Defining and calculating steric strain in the first step of the dissociative alkene metathesis mechanism using Grubbs 1-type catalysts

(1)

Defining and calculating steric strain in the

first step of the dissociative alkene metathesis

mechanism using Grubbs 1-type catalysts

LM Botha

21112886

Dissertation submitted in partial fulfilment of the requirements for the

degree

Magister Scientiae

in

Chemistry

at the Potchefstroom Campus

of the North-West University

Supervisor:

Dr CGCE van Sittert

Co-supervisor:

Prof HCM Vosloo

(2)

(3)

Summary

Defining and calculating steric strain in the first step of the dissociative

alkene metathesis mechanism using Grubbs 1-type catalysts

During the dissociative alkene metathesis mechanism, a phosphine ligand dissociates from the ruthenium metal of a Grubbs type catalyst leading to the active 16-electron Grubbs 1-type catalyst with an open coordination site (vacant space) on the ruthenium metal and a dissociated phosphine ligand. A large increase in the energy is observed during the dissociation of a phosphine ligand from a Grubbs 1-type catalyst. This increase in energy can be a consequence of various factors, which includes steric strain. Therefore, in order to understand the energy increase observed during the dissociative mechanism, the steric strain in the Grubbs 1-type catalysts needs to be calculated.

Steric strain is a consequence of electron interaction and repulsion between atoms in close approximation to one another. Therefore, steric strain will be minimized when the molecules are in their ideal geometry. Moreover, steric strain affects the bonding angles and the close contact distance between atoms, causing an increase in the energy. That is why the size of the dissociating ligand, the size of the vacant space and the close contact distance between the atoms in various Grubbs 1-type complexes needs to be calculated.

Moreover, the size of the dissociating ligand and the size of the vacant space is an indication of the amount of space available around the ruthenium metal and the dissociating ligand. For instance, the steric strain will lessen if the space in the molecule increases, lessening the electronic repulsions between atoms. The hypothesis is that a decrease in the steric strain of a dissociating ligand will lead to an increase in the amount of energy needed for ligand dissociation. Furthermore, the close contact distance between non-bonded atoms has a relation to the size of the dissociating ligand, the size of the vacant space and the total energy in the complex.

In order to fully understand the effect that the steric strain has on the various Grubbs 1-type complexes during the dissociation step, a dynamic technique is needed. A dynamic technique is a technique that calculates the change in the steric strain along the pathway of dissociation. Having said that, the techniques currently available (the Tolman cone angle, the solid angle

(14)

and the percentage buried volume) to calculate the steric strain are insufficient because they were designed to calculate the steric stain of isolated (stationary) complexes.

Consequently, this study focuses on the development of a dynamic technique to calculate the steric strain in the first step of the dissociative alkene metathesis mechanism with various Grubbs 1-type complexes (Figures 3.1 to 3.5). For this reason, a computer program named SterixLB that uses four modified techniques, namely the modified Tolman, the outer pocket, the inner pocket and the inner-inner pocket techniques were created to calculate the steric strain of the various Grubbs 1-type complexes. In addition, the program also calculates the close contact distances that may contribute to the increase in the observed dissociation energy of Grubbs 1-type complexes.

However, dynamic data is needed to calculate the steric strain observed in the dissociation mechanism. Therefore, potential energy surface (PES) scans, where the bond length between the phosphine of the dissociating ligand and the ruthenium was extended stepwise (25 steps) from the minimum energy optimised bond length to a bond length of 5 Å were performed. These PES scans were performed with Materials Studio 6.0 from Accelrys. Consequently, the stepwise Cartesian coordinate data of the various Grubbs 1-type complexes was obtained from the PES scans. Furthermore, Cartesian coordinate data from the PES scans was used in SterixLB to calculate the size of the dissociating ligand, the size of the vacant space and the close contacts.

A good correlation was found between the Tolman cone angle and the results obtained with the modified Tolman, the outer pocket and the inner pocket techniques. The results found showed that the outer pocket technique was the best modified technique to calculate the outermost size of the dissociating ligand, while the inner-inner pocket technique was the best modified technique to calculate the innermost size of the vacant space. Also, the computer program named Solid-G (developed by Guzei et al.[1]) that uses solid angle calculations correlated with SterixLB, even though the- program uses different mathematics and calculates different results.

The results obtained with SterixLB indicated that both the electronegativity and the Bondi van der Waals radii of the substituents on the ligand influenced the sizes of the dissociating ligand and the vacant space. In addition, the size of the group on the carbene carbon

(15)

influenced the sizes of the dissociating ligand and the vacant space on the Grubbs 1-type complex.

As a result, the size of the dissociating ligand increased after dissociation, which indicated less steric strain was present in the Grubbs 1-type complex that allowed the dissociating ligand to find a less strained geometry. On the other hand, a decrease in the vacant space size indicated that the steric strain in the Grubbs 1-type complex has decreased. The steric strain in the Grubbs 1-type complex decreases since the dissociated phosphine ligand left an unoccupied space around the metal that allows the groups/atoms around the metal to change positions/orientations to find a less strained geometry.

Furthermore, both the energy and the dissociating ligand size increased in preparation of dissociation. On the other hand, the vacant space size decreased with the increase in dissociation energy.

In conclusion, understanding the first step of the dissociative mechanism for various dissociating groups can lead to the design of better catalysts that need less energy for dissociation. Furthermore, the maximum size of the vacant space, to accommodate the incoming alkene, could be determined.

Keywords: steric strain, Grubbs 1-type complexes, SterixLB, Tolman cone angle, pocket

(16)

(17)

Preface

This dissertation is the original, unpublished, independent work by the author, LM Botha under the supervision of Dr CGCE van Sittert and Prof HCM Vosloo.

The dissertation is based on developing modified techniques (modified Tolman, inner pocket, outer pocket and the inner-inner pocket techniques) and a program (SterixLB) that utilises these modified techniques to calculate the steric strain in various Grubbs 1-type complexes, using Cartesian coordinate data obtained from molecular modelling. Furthermore, the steric strain is calculated during the dissociative mechanism of an alkene metathesis reaction, where a ligand dissociates, leaving an open coordination site on the ruthenium metal. The steric strain is calculated for both the dissociating ligand and the open coordination site (vacant space) before, during and after ligand dissociation. Furthermore, the close contact distance between non-bonded atoms was calculated to describe the steric strain in Grubbs 1-type complexes.

