Synthesis Computational Lab 1C - Daniel Spencer, 00736964
The aim of this experiment was to investigate the asymmetric epoxidation of alkenes, along with the characterisation of the epoxide formed via NMR spectroscopy and chiroptical measurements, using computational techniques. Here, the epoxidations of styrene and 1,2-dihydronaphthalene were considered, using both the Shi and Jacobsen catalysts - a good agreement with literature was found in terms of chiroptical properties; however, prediction of the preferred enantiomeric product for each case, by examination of transition state energies, resulted in significant differences to literature, for both the preferred enantiomer and the enantiomeric excess.
The aim of this computational investigation is to study the asymmetric epoxidation of alkenes and the characterisation of the epoxides formed using computational methods. In order to achieve this, general methods for computational analysis and spectroscopic simulation will be considered, followed by a direct consideration of asymmetric epoxidation of two alkenes using two different catalysts - the Shi fructose-derived catalyst[1] and the Jacobsen salen-derived catalyst.[2] These catalysts are shown in Figures 1 & 2 respectively. Notably, the Shi catalyst shown is produced from the stable precursor of the corresponding ketone (in place of the dioxirane)[3]. Similarly, the Jacobsen catalyst is prepared from a precursor featuring a Mn-Cl bond, as opposed to the Mn=O bond in the catalyst.[2] These catalysts are known to have differing degrees of selectivity according to the alkene under consideration. The two alkenes selected here are styrene and 1,2-dihydronaphthalene, shown in Figures 3 & 4 respectively.
Computational Methods
In order to conduct this investigation, two primary computational methods will be employed: Molecular Mechanics(the MM2/MMFF94 methods) and Density Functional Theory. Here, the MM2/MMFF94 methods will be used to study the initial systems (the cyclopentadiene dimer and the atropisomerism of a Taxol intermediate) and for quick geometric optimisations, whilst DFT (employed at the B3LYP/6-31G(d,p) for the most part) will be used in the simulation of spectra and the examination of asymmetric epoxidation transition states. Molecular Mechanics techniques are utilised here as a consequence of their short computation times, even for large systems, when compared to DFT (for which some of the calculations performed here can be extremely time-consuming).
Asymmetric Epoxidation
Asymmetric epoxidation, first investigated by Katsuki and Sharpless,[4] refers to epoxidation techniques with a degree of selectivity for which face of a specified alkene is 'attacked', resulting in preference for the formation of a particular enantiomeric product. Given the significance of this work, numerous catalytic systems have been prepared for use in this reaction, such as those of Sharpless[4], Shi[1] and Jacobsen.[2]
Through a computational investigation of the Shi and Jacobsen catalysts for asymmetric epoxidation, enantiomeric preference will be examined, with comparison to experimental literature for a consideration of the obtained results and their validity.
Table 1: Dihydro derivative energies (all energies in kcal mol-1)
Energy Contribution
Derivative (1)
Derivative (2)
Stretching (str)
1.2329
1.0962
Bending (bnd)
18.9345
14.5235
Stretch-Bend
-0.7603
-0.5493
Torsion (tor)
12.1313
12.4971
1,4 van der Waals (vdw)
5.2799
4.5133
non-1,4 van der Waals
-1.5048
-1.0693
Dipole/dipole or Electrostatic
0.1631
0.1406
Total
35.9266
31.1520
The dimerisation of cyclopentadiene is an example of a Diels-Alder reaction - the reaction is classified as a pericyclic π4s + π2s process, with a transition state of Hückel topology.[5][6] The dimerisation can lead to either an exo or an endo product - these are shown in Figures 5 & 6. To computationally justify the observation that the endo dimer is formed when this reaction occurs, the two possible products were studied, using MM2 molecular mechanics within ChemBio3D, with a MMFF94(s) force field. To achieve this, the endo/exo isomers were first drawn and then optimised using the method defined.
The optimised structure of the exo dimer is presented in Figure 7 - the energy of this product was calculated to be 55.3740 kcal mol-1. For comparison, the optimised structure for the endo dimer is shown in Figure 8 - the energy of this product was found to be 58.1930 kcal mol-1. Clearly the endo product has a higher steric energy than the exo product - this observation suggests that the dimerisation occurs under kinetic conditions as the reaction is known to give the endo product.[6] The reaction proceeds in this way as a consequence of the lower transition state energy for the endo product (a consequence of stabilising secondary orbital interactions in the endo transition state and a steric interaction between methylene hydrogen atoms in the exo transition state)[6], leading to it being the kinetic product of the dimerisation.
Regarding the hydrogenation of the endo product formed, it is clear that there are two possibilities, based on which carbon-carbon double bond undergoes hydrogenation (labelled A and B in Figures 5 & 6 for clarity). These two dihydro derivatives are presented in Figures 9 & 10, where Figure 9 shows hydrogenation at double bond B, whilst Figure 10 shows hydrogenation at A. The derivative labelled (1) has an energy of 50.7259 kcal mol-1, whilst that labelled (2) has an energy of 41.2620 kcal mol-1. The energies of these dihydro derivatives can be broken-down into individual components to allow a more in-depth consideration of the products - these data are presented (having been calculated using MM2 optimisation) in Table 1.
A comparison of the total energies obtained using MMFF94s and MM2 reveals that the total energy of dihydro derivative (2) is lower than that of (1) in both cases. Examining the MM2 data in Table 1, it is clear that derivative (2) has lower stretching, bending, van der Waals and electrostatic energy components, with a slightly larger torsion component, when compared to the energies recorded for derivative (1). The larger stretching contribution for derivative (1) indicates a greater deviation from the 'natural' bond length, providing evidence for greater strain compared to derivative (2). Hence, it can be concluded that derivative (2) is the thermodynamic product of the hydrogenation of the endo dimer of cyclopentadiene.
Literature is available for comparison of this observation - it has been reported that the double bond labelled A in this investigation undergoes hydrogenation significantly faster than that labelled B.[7] As the rate of hydrogenation is faster for the reaction at double bond A than B, and, as observed in this investigation, the product resulting from hydrogenation at A is lower in energy than that for B, it is clear that derivative (2) is both the kinetic and thermodynamic product of the hydrogenation reaction.
Atropisomerism in an Intermediate Related to the Synthesis of Taxol
Figure 16: Taxol intermediate isomer (2) in twist-boat conformation
Figure 17: Taxol intermediate isomer (2) in half-chair conformation
In the synthesis of Taxol, an important intermediate can be drawn in two different isomeric forms - these are shown in Figures 11 & 12. Structurally, the two isomers differ in the positions of their carbonyl group - isomer (1) has the carbonyl pointing 'up' (if the molecule is arranged such that the upwards pointing carbonyl is at the front, the cyclohexane ring appears on the left-side of the carbonyl), whilst isomer (2) has the carbonyl pointing 'down'. These structures were drawn in ChemBio3D and then optimised using MMFF94s - the optimised structure of isomer (1) is shown in Figure 13. The energy this isomer was found to be 70.5608 kcal mol-1 - the structure can be observed to have the cyclohexane ring in a chair conformation, as would be expected for an energy minimum. However, the calculated energy can be lowered by manual adjustment to a second chair conformation - this newly optimised structure is presented in Figure 14. The energy of this alternative structure was calculated to be 60.6561 kcal mol-1, which is substantially lower than that previously found.
