Molecular Mechanics (MM) is a useful tool to understand three dimensional effects in chemical reactivity, particularly with respect to the different conformations that some larger molecules can adopt. In this project, we will investigate a series of organic molecules that can adopt various conformations in three dimensions and determine using Molecular Mechanics via ChemBio3D Ultra 12.0 the conformation that will have the highest stabilisation energy and interpret the implications of these results.
Note that there are limitations in the molecular dynamics approach, in the sense that it is interpolative rather than extrapolative, so that the approach will only be useful if the molecule under investigation is similar to well-documented chemistry. This should be taken into account in the following calculations, but the compounds under investigation in this project are well known with properties that have been extensively reported.
The Hydrogenation of Cyclopentadiene Dimer
The hydrogenation of cyclopentadiene dimer is a common reaction. However the regioselectivity with respect to the hydrogenated alkene bond is not trivial, and analysis of their respective energies must be examined to determine the most thermodynamically favourable reaction.
Figure 1 - Cyclodimersation to form exo (1) and endo (2) isomer
In the cyclodimersation (4+2 cycloaddition) of cyclopentadiene, both the exo (1) and endo (2) isomer can be formed (w.r.t. to the stereochemistry of the indicated hydrogens. The MM2 and MMFF94 force field options were used to optimise the structure on ChemBio3D, i.e. minimise the energy. The results of the two are presented and compared in Table 1 below:
Table 1 - Comparison of optimised structure of exo and endo isomers (1&2)
We can see that the total energy of the two isomers are relatively similar, as are their relative contributions. Nonetheless it can be seen that the exo-isomer is 2.12 kcal/mol lower in energy than the endo-isomer, suggesting that it is thermodynamically the more favourable isomer and under thermodynamic conditions it will be the major product.
However, as seen in literature [1], the endo-selectivity is seen as the major product uner kinetic conditions. This was proposed by Alder et al. to be the consequence of the consequence of a plane-to-plane orientation of diene and dienophile with "maximum accumulation of double bonds". Thus the complex formation between the reactants was interpreted to be the reason for the endo preference.
Hydrogenation of endo dimers
The cyclodimerisation of cyclopentadiene allows for an interesting follow-up observation of the hydrogenation of the endo-dimer which results in two possible isomers (3) and (4) (as presented in Figure 2). The selectivity for either one of the alkenes can be determined by calculating their respective minimum energies using MM2 forcefield on ChemBio3D. The results are presented in Table 2 below.
Figure 2 - Hydrogenation of Endo-isomer to form (3) and (4)
Table 2 - Comparison of optimised structure of dihydro derivatives 3 and 4 using MM2
Again looking at the total energies, we see that the hydrogenated derivative, isomer 4 is more stable than isomer 3 by approximately 4.5 kcal/mol, indicating that under thermodynamic conditions, isomer 4 will be the major product. It is worth mentioning as an additional note that the difference in energy is almost entirely resulting from the respective contributions of bend energy. Isomer 3 is higher in energy by 5.3 kcal/mol because the C-C=C bond is at a more strained angle compared to the C-C=C bond in isomer 4. This is illustrated in the table below:
Table 3 - Comparison of C-C=C bond angle between isomer 3 and 4
Isomer 3
Isomer 4
C-C=C Bond Angle: 108o
C-C=C Bond Angle: 112o
As it can be seen, the the C-C=C bond is strained at a smaller angle in isomer 3 (108 degrees) compared to isomer 4 (112 degrees). Although both cases deviate quite significantly from ideality (120 degrees for sp2), the higher strain of isomer 3 will result in an overall more destabilised structure. This interpretation is in good agreement with literature, suggesting that the first hydrogenation occurs at the norbornene ring (isomer 4) at a much higher reaction rate compared the hydrogenation on the cyclopentene ring (isomer 3)[2].
Atropisomerism related to the Synthesis of Taxol
Another interesting compound that shows interesting properties in terms of its three dimensional components is a key intermediate in the synthesis of Taxol which can take two forms, referred to in this paper as atropisomers 9 and 10 (illustrated below). It can be seen that the sole difference between the two isomers is the orientation of the carbonyl bond which is referred to as "up" in 9 and "down" in 10. By using molecular mechanics, we will similarly to previous procedures, determine the most stable atropisomer thus suggesting the dominating compound in the oxy-cope rearrangement.
