are usually not distorted, this term is typically destabilizing. In principle, the strain term can also incorporate excitations to electronic configurations that are better suited or required for the interaction studied, but most often the reference fragments are chosen as already having the correct valence configuration. The strain term can readily be split further into separate contributions from each reactant.

The interaction energy ΔE‡int accounts for all chemical interactions as they arise when the structurally deformed reactants are brought from infinity to their positions in the transition state geometry and allowed to interact. It comprises all the energetic effects that result from combining and mixing the charge distributions of the fragments. Often, this term is further dissected using an energy decomposition scheme, of which many are available. We will elaborate on this term in the next section.

The activation strain model can be generalized to any point along an energy profile. The relative energy ΔE, as well as its components, then becomes functions of the reaction coordinate ζ and Eq. (1) generalizes to

When applied to an energy profile of a chemical reaction with a central reaction barrier, all terms start at a value close to zero, but not necessarily at zero. This is because a reaction (in the gas-phase) typically starts from a precursor complex, in which the fragments are slightly distorted (small ΔEstrain) and interact only weakly (small ΔEint). From there on, the reactants become increasingly deformed along the reaction coordinate, leading to a continuously increasing strain energy ΔEstrain. Concomitantly, the interaction between the fragments usually strengthens, which leads to the interaction energy ΔEint becoming more stabilizing along the reaction profile. At the point where the destabilization from the strain term increases at the same rate as the stabilization from the interaction energy term strengthens, that is, dEstrain(ζ)/dζ = −dEint/dζ, the derivative of the total energy profile with respect to the reaction coordinate is zero (dE/dζ = 0). At this point, the energy profile achieves either a maximum (the reaction barrier), where the transition state occurs, or a minimum.

It follows that, to elucidate heights of reaction barriers or stability of stationary points, one should not only consider the rigidity of the fragments and the strength of their mutual interaction, but also their position along the reaction coordinate, and therefore the slopes of the strain and interaction terms.2 Depicted in Figure 2 is a comparison of the activation strain analyses of two generic chemical reactions to exemplify this. In this comparison, the strain curves ΔEstrain(ζ) resulting from both reactions are chosen to be equal, while the interaction curves ΔEint(ζ) are different. From Figure 2, it is easily concluded that, upon going from the first reaction (black lines) to the second reaction (red lines), the energy profile ΔE(ζ) is shifted up in energy, due to a weaker interaction between the fragments. The result is a higher reaction barrier, which is shifted to the product side because the interaction energy curve is descending less steeply. This is in agreement with the Hammond postulate,99 which indeed follows naturally from the activation strain model. Note, however, that an analysis at the transition state geometries only, as indicated by the dashed lines, can be misleading, as in this case one would conclude that the reaction barrier becomes higher due to a significant increase in strain energy, and even despite (!) a slightly more stabilizing interaction between the fragments.

FIGURE 2

Comparison of generic activation strain analyses of two different reactions. The dashed lines indicate the results from a single-point analysis at the transition state geometry.

Although analyses along full reaction paths (or critical sections thereof) are more insightful than single-point analyses at the transition state only, there are still a number of important factors to take into account in order to avoid misleading results. First, analyses of two similar reactions are more readily compared when the energy profiles are projected onto a critical geometrical parameter. Again, one can choose freely, but to arrive at insightful results, the parameter should be well-defined along the reaction profile and be sufficiently descriptive for the overall reaction process, as well as undergo considerable changes in the transition state region.100 Second, the total energy profile is the sum of two contributions that are not orthogonal and thus influence each other. The strain term, for example, is almost always positive as a consequence of its very definition. Eliminating the strain term by freezing the geometries of the fragments and pushing them toward each other, however, would not lead to a lower total energy profile. Instead, the interaction term would weaken and, as it includes a repulsive component as well, eventually become repulsive, likely raising the relative energy profile to higher values than the initial energy profile that was obtained with relaxed geometries. Thus, a significant part of the interaction energy that is built up during the reaction, requires a certain amount of geometrical deformation, and thereby strain energy. This balance between mutually dependent terms in itself is not problematic, but should be kept in mind when applying any model that contains interacting and opposing components. To get insight into the importance of this interplay, it can be useful to use additional analyses where (part of) the geometries are fixed.10,11,44,87,101–104 This prevents any perturbation stemming from geometry changes of the fragments that could easily hide a more clear picture of the electronic interactions.

In this review, we will focus exclusively on bimolecular processes, but it is worth mentioning that the activation strain model can also be used to study unimolecular processes.51,67,86,101 When chemically meaningful fragments can be discerned, the relative energy ΔE of the molecular structure is given by the sum of the change, during the process, in strain energy within the fragments and the change in interaction between these two fragments: ΔE(ζ) = ΔΔEstrain(ζ) + ΔΔEint(ζ).

Furthermore, it should be noted that an understanding of the energy profile of a certain process does not necessarily lead to a complete picture of the chemical phenomenon being investigated. Obviously, the activation strain model does not aim to explore, let alone explain, dynamic effects on the reactivity. Nor does it render unnecessary the search for alternative reaction pathways, or competitive reactions.