This dissertation consists of a table of contents, a list of abbreviations and symbols, a summary, five chapters and an appendix.

Chapter 1 of the dissertation provides a general overview of alkene metathesis with Grubbs 1-type catalysts. In addition, the aims, objectives and methodology of the study are discussed. A literature study is included in Chapter 2. The literature study consists of a detailed introduction into alkene metathesis. An overview of the Tolman cone angle, the solid angle, the pocket angle and the percentage buried volume techniques is given. In addition, the computer program named Solid-G that calculates the solid angles of complexes is discussed. The Materials Studio settings used for the study, the discussion and equations of the modified techniques and the development of the computer program created with Matlab are discussed in Chapter 3.

In Chapter 4, the results obtained from calculating the steric strain using the Cartesian coordinate data obtained from molecular modelling software are shown. The modified techniques are evaluated by comparing them to the Tolman cone angle, solid-G and the percentage buried volume techniques for nickel complexes. Thereafter, the modified techniques are used to calculate the steric strain and close contact distances of various Grubbs

(18)

1-type complexes before, during and after the first step of the dissociative mechanism of an alkene metathesis reaction.

The conclusions derived from the knowledge obtained from the literature and from the results obtained are provided in Chapter 5.

1. Guzei, I.A. and M. Wendt, An improved method for the computation of ligand steric

(19)

Chapter 1: Introduction and objectives

1.1 Background

Alkene metathesis is a chemical reaction where the double bonds between two carbon atoms break and new double bonds are formed in the presence of a metal carbene catalyst.[1]The first well-defined metal carbene catalyst systems for alkene metathesis reactions are the ruthenium-based Grubbs catalysts (Figure 1.1).[1-7]

Cl P Ru P Cl C H

Figure 1.1: Grubbs 1st generation[1]

The alkene metathesis reaction can take place according to various mechanisms. However, in this study, the focus is only on the dissociative mechanism for alkene metathesis, as shown in Figure 1.2. The first step of the dissociative mechanism is where the metal carbene undergoes the dissociation of a ligand and therefore leaves an open coordination site on the metal. The second step of the dissociative mechanism is where an incoming alkene, in turn, coordinates to the open coordination site on the metal to form a metallacyclobutane intermediate.[1] During the first step, a large amount of energy[8] is needed for the dissociation of the ligand, as shown in Figure 1.3. This amount of energy correlates to the activity of the metal carbene catalyst and this activity is influenced by various factors. The most studied factors are the electronic effects, which arise from different electron distributions in the molecule,[9]and the steric effects.

(20)

Figure 1.2: Dissociative alkene metathesis mechanism[1]

Figure 1.3: Energy profile of the activation steps of the metathesis of ethene and 1-octene using RuCl2(PCy3)2(=CHPh)[8]

M C R2 H R1 X -X +X M C R2 H R1 _C _C H R3 H H C C H R3 H H C H R3 C R2 H C H H M R1 + C C H R3 H R2 -M C H H R1 M C R2 H R1

(21)

Steric effects can be classified into various types, depending on the structure of the molecule and the environment around the molecule.[10, 11] These effects influence the type of products that could form and the rate at which reactions occur. In the case where the reactions between the environment and the molecule is prevented by a large group in the structure of the molecule, it is called steric shielding. On the other hand, when repulsion occurs between non-bonded atoms in a molecule it is steric repulsion. However, steric attraction is when molecules are optimised in specific orientations or conformations to ensure reaction between the molecules.

Steric inhibition of resonance occurs in benzene rings, where the group on the ortho position (in benzoic acid) forces the carboxylic acid out of the plane. Whereas, in steric inhibition of protonation, a group cannot be protonated since it is removed from the benzene ring plane by the bulky COOH group.

If the change in conformation of a molecule induce the crossing of chains or rings passing through itself or other rings or chains, the steric effect is called chain crossing.

Steric strain is the steric effect that occurs as a consequence of two atoms in close proximity so that their electron clouds start to repulse one another, restricting the free movement of atoms.[10, 11] This restriction of free movement restricts possible bond formation, conformers and reactions, which can occur. The repulsion between the electrons of non-bonded atoms results in an increase of potential energy, which is an indication of increasing steric strain.[8, 12] In addition to steric strain, angular and torsional strain can also occur in complexes.[10, 11]

In literature, there are a number of techniques available to calculate the steric strain of ligands in organometallic chemistry. Among these available techniques are the Tolman cone angle[9], the solid angle[13], solid-G[14], pocket angle[15] and the percentage buried volume[16]techniques.

The Tolman cone angle[9] is known as the simplest and most used technique to calculate the steric strain of phosphine ligands, although the technique has numerous limitations. The first limitation is that the technique does not take the lowest energy ligand conformations into consideration. The second limitation is that the ligands have a tendency to mesh together (overlap of atom spheres) when they are forced into a cone.[9]Therefore, the solid angle[13] technique was developed as a means to address the limitations of the Tolman cone angle

(22)

technique through the use of mathematical calculations. The solid angle[13] technique calculates the size of a ligand projection on the surface of a sphere if the metal was replaced by a light source. However, because of the shadow projection, the solid angle technique has the disadvantage of being conformation specific (only allowing certain conformations).[13] On the other hand, the pocket angle [15]technique was designed to calculate the size of the interior angle in chelating phosphine catalysts in order to determine the structure activity of a palladium catalyst.[15]

However, none of the above mentioned techniques (especially the Tolman cone angle) could effectively describe the steric properties of NHC (N-heterocyclic carbene) ligands. The reason for this is their C2-symmetric shapes.[16] Therefore, the percentage buried volume

technique[16]was developed to calculate the steric strain of NHC ligands. The percentage buried volume technique calculates the amount of space occupied by NHC or PR3 ligands in

order to obtain steric strain data.[16]

Also, all of the above-mentioned techniques can only measure the steric strain of a complex at one point in time. Therefore, in order to obtain insight into the large amount of energy needed for the dissociation of the ligand in the first step of the dissociative mechanism, the change in steric strain must be calculated. This must be done continuously (dynamic studies) by calculating the size of the dissociating ligand, the size of the vacant space and the close contact distance as the ligand dissociates from the metal carbene complex.