A similar process was conducted for isomer (2), that with the carbonyl group pointing 'down' - the structure of this isomer is presented in Figure 15, where adjustment of the cyclohexane ring was undertaken in order to minimise the energy. The energy of this isomer was calculated to be 60.0461 kcal mol-1. A comparison of these energies indicates that isomer (2), with the downwards pointing carbonyl group, is the more stable isomer, as reported in literature.[8]
As the calculations indicate, isomer (2) is more stable; as a consequence of this energy difference, the molecule may spontaneously isomerise such that only a single isomer is present, dependent upon the barrier to rotation. This conversion could be further investigated through a consideration of the transition structure of the atropisomerism, which would allow an investigation into the activation energy.[9]
With regards to the observation that the alkene undergoes a slow reaction, one can consider hyperstable alkenes and Bredt's rule.[10][11] The notable feature of the alkene in the Taxol intermediate is that it appears at a bridgehead (a ring-joining position) - Bredt's rule predicted that such alkenes would not be formed. Maier et al.'s proposed class of hyperstable alkenes (also referred to as anti-Bredt compounds[12] is defined by an alkene being less strained then its corresponding saturated molecule and having a negative "olefinic strain" - such alkenes would be predicted to be extremely unreactive. Whilst this is likely to be the cause of the stability of the alkene under consideration, a calculation of its "olefinic strain" would be required for confirmation. Notably, the study proposing these hyper stable alkenes made use of molecular mechanics MM1 calculations; however, it has been noted that molecular mechanics is not particularly applicable to the study of such alkenes and hence, further investigation using a G3/B3LYP method has been conducted by Novak.[12] These results also supported the stability of such hyperstable alkenes, further supporting this being the cause of the stability of the Taxol intermediate alkene investigated here.
With regards to finding the lowest energy conformation of the molecules under consideration here, it is evident that the conformation of the cyclohexane ring is important, as evidenced by the observed decrease in energy upon changing of the chair conformation. As a consequence of this observation, further conformations of the cyclohexane ring were considered - Figures 16 & 17 show twist-boat and half-chair conformations respectively. The MMFF94s optimised energies of these conformations were found to be 66.3035 kcal mol-1 for the twist-boat and 299.921 kcal mol-1 for the half-chair. These energies reinforce the identification of the chair conformation shown in Figure 14 as the minimum energy conformation of the molecule, whilst supporting the half-chair as the highest energy conformation of the cyclohexane substructure. This could be further confirmed by optimisation of a boat conformation; however, this was not achieved as attempts resulted in a twist-boat conformation.
Spectroscopic Simulation using Quantum Mechanics
Spectroscopy of an Intermediate Related to the Synthesis of Taxol
Figure 18: Taxol intermediate 18
Figure 19: Taxol intermediate structure for spectroscopic simulation (MMFF94s minimised)
Table 2: Literature 1H NMR data - taken from Paquette et al.[8]
Chemical Shift, δ, ppm
Assignment (multiplicity, number of 1H, J-value where reported)
Table 4: Computational 13C NMR data and Literature data[8]
C environment, x-C
Chemical shift, δ, ppm (Computational)
Chemical shift, δ, ppm (Literature)
Deviation
14
204.66
211.49
-6.83
9
136.86
148.72
-11.86
8
111.06
120.90
-9.84
3
82.58
74.61
7.97
1
50.07
60.53
-10.46
2
49.27
51.30
-2.03
12
46.36
50.94
-4.58
19
39.69
45.53
-5.84
4
38.93
43.28
-4.35
17
35.74
40.82
-5.08
18
34.16
38.73
-4.57
6
33.32
36.78
-3.46
13
27.68
35.47
-7.79
30
21.82
30.84
-9.02
10
18.56
30.00
-11.44
21
16.61
25.56
-8.95
11
15.62
25.35
-9.73
5
15.20
22.21
-7.01
20
13.21
21.39
-8.18
27
12.55
19.83
-7.28
Figure 22: Atom numbers for NMR analysis of Taxol intermediateFigure 23: Computationally simulated IR spectrum
In order to simulate the 1H and 13C NMR spectra of the molecule shown in Figure 18 (an intermediate related to the synthesis of taxol), the structure was first drawn in ChemBio3D, using the observations made in the previous section to assist in energy minimisation. The structure under consideration is clearly a derivative of isomer (2) - that with the carbonyl pointing downwards. This minimisation using MMFF94s gave an energy of 97.1177 kcal mol-1 - the structure is shown in Figure 19.
After simulation of spectra, the results obtained will be compared to literature and hence, the solvent was set according to that used in the literature. In this case, the solvent used in the literature is deuterated benzene[8] - as this was not available as an option within the Gaussian interface of ChemBio3D, benzene was used as an alternative. As the calculation will account for the presence of the solvent, signals corresponding to the carbon and hydrogen atoms of benzene will not appear in the computed spectra; however, the chemical shifts calculated may be somewhat erroneous compared to literature. The structure was set-up for optimisation at the DFT UB3LYP/6-31G(d,p) level of theory. To account for solvent effects, SCRF(CPCM,Solvent=benzene) was added to the Gaussian job file, along with Freq, NMR and EmpiricalDispersion=GD3. This job file was then submitted to the HPC - the results were then published on DSpace DOI:10042/146561 .
The 1H NMR data from literature are summarised in Table 2, with the corresponding 13C NMR data similarly summarised below.[8] 13C NMR (ppm): 211.49, 148.72, 120.90, 74.61, 60.53, 51.30, 50.94, 45.53, 43.28, 40.82, 38.73, 36.78, 35.47, 30.84, 30.00, 25.56, 25.35, 22.21, 21.39, 19.83.
The spectrum simulated using DFT calculations in this experiment is presented in Figure 20 for 1H NMR in an scalable format. Were coupling constants also simulated, it would be possible to predict the experimental spectrum of the molecule using gNMR - this was not undertaken due to time limitations. Figure 21 gives the spectrum computed for 13C NMR. It should be noted that both of these spectra use a reference of TMS B3LYP/6-311+G(2d,p) GIAO - this is somewhat different to the calculation settings used in the input file and hence, this may explain some errors in the results obtained. The 1H NMR data seen in the spectrum is given in Table 3 for comparison to literature; note that the hydrogen atom labels given correspond to those observed in Figure 22. A direct comparison to literature for the 1H NMR data simulated in this investigation is difficult as a consequence of not accounting for J-coupling - this could be rectified by further calculation of the JH-H coupling constants; however, this was not undertaken due to time constraints. A simpler comparison is achieved through mathematical comparison of the 13C NMR data computed to the literature values presented above - these data are presented together in Table 4, along with an expression of the deviation of the computational results from the experimental literature. Note that atom labels are, again, those shown in Figure 22.