MOLECULAR ORBITAL THEORY AND INTERACTION ENERGY DECOMPOSITION

The activation strain model provides great insight into relative energies of transition states and even entire reaction energy profiles, as it enables us to ask the very relevant question why a certain geometrical deformation leads to an energetic destabilization, or why molecular fragments can build up a particular mutual interaction. Thus, to achieve a genuine explanation of the phenomena of interest, the reasons behind these changes in strain and interaction energy can be subjected to further investigation using quantitative (Kohn–Sham) MO theory.94–97,107,108

As discussed in the previous section, the strain energy of a fragment is the energy needed for the geometric deformations of the fragments, relative to a reference geometry. As this reference geometry is typically the equilibrium geometry of the fragment, the amount of strain energy is often directly related to the amount of geometrical distortion, and can be readily linked to the extent to which, for example, bonds are stretched or angles have changed. Further explanation is therefore not always needed. However, when required, MO theory can help to understand why a certain geometric deformation leads to a less stable molecular species. This is because changes in the total energy of the molecular fragment tend to parallel the changes in the sum of its orbital energies, and the orbital energies are again altered by changes in the molecular geometry. Thus, by investigating the dependence of the orbital energies on a geometrical parameter of interest, as is done in Walsh diagrams, one can explain why a certain molecular deformation leads to a destabilization of the molecular fragment. In a subsequent step, one can of course divide the fragment itself into smaller fragments, and provide an explanation for, for example, a rise in the orbital energy in terms of a decreased in-phase or increased out-of-phase overlap of the orbitals of the smaller fragments. This process can be repeated until the explanation is provided in terms of atomic orbitals, which no longer have any geometry dependence. Luckily, a satisfactory level of understanding is usually achieved at an earlier stage, based on the transferability of properties of common functional groups.

To obtain insight into the interaction energy ΔEint, it is usually split into separate terms arising from different types of interactions, in order to get a quantitative idea of their contributions to the total interaction energy. Such an interaction energy decomposition scheme is a useful tool to get insight into the relative importance of the different types of contributing interactions, and many different varieties have been developed.109–114 Any decomposition of the interaction energy into separate terms is artificial, so in principle one cannot go wrong when choosing one (Box 1). Some schemes, however, may be more suitable than others for certain applications. In the following, we will discuss the energy decomposition analysis (commonly abbreviated EDA) as implemented in the Amsterdam Density Functional (ADF) software package.107,115,116 We choose this scheme for its transparent, easy-to-understand nature, as it dissects the interaction energy into terms that directly correspond with a causal bonding mechanism in MO theory. This last step is of crucial importance, because interaction energy decomposition schemes are just quantitative tools, and numerical data provided by such schemes should not be interpreted as the final answer to a question, nor be presented as such.

BOX 1 THEORIES, MODELS, AND THE SCIENTIFIC METHOD. A SEMANTIC INTERMEZZO

Here we shortly discuss the meaning of the terms ‘theory’ and ‘model’, which feature prominently in the title of this review, and appear regularly throughout its contents. The following is a word of caution to keep in the back of one's mind, and certainly not an attempt to provide a definitive view on theories and models and their application within the scientific method.

Colloquially, the word ‘theory’ often refers to a hypothesis, or just a speculative idea.105 But in science, a ‘theory’ is more aptly described as ‘a well-substantiated explanation of some aspect of the natural world, based on a body of facts that have been repeatedly confirmed through observation and experiment’.106 Clearly, the term gets an entirely different meaning when changing context! This is why it is nonsensical to state that, for example, ‘the theory of evolution is just a theory’, as it fallaciously equivocates on the two different meanings. But how does this relate to the field of ‘theoretical’ chemistry, or density functional ‘theory’, or molecular orbital ‘theory’? Here, we encounter a third meaning of the term, where it refers to a mathematical framework, derived from a set of postulates, which is intended to make predictions of physical results.

The data obtained using, for example, density functional theory, can be interpreted using models. A model, within scientific context, can be described as an ‘idealized description of a particular system, situation, or process, often in mathematical terms, that is put forward as a basis for theoretical or empirical understanding’.105 Within the field of theoretical chemistry, there are many discussions centered around the use of models. Given that a model is an idealized, often simplified, description, and therefore inherently ‘false’, it should be judged on the basis of its usefulness: the quality of its predictions, general applicability, ease of understanding, and revelation of causal relationships, etc. Unfortunately, however, many of the discussions focus on particular individual components of a model. The main cause of the debates is therefore often not the model itself, but rather the overinterpretation of its results.

The EDA approach is based on that of Morokuma117,118 and the extended transition state (ETS) method developed by Ziegler and Rauk.119–121 Within this scheme, the interaction energy ΔEint is decomposed into three terms, that can be interpreted physically meaningfully and quantitatively accurately in the framework of the MOs arising from Kohn–Sham density functional theory:

Similar to the generalization of the ΔEint term to any point along an energy profile (as discussed in the previous section, see Eqs (1) and (2)) also this equation can be generalized to the entire reaction profile, again making each term a function of the reaction coordinate ζ.

For the discussion of the individual terms contributing to the interaction energy ΔEint, the formation of AB is considered from two fragments, A and B, which, as discussed in the previous section, already have the geometry and electronic configuration corresponding to the combined complex AB. These fragments have electronic densities ρA and ρB, with corresponding wavefunctions ΨA and ΨB and energies EA and EB. The first term, ΔVelstat, is the classical electrostatic interaction between the fragments as they are brought from infinity to their positions in the complex AB, giving rise to the sum density ρA+B = ρA + ρB, and corresponding Hartree product wavefunction ΨAΨB. It consists of the Coulombic repulsion between the nuclei α and β (at positions R, with charge Z) of the fragments A and B, respectively, as well as the repulsion between their unperturbed electron densities ρA and ρB, and the attractive interactions between the nuclei of one fragment with the electron density of the other fragment:

It is known from elementary electrostatics that two interpenetrating charge clouds have a repulsion that is smaller than the one between point charges at their centers, from which follows that fragments consisting of electronic densities around positive nuclei will typically experience a net attraction. Thus, ΔVelstat is usually attractive for molecular fragments at chemically relevant distances. It is important to remember that this term is computed from frozen electron densities ρA and ρB, obtained by optimizations in absence of the other fragment.