In this section, we will also experiment with other molecular mechanics systems, namely the MMFF94 force field. Thus Table 4,showing the results of the molecular mechanics calculations on ChemBio3D, we will present the results obtained from both the MM2 and MMFF94 force fields.
Figure 3 - Oxy-Cope rearrangement to form atropisomers 9 and 10, key intermediates in synthesis of Taxol
Table 4 - Comparison of optimised two atropisomers (9 and 10) using MM2 and MMFF94
In the comparison of the two atropisomers 9 and 10, we see that the total energy difference based on the MM2 force field calculations is ~4.41 kcal/mol. Thus we can deduce that thermodynamically, atropisomer 10 is the more stable isomer and over time the rearrangement will equilibriate to favour the isomer. This result is supported by the MMFF94 calculation, which albeit giving higher energies, shows the same trend, with the 10 isomer being more stable by ~5.76 kcal/mol.
In comparing the two different force fields (MM2 and MMFF94) it is quite obvious that they do not yield similar results. The MMFF94 gives significantly larger values and also optimises to a different shape (MMFF94: Atropisomer 9 and Atropisomer 10). However this does not necessarily mean that it is a less accurate tool, but simply that it is based on different parameters which account for more/different factors. Nonetheless the total energy difference is relatively similar to the MM2 calculation and cements the interpretation of atropisomer 10 being the more stable.
Furthermore it is worth noting that optimisation that leads to the minimum energy involves the hexane ring adopting a chair conformation (which coincides with our understanding of the chair as the most stable conformation). Optimisations that involved the twist-boat conformation were higher in energy by approximately 3-5 kcal/mol.
Hydrogenation of Atropisomer 9 and 10 (Hyperstable alkenes)
It was noted in the report of the synthesis of Taxol that in further functionalisations of the alkene in the atropisomers, an abnormally slow reaction was observed. This was investigated by determining the energies of the hydrogenated derivatives of the atropisomer 10 (again with both MM2 and MMFF94) and comparing it with the energy of the original isomer. The results are presented in Table 6 below.
Table 6 - Comparison of optimised atropisomers 9 and its hydrogenated derivative using MM2 Forcefeld and MMFF94 Forcefeld
The observation of the slow alkene reaction results from the concept of hyperstability. Specifically the hyperstability of alkene is observed when an alkene is positioned at a bridgeheads. Where destabilisation is expected due to the added strain on the bonds, it is observed to be more preferred in larger polycyclic systems such as the atropisomerk in question. This unexpected stabilisation is referred to as the hyperstability of alkene, and with this added stabilisation, a slower reaction will be expected [3].
This interpretation is directly supported by the calculations presented above, where the energy of the hydrogenated derivative of atropisomer 10 is significantly higher compared to the original by ~14.53 kcal/mol. Thus with the product of the hydrogenation being higher in energy than the reactant it stands to reason that the reaction will be unfavourable and thus slow.
Modelling Using Semi-empirical Molecular Orbital Theory
In the first section we demonstrated some of the limitations of the molecular mechanics approach, where the main weakness that was highlighted was that it does not account for secondary orbitals and their contribution to the energy of stabilisation. So in this section we will employ the semi-empirical molecular orbital theory as a modelling tool to account for orbital interactions. By using a variety of methods we will highlight the electrons influence on bonds and other properties in a series of molecules.
Regioselective Addition of Dichlorocarbene to a diene
First we will look at a diene molecule which will be referred to in this paper as compound 12, and its reactivity with electrophilic reagents such as dichlorocarbene in relation to orbital control. The MOPAC electronic structure method provides an optimised geometry that accounts for any stereoelectronic effects in a molecule. Thus by comparing the molecular mechanics method MM2 and MOPAC, we can illustrate the transition from a purely classical mechanics treatment of a molecule to a quantum mechanics treatment including wave description of the electrons.
The reaction of compound 12 with dichlorocarbene is given in the illustration below, involving attack of the carbene on the endo face to avoid steric hindrance. The MM2 force field was employed on the molecule to clean the geometry and then the MOPAC/PM6 method was used on the minimised geometry of compound 12. The results were obtained and presented in Table 6 below.
