Tautomerization of Vinyl Alcohol
In this post we wish to determine the reaction energy, barrier, and transition state structure for the tautomerization reaction of vinyl alcohol converting to acetaldehyde using DFT. This reaction is shown below in Fig.1. Tautomers are constitutional isomers; having the same number of each atom, but with a different connectivity. Most often, as in this case, tautomerization reactions are intramolecular proton exchange reactions.
Fig. 1 Tautomerization reaction of vinyl alcohol (left) to acetaldehyde (right).
It is known that acetaldehyde is more stable than vinyl alcohol at ambient conditions [1]. Thus, we should expect to find that the reaction energy is negative (forward direction of Fig. 1), which we can verify after we have performed our calculations.
Setting Up the Calculation
We will use DMol3 to geometry optimize the reactant and product state as well as locate the transition state. In addition, we will calculate the vibrational modes and frequencies of each the reactant, product, and transition state for reasons we will return to later.
In order to begin a transition state search in DMol3, the user must supply initial geometries for the reactant and product. With DMol3, the transition state procedure includes a routine for optimizing the reactant and product states before finding the TS, so these initial structures don’t need to be very accurate. Below are figures showing the initial reactant and product structures used for this example.
Fig. 2 Initial geometries of reactant (left) and product (right) to DMol3.
As TS searches follow atoms through spatial translations, it is important to tell DMol3 which atoms from the reactant and product are “equivalent”. This can be done using the reaction preview tool. Once a reaction preview has been made between the initial and final states, we can enter the calculation parameters we wish to use.
Calculation Details
Calculations were performed using the DMol3 Module as available in Materials Studio. The TS search method used was the Complete LST/QST Method with conjugate gradient optimization of reactants and products using a RMS convergence of 0.00185 Ha/Å and a maxinum of 5 QST steps. Exchange and correlation were treated within the generalized gradient approximation using the PBE functional with the DNP basis set and explicit core electrons (all electron calculation) [2-5]. The SCF tolerance was set to 1.0E-5 Ha with linear smearing using a value of 0.005 Ha.
Results and Discussion
First we present the optimized structures and compare the geometries to experimental values available from NIST [6,7].
Fig. 3 Optimized reactant and product from DMol3.
| Calculated | Experimental | |
|---|---|---|
| C1-C2 | 1.336 | 1.326 |
| C2-O3 | 1.371 | 1.372 |
| C2-H7 | 1.091 | 1.097 |
| C1-H5 | 1.088 | 1.079 |
| C1-H6 | 1.093 | 1.086 |
| O3-H4 | 0.974 | 0.960 |
| C1-C2-O3 | 126.7 | 126.2 |
| C1-C2-H7 | 122.9 | 129.1 |
| C2-C1-H6 | 122.0 | 121.7 |
| C2-C1-H5 | 119.9 | 119.5 |
| C2-O3-H4 | 108.6 | 108.3 |
| Calculated | Experimental | |
|---|---|---|
| C1-C2 | 1.502 | 1.501 |
| C2-O3 | 1.218 | 1.216 |
| C1-H4 | 1.097 | 1.086 |
| C2-H7 | 1.121 | 1.114 |
| C1-C2-O3 | 124.7 | 123.9 |
| H4-C1-H6 | 110.3 | 108.3 |
| C1-C2-H7 | 114.9 | 117.5 |
Below is a plot showing the evolution of the TS search. This plot includes the LST, QST, and GC (conjugate gradient) paths along the reaction coordinate. Note the transition state is also labeled on this plot–indicating a transition state has been found with the calculation parameters.
Fig. 4 DMol3 TS Search Energy Trajectories
DMol3 also outputs the geometry of the TS. One should confirm that the TS found (by computational means) matches intuition about how the reaction proceeds. Below the TS is shown with atom labels.
Fig. 5 TS structure found using DMol3.