The Pauli repulsion, ΔEPauli, is the energy change that occurs upon going from the product wavefunction ΨAΨB to an intermediate wavefunction Ψ0, that, after antisymmetrization by an operator  and renormalization by a constant N, properly obeys the Pauli principle: Ψ0 = N {ΨAΨB}. This intermediate state, with density ρ0, has energy E0, such that ΔE0 = E0 − EA − EB = ΔVelstat + ΔEPauli and ΔEPauli = ΔE0 − ΔVelstat. The Pauli repulsion comprises the repulsive interaction between electrons having the same spin. It is responsible, for example, for the four-electron destabilizing interactions between doubly occupied orbitals from the different fragments. This is the origin of steric repulsion: when two occupied valence orbitals from different fragments overlap, antisymmetrization results in a nodal plane. The large gradients in this region significantly increase the kinetic component of the orbital energies. In the past, the Pauli repulsion term has been criticized for being based on an ‘arbitrarily chosen, nonphysical reference state’, and resulting from ‘first violating the Pauli principle and then imposing it’.122,123 Such statements may, at times, be relevant reminders for the overenthusiastic user of the decomposition scheme, but serve no further purpose as they themselves are examples of overinterpretation: all terms are defined as they are, and produce the numbers that follow from these clear and transparent definitions (Box 1).

In the final step of the bond formation between fragments A and B, the system is allowed to relax from Ψ0, and corresponding ρ0, to the final ΨAB and optimized density ρ of the molecular complex AB. The accompanying energy change is the orbital interaction term: ΔEoi = EAB − E0. This term is by definition stabilizing, because it involves an optimization. More specifically, it allows the virtual orbitals on the fragments to be mixed in, and therefore includes HOMO–LUMO (highest occupied molecular orbital and lowest unoccupied molecular orbital, respectively) interactions. As a result of this mixing, the orbital interaction component contains the stabilizing contributions from polarization of the fragments A and B, as well as charge transfer between the fragments. It is hard, if not impossible, to rigorously distinguish polarization from charge transfer. This is therefore not attempted in this interaction energy decomposition scheme (in contrast to the scheme of Morokuma117,118). However, when using a fragment-based approach and further orbital analyses, it is possible to get insight into the two individual contributions at least partially. Both polarization and charge transfer will show up as occupied-virtual orbital mixing, but in the case of polarization, the occupied and virtual orbitals will be localized on the same fragment, whereas charge transfer will show up as the mixing of occupied orbitals on one fragment with unoccupied orbitals on the other fragment. Besides detailed orbital analyses, additional electron density analyses (based on atomic charge analyses, the deformation density, etc.) can also contribute to distinguish between polarization and charge transfer.

Furthermore, it follows from group theory that only orbitals of the same symmetry, that is, the same character under the available symmetry operations, can interact and mix. This allows for a further decomposition of the total orbital interaction energy ΔEoi into contributions from each irreducible representation Γ of the point group to which the molecular system belongs120:

Finally, when the functional is augmented with an explicit correction to account for dispersion interactions, the contribution ΔEdisp from this correction term is simply added to Eq. (3) as an additional component of the interaction energy ΔEint. In Figure 3, schematic orbital interaction diagrams are shown for the different types of orbital interactions that play a main role throughout this review.

FIGURE 3

Orbital interaction diagrams for the most commonly appearing interactions.

After this rigorous numerical treatment, explanations are often presented schematically, using generic hydrogen-like atomic orbitals and a few core concepts from MO theory, further aided by symmetry considerations derived from group theory. A number of elementary atomic properties, such as electronegativity and atomic radii, can easily be taken into account, using what is in essence a perturbative treatment. Thus, although the results are derived from state-of-the-art density functional theory computations and detailed analyses, the final explanation often shows the essence of the phenomenon of interest in a pictorial manner that is easy to memorize, to communicate, and to apply to new situations. Such representations nevertheless accurately account for the observations made, and provide important insights. In the past, this approach, and MO theory in general, has proven to be very powerful for explaining many observations made in a variety of fields in chemistry, limited not only to molecules, but also including solids.94–97,124

SELECTED APPLICATIONS

In this section, we take the reader through some representative examples from the scientific literature, in which the activation strain model plays a prominent role. We will draw parallels between different chemical processes, and, when appropriate, focus on important related aspects, such as a more detailed view on the structural properties of the transition metal catalysts applied in oxidative addition.

Palladium-Mediated C–H Activation

We start our discussion of applications with the oxidative addition of a methane C–H bond to an uncoordinated Pd metal center. This model reaction represents general Pd-catalyzed bond activation reactions that are of paramount importance in modern synthesis.125 Palladium-catalyzed methane activation starts from a weakly bound reactant complex, and proceeds via a reaction barrier in which the C–H bond is stretched, to a product complex where the C–H bond is effectively broken, and new Pd–C and Pd–H bonds are formed35,47,126 (see also Figure 1(a)). The energy profile is shown in Figure 4 in black lines (we return to the blue and red lines later). Here, the fragments are the reactants Pd and CH4, and the stretch of the activated C–H bond is chosen as the reaction coordinate.100

FIGURE 4

Activation strain analyses for the oxidative addition of methane to Pd (black), PdPH3 (blue), and Pd(PH3)2 (red). A dot designates a TS. Energies and bond stretch are relative to reactants.

It becomes clear from this graph that, as the reaction proceeds, the strain term ΔEstrain becomes monotonically and increasingly destabilizing. This is not surprising, as the reaction involves rupture of a chemical bond and the energy is projected on the amount by which this bond is stretched. The strain term in this case matches the methane C–H dissociation curve closely, because the only significant contribution to ΔEstrain stems from H3C–H stretching, which destabilizes the bonding σC–H orbital due to a decreased in-phase overlap of the CH3• sp3 lobe (note that we use the ‘sp3’ moniker only as a general description of this orbital's shape, not its exact composition) with the H• 1s orbital. The true dissociation curve is slightly lower in energy than the strain curve, because the CH3• moiety has more freedom to relax, whereas during the reaction there are additional small geometrical changes in the CH3• moiety induced by the presence of the Pd center, which occur to lower the Pauli repulsion between the fragments. The contribution to the strain term from the monatomic Pd fragment itself, however, is zero, as follows from the definition.