Figure 4 - The electrophilic addition of dichlorocarbene on Compound 12
Table 6 - Comparison of MM2 and MOPAC optimised structures of compound 12
From the total energy, we see that there is a rather significant difference between the two methods. To obtain a clearer comparison, the two structures of the molecules were superimposed on each other, and the overlay image is presented below along with some of the deviation distances. It can be seen that based on the structure there is very little deviation, most of which are limited to below 0.1 Å. However if we investigate the MO orbitals that were obtained from Gaussview 5.0, we might shed light on some additional factors that are included in the MOPAC/PM6 method.
Figure 5 - Overlay of MM2 and MOPAC optimised molecule
Table 6 - MO orbitals generated from Gaussview of HOMO and LUMO orbitals of Compound 12
MO 47 (HOMO -1)
MO 48 (HOMO)
MO 49 (LUMO)
MO 50 (LUMO +1)
MO 51 (LUMO +2)
-0.24519
-0.23588
0.01555
0.02025
0.03613
Having obtained the molecular orbitals (HOMO, HOMO -1, LUMO, LUMO +1, LUMO +2) it becomes obvious that the orbitals differentiate very significantly for the two alkenes. It is particularly with respect to the HOMO orbital that will be the point of focus in the interpretation. From the HOMO π-bonding orbitals it is evident that the π orbitals on the endo-alkene are significantly larger than exo-alkene signifying a larger electron density, and the more preferable site of attack from the electrophilic dichlorocarbene component.
Note that in the LUMO orbital, the high electron density is shifted to the exo-alkene, but as this is an antibonding π orbital interaction, it still supports our interpretation of the endo-alkene being the favoured site of attack.
It is interesting and worth noting that the LUMO +2 MO depicts the σ* C-Cl orbital, and shows a large electron density around the region, resulting in the electron density being entirely drawn from the endo-alkene. This again cements our interpretation and justifies the endo-alkene being the favoured position of electrophilic attack.
Table 7 - Major IR frequencies of compound 12
Stretch
Frequency / cm-1
Intensity / 10-40 esu2cm2
Vibration
C-Cl
805
20.45
Exo C=C
1735
4.23
Endo C=C
1748
2.89
The three major vibrations related to the investigation of difference in the exo- and endo-alkene in compound 12 are highlighted in Table 7 above. Several C-Cl stretches were observed, and the one with the highest intensity was selected for discussion. It can be seen that the stretching frequency is significantly smaller than the alkene stretches, which reflects the size of the the chlorine atom and the result of a weaker bond c.f. the C=C bond.
The most interesting observation from the table is the difference in stretching frequency of the exo- and endo-alkene. This supports the interpretations made based on molecular orbitals. The exo-alkene has a lower frequency than the endo-alkene (albeit small difference of ~13 cm-1) suggesting that the bond is slightly weaker, which following previous interpretations must result from a smaller electron density and thus less prone to electrophilic attack by the dichlorocarbene.
Monosaccharide chemistry and the mechanism of glycosidation
Another reaction that can adopt different conformations in reaching the product is the glycosidation reaction, which involves the replacing of group X (e.g. halides, esters etc.) with a nucleophile Nu as shown in the figure below. The reaction is highly stereospecific, where the stereochemistry of the C-OAc bond (shown in red) controls the stereochemical outcome of the C-Nu bond by nucleophilic attack at the anomeric carbon. This effect is referred to as the anomeric effect which results from participation of the neighbouring acetyl group, as seen in intermediate A(1,2,3 and 4) forming the second-intermediate oxenium cation B(1,2,3 and 4 respectively) which can either allow for top- or bottom-face attack from the incoming nucleophile. With long reaction times the most thermodynamically stable product will result, owing to the anomeric effect. In this section we will investigate using the techniques implemented above to determine the most thermodynamically stable product.
Figure 5 - ChemDraw illustration of all possible conformations of glycosidation leading to either the α- or β-anomer product
As previously, the structures of the four different conformations of the anomer (as both intermediates A and B) were drawn on ChemBio3D and optimised with both MM2 and MOPAC calculations. It should be noted that different conformations arise both from the orientation of the acetyl group -OAC (axial/equatorial) and the orientation of the carbonyl group on the acetyl -CO (up/down), such yielding four different possible anomers for intermediates A and B. The results are presented below in Table 8 and 9, displaying both the total energy obtained from MM2 and the heat of formation given by the MOPAC optimisation.
