Design of Asymmetric Catalysts with Reaction Network Enumeration Profiler

Tutorial Created with Software Release: 2025-2
Topics: Catalysis & Reactivity, Energy Capture & Storage
Methodology: Molecular Quantum Mechanics
Products Used: AutoTS, Jaguar, MS Maestro, MS Reactivity

Tutorial files

371 MB
This tutorial is written for use with a 3-button mouse with a scroll wheel.
Words found in the Glossary of Terms are shown like this: Workspacethe 3D display area in the center of the main window, where molecular structures are displayed

 

Tip: You can hover over a glossary term to display its definition. You can click on an image to expand it in the page.
Abstract:

 

In this tutorial, we will learn to leverage the Reaction Network Enumeration Profiler workflow towards the design of asymmetric catalysts.

 

Tutorial Content

  1. Introduction to RxnEnumProfiler

  1. Creating Projects and Importing Structures

  1. Preparing the Reference Reaction Network for % ee Prediction

  1. Running and Analyzing the Reference Reaction Network for % ee Prediction with RxnProfiler

  1. Running RxnEnumProfiler for % ee prediction and analyzing results

  1. Preparing the Reference Reaction Network for TOF prediction

  1. Running and Analyzing the Reference Reaction Network for TOF prediction

  1. Running and Analyzing the Reaction Network Enumeration Profiler for TOF Prediction

  1. Conclusion and References

  1. Glossary of Terms

1. Introduction to RxnEnumProfiler

Virtual high-throughput screening of reaction networks is important for various applications:

  • Homogeneous (molecular) catalyst design
  • Maximizing or minimizing activity and selectivity in non-catalytic reactions
  • Minimize or maximize side or competitive reactions

In previous tutorials, we have learned to perform geometry optimization for equilibrium geometries and transition states (minima and 1st-order saddle points on the potential energy surface, respectively). For review, visit: Introduction to Geometry Optimizations, Functionals and Basis Sets, Locating Transition States: Part 1 and Part 2. We have also introduced the Reaction Network Profiler (RxnProfiler) panel in Materials Science (MS) Maestro, which accepts any reaction network as input and converts it into a (lowest-energy G or Boltzmann-averaged GBA) free energy profile (FEP) based on a specified quantum-mechanical level of theory.

The RxnProfiler supports any type of reaction network ranging from general case A), and including special cases B) and C) as input such as:

This tutorial introduces the Reaction Network Enumeration Profiler (RxnEnumProfiler) panel. RxnEnumProfiler expands on the capabilities of RxnProfiler by enabling the user to enumerate the reference reaction network from the user-defined list and calculate corresponding Boltzmann-averaged (or lowest-energy) FEPs based on a specified quantum-mechanical level of theory. Key kinetic and thermodynamic properties (e.g. selectivity, rate constant, turnover frequency) are then extracted from each FEP.

Application of RxnEnumProfiler for homogeneous catalyst design: starting from a reference reaction network, user-defined library for enumeration (with N components) and specified quantum mechanics (QM) level, a set of N reaction networks is created and corresponding (lowest G or Boltzmann-averaged GBA) N free energy profiles (FEPs)  are computed on-the-fly. Key metrics for catalyst design (selectivity, TOF) are then extracted from the FEPs

This tutorial demonstrates the application of RxnEnumProfiler for asymmetric molecular catalyst design. Two key metrics in catalyst design are enantiomeric excess (% ee), which reflects enantioselectivity, and turnover frequency (TOF), which measures the efficiency of the catalytic cycle.

We will explore an experimentally studied reaction for enantioselective imine (im) hydrogenation, as described in two publications (see the works by Hamza et al. and Chen et al. cited in the References). The reference catalyst (ref. cat) delivers the R-enantiomer of the product with 79% ee.

Reference reaction of interest studied in this tutorial

The primary goal is to identify and optimize a catalyst that delivers the desired product (am) with enhanced enantiomeric excess (% ee). This is accomplished by evaluating all catalyst candidates within a user-defined virtual library. The secondary objective is to select a catalyst from the resulting pool that exhibits improved or comparable turnover frequency (TOF) relative to a reference catalyst. The overall strategy can be viewed as comprising two sequential computational funnels—an initial one focused on selectivity, and a subsequent one on activity.