Whilst the simulation computes a chemical shift for each nucleus of the specified type, these shifts can also be assigned manually by a consideration of their magnitude. Whilst many of the observed shifts are typical (such as alkyl environments), some characteristic chemical shifts of the compound can be considered, such as those adjoined to the sulphur heterocycle and the carbonyl. With the highest 1H NMR chemical shift, it is clear that the signal at 6.11 ppm can be attributed to the vinylic proton (atom 7) - this assignment is of interest due to the significant deviation from literature in this value (0.90 ppm). On the scale of 1H NMR, this difference is substantial, making it possible to note the difficulty involved in comparing the 1H NMR data generally. Regarding 13C NMR, the interesting signals are those at 204.66 ppm, 136.86 ppm, 111.06 ppm and 82.58 ppm. Here, the largest can be easily attributed to the carbonyl carbon, due to the large deshielding effect of the electronegative oxygen atom. The chemical shifts of 136.86 and 111.06 ppm can be assigned to the vinyl carbon atoms and that at 82.58 ppm corresponds to the carbon atom bonded to the two heterocyclic sulphur atoms. With the latter, it is interesting to observe that atoms 17 and 18, bonded to a single sulphur atom, have substantially smaller shifts (~35 ppm). A consideration of these assignments, alongside those computationally assigned, suggests that little difference would be observed in the NMR spectra for the 'up' atropisomer of the Taxol intermediate.
A consideration of the data presented in Table 4 allows for a numerical analysis of the computational and literature data, giving the deviations shown. Whilst these deviations can be presented graphically against the carbon environment number, it was concluded that this did not provide much valuable information regarding the comparison. Instead, the standard deviation for the deviations was calculated - this was found to be 4.25 ppm, which indicates that whilst the data are reasonably comparable to literature, there is a significant error when compared to literature. Some of this deviation may be a consequence of the requirement of correction factors for 13C NMR shifts when halogen or sulphur atoms are bonded to carbon atoms in the compound under consideration; however, such a correction factor has yet to be determined for C-S environments.[13]
As the molecule also contains oxygen and sulphur nuclei, 17O and 33S NMR could also be considered - these are computed as part of the same NMR computation. However, due to the low abundance of these spin-active isotopes, comparison with experimental literature would be difficult and for these molecules, these NMR spectra would not reveal much useful information.
Using the same calculation, the vibrational frequencies were also computed - these are summarised in the IR spectrum presented in Figure 23. Literature data for the IR spectrum were also recorded, as below.[8] Note that these were recorded in chloroform, ergo there will be some degree of error when compared to those frequencies predicted in this investigation.
IR (cm-1): 3070-2800, 1675, 1470, 1390, 1345, 1275, 1260-1200, 1170, 1050.
A comparison of these literature data with the spectrum in Figure 23 shows that whilst many of the signals experimentally observed have been simulated, there appears to be a systematic error in the computational data - these is most clearly observed with the peaks in the range 3200-3000 cm-1, for which the literature has the range 3070-2800 cm-1. This suggests a considerable error of ~200 cm-1
The .log file produced by the calculation can also be considered, as it provides thermochemical data - in this investigation, the Sum of Electronic and Thermal Free Energies is of interest as it corresponds to the free energy of the molecule (ΔG). For the intermediate in the synthesis of Taxol considered here, the free energy was found to be -1651.464393 hartree - this free energy could be used to compare the isomer studied here with that involving the carbonyl group pointing upwards.
Analysis of the Properties of Styrene and 1,2-Dihydronaphthalene Epoxides
Crystal Structures of Shi and Jacobsen Catalysts
Figure 24: Crystal structure of Shi catalyst with anomeric bond lengths
Table 5: Shi catalyst anomeric centre bond lengths
Anomeric centre
Bond length 1, Å
Bond length 2, Å
A
1.437
1.435
B
1.427
1.424
C
1.434
1.433
Figure 25: Crystal structure of Jacobsen catalyst with tert-butyl distancesFigure 26: Crystal structure of Jacobsen catalyst with tert-butyl distances
In order to investigate the crystal structure of the Shi catalyst, the precursor of the catalyst shown in Figure 1 (where the dioxirane is replaced by a ketone) was searched in the PubChem database (identified as ST090500). The structure obtained was then imported as an SDF file into Mercury. The crystal structure obtained is shown in Figure 24, with the C-O bond distances at the anomeric centres labelled for examination. The catalyst structure features three anomeric centres, for which the bond lengths are tabulated in Table 5. It is notable that the average C-O bond length is 1.43 Å.[14] However, the bond lengths observed are consistent with the anomeric effect - the shortening of C-O acetal bond lengths according to the pKa of the 'conjugate' alcohol, formed by donation of the oxygen lone pair into the adjacent C-O σ* orbital.[15] The anomeric centre designated B is interesting, due to its lower bond lengths.
To similarly investigate the Jacobsen catalyst, the precursor (with an Mn-Cl bond in place of the Mn=O bond observed in Figure 2) was searched for in the Cambridge Crystal Database (CCDC) using ConQuest. The obtained crystal structure (TOVNIB03) is presented in Figure 25, with the distances between two close-approaching tert-butyl groups (on adjacent phenyl rings) labelled for a consideration of sterics. The distances measured here were those between the nearest approaching carbon atoms in each tert-butyl group - these distances were measured to be 3.994 Å and 3.695 Å. At these short distances, a significant steric interaction between the tert-butyl groups is expected. The close steric bulk of these groups means that the back face of the manganese is sterically protected, thus forcing the alkene to approach at the Mn=O end in the catalyst. Additionally of note here are the H-H distances of these ter-butyl groups - these distances range from 2.375 Å to 3.279 Å for the closest approaching H atoms. However, only the shortest of these differences are shorter than twice the van der Waals radius of hydrogen (2.40 Å),[16] indicating that these steric interactions may not be large. An alternative crystal structure was also available (TOVNIB01) - this is shown in Figure 26. Here, a shorter H-H distance between adjacent tert-butyl groups can be observed, with a distance of 2.081 Å; being shorter than twice the van der Waals radius of hydrogen,[16] it is clear that a greater steric interaction would be expected in this structure than for that considered previously.
Also of note with regards to the crystal structure of the Jacobsen catalyst is the bimolecular arrangement shown (which increases in size substantially when a whole unit cell is visualised). From Figure 25, it is easy to observe that the central Mn atom has a square-based pyramidal arrangement of ligands, which results in steric bulk for the bottom face (opposite the Mn-Cl/Mn=O bond), encouraging approach of the alkene substrate at the top face, where catalytic epoxidation is possible. Finally, it can be observed that, as expected from previous considerations of cyclohexane conformations, the cyclohexane fragment of the catalyst adopts a chair conformation in the crystal structure, resulting in a minimum energy structure.