The interaction energy curve starts just below zero, because the moderate Pauli repulsion is compensated for by the sum of the electrostatic attraction and the orbital interactions. At this stage of the reaction, the stabilizing orbital interaction term originates primarily from electron donation from the occupied orbitals on the methane substrate to the empty 5s acceptor orbital on Pd. As the reaction proceeds, the Pd intrudes into the space of the CH4 fragment, strengthening all components of ΔEint. Firstly, ΔVelstat becomes more strongly stabilizing simply because the electronic densities of the fragments get closer to the nuclei of the other fragment. At the same time, the overlap between the occupied orbitals on both fragments increases, leading to a stronger Pauli repulsion ΔEPauli. The orbital interaction ΔEoi becomes more strongly stabilizing (more negative), primarily due to a different donor–acceptor interaction that comes into play. As the C–H bond is stretched and the σ*C–H LUMO is stabilized, the Pd fragment migrates to a position where one of its occupied 4d orbitals has favorable overlap with this σ*C–H orbital. This provides a backbonding interaction, where charge is being donated from the metal to the substrate, in accord with the oxidative nature of this reaction. Population of the σ*C–H orbital leads to C–H bond breaking, and the in-phase overlap of the Pd 4d orbital with this σ*C–H orbital results in the newly formed Pd–C and Pd–H bonds.

Proceeding from this starting point, we will introduce modifications to the reaction system, and compare the results to see how these modifications influence the different energy terms, and thereby the overall energetics of the reaction. We start with a brief study on ligand effects by investigating the oxidative addition of the methane C–H bond to PdPH3.

Ligand Effects on Palladium-Mediated C–H Activation

The energy profile of oxidative methane addition to monocoordinated PdPH3 is shown in Figure 4. The reaction profile starts from a similar reactant complex as for the bare Pd catalyst, but becomes progressively higher in energy (compare black and blue lines).47 Activation strain analyses reveal that this difference is caused by a weaker catalyst–substrate interaction, while the structural deformations are very similar. The latter is a straightforward result, as the substrate goes through the same geometrical transformation, and this is still the main source of strain energy ΔEstrain. The contribution from the catalyst is close to, but no longer exactly zero, as there is a minor contribution originating primarily from a slight elongation of the Pd–PH3 bond length, caused by a small donor–acceptor interaction between occupied orbitals on CH4 to the LUMO of PdPH3, which has antibonding character along the Pd–PH3 bond.

The more important consequence of introducing the PH3 ligand is the weaker ΔEint, which translates directly into a higher reaction barrier and less stable product complex. Note that the less steeply descending interaction energy curve translates not only in a higher reaction barrier, but also shifts the transition state to the product side. The weakening of the interaction energy is the net result of a weakening of all its components (results not shown), which can be explained by the effect of the PH3 ligand on the electronic structure of the Pd center. PH3 is a moderately π-accepting ligand, and as such depletes some of the electron density from the Pd center in PdPH3 via π backdonation. This not only leads to less electrostatic attraction between the PdPH3 fragment and the substrate, but also decreases the overlap of occupied orbitals on both fragments, which reduces Pauli repulsion and compensates for the decreased electrostatic attraction. However, the π backbonding from Pd to PH3 also stabilizes the d orbitals on Pd, most notably the dxz and dyz (with the Pd–P bond being the z-axis). This includes the d orbital that donates into the substrate σ*C–H orbital, which is one of the main contributing factors to the orbital interactions ΔEoi. Lowering the energy of the donating orbital leads to less stabilization when this orbital is mixed in with the virtuals of the substrate. The result is a diminished orbital interaction term, raising the energy profile. In more chemical terminology: stabilizing this occupied orbital leads to a smaller amount of charge donation into the antibonding σ*C–H orbital, thereby hampering the bond-breaking process.

Without going into details here, we like to add for completeness that this effect of π-accepting ligands is only valid for catalysts that primarily interact with the substrate via donation from their d orbitals. This is true for most Pd-based catalysts, and we have designated this the d regime, as opposed to the s regime. In the latter, catalyst activity depends primarily on the s-accepting capability of the catalyst, and ligands can have completely opposite effects.47

In practice, the catalyst is often a dicoordinated species, and therefore the addition of the methane C–H bond to Pd(PH3)2 has also been subjected to activation strain analyses (red lines in Figure 4). Introduction of the second ligand leads to significant changes in the energy profile: immediately in the beginning it raises approximately 20 kcal mol−1. After this initial increase, the profile is rather flat, but the reaction barrier is nevertheless much higher than for Pd and PdPH3, and the overall reaction is more endothermic as well. Note also that it occurs more to the reactant side than the transition state for PdPH3, that is, it shows anti-Hammond behavior. Again, we will discuss the strain and interaction energy components of the energy profile individually.

Starting with the strain energy, Figure 4 clearly shows that herein lies the origin of the sudden rise of the energy profile in the beginning of the reaction. Dissecting this term into individual contributions from the catalyst and the substrate reveals that the substrate deformation is rather similar for all reactions, as the character of its geometrical transformation is still very much unchanged. However, for the addition to Pd(PH3)2 a significantly larger catalyst contribution is found. Visual inspection of molecular geometries along the energy profile reveals that, in the beginning of the reaction, the phosphine ligands bend away to make room for the approaching substrate. This bending of the catalyst's bite angle is the origin of the increase in the strain term, and can be further rationalized using the Walsh diagram shown in Figure 5(a) for the general case of bisligated d10-ML2 transition metal complexes. Decreasing the L–M–L angle turns on the antibonding overlap of the ligand lone pairs with the dxz orbital, pushing it up in energy, while other orbitals are less strongly affected. The high-energy b1 orbital also has a second, smaller effect on the energy profile, which shows up in the interaction energy term. This curve descends more steeply for Pd(PH3)2, because of the better electron-donating capabilities of this destabilized orbital. Combined, these two effects reveal the origin of the observed anti-Hammond behavior; compared with PdPH3, the steeper descending interaction energy curve for Pd(PH3)2 shifts the transition state to the left, but it nevertheless occurs at a higher energy because the entire energy profile is, already from the beginning of the reaction, pushed up due to the additional strain energy that is required for bending away the ligands.