In principle, RxnEnumProfiler allows evaluating both catalyst metrics for any user-defined pool of candidates at once. However, in practice, in order to save computational resources and not run unnecessary calculations, the evaluation is based on first selectivity and after TOF computational funnels. This order is chosen since it is cheaper to evaluate % ee than TOF.

Two reference reaction networks and the virtual library for enumeration are shown:

Reference network for enatioselectivity prediction

Reference network for TOF prediction

Library for single R-group enumeration (sphere indicates dummy atom)

The Cartesian coordinates for two stereoselectivity-determining transition states and catalytic cycle species that form the reference network for enantiomeric excess and TOF prediction, respectively are available in the literature (see the work by Hamza et al. cited in the References). We compiled the corresponding Cartesian Coordinates into MS Maestro format for convenience. The virtual library of fragments to catalyst modification for enumeration is generally up to the user’s choice and imagination; here we will select those fragments that lead to synthetically accessible catalysts (see the work by Hamza et al. cited in the References). For more information on generating an enumerated virtual library see the Elemental Enumeration tutorial. Conformational sampling is included and performed with Monte Carlo using the OPLS4 force field (Macromodel). For the QM level of theory, we will choose ωB97X-D functional, C-PCM implicit solvent model (toluene) and 6-311G++**//6-31G* discretization level:

Revisit the Introduction to Geometry Optimizations, Functionals and Basis Sets tutorial for a discussion of functionals and basis sets. 

The overall project will proceed as follows, with four separate job submissions: 1) % ee calculations for the reference catalyst 2) % ee calculations for the defined library of catalysts 3) TOF calculations for the reference catalyst 4) TOF calculations for the pool of catalyst identified in 2).

For additional background on the panels used in this tutorial, visit the corresponding help documentation: Create Reaction Network Profiler Input Structures, Reaction Network Profiler, R-Group Creator, Reaction Network Enumeration Profiler. For additional reference material, see the Conclusion and References section.

2. Creating Projects and Importing Structures

At the start of the session, change the file path to your chosen Working Directorythe location where files are saved in MS Maestro to make file navigation easier. Each session in MS Maestro begins with a default Scratch Projecta temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project, which is not saved. A MS Maestro project stores all your data and has a .prj extension. A project may contain numerous entries corresponding to imported structures, as well as the output of modeling-related tasks. Once a project is saved, the project is automatically saved each time a change is made.

Structures can be built in MS Maestro or can be imported using File > Import Structures (or drag-and-dropped), and are added to the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data. The Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion is located to the left of the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed. The Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data can be accessed by Ctrl+T (Cmd+T) or Window > Project Table if you would like to see an expanded view of your project data.

  1. Double-click the Materials Science icon

Figure 2-1. Change Working Directory option.

  1. Go to File > Change Working Directory
  2. Find your directory, and click Choose
  3. Pre-generated files are included for running jobs or examining output. Download the zip file here: schrodinger.com/sites/default/files/s3/release/current/Tutorials/zip/rxnenumprofiler.zip
  4. After downloading the zip file, unzip the contents in your Working Directory for ease of access throughout the tutorial

Figure 2-2. Save Project panel.

  1. Go to File > Save Project As
  2. Change the File name to rxnenumprofiler_tutorial, click Save
    • The project is now named rxnenumprofiler_tutorial.prj

Figure 2-3. Import the starting structures.

  1. Go to File > Import Structures
  2. Choose Structures_for_tofprediction.maegz and Structures.maegz from the provided tutorial files
  3. Click Open

As was discussed briefly in the introduction, the Cartesian coordinates for two stereoselectivity-determining transition states are available in the literature (see the work by Hamza et al. cited in the References). Once imported in MS Maestro, these structures (Structures.maegz) can be used directly for further prediction of enantiomeric excess without any operations in MS Maestro. In general, make sure that chemical bond pattern is properly drawn in MS Maestro. For atoms involved in the transition state mode the bonding pattern should be drawn as corresponding to reagent or product. In the present case it corresponds to the reagent. Likewise, the Cartesian coordinates for the catalytic cycle can also be extracted from the same literature (Structures_for_tofprediction.maegz).

Figure 2-4. Visualizing the structures.

The imported structures are added to the entry list. Please proceed to visualize the structures in the workspace.

3. Preparing the Reference Reaction Network for % ee Prediction

Let’s first prepare the reference reaction network for a % ee prediction.

Figure 3-1. Selecting and including TS-2-R1 and opening the Create Reaction Network Profiler Input Structures panel.