The epoxides to be considered in this investigation are those produced through asymmetric epoxidation of styrene and 1,2-dihydronaphthalene - as a consequence, there are a number of different epoxides to be considered: (R)-styrene oxide, (S)-styrene oxide, (R,S)-1,2-dihydronaphthalene oxide and (S,R)-1,2-dihydronaphthalene oxide - these epoxides are shown in Figures 27-30. The 1H and 13C NMR of these epoxides were simulated using the DFT method previously implemented for the intermediate in the synthesis of Taxol - note that all of the spectra were simulated in chloroform, as this allowed for more similar comparison to literature. The results of the calculations were published on DSpace: (R)-styrene oxide DOI:10042/147252 ; (S)-styrene oxide DOI:10042/147256 ; (R,S)-1,2-dihydronaphthalene oxide DOI:10042/147267 ; (S,R)-1,2-dihydronaphthalene oxide DOI:10042/147264 . Note that for brevity, tables of chemical shifts (as for the Taxol intermediate discussed previously) have not been included.
The 1H and 13C NMR spectra of (R)-styrene oxide are presented in Figures 31 & 32 respectively, whilst those for (S)-styrene oxide are similarly presented in Figures 33 & 34. The 13C NMR chemical shifts for (R)-styrene oxide can be compared to literature values of[17]: 13C NMR (100 MHz, CDCl3), ppm: 137.6, 128.8, 128.2, 125.5, 52.4, 51.3.
This comparison yields a standard deviation in the difference of the computational values to the literature values of 3.539 ppm, which indicates a reasonable correlation between the data. Similar comparisons can be made for (S)-styrene oxide, using the literature data[18]: 13C NMR (CDCl3), ppm: 137.63, 128.54, 128.22, 125.52, 52.40, 51.25.
Thus, the standard deviation for the (S)- enantiomer chemical shifts was calculated to be 3.540 ppm, indicating a good reasonably good match with the literature data.
Figure 31: Computationally simulated 1H NMR spectrum of (R)-styrene oxide
Figure 32: Computationally simulated 13C NMR spectrum of (R)-styrene oxide
Figure 33: Computationally simulated 1H NMR spectrum of (S)-styrene oxide
Figure 34: Computationally simulated 13C NMR spectrum of (S)-styrene oxide
As for the styrene oxides, the 1H and 13C NMR spectra of (R,S)-dihydronaphthalene oxide are given in Figures 35 & 36, whilst those for (S,R)-dihydronaphthalene oxide are presented in Figures 37 & 38. The 13C NMR chemical shifts for (R,S)-dihydronaphthalene oxide can be compared to literature values of[19]: 13C NMR (CDCl3), ppm: 136.7, 132.6, 129.5, 128.4, 128.4, 126.1, 55.1, 52.7, 24.4, 21.8.
This comparison gives a standard deviation of 3.293 ppm. A comparison can also be made for the computed 13C chemical shifts of (S,R)-dihydronaphthalene oxide - the literature values are[20]: 13C NMR (CDCl3), ppm: 137.1, 132.9, 129.9, 129.8, 128.8, 126.5, 55.5, 53.2, 24.8, 22.2.
Here, the standard deviation was calculated to be 3.392 ppm. Given that the calculated data are the same for each enantiomer (identical chemical shift values), the difference between the two standard deviations is solely a consequence of differing literature values; however, these differences are unlikely to be significant enough to allow assignment of absolute configuration for the epoxides.
Figure 35: Computationally simulated 1H NMR spectrum of (R,S)-dihydronaphthalene oxide
Figure 36: Computationally simulated 13C NMR spectrum of (R,S)-dihydronaphthalene oxide
Figure 37: Computationally simulated 1H NMR spectrum of (S,R)-dihydronaphthalene oxide
Figure 38: Computationally simulated 13C NMR spectrum of (S,R)-dihydronaphthalene oxide
A simple qualitative consideration of the spectra for the different enantiomers for each epoxide indicates that there is very little difference observed in the 1H and 13C NMR spectra; consequently, it can be concluded that such spectra are not adequate for the assignment of absolute configurations, but they can be used to ensure that the expected epoxide is formed. This can also be shown quantitatively - the standard deviation of the differences between the (R)-styrene oxide 13C NMR chemical shifts and the (S)-styrene oxide shifts was calculated to be 0.015 ppm, indicating an extremely good match in the spectra - the spectra cannot thus be used to distinguish the configurations. Similarly, for the (R,S)-dihydronaphthalene oxide 13C chemical shifts and the (S,R)-dihydronaphthalene oxide shifts, the standard deviation was found to be 0.000 ppm, indicating that the simulated spectra are an exact match. This comparison thereby reinforces the conclusion that the NMR spectra cannot be used to distinguish the configurational isomers.
Assignment of Absolute Configurations
Computed Chiroptical Properties of Styrene Epoxides
Table 6: Optical rotations of R- and S- styrene oxides (at 589 nm)
Epoxide
Optical rotation, °
R-
-30.43
S-
30.95
Figure 39: ECD spectrum of (R)-styrene oxideFigure 41: VCD spectrum of (R)-styrene oxideFigure 40: ECD spectrum of (S)-styrene oxideFigure 42: VCD spectrum of (S)-styrene oxide
In all of the following calculations, chloroform was used as the solvent to allow comparison to literature.
To consider the chiroptical properties of (R)-styrene oxide and (S)-styrene oxide, the optical rotations were first computed - these are given in Table 6. Here, the calculation was set-up using CAM-B3LYP/6-311++g(2df,p), using the keyword polar(optrot) to calculate the optical rotations. These data were also published to DSpace: (R)-styrene oxide DOI:10042/147359 ; (S)-styrene oxide DOI:10042/147356 . A consideration of these values indicates that they are roughly equal in magnitude, with the anticipated difference in sign - the enantiomers can be concluded to be (R)-(-)-styrene oxide and (S)-(+)-styrene oxide. Literature values for R- and S- styrene oxide in chloroform are -24° (for R-)[21] and 24° (for S-)[21], where both were measured at 589 nm. Comparison of the computational values with those from literature indicates that whilst the direction of the optical rotation is the same, there is a considerable difference of ~6° (a percentage error of 26.79 % for R- and 28.96 % for S-). Hence, it is clear that the calculation results in a significant error when compared to literature - this could be reduced through an investigation of alternative computational methods. For example, a smaller percentage error in the computed values for the optical rotation could be obtained through the employment of the aug-pcS-2 basis set, which is likely to result in values closer to literature, allowing for a more accurate assignment of absolute configuration.[22]
Similarly, electronic circular dichroism (ECD) spectra can also be considered - these were calculated using CAM-B3LYP/6-311+G(d,p), with the additional keyword td(NStates=20). As before, the data were published to DSpace: (R)-styrene oxide DOI:10042/147388 ; (S)-styrene oxide DOI:10042/147394 . The ECD spectra of the R- and S- styrene epoxdies are given in Figures 39 & 40 respectively. Notably, the two can be seen to be related by reflection in the x-axis. Generally, ECD is a useful technique for the study of chirality; however, it requires multiple chromophores to examine their geometric orientation.[23] As the epoxide does not have a chromophore, the technique will not allow for assignment of the absolute configuration of the epoxides considered here.