In order to determine the dynamic change in steric strain during the dissociative mechanism with various Grubbs 1-type complexes, modifications were made to available techniques. As a result, the techniques created include the modified Tolman, the outer pocket, the inner pocket and the inner-inner pocket techniques. These modified techniques were automated with a computer program (SterixLB) to calculate the size of the dissociating ligand, the size of the vacant space and close contacts, which are indications of the change in the steric strain during ligand dissociation from the various Grubbs 1-type complexes.

Moreover, through a better understanding of the dissociative mechanism, it is possible to design dissociating ligands that can easily be removed without needing large amounts of energy, leading to the design of better catalysts.

(23)

1.2 Aim of study

The aim of this study is to create a computer program to accurately calculate the change in the ligand size and vacant space size, which is an indication of the steric strain, during the first step of the dissociative alkene metathesis mechanism using the output from molecular modelling software.

1.3 Objectives

The objectives of the study are to:

• Create a computer program that calculates the size of the dissociating ligand, the size of the vacant space and the close contacts dynamically by using potential energy surface (PES) scans of the dissociation of PX3 (X = H, F, Cl, Br, I and Cy) from the Grubbs

1-type complex as input data. The PES scans are obtained with Materials Studio 6.0 from Accelrys.[17]

• Evaluate the modified techniques within the computer program created by comparison with results obtained with currently available techniques used in literature.

• Use the computer program to investigate the size of the dissociating ligand, the size of the vacant space and close contact within various Grubbs 1-type complexes during the first step of the dissociative mechanism.

1.4 Methodology of study

The methodology of this study consisted of five sections, namely geometry optimisation, PES scans, development of a computer program, evaluation of the modified techniques within the computer program by comparison with existing techniques from literature, and the application of the computer program.

1.4.1 Geometry optimisation

The Cartesian coordinate data was obtained for various Ni(CO)3PX3 complexes (Figure 1.4)

and various Grubbs 1-type complexes (Figure 1.5) by geometrically optimising the complexes with density functional theory (DFT) within Materials Studio software.[17]

(24)

Ni P X X X C C C O O O

Figure 1.4: The nickel complexes where X = F, Cl, Br or H

Cl PX₃ Ru L Cl C H R -Cl Ru Cl C H R L PX3 +PX3

Figure 1.5: The Grubbs 1-type complex where G complex (L = PCy3, R = Ph), the B

complex (L = PX3, R = Ph) and A complex (L = PX3, R = H) where X = F, Cl, Br, I, H or Cy

1.4.2 PES scans

The geometrically optimised structure was submitted to potential energy surface (PES) scans using Materials Studio software. The Cartesian coordinate data and the related change in energy data during the dissociation step were obtained.

1.4.3 Evaluation of techniques

In this study, the available techniques, namely Tolman cone angle and pocket angle[15], were modified. These modified techniques were the modified Tolman, the outer pocket, the inner pocket and the inner-inner pocket techniques. These techniques were evaluated by comparing them to the Tolman cone angle[9], solid-G[14] and the percentage buried volume[16] techniques for the nickel and the Grubbs 1-type complexes.

1.4.4 Development of computer program

A computer program (SterixLB) was designed in Matlab that was capable of implementing modified techniques created in this study.

(25)

1.4.5 Application of computer program

The Cartesian coordinate data of the G, B and A complexes was imported from the PES scan data files and the sizes of the dissociating ligand, vacant space and the close contact distances between non-bonded atoms before, during and after the dissociation were calculated with the modified Tolman, the outer pocket, the inner pocket and the inner-inner pocket techniques developed in this study.

1.5 References

1. Dias, E.L., S.T. Nguyen, and R.H. Grubbs, Well-defined ruthenium olefin metathesis

catalysts: mechanism and activity. Journal of the American Chemical Society, 1997.

119(17): p. 3887-3897.

2. Trnka, T.M. and R.H. Grubbs, The development of L2X2Ru CHR olefin metathesis

catalysts: an organometallic success story. Accounts of Chemical Research, 2001.

34(1): p. 18-29.

3. Grubbs, R.H. and S. Chang, Recent advances in olefin metathesis and its application

in organic synthesis. Tetrahedron, 1998. 54(18): p. 4413-4450.

4. Chatterjee, A.K., et al., Synthesis of functionalized olefins by cross and ring-closing

metatheses. Journal of the American Chemical Society, 2000. 122(15): p. 3783-3784.

5. Schwab, P., R.H. Grubbs, and J.W. Ziller, Synthesis and applications of RuCl2 (CHR

‘)(PR3) 2: the influence of the alkylidene moiety on metathesis activity. Journal of the

American Chemical Society, 1996. 118(1): p. 100-110.

6. Scholl, M., et al., Synthesis and Activity of a New Generation of Ruthenium-Based

Olefin Metathesis Catalysts Coordinated with 1, 3-Dimesityl-4, 5-dihydroimidazol-2-ylidene Ligands §. Organic Letters, 1999. 1(6): p. 953-956.

7. Lehman Jr, S. and K. Wagener, In Handbook of Metathesis; Grubbs, RH, Ed. Wiley-VCH: Weinheim, Germany, 2003. 3: p. 283-353.

8. Jordaan, M., et al., Experimental and DFT investigation of the 1-octene metathesis

reaction mechanism with the Grubbs 1 precatalyst. Journal of Molecular Catalysis A:

Chemical, 2006. 254(1): p. 145-154.

9. Tolman, C.A., Steric effects of phosphorus ligands in organometallic chemistry and

homogeneous catalysis. Chemical Reviews, 1977. 77(3): p. 313-348.

10. Pophristic, V. and L. Goodman, Hyperconjugation not steric repulsion leads to the

(26)

11. Weinhold, F., Chemistry: A new twist on molecular shape. Nature, 2001. 411(6837): p. 539-541.

12. Hardinger.S, Steric strain notes. 2006.

13. White, D., et al., Solid angles III. The role of conformers in solid angle calculations. Journal of organometallic chemistry, 1995. 495(1-2): p. 41-51.