Table 8 - Optimisation results of the intermediate oxonium cation A using MM2 and MOPAC
At this point it is worth making a brief comment regarding the the molecular mechanics approach (MM2) and the quantum mechanics approach (MOPAC/PM6), in terms of how they differ in energy. As we know MM2 deals via a purely classical mechanics approach and thus has difficulty calculating for the non-classical oxenium cation as well as any particular orbital interactions. However when comparing with the trend observed in the MOPAC/PM6 results, which accounts for the stabilising effect of the acetyl oxygen lone pair on the cationic intermediate, we see that MM2 is still capable of determining the relative stabilities of the different conformers.
From the tables above, we see that it is A2/B2 and A3/B3 that has the lowest energy (in both intermediates) in their respective acetyl orientation pairs. This is neither a result of the orientation of the acetyl group (axial/equatorial) nor a result of the trajectory of the nucleophile attack, but rather (as evident from the jmol images) the orientation of the carbonyl group relative to the anomer carbon and the close proximity of the two components in going from A to B. Essentially the higher stabilisation of conformer 2(A/B) and 3 (A/B) corresponds to the alignment of the carbonyl oxygen with the anomer carbon as intermediate A to allow for maximum orbital interaction. This interaction directly transfers over to the stability of intermediate B where 2 and 3 (now more comparable in energy) are significantly lower in energy compared to their respective counterparts (~ 10 kcal/mol). Furthermore it is worth noting that their respective anomer products have the acetyl group and the nucleophile in trans- geometry which again connects back to the particular carbonyl oxygen and anomer carbon bonding which allows for the nucleophile to attack at the face which is trans- to the acetyl group.
Finally attention can be drawn to the particular angle at which the carbonyl bond oriented to the oxenium cation-carbon bond takes when forming intermediate B, as these results help to cement the interpretation above. From Table 10 below, it can be seen that the distance between the carbonyl oxygen on the acetyl group and the anomer carbon significantly shorter in 2A and 3A compared to 1A and 4A. This already gives some indication of the higher stability of 2A and 3A which allows for better orbital interaction. To further support the argument we can direct our attention at the bonding angle in which the 2A and 3A conformation are close to the Burger-Dunitz angle of 107˚, which indicates the optimised angle that a nucleophile can approach via collision to an unsaturated site, hence resulting in a lower energy. Note that conformation 4A is also close the the Burger-Dunitz angle. However its closeness to ideality in the bonding angle is not enough to compensate for the large distance (larger than 1A) and is thus higher in energy c.f. 2A and 3A.
Table 10 - Comparison of bond angle and bond length of conformers
Mini-Project: Simulation of Spectroscopic Data for a Literature Molecule
Useful synthetic routes developed by chemists to obtain a desired compound will most often result in a mixture of products being formed. To be able to identify the desired compound amongst the products, useful techniques have been devised to allow the accurate identification of a complex molecule. For solids, X-ray crystallography, AFM and neutron diffraction prove to be useful tools, but in the case of liquids (or solutions) a variety of spectroscopic techniques can be used ranging from UV-Vis spectroscopy, to IR to NMR spectroscopy. When chemists obtain a new compound, it is often compared to the spectroscopic information of very similar known compounds. However computational modelling proves to be a powerful tool in predicting the spectroscopic data of any compound.
In this section we will aim to investigate the product yielded in a reaction cited in a literature source, and use various techniques in computational modelling to predict the spectroscopic properties associated with the compound.
Spectroscopy of an intermediate related to the synthesis of Taxol
Before we look into a reaction quoted in literature, it is useful to employ the aforementioned modelling techniques to obtain the 1H NMR and 13C NMR spectra of an intermediate related to the synthesis of Taxol, referred to in this section as Compound 18. It is worth noting that the intermediate can exist in two different conformations, the other conformer being where the carbonyl oxygen is orientated in the opposite direction. The compound is presented below both in 2-D and 3-D, both labelled by their carbons. The labelled numbers will have a significance in the interpretation of the NMR spectra predicted by computational modelling.