Finally, the vibrational circular dichroism (VCD) spectra can be computed. The VCD technique has been found to be extremely useful in assignment of absolute configuration, with an advantage compared to ECD of accounting for all of the vibrational modes of a molecule.[24] These VCD spectra were previously generated during the NMR calculations previously conducted on these epoxides - spectra of the R- and S- styrene epoxdies are given in Figures 41 & 42 respectively. Here, the two VCD spectra can be considered to be related by reflection in the x-axis, as would be expected from previous examinations of optical rotation and ECD spectra. Additionally, the VCD spectra of (R)- and (S)- styrene oxide can be compared to literature spectra - the literature data available is in the range 1000-1600 cm-1(literature Fig. 9).[25] The computational spectra presented here compare well with the computational data in the literature; however, some discrepancies appear in comparison with the experimental spectra. Primarily, this difference occurs in a non-fixed baseline experimentally, but in terms of peak wavenumbers, the spectra are reasonably comparable.
Computed Chiroptical Properties of 1,2-Dihydronaphthalene Epoxides
Table 7: Optical rotations of R,S- and S,R- dihydronaphthalene oxides (at 589 nm)
Epoxide
Optical rotation, °
R,S-
155.80
S,R-
-155.83
Figure 43: ECD spectrum of (R,S)-dihydronaphthalene oxideFigure 45: VCD spectrum of (R,S)-dihydronaphthalene oxideFigure 44: ECD spectrum of (S,R)-dihydronaphthalene oxideFigure 46: VCD spectrum of (S,R)-dihydronaphthalene oxide
To compute the optical rotations of the dihydronaphthalene oxides, the calculations were run as for the styrene oxides - these results are given in Table 7. As before, chloroform was used as the solvent. The calculations performed were published to DSpace: (R,S)-dihydronaphthalene oxide DOI:10042/149588 ; (S,R)-dihydronaphthalene oxide DOI:10042/149594 . As for the styrene epoxides considered previously, the magnitudes of the optical rotations are approximately equal, whilst the values differ in sign. The optical rotations recorded at 589 nm can be compared to literature values of 133° (for R,S)[26] and -138° (for S,R)[26]. The two enantiomers of dihydronaphthalene oxide can, therefore, be identified to be (R,S)-(+)-dihydronaphthalene oxide and (S,R)-(-)-dihydronaphthalene oxide. A comparison to the literature values given reveals, as before, an agreement in the sign of the optical rotations of computations and experimental literature, whilst there is a substantial difference in magnitude (a percentage error of 17.14 % for (R,S) and 12.92 % for (S,R)).
As for the styrene oxides, the ECD spectra were calculated for the two dihydronaphthalene oxides - these are presented in Figures 43 & 44. The calculations performed were published to DSpace: (R,S)-dihydronaphthalene oxide DOI:10042/149734 ; (S,R)-dihydronaphthalene oxide DOI:10042/149740 . As discussed for the styrene oxides, ECD is not a useful technique for the assignment of the absolute configuration of the epoxides considered here.
From the NMR calculations previously conducted, the VCD spectra were also obtained - these are given in Figures 45 & 46. Unfortunately, no literature data were available for comparison here, but the two spectra can again be observed to be reflections of each other.
Analysis of the Transition States for Shi Epoxidation of Styrene
Table 8: Energies of the (R)-styrene oxide transition states 1-4 (Shi)
Transition state
Total energy, hartree
1
-1304.117596
2
-1304.115848
3
-1304.122874
4
-1304.124394
Here, Figure 46 shows a transition state in which the phenyl group of styrene is oriented towards the catalyst (endo), whilst Figure 47 shows a state where this phenyl group is oriented away from the catalyst (exo). The transition states identified as 1 and 2 have a commonality in that they both use the same oxirane oxygen for the epoxidation. Figures 48 & 49 use the opposite oxirane oxygen atom, where the transition state designated 3 has the phenyl group oriented away from the catalyst (exo), whilst 4 has the phenyl group oriented towards the catalyst (endo). The calculated total energies (sum of electronic and zero-point energies) for these transition states are given in Table 8 for consideration. Using these calculated energies, the lowest-energy transition state for the formation of the (R)-styrene epoxide can be deduced to be transition state 4 (shown in Figure 49), as this has the lowest total energy.
Table 9: Energies of the (S)-styrene oxide transition states 1-4 (Shi)
Transition state
Total energy, hartree
1
-1304.118747
2
-1304.110187
3
-1304.115130
4
-1304.123294
An examination of these transition states reveals that Figure 50, transition state 1, shows the phenyl group of styrene in an exo orientation, whilst Figure 51 (transition state 2) show the use of the same oxirane oxygen atom, but with the phenyl group in an endo orientation. Figures 52 & 53 (transition states 3 and 4) show the use of the alternative oxirane oxygen atom, where transition state 3 has an endo orientation, whilst transition state 4 has an exo orientation. The calculated total energies for the transition states are given in Table 9. An examination of these transition states reveals that transition state 4 (shown in Figure 53) is the lowest-energy transition state for the epoxidation of styrene to form (S)-styrene oxide.
Table 10: Free energies of the (R)- and (S)- styrene oxide minimum energy transition states (Shi)
Transition state
Free energy, hartree
(R)- 4
-1303.738044
(S)- 4
-1303.738503
Figure 54: Reaction coordinate diagram
Figure 55: Animation showing the asynchronous nature of the epoxide formation
Using the transition states computed to be the lowest in energy for the formation of (R)- and (S)- styrene oxides, an analysis of the preferred product using the Shi catalyst was conducted. Notably, the two transition states under consideration can be classified as diastereomers. The free energies of the transition states are given in Table 10 - the difference between the (R)- and (S)- transition states can be calculated to be 0.288 kcal mol-1, which indicates that the (S)-styrene oxide transition state is the lower in energy (a general reaction coordinate graph showing this difference between transition state free energies is given in Figure 54). Using this, the ratio of the concentrations of the two enantiomers, K, can be calculated using the equation
.
This gives a value for K of 1.626. The enantiomeric excess (ee) was then calculated using the equation[27]
,
giving a value of 23.8 % ee for the (S)- enantiomer. This can be compared to a literature value of 24 % for the (R)- enantiomer,[28] indicating an extremely good match in terms of the value of the enantiomeric excess; however, the calculations predicted (S)- to be the major enantiomer. Hence, it can be concluded that there was a significant error in the computations performed. This difference may be a consequence of the calculation method employed - it is possible that the calculations converged on structures which were not global minima for the R- and S- enantiomers, which may result in a comparison between structures which were only local minima. Further calculations could be conducted to attempt to identify the global minima for each case, allowing for a better comparison to deduce the minimum free energy transition state for the epoxidation reaction.