FIGURE 5

Simplified Walsh diagrams for bending ML2 complexes (a) without and (b) with π backbonding, as they emerge from Kohn–Sham MO analyses (+/− indicate bonding/antibonding). A more detailed scheme of the intermixing occurring for the a1 orbitals is available in the supporting information of Ref 87.

Bite Angles and Their Flexibility in Transition Metal-Mediated Bond Activation

An important class of ligands used is that of chelating phosphine ligands, where the coordinating phosphine sites are connected by a polymethylene bridge. By varying the length of this bridge, one can obtain catalysts with different P–Pd–P bite angles. It is well-known that a smaller bite angle leads to an improved rate for the oxidative addition step.127 Activation strain analyses on a series of chelating palladium-phosphine catalysts Pd[PH2(CH2)nPH2] have revealed43,44 the origin of this bite-angle effect. Selected results of these analyses are shown in Figure 6, where the results for Pd and Pd(PH3)2 are compared with the Pd complex with a chelating ligand, Pd[PH2(CH2)2PH2], having a bite angle of only 98°. From these analyses it can be concluded that the lower reaction barriers for complexes with smaller bite angles originate from a softer strain term. This is closely related to the effect on this component described for the Pd(PH3)2 catalyst in the previous section: when a catalyst complex with a smaller bite angle is applied, there is less need to bend away the phosphine ligands. The reduced need for catalyst deformation prevents the strain energy term to rise in the beginning of the reaction, as found for Pd(PH3)2. Note that the analyses clearly reveal that this geometric effect is the reason for the lower barriers, that is, the bite-angle effect on reaction barriers results from steric reasons. The stronger donation from the destabilized b1 orbital (see Figure 5) leads to a slightly improved catalyst–substrate interaction, but plays only a minor role.

FIGURE 6

Activation strain analyses for the oxidative addition of methane to Pd, Pd(PH3)2, and Pd[PH2(CH2)2PH2]. A dot designates a TS. Energies and bond stretch are relative to reactants.

Other studies have investigated the effect of using catalysts with metal centers other than Pd,42,47 such as the d10-ML2 catalysts Ni(PH3)2 and Pt(PH3)2 (see Figure 7). Although all three catalyst complexes are rather similar, activation strain analyses revealed intriguing differences in both the strain and interaction terms. First, the weaker interaction for the Pd catalyst results mainly from the electron-donating capabilities: Ni(PH3)2 is a better electron donor than Pd(PH3)2 due to its higher-energy d-derived orbitals, whereas Pt(PH3)2 has larger d-derived orbitals that provide better overlap with the substrate σ*C–H orbital. Second, the relatively low barrier for Ni(PH3)2 is also partly the result of a softer strain term, already in an early stage of the reaction, originating from the catalyst's contribution. Surprisingly, it appears that bending Ni(PH3)2 comes with a smaller energy penalty than bending the isoelectronic Pd(PH3)2 or Pt(PH3)2. Thus, there is a ‘bite-angle effect’, even though all three M(PH3)2 catalyst complexes have linear equilibrium geometries and their L–M–L angles are decreased to similar values in the course of the oxidative addition. This result indicates that the bite angle itself is not necessarily sufficient to predict the activity of the catalyst. Considering the bite-angle flexibility, that is the ease of decreasing the bite angle of a catalyst, gives better insights into catalyst activity.47

FIGURE 7

Activation strain analyses for the oxidative addition of methane to Ni(PH3)2, Pd(PH3)2, and Pt(PH3)2. A dot designates a TS. Energies and bond stretch are relative to reactants.

Yet, one may wonder why bending Ni(PH3)2 is more feasible than bending, for example, Pd(PH3)2. This issue has been addressed in a study on a large set of d10-ML2 catalyst complexes.46,87 The systematic approach taken revealed a number of interesting trends, which we will address using two representative series, namely Rh(PH3)2−, Pd(PH3)2, Ag(PH3)2+ in which the metal center is varied, and Pd(NH3)2, Pd(PH3)2, Pd(CO)2 in which the ligand is varied. It is found that the bite angle becomes less flexible when the metal center is varied from Ag+ to Rh−, and increasingly flexible from Pd(NH3)2 to Pd(CO)2. In fact, the most flexible complexes from these series even have nonlinear equilibrium geometries. Rh(PH3)2− was found to have an intrinsically bent L–M–L angle of 141.2°, whereas Pd(CO)2 has an L–M–L angle of 155.6°. Thus, as the metal center is varied from the poor electron-donating Ag+ to Pd and the excellent electron-donating Rh−, there is enhanced bite-angle flexibility. In addition, when the ligand is varied from the poor π-accepting NH3 to the moderate π-accepting PH3 to the excellent π-accepting CO, we find increased bite-angle flexibility. Detailed MO analyses revealed that π backbonding is indeed of decisive importance for the observed bite-angle flexibility. This follows from a close inspection of the Walsh diagrams shown in Figure 5. Figure 5(a) shows the Walsh diagram for bending ML2 complexes, as commonly found in textbooks.97 In this diagram, the ligands are considered pure σ donors. We have extended this with a second Walsh diagram (shown in Figure 5(b)), for the case of π-accepting ligands. In the latter, bending results in a significant π-backbonding stabilization of the lower a1 and b1 orbitals (indicated in red). If this effect is sufficiently strong, it compensates the destabilization of the dxz-derived b1 orbital, and results in nonlinear equilibrium geometries. This is the case for Rh(PH3)2− and Pd(CO)2, and indeed even more so for Rh(CO)2−, which has a ligand–metal–ligand angle of only 130.8°. For Ni(PH3)2 and Pd(PH3)2, the effect is not strong enough to lead to a nonlinear equilibrium geometry, but the additional stabilization of the lower a1 and b1 orbitals at smaller bite angles nevertheless softens the resistance against bending. This effect is stronger for Ni(PH3)2, because the nickel metal center (which, in the complex, has a d10s0 valence configuration) has energetically higher d orbitals, and therefore stronger π backbonding with the PH3 ligands.