Figure 6 - Compound 18 labelled on Gaussview 5.0Figure 7 - Compound 18 labelled on ChemDraw Pro 12.0
In order to obtain the NMR spectra (1H and 13C), the initial structure was minimised via MM2 force field. Then the Gaussian interface was implemented using the method DFT=B3LYP - 6-31 (d, p) with chloroform as solvent, which was submitted to HPC (DOI:10042/24260 ). The keyword of the output was altered (# mpw1pw91/6-31G(d,p) NMR SCRF=(CPCM,Solvent=chloroform)) and submitted to HPC again (DOI:10042/24261 ). The log file was extracted to finally give the NMR spectra. Both spectra were referenced against TMS mPW1PW91/aug-cc-pvtz CDCl3 GIAO to give the results given below.
Table 11 - 1H NMR chemical shifts from Gaussview 5.0 compared to literature values
Atom #
Degeneracy
Calc. chemical shifts (ppm)
Literature chemical shifts
26
1
6.01
5.21
32
1
3.40
3.00-2.70
42
1
3.31
3.00-2.70
41
1
3.22
3.00-2.70
33
1
3.15
3.00-2.70
43
1
2.98
3.00-2.70
44, 27
2,3
2.96
3.00-2.70, 2.70-2.35
30
1
2.63
2.70-2.35
38
1
2.58
2.70-2.35
36
1
2.42
2.70-2.35
35
1
2.29
2.20-1.70
34
1
2.24
2.20-1.70
25
2
2.23
2.20-1.70
31
1
2.10
2.20-1.70
24
2
2.07
2.20-1.70
46
1
2.02
2.20-1.70
29
1
1.86
1.58
28
1
1.66
1.50-1.20
40
2
1.63
1.50-1.20
47
3
1.59
1.50-1.20
37
4
1.58
1.10
53
5
1.55
1.10
50
6
1.53
1.10
39
1
1.47
1.07
52
1
1.23
1.07
51
1
1.08
1.07
45
1
1.02
1.03
48, 49
2,3
0.99
1.03
Table 12 - 13C NMR chemical shifts from Gaussview 5.0 compared to literature values
Atom #
Degeneracy
Calculated chemical shifts / ppm
Literature chemical shifts / ppm
Difference / ppm
8
1
219.81
211.49
8.32
2
1
148.03
148.72
-0.69
3
1
121.01
120.90
0.11
12
1
87.31
74.61
12.7
10
1
57.32
60.53
-3.21
9
1
55.83
51.30
4.53
6
1
55.73
50.94
4.79
7
1
49.90
45.53
4.37
17
1
46.23
43.28
2.95
13
1
44.02
40.82
3.2
18
1
43.77
38.73
5.04
4
1
37.84
36.78
1.06
15
1
36.66
35.47
1.19
1
1
28.71
30.84
-2.13
11
1
28.40
30.00
-1.6
22
1
26.33
25.56
0.77
20
1
24.80
25.35
-0.55
5
1
23.84
22.21
1.63
23
1
22.94
21.39
1.55
14
1
20.95
19.83
1.12
The tables above show each of the chemical shifts that was predicted by GaussView 5.0 for the 1H NMR and 13C NMR and both compared to literature [4]. Firstly it can be seen from the 1H NMR that several of the chemical shifts were quoted as multiplets in the literature, so it is difficult to have a clear image of the degree of deviation. But nonetheless, it can be seen that the predicted values are in relatively close range to the literature values, deviating by no more than ±0.5 ppm. It is worth noting that the chemical shifts generally range between 3.0-0.0 ppm rationalisation the aliphatic nature of the compound, where the only exception is proton atom 26 which is adjacent to the alkene functional group. Overall the predicted values are in good agreement with literature values and demonstrate its power in accurately modelling organic compounds and their spectroscopic properties.
Looking at Table 12, we can see the predicted values of the 13C NMR chemical shifts compared to the literature (same source as from above) chemical shifts. The difference was found between the two sets of values and it can be seen that there is relatively little deviation between the two, staying within the deviation of ±5.0 ppm with the exception of one chemical shift. This is seen in carbon atom 12, where the deviation is 12.7 ppm which is quite significant. This deviation has been interpreted to be associated with the two adjacent sulfur atoms to the carbon atom, affecting the realistic chemical shift. The modelling calculation on GaussView is not advanced enough to account for this accurately as it may possibly be related to additional spin orbit coupling, and thus overestimates the value.