In the transition structure deduced to be the lowest-energy pathway for the reaction, it is interesting to note the asynchronous nature of the bond formation of the epoxide - this can be visualised in the animation shown in Figure 55, which corresponds to the single negative vibrational frequency (the frequency corresponding to reaction). Numerical evidence for this is obtained through a consideration of the partially formed bonds in the transition state - these were found to be 1.96922 Å and 2.34858 Å. Given the large difference observed in these bond lengths (0.37936 Å), it is evident that the bonds are being formed in an asynchronous manner - this is an interesting feature of the transition states.
Analysis of the Transition States for Jacobsen Epoxidation of Styrene
As for the Shi epoxidation of styrene, the Jacobsen catalyst for epoxidation of styrene can also be considered. Here, there are four transition states to be considered, two for (R)-styrene oxide and two for (S)-styrene oxide. Alongside which enantiomeric epoxide is formed, there are two possibilities regarding the orientation of the alkene with respect to the catalyst - endo or exo. Firstly, the transition states for (R)-styrene oxide will be considered - these are presented in Figures 56 & 57 (DOI:10.6084/m9.figshare.860446 ; DOI:10.6084/m9.figshare.860449 ).
Figure 56: (R)-Styrene transition state 1 (endo)
Figure 57: (R)-Styrene transition state 2 (exo)
Table 11: Energies of the (R)-styrene oxide transition states 1-2 (Jacobsen)
Transition state
Total energy, hartree
1
-3344.624718
2
-3344.626384
Here, Figure 56 shows an endo arrangement of the substrate with respect to the catalyst, whilst Figure 57 shows an exo orientation. The total energies of the two transition states are given in Table 11 - a comparison of the two reveals that transition state 2 is the lower in energy of the (R)-styrene oxide transition states.
Table 12: Energies of the (S)-styrene oxide transition states 1-2 (Jacobsen)
Transition state
Total energy, hartree
1
-3344.632372
2
-3344.627650
An examination of the transition structures reveals that Figure 58 shows an endo orientation of substrate and catalyst, whilst Figure 59 presents an exo arrangement. As for the (R)- enantiomer, the transition state energies are given in Table 12 - it is clear from the data that transition state 1 is the lower in energy.
Table 13: Free energies of the (R)- and (S)- styrene oxide minimum energy transition states (Jacobsen)
Transition state
Free energy, hartree
(R)- 2
-3343.962162
(S)- 1
-3343.969197
Figure 60: Animation showing the asynchronous nature of the epoxide formation for Jacobsen epoxidation to (S)-styrene oxide
In order to determine the preferred enantiomer of the Jacobsen epoxidation of styrene, the free energies of the two transition states identified as the respective minima can be compared - these are given in Table 11. It is apparent from these data that the minimum (S)- transition state is the lowest energy transition state for the epoxidation of styrene using the Jacobsen catalyst. This can be further investigated through calculation of the difference in free energy between the two transition states, giving a value of 4.41 kcal mol-1. This can be used, as before, to find the K value for the two diastereomeric transition states, with a value of 1721.363. This can be used to find the enantiomeric excess using the equation stated previously - the enantiomeric excess for this epoxidation was calculated to be 99.9 % for the (S)- enantiomer. This result can be compared to that of experimental literature: an enantiomeric excess of 53 % for the R- enantiomer.[29] Clearly, the calculations performed in this computational investigation have resulted in a significant error when compared to literature, in both the value of the enantiomeric excess (a percentage error of 88.5 %) and the prediction of the major enantiomer formed in the epoxidation. As discussed with regards to the Shi epoxidation of styrene, the difference in the stereochemistry of the major enantiomer may be a consequence of convergence to local minima, as opposed to global minima in terms of transition states, resulting in inaccurate ordering of energies. The difference in enantiomeric excess may be due to experimental occurrences that are not predicted computationally, such as solvent effects, as the calculations use a mean field approximation, as opposed to individual interactions. Additionally, catalyst degradation could also occur, potentially resulting in achiral species which are still capable of epoxidising a substrate, resulting in a degree of non-asymmetric epoxidation.
As for the Shi epoxidation, the asynchronous nature of the epoxide bond formation can be observed in the imaginary vibrational mode of the transition state - this animation is shown in Figure 60. Here, the observed bond lengths were 1.80326 Å and 2.62009 Å, highlighting the substantial asymmetry to the transition state.
Analysis of the Transition States for Shi Epoxidation of 1,2-Dihydronaphthalene
As for the Shi epoxidation of styrene, eight transition states can be considered for the Shi epoxidation of 1,2-dihydronaphthalene (four for (R,S)- and four for (S,R)-) in order to determine which absolute configuration of the epoxide will be formed when using the Shi catalyst. First, the (R,S)-styrene oxide transition states will be considered using the calculation results provided - these are presented in Figures 61-64 (DOI:10.6084/m9.figshare.832492 ; DOI:10.6084/m9.figshare.832510 ; DOI:10.6084/m9.figshare.832511 ; DOI:10.6084/m9.figshare.832512 )
Figure 61: (R,S)-Dihydronaphthalene transition state 1 (endo)
Figure 62: (R,S)-Dihydronaphthalene transition state 2 (exo)
Figure 63: (R,S)-Dihydronaphthalene transition state 3 (exo)
Figure 64: (R,S)-Dihydronaphthalene transition state 4 (endo)
Table 14: Energies of the (R,S)-dihydronaphthalene oxide transition states 1-4 (Shi)
Transition state
Total energy, hartree
1
-1381.543819
2
-1381.547260
3
-1381.556204
4
-1381.549901
Here, Figures 61 & 62 show the use of one of the dioxirane atoms, with an endo and exo arrangement of substrate and catalyst respectively; Figures 63 & 64 show the use of the alternative dioxirane atom, with exo and endo orientations respectively. The total energies of each of these transition states 1-4 are given in Table 14 - it can be observed that transition state 3 (Figure 63) is the lowest energy transition state for the epoxidation to (R,S)-dihydronaphthalene oxide.
Figure 65: (S,R)-Dihydronaphthalene transition state 1 (exo)
Figure 66: (S,R)-Dihydronaphthalene transition state 2 (endo)
Figure 67: (S,R)-Dihydronaphthalene transition state 3 (endo)
Figure 68: (S,R)-Dihydronaphthalene transition state 4 (exo)
Table 15: Energies of the (S,R)-dihydronaphthalene oxide transition states 1-4 (Shi)
Transition state
Total energy, hartree
1
-1381.552801
2
-1381.536075
3
-1381.547613
4
-1381.558132
The transition structures presented in Figures 65 & 66 show exo and endo orientations of substrate and catalyst respectively, whilst Figures 67 & 68 show endo and exo arrangements, making use of the second (inequivalent) doctrine oxygen atom. The calculated total energies for the transition states are given in Table 15. An examination of these energies reveals that transition state 4 (Figure 68) is the lowest in energy for Shi epoxidation to (S,R)-dihydronaphthalene oxide.