Oxidative Addition of Different C–X Bonds to Palladium

So far, we have discussed examples in which the catalyst complex is modified, while keeping methane as the model substrate. Numerous studies have focused on variations in the substrate.2,35–37,41,128 We will discuss the addition of different bonds to bare Pd, that is, besides the methane C–H bonds, also the C–X bonds from the series of halomethanes CH3F, CH3Cl, and CH3Br.41 Initially, one might expect that the barriers for Pd-mediated C–H and C–F activation are rather similar due to the similar bond dissociation energies of these bonds, and that the barriers for activating the weaker CH3–Cl and CH3–Br bonds are lower. Thus, in terms of the activation strain model, these expectations would show up in the strain energy term. Figure 8 confirms these expectations, although only partly.

FIGURE 8

Activation strain analyses for the oxidative addition of methane and halomethanes to Pd. A dot designates a TS. Energies and bond stretch are relative to reactants.

Along the reaction coordinate, the strain energy curves are indeed lower for addition of the CH3–Cl and CH3–Br bonds, which is a direct consequence of these bonds being weaker than the CH3–H and CH3–F bonds, which have roughly similar homolytic bond dissociation energies. For all four reactions, the strain curve is reminiscent to the dissociation curve for the bond being activated. However, this does not directly translate into the C–H activation barrier being similar to that of C–F, and both being higher than the barriers for C–Cl and C–Br activation. In fact, the methane addition barrier is lower than that of fluoromethane and closer to those of chloromethane and bromomethane. As shown in Figure 8, this is because early during the reaction the interaction energy curve for C–H activation is more stabilizing than the interaction energies for carbon–halogen bond activation.

While the Pd–CH4 interaction gains in strength immediately at the beginning of the addition process, the interaction curves for the halomethanes ‘lag behind’, that is, are initially weaker but later on descend more steeply and arrive at equal, and even lower values. This lag is caused by the different orbital-electronic structure of the halomethane substrates as compared with methane. For each reaction, a significant part of the interaction results from donation from the Pd d orbitals to the antibonding σ* orbital of the bond being broken. Shown below in Figure 9 are the different overlap situations for a Pd d orbital with (a) a C–H σ* orbital and (b) a σ* orbital of C–X and a stretched C–X bond. For methane, the σ*C–H orbital consists of the out-of-phase combination of the CH3• sp3 lobe (again, ‘sp3’ serves merely as a general description of this orbital's shape and is not used in its strictly formal sense) with the H• 1s orbital, while for the halomethanes the σ*C–X orbital is the out-of-phase combination of the CH3• sp3 lobe with the halogen X• np orbital. The latter has an additional nodal plane centered at the halogen X, leading to partial cancelation of overlap with the Pd d orbitals in the beginning of the reaction, and a diminished (orbital) interaction energy. This delay in the build-up of stabilizing interactions causes the C–F activation barrier to be much higher than the C–H activation barrier, despite the similar bond strength. Later on, at larger C–X bond stretch (around 0.5 Å), the ΔEint curve for C–X activation catches up, as the Pd d orbitals then also favorably overlap with the σ*C–X orbital, leading to a strengthening of the interaction energy. This effect shows up not only when going from the C–H to the C–X bonds, but also along the C–X bonds from C–F to C–Cl to C–Br, which have increasingly stabilizing interaction energy curves. For the larger halogen atoms, less bond stretch is needed before good overlap with the Pd d orbital is achieved, and therefore the lag in the interaction energy term becomes less pronounced along this series. This is furthermore accompanied by the increased electron-accepting capability from the σ*C–F to the σ*C–Br orbitals, due to a lowering of the orbital energy along this series.

FIGURE 9

Different overlap situations for the metal d orbital with (a) a carbon–hydrogen bond and (b) a carbon–halogen bond.

Thus, using activation strain analyses, the variations among the reaction barriers for C–H and C–X activation are readily explained using a combination of the bond dissociation energies of the targeted bonds, as well as the orbital-electronic structure of the substrate. From these results, it is readily concluded that (and understood why!) it is often insufficient to consider only the bond dissociation energies of the bonds. For the series discussed here, this would give qualitatively correct results, that is, the order of the reaction barriers would be predicted correctly. However, it is clear that a quantitative prediction of the reaction barriers, based solely on the BDE of the targeted bonds, would fail rather dramatically. When comparing the methane C–H bond and the ethane C–C bond, the situation would be even worse, as a prediction based on the bond dissociation energies would also fail qualitatively: although the C–C bond is weaker, cleaving it via oxidative addition to Pd goes with a higher barrier.35 Also for this bond, with a σ*C–C orbital that is the antibonding combination of two CH3• sp3 lobes, this is due to the delayed build-up of stabilizing catalyst–substrate interactions. In a notable study on Pd-catalyzed cross coupling reactions involving a chloro aryl triflate as a coupling partner, Schoenebeck and Houk investigated the differences in electronic structure of C–Cl bond and C–OTf bonds and, combined with the ligand effects discussed above, were able to explain the regioselectivity observed.129

In the following, we will consider a different elementary chemical reaction in which halomethane C–X bonds are broken, namely, (backside) nucleophilic substitution, and draw a number of parallels derived from applying the activation strain model to both reactions and interpreting the results with MO theory. In principle, one could also describe the oxidative addition process as a frontside nucleophilic attack of the transition metal center at carbon, but for the sake of brevity we will not discuss this aspect here, and refer the reader to previous works dealing with this topic.3,5,35,37,130

Leaving-Group Ability and Nucleophilicity in SN2 Reactions

To arrive at a straightforward relationship between the electronic structure of the reactants and their SN2 reactivity, a number of energy profiles have been investigated while systematically varying one of the reactants.13 In Figure 10(a), we show the first of two representative series that we will discuss, namely the backside nucleophilic attack of Cl− on CH3X substrates, where the leaving group X in the substrate is varied along the halogens F to I. This gives the halomethanes CH3F, CH3Cl, CH3Br, and CH3I, thus resembling the series of C–X bonds in the last example of oxidative addition. Also in these SN2 reactions, the C–X bond is broken by populating the σ*C–X orbital, but now the electrons are being donated from the chloride nucleophile, which approaches the methyl moiety from the back, and not side-on like the Pd metal center in oxidative addition (the original study, Ref 13 also includes such frontside nucleophilic attacks).