The individual spectra are presented below in Table 14 to give a spectral illustration of each of NMR calculations. These can be compared to the experimental spectra and can possibly allow any coupling to be determined.
Table 14 - 1H NMR and 13C NMR of Compound 18 (Taxol derivative)
1H NMR
1C NMR
Literature molecule: 1,6-diazabicyclo[4.3.1]-decane
The molecule that will be investigated is the tertiary amine 1,6-diazabicyclo[4.3.1]-decane, where the two separate conformers that it can adopt are presented in the figure below. They are referred to as conformer A and B. It can be synthesised from 1,5-diazacyclononane and formaldehyde in toluene to give a significant yield of 73%. Two separate literature sources were consulted in the discussion of this molecule[5][6].
This molecule is an interesting example of how the observed conformer is opposite to the most energetically stable conformer as calculated from computational techniques. In this section we will predict the 13C NMR spectra and compare it to literature. Initially, the energy of the conformers are minimised using MM2 forcefield on ChemBio3D 12.0 and compared to each other in the table below.
Figure 8 - Illustration of tertiary amine compound conformers A and B
Table 14 - Jmol depiction of A and B
Conformer A
Conformer B
Conformer A
Conformer B
Table 14 - Optimisation results of 1,6-diazabicyclo[4.3.1]-decane A using MM2 and MOPAC
We see from the table above that according to the MM2 minimised energies, A is the most stable conformer with an energy lower by ~1.01 kcal/mol. However the MOPAC minimisation shows a contrasting result, indicating conformer B to be the more stable. As we have seen before, MOPAC is usually considered as the more accurate method as it accounts for orbital interactions by considering the quantum mechanics of the molecule. Thus it is worth obtaining the molecular orbitals of each individual conformer to see whether it can shed light on which conformer is seen as the major product in the reaction.
Table 15 - MO orbitals generated from Gaussview of HOMO and LUMO orbitals of conformer A and B
Conformer B
MO 42 (LUMO +2)
MO 41 (LUMO +1)
MO 40 (LUMO)
MO 39 (HOMO)
MO 38 (HOMO -1)
0.11147
0.09807
0.07548
-0.18395
-0.22389
Conformer A
MO 42 (LUMO +2)
MO 41 (LUMO +1)
MO 40 (LUMO)
MO 39 (HOMO)
MO 38 (HOMO -1)
0.11226
0.10464
0.07871
-0.18174
-0.21515
From the table above aligning the corresponding MO orbitals for easier interpretation, we can see that their orbitals, as expected, are indeed very different, as a result of the symmetry features associated with their conformation. Because of the difference in symmetry, determining the extent of orbital interaction is not trivial and thus not much can be concluded from interpreting the MOs. However it is worth noting that particularly with respect to the LUMO and HOMO MOs, more orbital interaction is observed for conformer B where orbitals interact across the entire molecule. In the LUMO illustrating, it can be seen that a large electron cloud (green) stretches across the molecule. Further to this, the HOMO in conformer 2 has an added orbital interaction in the two carbon atoms in the front of the picture. These interactions may possibly have a stabilising effect on the structure which the MM2 force field does not account for and thus support the argument that conformer B is in fact the major product.
13C NMR Analysis
At this point we realise the importance of obtaining the spectroscopic data of a product as it can help to ultimately determine the identity of the compound, and thus the molecule that is observed in the product. So again as previously the 1C NMR was obtained from GaussView 5.0, but using the method DFT=mpw1pw91 B3LYP with the basis set 6-31G (d, p) and chloroform as the solvent. The calculations were submitted to HPC (A:DOI:10042/24262 and B:DOI:10042/24263 ) and the log files were extracted to open the NMR spectrum. The results are presented in the table below and compared against each other.
Table 15 - 1C NMR chemical shifts from Gaussview 5.0 of conformers A and B
C Atom #
Degeneracy
A chemical shifts / ppm
B chemical shifts / ppm
Difference
6
1
65.38
64.96
0.42
3, 2
1, 2
50.57
54.23
-3.66
8, 7
1, 2
46.77
52.00
-5.23
9, 10
1,2
23.95
31.87
-7.92
1
1
11.95
23.23
-11.28
*The B conformer rapidly equilibrates with its mirror image twist conformer to average the two sides, so the reported shifts for C2,3, C8,7, and C9,10 are averaged values. This can be seen from the NMR spectrum later in the section.