Table 16: Free energies of the (R,S)- and (S,R)- dihydronaphthalene oxide minimum energy transition states (Shi)
Transition state
Free energy, hartree
(R,S)- 3
-1381.134059
(S,R)- 4
-1381.136239
Figure 69: Animation showing the asynchronous nature of the epoxide formation for Shi epoxidation to (S,R)-dihydronaphthalene oxide
In order to identify the preferred product of the Shi epoxidation of 1,2-dihydronaphthalene, the free energies of the minimum transition state for each of the two possible enantiomeric products were compared - these energies are given in Table 16. It is evident that the (S,R)- transition state has the lower free energy, making it the major product of the reaction. As before, the K value can be calculated (using a free energy difference of -5.724 kJ mol-1), giving a value of 10.063. This can then be used to find the enantiomeric excess for the reaction, which was calculated to be 81.9 % for the (S,R) enantiomer. Comparison to literature data of an enantiomeric excess of 32 % for the (S,R)- enantiomer[30] reveals that whilst there is a large difference in enantiomeric excess value (155.9 %), the major enantiomer found computationally matches that observed experimentally. Here, the substantial difference between computational and experimental enantiomeric excess values can be attributed to laboratory decomposition pathways of the Shi catalyst which are not modelled in the computational investigation. This could result in species capable of epoxidation, but without enantiomeric preference, resulting in lower enantiomeric excess in laboratory experiments.
As for the previous systems, the sychronicity of the bond-formation in the epoxide transition structure can be considered - the animation displaying the vibrational mode corresponding to reaction is presented in Figure 69. For this system, the bond lengths were found to be 2.06703 Å and 2.18948 Å. This distances indicate that the bond-formation is more synchronous than for the previous systems, where differences on the order of 0.8 Å were observed.
Analysis of the Transition States for Jacobsen Epoxidation of 1,2-Dihydronaphthalene
As for the Jacobsen epoxidation of styrene, the Jacobsen epoxidation of 1,2-dihydronaphthalene can also be investigated. Here, there are four transition states to be considered, two for (R,S)-dihydronaphthalene oxide and two for (S,R)-dihydronaphthalene oxide. Alongside which enantiomeric epoxide is formed, there are two possibilities regarding the orientation of the alkene with respect to the catalyst - endo or exo. Firstly, the transition states for (R,S)-dihydronaphthalene oxide will be considered - these are presented in Figures 70 & 71 (DOI:10.6084/m9.figshare.909346 ; DOI:10.6084/m9.figshare.907332 ).
Figure 70: (R,S)-dihydronaphthalene transition state 1 (endo)
Figure 71: (R,S)-dihydronaphthalene transition state 2 (exo)
Table 17: Energies of the (R,S)-dihydronaphthalene oxide transition states 1-2 (Jacobsen)
Transition state
Total energy, hartree
1
-3422.058301
2
-3422.059652
Figures 70 & 71 show endo and exo orientations of catalyst and substrate, respectively, for the (R,S)- enantiomer. The total energies of the two transition states are given in Table 17 - an examination of these energies shows that transition state 2 is the lower-energy path to the (R,S)-dihydronaphthalene oxide.
Figure 72: (S,R)-dihydronaphthalene transition state 1 (endo)
Figure 73: (S,R)-dihydronaphthalene transition state 2 (exo)
Table 18: Energies of the (S)-styrene oxide transition states 1-2 (Jacobsen)
Transition state
Total energy, hartree
1
-3422.068538
2
-3422.060548
Finally, Figures 72 & 73 present transition structures for Jacobsen epoxidation to (S,R)-dihydronaphthalene oxide, where Figure 72 shows an endo orientation of the substrate with respect to the catalyst, whilst Figure 73 presents the alternate exo arrangement. As for the (R,S)- enantiomer, the transition state energies are given in Table 18 - it is clear from the data that transition state 1 is the lower in energy.
Table 19: Free energies of the (R,S)- and (S,R)- dihydronaphthalene oxide minimum energy transition states (Jacobsen)
Transition state
Free energy, hartree
(R,S)- 2
-3421.359499
(S,R)- 1
-3421.369033
Figure 74: Animation showing the asynchronous nature of the epoxide formation for Jacobsen epoxidation to (S,R)-dihydronaphthalene oxide
An examination of the free energies presented in Table 19 shows that the (S,R)- enantiomer is the preferred product of the reaction - the free energy difference calculated was -25.032 kJ mol-1. This can be used to compute K for the epoxidation, giving a value of 24282.644. Finally, this value can be used to compute the enantiomeric excess - this was found to be 99.99 % for the (S,R) enantiomer. This result can be compared to literature data of an enantiomeric excess of 52 % for the R,S- enantiomer.[31] Evidently, issues have occurred in the computational investigation, resulting in a large difference in enantiomeric excess (92.3 %) and prediction of the incorrect major enantiomeric product. The cause of the identification of the S,R enantiomer, rather than R,S, as the major product is likely a consequence of the computations not converging to global minima and instead finding local minima, whose energies may have a different ordering to the global minima. The large difference in enantiomeric excess is, as previously discussed, most probably due to some experimental factors affecting the observed enantiomeric excess, such as catalytic decomposition and individual solvent interactions.
In the same manner as that employed for previous systems, the synchronicity of the partially-formed epoxide bonds can be considered - the animation demonstrating this is shown in Figure 74. The bond lengths observed here were 1.81301 Å and 2.61552 Å, indicating that the bond-formation is significantly asynchronous in nature.
Investigating the Non-Covalent Interactions (NCI) in the Active-Site of the Reaction Transition State
Figure 75: NCI surface for (S)-styrene oxide transition state 4, using the Shi catalyst
Figure 76: NCI surface for (S)-styrene oxide transition state 3, using the Shi catalyst
Figure 77: NCI surface for (S)-styrene oxide transition state 1, using the Jacobsen catalyst
In order to conduct a non-covalent interaction (NCI) analysis, the (S)-styrene oxide transition state 4, using the Shi catalyst, was considered. After generating a cube in GaussView (total density, fine grid), this was converted to an NCI surface (.xyz and .jvxl files) - the NCI surface generated is presented in Figure 75. Additionally, Figure 76 shows the NCI surface generated for transition state 3 for this system. For comparison, Figure 77 presents an NCI surface generated for the Jacobsen transition state 1 for (S)-styrene oxide.
A consideration of the exo transition structure in Figure 75 makes this exo arrangement more apparent than a simple geometric examination - the phenyl ring of the substrate can be observed to have no non-covalent interactions with the catalyst. This is in contrast to the structure shown in Figure 76, the corresponding endo transition structure, where a significant mildly attractive interaction between the ring and the catalyst is evident. Despite this lack of interaction in Figure 75, there is a mild attraction between the reacting alkene and the catalyst, which explains why the endo transition state does not have a lower free energy than the exo orientation. Of note in both of these NCI surfaces is the intramolecular H-bond interaction between oxygen and hydrogen atoms which are geometrically close - these strong interactions are visualised as a blue surface in the space separating the atoms.