The transition state structure is what we might expect: somewhere in between vinyl alcohol and acetaldehyde. Here we can see that at the TS, the C-C double bond hasn’t quite broken while the O-H bond hasn’t quite formed. This leads us to believe we may have found the correct transition state, however, we must verify by examining the vibrational modes/frequencies of the TS. Transition states occur at saddle points in the potential energy surface (PES) and, therefore, should have one negative (imaginary) frequency corresponding the the vibrational mode along the reaction path.
The vibrational analysis tool can be used on the TS to verify these requirements are met. Indeed, we find the TS has one negative vibrational frequency: -2000. To verify this mode is along the reaction path we may animate it and confirm visually. This animation is reproduced below as a series of frames.
Fig. 6 Imaginary mode for the TS found using DMol3.
Thus we can be confident that we have found a real TS corresponding to the tautomerization of vinyl alcohol. Of special importance are the reaction energy and reaction barrier which DMol3 calculates during the search procedure. The energy for this reaction was -11.2 kcal/mol while the barrier was 51.9 kcal/mol. Note that our prediction of a negative reaction energy was correct. We also notice this reaction has a relatively high barrier (inaccessible at room temperature).
So, we have found a TS structure for the reaction of vinyl alcohol to acetaldehyde and have calculated the reaction energy and barrier. Next we need to compare the structure and energies with literature values.
Juan Andres et. al. did a comprehensive ab inito study on this very reaction and their results are readily available online [8]. In their results they report that DFT methods found the reaction energy to be in the range 8.5-14.3 kcal/mol while the barrier was found to fall within the range 87.9-93.5 kcal/mol. Thus there is great agreement with our reaction energy but the barriers are in severe disagreement. This, however, is most likely a symptom of basis set selection and the differences in DFT methods for finding TS between our attempt and those of Andres.
To verify that our structures are the same we reproduce their table of geometry data and include our own. The labels in the tables below are consisted with Fig. 4.
Table 1 TS Geometry| This Work | Literature | |
|---|---|---|
| C1-C2 | 1.407 | 1.428 |
| C1-O3 | 2.228 | 2.299 |
| C1-H4 | 1.500 | 1.541 |
| O3-H4 | 1.311 | 1.305 |
| C1-C2-O3 | 111.093 | 110.500 |
| H4-C1-C2 | 66.737 | 67.200 |
| C2-O3-H4 | 75.827 | 78.600 |
| C1-C2-H7 | 130.421 | 131.100 |
| H5-C1-C2-O3 | 155.067 | 147.4 |
| H6-C1-C2-O3 | 63.976 | 77.400 |
From the table above we see there is great agreement in the TS geometry, with the greatest errors showing in the dihedral angles. This leads us to conclude that we have indeed found a TS and that the departures from literature values (especially the reaction barrier) are purely from differences in the level of theory/basis set used in the calculation.
References
[1] R.D. Johnson III. “CCCBDB NIST Standard Reference Database”. Retrieved 2014-08-30.
[2] B. Delley, J. Chem. Phys. 92, 508 (1990).
[3] B. Delley, J. Chem. Phys. 113, 7756 (2000).
[4] J. P. Perdew, K. Burke, and M. Ernzerhof. Generalized gradient approximation made simple. Phys. Rev. Lett., 77:3865, 1996.
[5] J. P. Perdew, K. Burke, and M. Ernzerhof. Erratum: Generalized gradient approximation made simple. Phys. Rev. Lett., 78:1396, 1997.
[6] Kuchitsu (ed.), Landolt-Bornstein: Group II: Atomic and Molecular Physics Volume 15: Structure Data of Free Polyatomic Molecules. Springer-Verlag, Berlin, 1987.
[7] H. Hollenstien, Hs. H. Gunthard, “Solid State and gas infrared spectra and normal coordinate analysis of 5 isotopic species of acetaldehyde” Spec. Acta 27A, 2027 (1971)
[8] Andrés, J. , Domingo, L. R., Picher, M. T. and Safont, V. S. (1998), Comparative theoretical study of transition structures, barrier heights, and reaction energies for the intramolecular tautomerization in acetaldehyde/vinyl alcohol and acetaldimine/vinylamine systems. Int. J. Quantum Chem., 66: 9-24.
