Building a TS Library
AARON relies on a library of pre-computed TS structures for a given reaction. These pre-computed TS structures are typically for a representative model catalyst/substrate. AARON then computes analogous structures for new catalysts/ligands or substrates as specified in the AARON input file, allowing the user to make predictions of selectivities, etc. for other ligands and substrates.
AARON comes with TS libraries for several reactions.
The first step in applying AARON to a new reaction is constructing a TS library. An ideal TS library should include all reasonable low-lying transition states for the reaction. This requires the user to manually explore different configurations and relative orientations of the substrate and catalyst. Note that even if a given configuration is relatively high in energy for a given catalyst does not mean it will be necessarily unimportant for other catalysts being screened.
The simplest way to see how to construct a TS library is to browse the files in $QCHASM/Aaron/TS_geoms. You can also browse them here.
For systems in which you do not have to manually specify any conformations (e.g. ring flips, etc), we can use the TS library here as an example.
1. Save XYZ files for all TS structures and name them tsXXX.xyz:
ts1.xyz ts2.xyz ts3.xyz ...
These can be numbered/named however you want, but the names must start with ts.
2. Re-order atoms in these XYZ files using the AaronTools convention: Substrate atoms, then metal center (if it exists), then ligand/catalyst atoms. It is recommended that you keep atom ordering as consistent as possible across different TS structures. It is mandatory that any atom that will be substituted has the same number in all TS structures. For instance, if you plan on replacing a Me group on the substrate, the carbon atom of that Me group (and preferably the hydrogens as well) must be the same atom number in all structures.
3. For each XYZ file, information about the structure must be provided on the comment line, including:
F: bonds that must be constrained during step 2 (this sometimes requires some trial and error) C: Reaction center is it exists. For transition metals, this is not required since the metal will be detected automatically. For non-metal catalyzed reactions that still have a metal-like atom to which the catalyst binds, specifying the central atom will generally make it easier for AARON to map new catalysts/ligands. Note that this atom should come directly before the ligand section. L: Atom range for the ligand atoms; for transition-metal catalyzed reactions this is not required. K: List of 'key atoms' that directly bind the substrate '''using the relative atom numbers for the ligand.'''
Note that these atom numbers for all but 'K' are 'absolute' numbers within the XYZ file. For the key atoms, the atom numbers are relative numbers for the ligand.
For example, the comment line for the structure below is as follows:
F:19-25;2-15 C:25 L:26-47 K:1,2
In particular, we are requesting that AARON freeze the 2-15 and 19-25 bonds and denoting that atom 25 (Si) is the central atom, the ligand is atoms 26-47 (absolute atom numbering), and the two key atoms are 1 and 2 (Ligand atom numbering).
For transition metal catalyzed reactions the only required information on the comment line of each XYZ file are the frozen bonds:
F:4-33;23-4;
All other information (central atom, key atoms, ligand atoms, etc.) will be automatically detected by AARON.
Once all XYZ files are prepared, create a directory in
$HOME/Aaron_libs/TS_geoms
corresponding to the reaction (i.e. allylation). This is the name you will specify as 'reaction_type' in the AARON input file.
Then create a subdirectory corresponding to the 'template'. For the above example, this would be NN-bipyridine-N-oxide. This is the name you will specify as 'template' in the AARON input file, and allows you to build different TS libraries for the same reaction based on different catalyst/ligand frameworks, for example.
Then put your XYZ files organized in whatever subdirectories you want under this template directory.
For enantioselective reactions, AARON will expect subdirecties called 'R' and 'S.' For example, for the example here we have the following directory structure:
R/ ts1.xyz S/ ts1.xyz ts2.xyz ts3.xyz
If you do not have a selective reaction, you can put the XYZ files directly under the template directory (as done here).
If you have multiple conformations of each TS (i.e. corresponding to ring flips, etc.), these should be saved as Cf1.xyz, Cf2.xyz, etc. under subdirectories specifying the ts number.
See an example here.
However you organize the structures in your library, AARON will replicate this directory structure for each ligand/substrate combination.