14. Guzei, I.A. and M. Wendt, An improved method for the computation of ligand steric

effects based on solid angles. Dalton Transactions, 2006(33): p. 3991-3999.

15. Koide, Y., S.G. Bott, and A.R. Barron, Alumoxanes as Cocatalysts in the

Palladium-Catalyzed Copolymerization of Carbon Monoxide and Ethylene: Genesis of a Structure− Activity Relationship. Organometallics, 1996. 15(9): p. 2213-2226.

16. Poater, A., et al., SambVca: A Web Application for the Calculation of the Buried

Volume of N‐Heterocyclic Carbene Ligands. European Journal of Inorganic

Chemistry, 2009. 2009(13): p. 1759-1766.

(27)

Chapter 2: Literature overview

2.1 Introduction to alkene metathesis

Alkene metathesis is a reaction where the double-bonded carbons in the alkenes will change places with one another.[1-6] The term alkene metathesis was first given by Calderon in 1967.[7-9] Although the alkene metathesis reaction was discovered in the mid-1950s, [7-9] the mechanism by which the interchange occurred was not understood even though countless experiments were done using alkene metathesis.[7-9] In 1970, a mechanism for alkene metathesis with a metal carbene catalyst was proposed by Yves Chauvin and his co-worker, Jean-Louis Hérrison.[10] Chauvin concluded that a metathesis reaction of alkenes, which produces two new alkenes, occurs in the presence of a metal-carbene complex that acts as a catalyst.[10] Chauvin's mechanism was supported by a number of experiments done by Katz.[11, 12] Therefore, today, the Chauvin mechanism (Figure 2.1) is generally accepted as the mechanism for alkene metathesis.

Cl PX3 Ru PX₃ Cl C H R -PX₃ PX₃ (a) [M] C H R C C H H H R₁ [M] C H R + -(b) (c) (d) C C H H H R1 M C H R C C H H H R₁ + H C C H R H R₁ [M] C H

(28)

The first step in the Chauvin mechanism is the PX3 ligand dissociation from the Grubbs

1-type catalyst (Figure 2.1(a)), leaving an open coordination site on the metal. An incoming alkene coordinates to the open coordination site on the metal of the metal-carbene complex to form a metallacyclobutane intermediate (Figure 2.1(b)). The metallacyclobutane intermediate is cleaved (Figure 2.1(c)) forming two new products, a metal alkylidene and a new alkene (Figure 2.1(d)).

However, it should be noted that the alkene metathesis can take place according to a dissociative, an associative or an interchange mechanism, as shown in Figure 2.2. In the dissociative mechanism, the metal-carbene complex loses a phosphine ligand from the metal centre to form an active 14-electron metal-carbene species.[13-15] During the associative mechanism, an 18-electron metal-carbene complex is formed when an incoming alkene coordinates to the metal-carbene complex.[13-15] In the interchange mechanism, the binding of the alkene and the loss of the phosphine ligand occurs simultaneously.[13-15]According to work done by Grubbs et al.,[13]the dissociative pathway accounts for 95% of the turnover in the alkene metathesis reactions, instead of the associative mechanism.

Dissociative Associative Interchange Ru C P P Cl Cl H H Ru C P P Cl Cl H H R R -P H Ru C P Cl Cl H H R Ru C P P Cl Cl H H R H Ru _C P Cl Cl H H R P Ru C P P Cl Cl H H R H Ru C P Cl Cl H H R P H Ru C P Cl Cl H H R H Ru P Cl Cl R P H Ru P Cl Cl R H Ru P Cl Cl R R R -P P

(29)

Breakthroughs in alkene metathesis were made when Richard R Schrock[17] and Robert H Grubbs [18, 19]developed well-defined ruthenium(II) and molybdenum/tungsten complexes for alkene metathesis, namely the Grubbs-type and Schrock-type catalysts. Since the development of Schrock and Grubbs catalysts, new industrial applications for alkene metathesis were discovered in petrochemicals, oleochemicals, polymers and specialty chemicals.[20] Applications of alkene metathesis in the petrochemicals field are the olefins conversion technology (OCT) process and the Shell higher olefins process (SHOP).[20]

The Schrock-type carbenes are classified as nucleophilic covalent (nucleophilic electron-sharing) carbenes, while the Grubbs-type catalysts are classified as electrophilic covalent (electrophilic electron-sharing) carbenes.[21] _{Schrock-type catalysts[17] are}

molybdenum-based catalysts, which are highly reactive and can be used with sterically demanding molecules. However, these catalysts are disadvantaged by having poor functional group tolerance, high sensitivity to air and moisture, and are thermally unstable. The Grubbs-type catalysts[22], on the other hand, are ruthenium-based catalysts that have the advantage of being stable in air, oxygen and have a tolerance toward many different organic functional groups.[13, 23]

Grubbs catalysts are used in various reactions, namely ring-closing metathesis[24], metathesis[25-27], acyclic diene metathesis polymerisation[28, 29], ring-opening cross-metathesis[30, 31], ring-opening metathesis polymerisation[32] and enyne metathesis reactions[33]. Recent studies on the application of Grubbs catalysts were done either on the modification or improvement of the catalysts.[34, 35] The modification and improvement focuses on finding more selective, active and recyclable catalysts.[36-38]

In computational studies of the self-metathesis (a special case of cross-metathesis) of ethene and 1-octene with RuCl2(PCy3)2(=CHPh), it was shown (Figure 2.3) that a large amount of

energy is required for the dissociation of the phosphine ligand.[16] The amount of energy required for the dissociation is dependent on the type of phosphine ligand that dissociates. A smaller amount of energy is needed for sterically bulky groups since phosphine dissociation is promoted, while with smaller groups, larger amounts of energy are needed for dissociation. [16] The amount of energy could be related to the stability, activity and selectivity of the catalyst.

(30)

Figure 2.3: Energy profile of the activations steps of the metathesis of ethene and 1-octene using RuCl2(PCy3)2(=CHPh)[16]

The activity, stability and selectivity of a catalyst can also be influenced by electronic effects, due to electron withdrawing and donating nature of the substituents. The steric and electronic effects are related to one another, and are difficult to separate. However, a practical separation through the use of electronic and steric parameters is possible.[39]

Steric and electronic effect is the consequence in the molecular properties when a part of the molecule is changed.[39]Electronic effects occur when one substituent, such as the OCH3 in

the P(p-C6H4OCH3)3 ligand, is replaced by Cl to obtain the (P(p-C6H4Cl)3) ligand. Therefore,

electronic effects is a result of transmission of electrons along chemical bonds. On the other hand, steric effects occur when the position of a group in a molecule is changed or when a large group is inserted into the molecule, since these groups will compete for coordination sites in the molecule.