Quite surprisingly the 1C NMR chemical shifts are significantly different for the two conformers, where deviations as large as ~11.3 ppm are observed. As expected the carbon 6 shows similar shifts in the two conformers as its environments are virtually the same. Carbon atoms 9 and and 10 also deviate very significantly from each other, but this is expected as the shape of the conformer w.r.t. these atoms are different (symmetry is lost). What is very surprising is that carbon atom 1 has the largest difference, despite the neighbouring atoms being in the same arrangement for both conformers. This is most likely associated with the loss of symmetry in the B conformer.
To determine the conformer that is observed in the product of the reaction, the chemical shift of the two conformers are compared to experimental values. The table below presents the chemical shifts of each as denoted below.
Table 16 - 1C NMR comparison of conformers A and B with literature
C Atom #
Degeneracy
A chemical shift / ppm
B chemical shift / ppm
Literature / ppm
6
1
65.38
64.96
66.3
3, 2
1, 2
50.57
54.23
54.8
8, 7
1, 2
46.77
52.00
52.8
9, 10
1,2
23.95
31.87
30.6
1
1
11.95
23.23
21.4
Comparison with experimental data clearly identifies the dominant conformer to be B, as it matches the experimental values much more than conformer A. In fact there is very little deviation between B and the experimental results, indicating very strongly that the major conformer is indeed B. In fact according to literature, the results most likely suggest that B is the only conformer that is observed in the reaction.
Table 17 - 13C NMR spectrum of conformers A and B
Conformer A
Conformer B
1H NMR Analysis
The proton NMR was also obtained to give a more comprehensive overview of the literature molecule, where the NMR information was extracted from the same file as for the 13C NMR and the same reference was applied. From the table below, it can be seen that the proton NMR values are significantly different, and because of loss of symmetry in conformer B, more chemical shifts are observed (most likely due to coupling because of differences in chemical environments). However it is difficult to make comparisons with literature, as the values are quoted with in large ranges, and both fall quite well into these ranges. Nonetheless it could be argued that the predicted NMR values for conformer B correlate better to the literature values (less deviation) and serve as an argument for the B conformer being the major product.
Table 18 - 1C NMR chemical shifts from Gaussview 5.0 of conformers A and B
H Atom #
Degeneracy
A chemical shifts / ppm
B chemical shifts / ppm
Literature
17
1
4.008
3.856
4.06-3.83
18
1
3.448
3.820
4.06-3.83
20
1, 2
3.273
3.422
3.29-2.69
22
2
3.273
3.245
3.29-2.69
14
1
3.076
3.064
3.29-2.69
16
2
3.076
2.950
3.29-2.69
19
3
2.861
2.837
3.29-2.69
21
5
2.861
2.835
3.29-2.69
13
1
2.831
2.699
3.29-2.69
15
2
2.831
2.689
3.29-2.69
12
1
2.674
2.674
1.99-1.69
24
1
1.946
2.026
1.99-1.69
26
2
1.946
1.920
1.99-1.69
23
1
1.752
1.848
1.99-1.69
25
2
1.752
1.785
1.99-1.69
11
1
0.436
0.773
1.10-0.93
Table 19 - 1H NMR spectrum of conformers A and B
Conformer A
Conformer B
Frequency Analysis
Although the NMR results clearly indicate conformer B being the major product, it is still worth analysing the predicted IR spectrum, to see whether it can allude to more properties of the conformers. Thus the frequency analysis was performed on both of conformers with B3LYP and basis set 6-31G (d, p). The calculations were submitted to HPC (A:DOI:10042/24264 B:DOI:10042/24267 ) and presented in the tables below.
Table 20 - Key frequencies of tertiary amine conformer A
Number
Vibrational Mode
Frequency/ cm-1
Intensity / 10-40esu2cm2
Brief description
71
3096
78
C-H stretching of the two "terminal" groups, where the protons on each group stretch out-of-phase, but the corresponding proton on each group stretch in a concerted fashion.