FIGURE 10

Activation strain analyses of the SN2 reaction profiles for variation of (a) the leaving group and (b) the nucleophile. A dot designates a TS. Energies and bond stretch are relative to reactants.

It follows from these analyses that the leaving-group ability is directly determined by the C–X bond strength. Again, as from CH3F to CH3I the C–X bond strength decreases, the strain terms become less destabilizing, but now this directly translates into lower reaction barriers and more exothermic reactions. The differences between the interaction energy curves that were observed for the oxidative addition of these bonds (due to lagging, see Figure 8 in the previous section) are not observed here, because the nucleophile approaches from the back and has comparable overlap with the back-lobe of the σ*C–X orbital for each substrate.

In Figure 10(b) the SN2 energy profiles are compared for the reactions of different nucleophiles (F−, Cl−, Br−, and I−) with CH3Cl. Along this series the leaving group is kept constant, which makes all strain curves essentially coincide. But, as the nucleophile is varied from F− to I− the interaction term becomes significantly less stabilizing, resulting in higher, more product-like reaction barriers, and less exothermic reaction energies. This is not only caused by the more diffuse charge distributions of the larger halogens, but also a direct consequence of the electron-donating capability of the nucleophiles. This capability is reduced because of the decreasing np orbital energies from F− to I−, and therefore larger HOMO–LUMO energy gap with the substrate σ*C–Cl orbital.

The analyses on the trends for leaving-group ability and nucleophilicity readily explain what makes for a good leaving group, or a good nucleophile, and how and why these properties translate into a lower SN2 reaction barrier. Combined, by using for example a good nucleophile (such as F−) and a substrate with a good leaving group (such as CH3I), the barrier even vanishes entirely. However, for most SN2 reactions Nu− + CH3X, a transition state occurs, due to the steric congestion that arises as five substituents try to bind to the central carbon atom.10,104 This steric congestion can be lowered by going to a larger central atom, such as silicon, which forms stable five-coordinate compounds. Increasing the steric congestion by introducing bulky substituents R in Cl− + SiR3Cl can cause SN2 reaction barriers to reappear.11,131 Similar results have been obtained for SN2 at phosphorus centers.10,132

From Nucleophilicity to Halophilicity and Protophilicity

Proceeding from the Cl− + CH3Cl reactions, one can also replace the central CH3• moiety with an isolobal halogen atom. This reduces the steric interactions with the incoming base substantially, and results in a disappearance of the reaction barrier. Thus, we arrive at a set of halogen bond formation reactions between halides X− and dihalides YZ, which form stable halogen-bonded trihalides [X–Y–Z]−. These halogen-bond formations have also been subject of detailed activation strain analyses and interpretation in terms of MO theory (X, Y, Z = F, Cl, Br, or I).133 Like oxidative addition and nucleophilic substitution, also halogen-bond formation can be described as a donor–acceptor interaction, stemming to a large extent from charge being transferred from a Lewis base to the halogenated compound.133–137 For the series studied, this stems from HOMO–LUMO interactions between the np lone pair on X− and the σ* orbital on YZ. Not surprisingly, the results from this study therefore reveal many parallels to those discussed for SN2 reactions between halides and halomethanes. Thus, stronger Lewis bases (i.e., better nucleophiles) form stronger halogen bonds due to stronger interactions, partly due to their better electron-donating capability. Also, halogen bonds X−···YZ become stronger as the YZ dissociation curves become softer, as this induces less deformation (lower ΔEstrain curve), while at the same time a larger stretch of this bonds leads to a greater stabilization of the σ*Y–Z orbital, and thereby enhances its electron-accepting capability.

One can go a step further and replace the central halogen atom with a hydrogen atom, to arrive at an analogous set of hydrogen-bond formations X− + HY→[X–H–Y]−. Again, an important driving force behind these reactions is the HOMO–LUMO interaction between the np lone pair on X− and the empty σ* orbital on HY, and the analyses revealed many parallels with SN2 reactivity and halogen bonds.133 The smaller hydrogen leads to even less steric repulsion than a larger central halogen. Interestingly, however, this does not necessarily mean that hydrogen bonds are stronger, because the σ*H–Y orbitals on hydrogen halides are higher in energy and therefore less good acceptor orbitals than those on dihalides. This leads to less stabilizing (orbital) interactions in the case of hydrogen bonds. Furthermore, as hydrogen halide H–Y bonds are much stronger than dihalide Z–Y bonds, the H–Y bonds typically stretch less, hampering the build-up of stabilizing interactions. Essentially, this comes down to a decreased leaving-group ability of the halogen Y when it is connected to a hydrogen, compared with the same halogen Y connected to another halogen atom.

Competition between SN2 and E2 Reaction Mechanisms

In a recent study,19 we compared the SN2 energy profiles for OH− + CH3CH2OH and H2O + CH3CH2OH2+ (see Figure 11, blue lines). This constitutes the effect of changing simultaneously from a model system with a good nucleophile and a poor leaving group (OH−) to a model system with a poor nucleophile and an excellent leaving group (H2O), by simply protonating both moieties. In agreement with the results discussed above, going from the OH− to the poorer H2O nucleophile weakens the interaction energy significantly, but this effect is almost entirely compensated for by the softer strain energy term, as simultaneously the leaving-group ability is enhanced from CH3CH2OH to CH3CH2OH2+. The net result is only a small difference in reaction barrier, with the barrier for the protonated reaction being a few kcal mol−1 lower.