According to Brown[40], ligand steric effects are a consequence of non-bonded repulsion between substituents. These occur when atoms come in close contact distance to one another, the electron clouds around each atom start to repel one another. In addition, the steric effect

(31)

of a ligand is not only dependent on the ligand size, shape and conformation, but also on the coordination centre and other ligands present in the system.[41]

2.2 Steric strain (steric hindrance)

There are three different types of strains, namely angular strain, torsional strain, and steric strain. The torsional strain is caused by repulsion between atoms when they pass each other as there is a rotation around a bond in a molecule. The angle strain is caused when the angle between bonds is different from the ideal bond angles of their geometry.[42, 43]

The steric strain in a molecule can be determined by calculating the close contact distance between the non-bonded atoms and the cone angle of ligands in the molecule.

If the close contact distance between the non-bonded atoms is small, the molecule is sterically strained. Therefore, an increase in the close contact distance between the non-bonded atoms will lead to a decrease in the steric strain. Bulky molecules or molecules with bulky groups are more likely to be sterically strained.

The size of the substituents on a ligand (measured by determining the cone angle) and the conformation of a ligand (which influences the size of the cone angle) are both examples of factors that may influence the steric strain and consequently the rate of ligand dissociation from a metal complex in a dissociative mechanism. Steric strain in a molecule affects a wide range of the molecules properties. These properties include the molecules, acidity or basicity, the reaction pathway, general reactivity, increased potential energy of a chemical reaction[5] and the conformation of molecules.[44, 45]

2.2.1 Size and electronic influences of substituents on dissociation of phosphine ligands

The steric strain between the substituents on the metal and the phosphine ligands has an influence on the rate of dissociation and the relative activity of the catalyst.[13]Grubbs et al. varied the ligand sphere around the ruthenium catalyst to determine how the electronic and steric properties of the ligands affect the catalyst activity. They used four different phosphine group catalysts (PR3 = PCy3, PCy2Ph, PiPr3 and PiPr2Ph), where the halogens on the

ruthenium atom were also changed (X = Cl, Br and I) as shown in Figure 2.4. Consequently, the changes in the catalyst activity were explored as the phosphine ligand and halogen ligands

(32)

were systematically varied. Furthermore, the Tolman cone angle technique was used to measure the cone angle in these structures.[13]

Ph Ph PR₃ Ru PR₃ X X

Figure 2.4: Various substituents on phosphine ligands[13]

One notable result was that if the R-group is Cy, the Grubbs catalyst was more active than if the R was Ph. Furthermore, the catalyst retained its stability towards air.[46]

The results of varying the phosphine ligand were to determine what effects the size of the phosphine ligand have on the activity of the catalyst.[13] Therefore, the results showed that by increasing the size of the phosphine (increasing the cone angle) and the electron donating ability, an increase in the catalyst activity was observed. The reason for this is that as the cone angle in the phosphines increases, the dissociative mechanism will be favoured due to the sterically crowded complex.[13] Therefore, catalysts with PCy3 ligands are more active

than the PCy2Ph ligands. Moreover, varying the nature of the phosphine ligands resulted in

changes in the catalyst activity.

It was noted that larger and more electron donating phosphines resulted in more active catalysts, while smaller and more electron withdrawing phosphines lead to less active catalysts.[13]

2.2.2 Coordination of alkene to the active ruthenium catalyst

Ulman et al.[47]investigated the coordination of alkenes with varying steric bulk, electronic properties and geometries to the active ruthenium catalyst, after dissociation of the phosphine ligand. The factors that affect the rate of coordination of the alkene to the active complex were studied. They studied three conformations that are shown in Figure 2.5. Figure 2.5(a) is an alkene complex with a cis-alkene and Figure 2.5 (b, c) two trans-alkene conformations. They concluded that the rates of metathesis were affected by the conformation of the alkenes, since the active benzylidene species react twice as fast with cis-3-hexene than with trans-3-hexene.[47]

(33)

PCy3 Cl Ru Cl C H Ph R R PCy3 Cl Ru Cl C H Ph R R PCy3 Cl Ru Cl C H Ph R R (a) (b) (c)

Figure 2.5: Coordination of cis- and trans- alkenes to an active ruthenium catalyst[47] In addition, the conformation of the metal complex during an alkene metathesis reaction determines the size of the open coordination site.[48] Furthermore, the size of the open coordination site has an effect on the rate and selectivity of metal catalysed reactions.[49]

2.3 Techniques to calculate the steric strain in organometallic chemistry

Numerous attempts to quantify ligand steric strain with the intention of establishing quantitative relationships between the reaction rate, the size of the ligand, equilibrium constants and the chemical and physical properties of organic and organometallic systems,[50] have been attempted. Data obtained from the steric strain techniques enables chemists to understand how the steric strain influences the course of a reaction and to modify reaction conditions appropriately for expected results.[50]

Most of the studies performed with these above techniques focused on the behaviour of individual ligands in certain systems without simultaneous consideration of all inter-ligand steric strain in the molecules. However, the studies performed by Coville et al.[51-54]and Fischer and Li[51], who used the solid angle technique, are exceptions to this.

2.3.1 Tolman cone angle (θ)

The Tolman cone angle, defined by Tolman,[39] is the simplest and most used technique to measure the steric strain of phosphine ligands. The Tolman cone angle measures the amount of steric strain through the use of a cone angle. The Tolman cone angle is the angle between the metal and the substituents on the phosphine atom that just touches the van der Waals radius of the substituents. Notably, Tolman used a constant P-M (phosphorus-metal) bond length of 2.28 Å, which is the standard P-Ni bond length for [Ni(CO)3(L)] complexes (where

L = H, F, Cl, and Br).[39]Figure 2.6 is a representation of the Tolman cone angle technique for symmetrical molecules where the substituents are the same. There is no mathematical

(34)

equation to measure the Tolman cone angles, since Tolman measured the cone angles using space-filling CPK models, a protractor and a ruler.