  1. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries and includethe entry is represented in the Workspace, the circle in the In column is blue the TS-2-R1 entry in the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and workspacethe 3D display area in the center of the main window, where molecular structures are displayed
  2. Go to Tasks > Materials > Quantum Mechanics > Reaction Network > Create Reaction Network Profiler Input

The Create Reaction Network Profiler Input Structures panel sets up structures for the reaction network.

The general purpose of the panel is to specify various reactants, transition states, intermediates, and products, and relationship between them. The panel also offers the possibility to include conformational and configurational space associated with each stationary point. In this example, the reaction network will only be based on two stereodetermining transition states.

In addition, we use this panel to specify restraints which will be used to pre-optimize transition states in one of few workflow stages.

Visit the help documentation for a complete overview of the panel.

Figure 3-2. Loading TS-2-R1 into the panel.

  1. Click Import Selected Entries
    • The panel will update to show (1 structures imported)
  2. Change the Conformer group name to TS_R
  3. Change the Sibling group name to TS_R
    • Note that you can input your own Sibling group name by typing directly into the input box rather than using the dropdown
  4. For Indices of atoms to restrain, input 1,61,62
  5. For Atom index pairs of distances to restrain, input (1,61);(61,62)
  6. For Atom index triples of angles to restrain, input (1,61,62)
    • These restraints are needed for either geometry preoptimization and/or TS search algorithm, which is based on the user-defined eigenvector

Atoms 1, 61 and 62 refer to the B, H, and C, respectively, associated bond formation/breaking within the transition state. When hovering over an atom in the workspace, the atom number will appear at the bottom of the MS Maestro application window.

  1. Check Transition state
    • If the structure is a minimum (e.g., reactant, intermediate or product), then leave this unchecked. For transition states, as in this case, be sure to check the checkbox.

Figure 3-3. Loading TS-2-S1 into the panel.

Now we will proceed to load TS-2-S1 into the panel.

  1. Click Add Group
    • The panel will update to show a second group, designated by (2)
  2. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries TS-2-S1
  3. Click Import Selected Entries
    • The panel will update to show (1 structures imported)

Figure 3-4. Settings for TS-2-S1 in the panel.

  1. Change the Conformer group name to TS_S
  2. Change the Sibling group name to TS_S
  3. For Add parent group name, be sure that TS_R is shown in the dropdown, and click Add
  4. For Indices of atoms to restrain, input 1,61,62
  5. For Atom index pairs of distances to restrain, input (1,61);(61,62)
  6. For Atom index triples of angles to restrain, input (1,61,62)
  7. Check Transition state
  8. Click View Energy Level Diagram

Note: In this example the atom numbering between the two transition state structures was the same, but this may not always be the case. It is best to check the atom numbering for the constraints for each structure being used. Additionally, you can also click on "View Energy level Diagram" to visualize the imaginary energy profile for the created reaction network.

Figure 3-5. Energy Level Diagram Prediction.

The Energy level Diagram visualizes the difference in the Gibbs free energy (at 298 K and 1 atm) for the two enantiomers of the TS..

  1. Close the Energy Level Diagram panel

 

Note that identifying Indices of atoms to keep and Indices of atoms to restrain can be facilitated with using the Atom Selection panel that can be opened by clicking Define next to the relevant text box. The user can then select atoms in the workspace or open the Atom Selection Dialog Box to make the appropriate selections.

Similarly, indices of superimposable atoms and selecting the atoms involved in adding a distance, angle, and/or dihedral restraint can be directly selected from the workspace using the Pick functionality.

Figure 3-6. Running the Create Reaction Network Input Structures panel.

  1. Check Ensure mass is conserved
    • Doing so is optional, but this is a good check to be sure that mass is conserved in any specified reactions. Here, where we are only looking at the two transition states, the setting is not particularly relevant
  2. Click Run
    • The panel does not require any host. The job will finish immediately.
  3. Close the Create Reaction Network Profiler Input Structures panel

Figure 3-7. The generated structures.

When the job finishes, a new entry group is added to the entry list, as shown in the Figure. The structures themselves are the same as the initial structures, but now they are now with the necessary information for running the RxnProfiler workflow in the next section.

The structures can be tiled in the workspace for ease of visualization, and the properties defined with the Create Reaction Network Profiler Input Structures panel are now available in the Project Table.