64
3046
62
C-H stretching of two the "terminal" groups, where the protons on each group stretch in-phase, but the corresponding proton on each group stretch out-of-phase.
22
966
9
Wagging motion of "terminal" groups out-of-phase with respect to one another.
25
1045
51
Wagging motion of bridging CH2 group.
Table 17 - Key frequencies of tertiary amine conformer B
Number
Vibrational Mode
Frequency/ cm-1
Intensity / 10-40esu2cm2
Brief description
66
3061
66
C-H stretching of the two "terminal" groups, where the protons on each group stretch out-of-phase, but the corresponding proton on each group stretch in a concerted fashion.
69
3075
69
C-H stretching of two the "terminal" groups, where the protons on each group stretch in-phase, but the corresponding proton on each group stretch out-of-phase.
23
970
23
Wagging motion of "terminal" groups out-of-phase with respect to one another.
25
1039
57
Wagging motion of bridging CH2 group.
The same corresponding stretches were found for each conformer and presented in the table above. It can be seen that some of the key stretches associated with the conformational change, that there are significant differences, thus allowing differentiation of the two conformers. Experimental values are not available for the IR spectrum, so comparison to literature is not possible, but by comparing the two, there are some key differences worth noting.
It can be seen that the C-H stretch of the two "terminal" groups differ between the two conformers by ~30 cm-1 which is significant compared to the reference vibration (i.e. expected to be same for two conformers) which only differed by 6 cm-1. This is attributed to the position of the protons on the respective carbon atoms, where the protons in conformer B are not "aligned" due to their relative positioning, and thus results in a different stretching frequency. Despite this, the third vibrational mode listed in the tables, the wagging motion of the terminal CH2 groups, the two conformers have more or less the same frequency, differing by 4 cm-1. But looking at how the other bonds are affected by this stretch, it can be seen that the same bonds are affected by the same extent, rationalising the similarity in the frequency. However a detailed reason requires further investigation.
The IR spetra presented below show that the spectra of the two conformers are very similar and it is only small differences that distinguish between the two. However it is worth recognising the general peaks of the tertiary amine that may be useful in future investigations. The free energies of the conformers were also presented along with the spectra, and presented both in the units Hartree and kcal/mol. It can be seen that the energies are very similar differing by only a matter of a few kcal/mol (~6 kcal/mol). Nonetheless the results still support the findings in this paper, and cement the determination of conformer B being the major product and the observered conformer.
Table 21 - Infrared spectrum of conformers A and B and their respective free energies (ΔG)
Conformer A
Conformer B
423.768311 Hartree
-423.775704 Hartree
-265913.99 kcal/mol
-265919.25 kcal/mol
Conclusion
This paper demonstrated the power of computational modelling tools such as ChemBio3D and GaussView, and has shown to accurately predict energies, structures and spectroscopic data of a wide range of compounds with conformational ambiguity. Some of the weaknesses of modelling methods such as MM2 force field were also highlighted and more advanced methods such as MOPAC/PM6 were presented as alternatives that take more factors into account, such as orbital interaction. In the mini-project, a reaction resulting in two possible conformers of 1,6-diazabicyclo[4.3.1]-decane was presented to demonstrate the possibility of using computational tools to predict spectroscopic data such as NMR spectroscopy. Overall the values obtained from computational predictions and calculations were in good agreement with experimental values cited in literature, and thus justifies the use of computational modelling as a powerful tool in chemistry.
References
↑ K. Alder, G. Stein , Angew. Chem. , 1937 , 50 , 510
↑D. Skala and J. Hanika, Petroleum and Coal, 2003, 45(3-4), 105
↑A. B. McEwan and P. von Rague Schleyer, J. Am. Soc., 1986, 108, 3951
↑L. Paquette, N. A. Pegg, D. Toops, G. D. Maynard, R. D. Rogers, J. Am. Chem. Soc., 1990, 112, 277-283
↑R. W. Alder, E. Heilbronner, E. Honegger, A. B. McEwen, R. E. Moss, E. Olefirowicz, P. A. Petillo, R. B. Sessions, G. R. Weisman, J. M. White and Z. Yang, J. Am. Chem. Soc., 1993, 115, 6580
↑ A. B. Sebag, D. A. Forsyth and M. A. Plante, J. Org. Chem., 2001, 66, 7967