Figure 2.6: Tolman cone angle for symmetrical molecules[39]

However, the phosphine ligand could contain different substituents, leading to an unsymmetrical molecule. The Tolman cone angle for an unsymmetrical molecule is calculated as the average of the individual substituent angles using the equation shown in Figure 2.7.

𝜃𝜃 = 2_{3 �}𝜃𝜃₂i

3

𝑖𝑖=1

Figure 2.7: Tolman cone angle for unsymmetrical molecule[39]

The space-filling CPK models that Tolman built to measure the cone angles with, do not depict attractive and repulsive forces that occur between substituents in ligands or ligands in molecules.[5, 44] Therefore, a minimum energy conformation was not taken into account when Tolman calculated the steric strain of various phosphine ligands. Consequently, the Tolman cone angle is not an adequate measure of steric crowding. It also allows substituents or ligands to mesh (overlap of substituents or ligands, where the physical size of the ligand is not taken into consideration) together to relieve steric strain.[51]

(35)

In an attempt to address the limitations of the Tolman cone angle technique, the solid angle technique, which is capable of calculating the steric strain of all ligand types in both organic and organometallic chemistry, was developed.[53]

2.3.2 Solid angle (Ω)

The solid angle technique was introduced by Hirota and co-workers[53] and later improved by White, Taverner and Coville.[51-54]The solid angle is defined using the projection of the substituents as shadows on the inside of a sphere, if the metal is replaced by a light source. The steric size of the ligands is then calculated by measuring how much of the surface is covered by the projection, as shown in Figure 2.8.[54]

Figure 2.8: Solid angle representation[54]

If the ligand covers the entire unit sphere, then the solid angle is 4π sr. The steric size of the ligand as the fraction of unit sphere occupied is calculated with Equation 2.1.[54] The solid angle (Ω) in steradians can be converted to solid angle in degrees (Ω) with Equation 2.2.[54]

Ωs= _4πΩ ... (2.1) Ω = 2arccos �1 −_2𝜋𝜋Ω� ... (2.2) The minimum steric influence of a substituent is measured with the solid angle technique, whereas the Tolman cone angle technique measures the maximum amount of steric strain of the substituents.

(36)

Figure 2.9 is a representation of the difference between the solid angle technique and the Tolman cone angle technique. The solid angle is the area occupied by the shadow, while the Tolman cone angle is a cone around the ligand.

Figure 2.9: Comparison of Tolman cone angle and solid angle

In some types of ligands, the Tolman cone angle technique forces atoms without any steric strain, that should have a 180 ° cone angle, into a 90 ° cone. On the other hand, the solid angle technique does not force the ligands into a cone, since the shadow is projected onto a sphere and then the shadow is the measure of the steric size of the object.

The advantage that the solid angle technique has over the Tolman cone angle technique is that ligand meshing is taken into consideration and only the space occupied by the ligand is calculated.[55] In addition, the values obtained with the solid angle technique contain information about the shape of a substituent.

One disadvantage of the solid angle technique is that it is sensitive to the conformation of the ligand. The solid angle is measured through the size of the projected shadow on the sphere, while the size of the shadow is determined through the position of the substituents within the ligands. White and co-workers[54] used Brown's approach[40] and generated single conformer (energy minimum) solid angles for a range of organometallic ligands. For comparative purposes, they used both the SYBYL[56] and MMP2[57] force fields. The solid angle technique correlates well (linear correlation of R2_{= 0.93) with the Tolman cone}

angles[39]for the [Ni(CO)3(L)] complexes (where L = PH3, PF3, PCl3or PBr3).[39, 58]

This is in agreement with the observation that the amount of space that the ligand takes up in the solid angle technique is similar to the Tolman cone angle technique when a cone is drawn around the ligand (Figure 2.9).[55]

(37)

In solid angle calculations, it is possible for different ligands in a complex to shield the same regions of the surrounding sphere, as shown in Figure 2.10. Therefore, the shadows of the ligands overlap with each other.

Figure 2.10: Shadow overlap for Grubbs 1-type catalyst

It was suggested by Taverner et al.[51]to use Equation 2.3 as a measure of steric congestion between ligands in order to calculate the shadow overlap.[41] The solid angle overlap (Γ) can be determined by knowing the difference between the sum of the solid angle (Ω) of two the overlapping spheres, A and B and the solid angles of the non-overlapping spheres.

Γ = (ΩA+ ΩB) - ΩAB ... ...(2.3)

Recently, Guzei et al. [41]developed an improved technique for the calculation of the solid angle, called the G-parameter. The expression for the G-parameter is shown in Equation 2.4.[41]

G = 100 Ω

4𝜋𝜋 ... (2.4)

In addition, they created a computer program named solid-G[41], which can be used to calculate the percentage of the metal sphere that is shielded by the ligand from molecular modelling coordinate data or from experimental single-crystal X-ray diffraction analysis data of a metal complex. The solid-G program does not use the van der Waals radii of Bondi,[59] but instead the G-parameter calculates the atomic Rz radius (radii of “hard spheres”) [41]

mathematically from the Morse potential equation. Furthermore, there are different values for the van der Waals radii that is shown in Appendix A, Table A.1.

A measure of unfavourable contacts (steric congestion) was also developed using a similar approach to that of Taverner et al.[51] This measure is known as the Gγ(complex) that defines all of the areas in the complex that are simultaneously shielded by several ligands in

(38)

order to quantify the unfavourable contacts (in terms of ligand shadow overlap as seen in Figure 2.11). [41]

Figure 2.11: Shadow overlap on sphere to obtain unfavourable contacts, where c = solid overlap of atomic spheres, VG = volume of solid overlap and Gγ = areas shielded by ligands

[41]

The equation used to calculate the Gγ(complex) parameter is shown in Equation 2.5, where GM(complex) is the amount of metal coordination sphere shielded by all ligands, while GM(L)

is the amount of the metal shielded by a ligand. However, the Gγ(complex) parameter does not quantify inter-ligand interactions. [41]

Gγ(complex) = ∑iGM(Li)- GM(complex)... (2.5)

A problem that was found with the Gγ(complex) parameter was that it did not differentiate between ligand shadow overlaps and unfavourable ligand-ligand interactions. Therefore, Guzei et al.[41] developed an improvement to the Gγ(complex) parameter to calculate the unfavourable close contacts in ligands. This new parameter is known as the GU(complex)

parameter.[41]

According to Guzei et al.[41], all non-coordinated atoms are considered to be spheres with radius Rz. If two atoms from different ligands are closer than the sum of their (Rz) radius

overlap may occur, this overlap indicates that unfavourable inter-ligand interaction occurs. However, it should be noted that not all overlaps indicate unfavourable inter-ligand interactions, since the calculations may include a possible hydrogen bonding interaction that does not signify an unfavourable contact.