FIGURE 11

Activation strain analyses of the model substitution and elimination reaction profiles under (a) basic (OH− + CH3CH2OH) and (b) acidic (H2O + CH3CH2OH2+) conditions. A dot designates a TS. Energies and bond stretch are relative to reactants.

The main reason to include this study here, however, is to provide an example on how the activation strain model is applied to compare two different reaction mechanisms that are in mutual competition. This study aimed specifically at revealing the origin of the often observed shift from elimination to substitution mechanism when the reaction conditions are changed from basic to acidic.

We have compared the model SN2 reactions between OH− and CH3CH2OH and between H2O and CH3CH2OH2+ with the competitive base-induced 1,2-elimination reactions for the same pairs of reactants. The resulting four reactions resemble the substitution and elimination reactions under extremely basic (OH− as nucleophile or base) and acidic conditions (H2O as nucleophile or base). The results revealed indeed the experimentally observed shift from an elimination to a substitution pathway when changing from basic to acidic conditions. From an activation strain model perspective, the elimination pathway is generally expected to go with higher reaction barriers, due to the greater deformation of the reactants. In both SN2 and E2 reactions, the C(α)–LG bond is broken, but E2 requires the additional rupture of a C(β)–H bond, where α and β refer to the positions of the carbon atoms relative to the leaving group LG. Only when the interaction energy is strong enough to overcome the extra strain energy, the elimination pathway (via an E1cb mechanism) becomes favored. For the strong base OH− this is the case, but not for the much weaker base H2O, as shown in Figure 11. Owing to the significant loss of stabilizing interactions (ΔEint curves) upon going from the elimination pathway under basic conditions to the acidic conditions, the preference shifts from protophilic attack for OH− to nucleophilic attack for H2O.

There is, however, yet another reason behind this shift in preferred mechanism, caused by the different nature of the leaving groups present under both circumstances. This can be derived from a careful investigation of the electronic structure of the CH3CH2OH and CH3CH2OH2+ substrates. In Figure 12, schematic MO diagrams are shown for these substrates, constructed from the fragments CH3• and •CH2OH or •CH2OH2+.

FIGURE 12

Schematic substrate LUMO composition for (a) CH3CH2OH and (b) CH3CH2OH2+.

In each reaction, the donor–acceptor interaction between an np HOMO on the nucleophile/base and the substrate LUMO plays a key role. We will therefore focus exclusively on the LUMO for our discussion of the electronic structure of the substrates. First, note that for CH3CH2OH the LUMO is, as commonly encountered, the π-bonding combination of the lowest unoccupied orbitals on both fragments, that is, the e*(β) orbital on CH3•, which has antibonding character along the C–H bond targeted by protophilic attack, and the e*(α) orbital on the •CH2OH fragment, which is antibonding along the carbon-leaving group bond. From here, it is easily seen that small variations in the leaving group will lead to small variations in the e*(α) orbital on the C(α) side of the substrate, and thereby change the relative amplitudes of the e* fragment orbitals in the overall LUMO. Thus, by stabilizing the e*(α) orbital, the LUMO gains amplitude on this side and reactivity is shifted toward SN2.1 However, this perturbative treatment breaks down when the e*(α) orbital is strongly stabilized due to the positive potential induced by the additional proton after protonation. As shown in Figure 12(b), the much lower e*(α) fragment orbital on •CH2OH2+ interacts no longer with the CH3• e*(β), but instead with the bonding e(β) orbital. The resulting substrate LUMO is the π-antibonding combination, which has bonding character along the C–H bond targeted in the elimination pathway. Obviously, populating this bonding orbital does not contribute to β-proton transfer to the base, as required in the elimination pathway. The two extreme situations sketched above constitute an orbital electronic basis for an E2H–SN2 mechanistic spectrum that may have E2C mechanisms somewhere in between.1,138,139

CONCLUSION

When, after 7.5 million years of computing the supercomputer ‘Deep Thought’ finally gives Loonquawl and Phouchg the answer to ‘The Great Question of Life, The Universe and Everything’, the two men are perplexed as the computer simply says ‘42’. So goes the story of Douglas Adams' classic novel ‘The Hitchhikers Guide to the Galaxy’.140 Many computational chemists will, at least to a certain extent, be familiar with such situations. Nowadays it is relatively easy to run advanced computer simulations of chemical processes, and collect numerical data. The harder, but more interesting task of a computational chemist, is to give meaning to the numbers produced.

The activation strain model of chemical reactivity provides both qualitative and quantitative insight into energy profiles (ΔE) of chemical processes by splitting the energy into a term stemming from geometrical deformation of the reactants (ΔEstrain) and a term accounting for the mutual binding capabilities between the reactants (ΔEint). We have described, for example, how this approach has helped to separate the electronic effect from the steric effect of ligands in transition metal-catalyzed oxidative addition reactions, and understand both. This has lead to uncover that the bite-angle effect for these reactions is steric in nature. Furthermore, by comparing the activation of different bonds, it was shown that the barrier height for this process depends to a great extent on the catalyst's bonding capability, thereby explaining why the activity of a catalyst toward a certain bond is poorly predicted when only the strength of the targeted bond is considered. Interestingly, for bond rupture via an SN2 mechanism, the bond strength can be used to predict reaction barrier heights, because these reactions are driven by a different orbital-electronic mechanism. Furthermore, applying the same methodology relates nucleophilicity of chemical species directly to basicity, and also to the stability of hydrogen- and halogen-bonded complexes in which these species participate.

Thus, these examples demonstrate how the activation strain model, combined with MO theory, provides insight into the essence of chemical reactivity and, importantly, how this yields powerful unifying concepts that enable rational tuning and design of molecular transformations across all branches of chemistry.