For comparison, an NCI surface for a Jacobsen transition state is also presented in Figure 77. Here, interactions between the phenyl rings of the substrate and the catalyst are immediately evident - these mildly attractive interactions have a large surface area between the two rings. Additionally significant here is the presence of the chlorine atom and the mildly/strongly attractive interactions between the catalyst and this atom - these make it clear that full dissociation has not occurred in the transition state. The position of this atom also has implications for the epoxidation mechanism - it is apparent that the the Mn=O bond is produced in situ by substitution of the Mn-Cl bond at the opposite face of the square-based pyramidal structure.
Investigating the Electronic Topology (QTAIM) in the Active-Site of the Reaction Transition State
Figure f: QTAIM analysis of (S)-styrene oxide transition state 4, using Shi catalyst
To investigate the electronic topology of transition state 4 for the Shi epoxidation to (S)-styrene oxide, a QTAIM analysis was conducted on this transition state using Avogadro2, allowing for the identification of bond critical points, BCPs. The molecular graph produced in this QTAIM analysis is presented in Figure 78.
This QTAIM analysis further supports the previous observation of asynchronous bond formation in the transition states - this can be observed due to the presence of a BCP between one of the dioxirane oxygen atoms and the terminal carbon atom of styrene. This BCP is clearly associated with the partially formed covalent bond between the two atoms and is approximately centred between the two; notably, there is not a BCP between this oxygen atom and the second vinyl carbon atom of styrene. As in the NCI analysis of such surfaces, H-bond interactions are also reflected in the BCPs - these can be observed between oxygen and hydrogen atoms, with approximately central BCPs (these are not associated with covalent bonds). These interactions can be noted between the catalyst and the allyl hydrogen atoms of the substrate (as well as intramolecularly in the catalyst), potentially resulting in stabilisation of the transition state and explaining the computational observations.
Suggesting New Candidates for Investigation
Figure 79: 2-StilbazoleFigure 80: Shi catalyst derivative
In addition to those alkene epoxides considered in this computational investigation, further epoxides could be considered in this manner. Such suitable epoxides were searched for using Reaxys, with an ORP.ORP>'300' (optical rotatory power) and an IDE.MW<'200.0' (molecular weight). Of the results, the epoxide of 2-stilbazole (Reaxys Registry Number: 1619667), shown in Figure 79 was deemed to be appropriate due to its optical rotatory power of 326°,[32] with a molecular weight of 197.236 g mol-1. Additionally, the corresponding alkene 2-stilbazole was found to be commercially available via Sigma-Aldrich, with a product number of S439843.[33]
Through analysis of the chiroptical properties of the 2-stilbazole epoxide, conducted as for the styrene and 1,2-dihydronaphthalene epoxides in this experiment, computational properties could be obtained for comparison with literature, allowing for a computational assessment of the literature value. In evaluating the computed chiroptical properties in this investigation, it was decided that an alternative method could be employed for better results (the aug-pcS-2 basis set) - this should allow for a more accurate comparison to literature.
In addition to the investigation of this further alkene, different catalytic systems could also be considered, such as the derivative of the Shi catalyst shown in Figure 80, which has been found to give a high enantiomeric excess for cis-substituted alkenes (such as 1,2-dihydronaphthalene).[28] In analysing the transition states of such substrate-catalyst systems, the locating of a greater number of possible transition states should be attempted, with the aim of identifying the global minimum. This will hopefully enable the correct prediction of the preferred product and the enantiomeric excess values should be closer to literature.
References
↑ 1.01.1O. A. Wong , B. Wang , M-X Zhao and Y. Shi, J. Org. Chem., 2009, 74, pp. 6335-6338. DOI:10.1021/jo900739q
↑ 2.02.12.2E. N. Jacobsen, W. Zhang, A. R. Muci, J. R. Ecker, L. Deng, J. Am. Chem. Soc., 1991, 113, pp. 7063-7064. DOI:10.1021/ja00018a068
↑A. Burke , P. Dillon , Kyle Martin and T. W. Hanks, J. Chem. Educ., 2000, 77, pp. 271-272. DOI:10.1021/ed077p271
↑D. C. Braddock, H. S. Rzepa, J. Nat. Prod., 2008, 71, pp. 728-730. DOI:10.1021/np0705918
↑F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen, R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1987, 12, pp. S1-S19. DOI:10.1039/P298700000S1
↑A. J. Briggs, R. Glenn, P. G. Jones, A. J. Kirby, P. Ramaswamy, J. Am. Chem. Soc., 1984, 106, pp. 6200-6206. DOI:10.1021/ja00333a014
↑K. Huang, H. Wang, V. Stepanenko, M. De Jesús, C. Torruellas, W. Correa, M. Ortiz-Marciales, J. Org. Chem., 2011, 76, pp. 1883-1886 (S1-S126). DOI:10.1021/jo102294j
↑K. Smith, C-H. Liu, G. A. El-Hiti, Org. Biomol. Chem., 2006, 4, pp. 917-927. DOI:10.1039/B517611P
↑M. W. C. Robinson, K. S. Pillinger, I. Mabbett, D. A. Timms, A. E. Graham, Tetrahedron, 2010, 66, pp. 8377-8382 (S1-S6). DOI:10.1016/j.tet.2010.08.078
↑ 21.021.1D. C. Forbes, S. V. Bettigeri, S. A. Patrawala, S. C. Pischek, M. C. Standen, Tetrahedron, 2009, 65, pp. 70-76. DOI:10.1016/j.tet.2008.10.019
↑E. D. Hedegard, F. Jensen, J. Kongsted, J. Chem. Theory Comput., 2012, 8, pp. 4425-4433. DOI:10.1021/ct300359s
↑Comprehensive Chiroptical Spectroscopy, Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, ed. N. Berova, P. L. Polavarapu, K. Nakanishi, R. W. Woody, John Wiley & Sons, Inc., Hoboken, New Jersey, 2011, pp. 318-319
↑F. L. Cherblanc, Y-P. Lo, W. A. Herrebout, P. Bultinck, H. S. Rzepa, M. J. Fuchter, J. Org. Chem., 2013, 78, pp. 11646-11655. DOI:10.1021/jo401316a
↑S. S. Thakur, S-W. Chen, W. Li, C-K. Shin, S-J. Kim, Y-M. Koo, G-J. Kim, J. Organomet. Chem., 2006, 691, pp. 1862-1872/ DOI:10.1016/j.jorganchem.2005.12.044
↑ 26.026.1D. R. Boyd, N. D. Sharma, R. Agarwal, N. A. Kerley, R. A. S. McMordie, A. Smith, H. Dalton, A. J. Blacker, G. N. Sheldrake, Chem. Commun., 1994, 14, pp. 1693-1694. DOI:10.1039/C39940001693
↑S. T. Schneebeli, M. L. Hall, R. Breslow, R. Friesner, J. Am. Chem. Soc., 2009, 131, pp. 3965-3973. DOI:10.1021/ja806951r