(39)

In addition, the sum of all unfavourable interatomic interactions between two ligands (L1 and

L2) is designated as GU(L1-L2). The unfavourable interactions in a complex (GU) is the sum of

all pairs of ligands in the complex GU(LX-LY), where LX and LY are different ligands. [41]

Solid-G expresses the unfavourable close contacts in terms of volumes, since volumes were more efficient to describe the unfavourable close contacts. Therefore, VG (as shown in

Figure 2.11) is used to characterise interactions between separate ligands VG(L1-L2) and

within a complex as a whole VG(complex), instead of the GU parameter (Figure 2.11).

However, the close contact data provided by solid-G is both in terms of bond distance (Å) and volume (Å3_).

Table 2.1 is a legend of what the solid-G program calculates for the ligand. Solid-G calculates both the solid angle and the cone angle. However, the cone angle is not the Tolman cone angle, since the cone angle (ECA) corresponds to the omega(L) or solid angle calculations. Furthermore, solid-G calculates the unfavourable close contacts between atoms.

Table 2.1: Solid-G ligand angle calculations legend Ligand The ligand number.

Omega(L) The solid angle of the ligand (Ω or %).

G(L) The percentage of the sphere shielded by the ligand (Ω or %).

ECA The cone angle (°) corresponding to the Omega(L) and not Tolman's cone angle for the ligand.

SUM(G(L)) The sum of all individual ligand G(L) values (Ω or %).

G(complex) The G value for the complex, all ligands treated as one (Ω or %).

G(gamma) The percentage of the sphere shielded by more than one ligand (Ω or %). G(Ru3 ) The percentage of metal's surface shielded by the ligated atoms only (Ω or

%).

S(Ru3 ) The percentage of metal's surface "in contact" with the ligated atoms (Ω or %).

GU(complex) The percentage of the metal shielded by all regions of unfavourable

(40)

Lastly, the relationship between the Tolman cone angle and the solid angle calculated with solid-G is shown as a non-linear graph[41] in Figure 2.12, where Ω = 2𝜋𝜋 �1 - cos 𝜃𝜃

2�.

Figure 2.12: The non-linear relationship between the solid angle and the Tolman cone angle[41]

2.3.3 Pocket angle

Koide and co-workers[49] developed the pocket angle as an aid to understand the steric effects of chelating phosphines. In order to determine how the steric properties of the phosphine ligand affect the catalytic activity they varied the chelate ring size and the substituents. The pocket angle is defined as the interior cone angle of a chelating phosphine that calculates the available space for substrates in a complex containing bidentate ligands. The steric bulk of monodentate phosphines is readily estimated by Tolman's cone angle[39]; however, there is no simple method to estimate the steric effects of bidentate phosphines. The concept of bite angles was used to calculate the interior angle.[49]

There are two different types of pocket angle measurements relative to the PPd plane; the parallel pocket (θ||) and the perpendicular pocket (θ┴), as shown in Figure 2.13. The parallel pocket angle is defined as the angle between two planes that are perpendicular to the PdP2

plane and bisects at the metal.[49] The perpendicular pocket angle is defined as the angle subtended between two planes that are parallel with the P2 vector and bisects at the metal.[49]

(41)

(a) (b)

Figure 2.13: (a) Parallel pocket angle and (b) Perpendicular pocket angle[49]

In particular, the size of the interior cone angle determines the catalyst activity. Therefore, complexes that have large pocket angles resulted in sterically less strained complexes. Moreover, when a bond length of 2.28 Ǻ for the P-metal bond is used in the pocket angle calculations, the values obtained are similar to the values obtained from the Tolman cone angle technique.[49]

2.3.4 Percentage buried volume (%VBur)

The percentage buried volume (%VBur) technique measures the steric parameter of ligands by

the amount of space occupied by a ligand in an organometallic compound.[31, 32]

The percentage buried volume technique was created to calculate the steric strain in the NHC ligands in catalysts, since the Tolman cone angle technique was unable to describe the C2

symmetry of the NHC ligands. In addition, the Tolman cone angle technique is difficult to use with large bulky molecules, whereas the percentage buried volume can be used with both small and large molecules.[58]

The percentage buried volume technique is shown in Figure 2.14, where the PX3 ligand is

buried in a sphere. The metal is in the centre of the sphere, the distance to phosphorus is d and the midpoint of the X atoms to the P atom is XC.[58]

(42)

Figure 2.14: %VBur of phosphine ligand where the sphere radius is 3.5Å[58]

The percentage buried volume can be calculated with Equation 2.6. The percentage buried volume indicates what amount of a sphere is occupied by the ligand. Therefore, the percentage buried volume size indicates steric strain in ligands. The steric strain will increase with bulky ligands as a result of atoms being forced closer to one another.

%VBur= 100*VBur

VSphere ... (2.6)

Notably, in 2009, Nolan[60] compared various Tolman cone angle values against the buried volume values of phosphine ligands. He used imported crystal structures of phosphines into a web application (program) named sambVca[58] to obtain the buried volume data. Nolan concluded that the Tolman cone angle and the buried volume techniques had good correlation with one another in view of a linear correlation of R2 = 0.96.[60]

2.3.5 Summary

A large amount of energy is needed to dissociate the phosphine ligand from the Grubbs 1-type catalyst in a dissociative alkene metathesis reaction. This increase in energy can be a consequence of various factors, which include steric strain. The steric strain can be derived from ligand sizes, vacant space sizes and close contact distances. Various techniques have been created and proven to be accurate in the calculation of the steric strain in organometallic chemistry. The steric parameters obtained from these techniques could enable scientists to gain a deeper understanding of how steric strain determines the shape of the molecule and the outcome of a reaction. A limitation of these techniques, however, is that the steric strain

(43)

understand the dissociation of a phosphine ligand and explain the increase in energy, a dynamic technique that follows the dissociation, which includes rearrangement of atoms in molecules during dissociation, is needed.

