3PHY

Back to Main - https://www.ch.ic.ac.uk/wiki/index.php/AndyForester

First thoughts

The 3PHY NMR structure was used as an initial structure and processed like 1W7S. NMR experiments like this lead to a huge amount of information about through-space distances between atoms and also bonded distances, but the sheer size of the data set means it can be fit lots of ways (fitting is done iteratively starting from a guess stucture). The result is a set of possible structures, in this case there are 26 different possible stuctures - im not sure which one the wiki starts with!

The file is slightly different to Mikes example... overall -4 charge due to different protonation state (pH7) but charges are the same except for around the chromophore. I included the chromophore + connecting residue when fitting charges since the chromophore is bound to a residue in the primary sequence via a sulphur in is sidechain - this will change the charge distribution of the residue so should be refitted with the chromophore attached. The file you provided uses standard amber charges for this residue and some estimated charges from ESP calculations (+ some fudging to get an integer charge!) for the chromophore.

The Plan

Find and Characterise a Ground State minimum
Vertical excitation to the S1 PES - frequency analysis of imaginary modes
IRC along S1 PES to S1 minimum.

Ground State calculations

A ground state minimum was found using ONIOM B3LYP 6-31G* : Amber with Electronic Embedding and characterised with a frequency calculation.

Ground state ONIOM optimisation and frequency analysis INPUT - Media:3PHY_ONIOM_Optimisation_Frequency_EE.gjf

S1 Frequency and IRC calculations

The aim of this part of the project is to follow an excited state reaction path using an IRC having simulated a vertical excitation from S0 → S1 (geometry is the same). To do this we cannot use a ground state theory like DFT cf. the density is the ground state density. Methods that can do excited states are TD-DFT, CIS and CASSCF (there are many - we are interested in these). So we will be using the ONIOM method with an excited state-capable theory on the model high level region and looking at the effects of Electronic embedding on the IRC of the S1 surface.

The IRC keyword requests that a reaction path be followed by integrating the intrinsic reaction coordinate, ONIOM calculations use the first-order Euler integration for the predictor step along with the HPC corrector step. It is a practical choice for calculations on large molecules.

It is assumed that the IRC will be started from a transition state structure (1st order saddle point) and thus have only 1 imaginary frequency. The IRC can then follow this transition vector either in the forward or backward direction. Vertical excitation from a S0 minimum will be highly unlikely to lead to a 1st order saddle point on the S1 PES and will almost certainly result in being on the "side of a hill" on the S1 PES - this leads to complications with the following IRC calculation.

There will be multiple imaginary frequencies at this non-stationary point on the S1 surface, the IRC will follow the first one it finds which is not necessarily the vector we want to follow! We need to explicitly define a specific vector for the IRC to follow which represents the reaction path we are investigating. The best way to do this is to run a frequency calculation at this point and visualize the imaginary frequencies to identify the imaginary frequency which represents the reaction, the frequency will be accompanied by a vector which can be read into the IRC calculation.

The standard frequency route will be meaningless on the S1 surface since the calculation assumes a harmonic function (untrue on the side of a hill). This is overcome by using freq=project in the route section. This invokes computation of the projected frequencies for vibrations perpendicular to the IRC path.

To read the vector into the IRC calculation use IRC=ReadVector in the route section with the vector listed after the molecule specification in rows containing 8 values per row with 10 characters (inc spaces) making up each value.

IRC calculations also require initial force constants to proceed. You must provide these to the calculation in some way, options defined in the route section:

IRC=RCFC reads the force constants from the checkpoint file from the preceding frequency calculation (used to identify the required frequency corresponding to the reaction coordinate)
IRC=CalcFC recalculates the force constants at the beginning of the IRC calculation

S1 Frequency Calculations

Using TD-DFT with the route section stating:

Electronic Embedding - #p freq=projected oniom(b3lyp/6-31g(d) TD=(Root=1):amber=softfirst)=embed geom=connectivity
Mechanical Embedding - #p freq=projected oniom(b3lyp/6-31g(d) TD=(Root=1):amber=softfirst) geom=connectivity

These calculations require minimum 14.5Gbytes memory to run - they are also extremely time consuming, taking xxx to run.