2.4 References

1. Anderson, A. and N. Merckling. Patent 2 721 189, 1955. in Chem. Abstr. 1955. 2. Lehman Jr, S. and K. Wagener, In Handbook of Metathesis; Grubbs, RH, Ed.

Wiley-VCH: Weinheim, Germany, 2003. 3: p. 283-353.

3. Ivin, K.J. and J. Mol, Olefin metathesis and metathesis polymerization. 1997: Academic Press.

4. Grubbs, R.H., Olefin metathesis. Tetrahedron, 2004. 60(34): p. 7117-7140.

5. Trnka, T.M. and R.H. Grubbs, The development of L2X2Ru CHR olefin metathesis

catalysts: an organometallic success story. Accounts of Chemical Research, 2001.

34(1): p. 18-29.

6. Fürstner, A., Olefin metathesis and beyond. Angewandte Chemie International Edition, 2000. 39(17): p. 3012-3043.

7. Calderon, N., HY Chen u. KW Scott. Tetrahedron Lett, 1967. 3327.

8. Calderon, N., et al., Olefin metathesis. I. Acyclic vinylenic hydrocarbons. Journal of the American Chemical Society, 1968. 90(15): p. 4133-4140.

9. Mol, J., J. Moulijn, and C. Boelhouwer, Carbon-14 studies on the mechanism of the

disproportionation of propene. Chemical Communications (London), 1968(11): p.

633-633.

10. Herisson, J. and Y. Chauvin, Transformation catalysis of olefins by tungsten

complexes. 2. Telomerization of cyclic olefins in presence of acyclic olefins.

Makromolekulare Chemie, 1971. 141(FEB 9): p. 161-&.

11. Katz, T.J. and S.J. Lee, Initiation of acetylene polymerization by metal carbenes. Journal of the American Chemical Society, 1980. 102(1): p. 422-424.

12. Katz, T.J., The olefin metathesis reaction. Advances in Organometallic Chemistry, 1977. 16: p. 283-317.

13. Dias, E.L., S.T. Nguyen, and R.H. Grubbs, Well-defined ruthenium olefin metathesis

catalysts: mechanism and activity. Journal of the American Chemical Society, 1997.

119(17): p. 3887-3897.

14. Wilson, M.R., et al., The quantitative analysis of ligand effects (QALE). The aryl

(44)

15. Nelson, D.J., et al., Key processes in ruthenium-catalysed olefin metathesis. Chemical Communications, 2014. 50(72): p. 10355-10375.

16. Jordaan, M., et al., Experimental and DFT investigation of the 1-octene metathesis

reaction mechanism with the Grubbs 1 precatalyst. Journal of Molecular Catalysis A:

Chemical, 2006. 254(1): p. 145-154.

17. Schrock, R.R., Olefin metathesis by molybdenum imido alkylidene catalysts. Tetrahedron, 1999. 55(27): p. 8141-8153.

18. Schwab, P., R.H. Grubbs, and J.W. Ziller, Synthesis and applications of RuCl2 (CHR

‘)(PR3) 2: the influence of the alkylidene moiety on metathesis activity. Journal of the

American Chemical Society, 1996. 118(1): p. 100-110.

19. Grubbs, R.H., P.L. Burk, and D.D. Carr, Mechanism of the olefin metathesis reaction. Journal of the American Chemical Society, 1975. 97(11): p. 3265-3267.

20. Mol, J., Industrial applications of olefin metathesis. Journal of Molecular Catalysis A: Chemical, 2004. 213(1): p. 39-45.

21. Occhipinti, G. and V.R. Jensen, Nature of the transition metal–Carbene bond in

Grubbs olefin metathesis catalysts. Organometallics, 2011. 30(13): p. 3522-3529.

22. Nguyen, S.T., R.H. Grubbs, and J.W. Ziller, Syntheses and activities of new

single-component, ruthenium-based olefin metathesis catalysts. Journal of the American

Chemical Society, 1993. 115(21): p. 9858-9859.

23. Scholl, M., et al., Synthesis and Activity of a New Generation of Ruthenium-Based

Olefin Metathesis Catalysts Coordinated with 1, 3-Dimesityl-4, 5-dihydroimidazol-2-ylidene Ligands §. Organic Letters, 1999. 1(6): p. 953-956.

24. Han, S.Y. and S. Chang, General Ring‐Closing Metathesis. Handbook of Metathesis: Catalyst Development, 2003: p. 5-127.

25. Grubbs, R.H. and S. Chang, Recent advances in olefin metathesis and its application

in organic synthesis. Tetrahedron, 1998. 54(18): p. 4413-4450.

26. Chatterjee, A.K., et al., A general model for selectivity in olefin cross metathesis. Journal of the American Chemical Society, 2003. 125(37): p. 11360-11370.

27. Connon, S.J. and S. Blechert, Recent developments in olefin cross‐metathesis. Angewandte Chemie International Edition, 2003. 42(17): p. 1900-1923.

28. Mutlu, H., L.M. de Espinosa, and M.A. Meier, Acyclic diene metathesis: a versatile

tool for the construction of defined polymer architectures. Chemical Society Reviews,

Defining and calculating steric strain in the first step of the dissociative alkene metathesis mechanism using Grubbs 1-type catalysts

Defining and calculating steric strain in the

first step of the dissociative alkene metathesis

mechanism using Grubbs 1-type catalysts

LM Botha

21112886

Dissertation submitted in partial fulfilment of the requirements for the

degree

Magister Scientiae

in

Chemistry

at the Potchefstroom Campus

of the North-West University

Supervisor:

Dr CGCE van Sittert

Co-supervisor:

Prof HCM Vosloo

Table of contents

Summary

Defining and calculating steric strain in the first step of the dissociative

alkene metathesis mechanism using Grubbs 1-type catalysts

Preface

Chapter 1: Introduction and objectives

Chapter 2: Literature overview