4. Running and Analyzing the Reference Reaction Network for % ee prediction with RxnProfiler

In this section, we will execute and analyze a RxnProfiler calculation on the reference reaction network prepared in the previous section. This is the first step in the process of designing asymmetric catalysts.

Figure 4-1. Selecting the reaction network input group and opening the Reaction Network Profiler panel.

  1. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries the output group, rxnwf_input (2)
  2. Go to Tasks > Materials > Quantum Mechanics > Reaction Network > Reaction Network Profiler Calculations
  3. Check Ensure mass is conserved on import
  4. Click Import
    • (2 structures imported) appears to indicate the input has been loaded into the panel

Figure 4-2. Setting parameters in the Reaction Network Profiler panel.

 

  1. Check Deduplicate structures using this RMSD 0.25 Å
    • This flag ensures proper thermostatistics by filtering conformers based on RMSD values
  2. Check All conformers with relative energies less than or equal to 20.00 kcal/mol
    • The subsequent option updates to Minimum numbers of conformers 5
    • The amount of conformers can be controlled with these settings
    • This energy corresponds to the force field energies
  3. Check Return conformational search job files
    • Optionally return the job files associated with the conformational search for later inspection

Ensure the remaining checkbox options match the Figure. Note that the panel will maintain settings from previous uses, but can be reset back to defaults at any time using the button

  1. Click Jaguar Options

Figure 4-3. Jaguar options.

We will perform the quantum mechanical calculations at the wB97X-D/6-31G* level of theory. Of course, other choices of functional and basis set will be appropriate depending on the study.

  1. Set the Theory to wB97X-D
  2. Set the Basis Set to 6-31G*

Note: An option for a machine learning force field (MLFF) is available as an alternative. Additional information regarding MLFF can be found in the help documentation, on our website, or the Reaction Network Profiler panel documentation.

Increasing the number of SCF iterations and geometry optimization steps makes the quantum mechanical calculations more robust.

  1. For SCF, change Maximum iterations to 150
  2. Check Switch to analytic integrals near convergence
  3. Set the Convergence criteria to Accurate
  4. Uncheck Use symmetry

We will perform the QM calculations with an implicit PCM toluene model.

  1. Set Additional keywords: isolv=7 solvent=toluene
    • The isolv keyword is specified to perform a solvation calculation. The PCM model is used here.
    • The solvent keyword specifies the solvent used for the solvation calculation.
    • More information about these keywords is available in the documentation
  2. Ensure the remaining settings match the Figure and then click OK

Figure 4-4. Additional parameters in the panel. 

Back in the panel:

  1. Set Initial Hessian for transition state searches to GFN2-xTB
  2. Set Tolerate negative (imaginary) frequencies greater than to -30.00 wavenumbers
    • If a TS optimization converges into second -or higher order saddle point, the workflow automatically takes out the corresponding structure. This setting allows to add an exception for a second order saddle point, if the second imaginary frequency is -30 cm-1 or above, we consider the structure as a TS.
  3. Check Return Jaguar job files
  4. Click Open next to Specify additional jobs using a Jaguar multistage workflow

Figure 4-5. The Input tab in the QM Multistage Workflow panel. 

We can specify the parameters for the single point quantum mechanical Jaguar jobs. Here, we will use the same functional from the original optimization, and a larger basis set:

  1. Set the Basis Set to 6-311G++**

Figure 4-6. The Solvation tab in the QM Multistage Workflow panel. 

  1. Go to the Solvation tab
  2. Set the Solvent model as PCM
  3. Change the Solvent to toluene

 

Figure 4-7. Adding the convergence keyword. 

  1. Scroll down within the stage, and in the Additional keywords input box, add nops=1
    • This turns the pseudospectral method off and calculates electronic energies analytically
  2. Click Append Stage
    • A 3rd stage is added to the multistage workflow

Figure 4-8 Adding an analysis stage. 

  1. Change the Stage type to Analysis
  2. For Property name, input G_(au)
  3. Next to Property terms, click Add

Note: G_(au) is the G_(kcal/mol)_298.15K_1.00E+00atm_stage_4 value in the results.pdf file. This energy represents the Boltzmann-average free energy.

Figure 4-9. Inputting the Property terms and saving the multistage workflow. 

  1. Use the Add and Property dropdowns to write the following formula:
    • 1.0*Total Free Energy (au) *K *atm(1) + -1.0*Solution Phase Energy (au)(1) + 1.0*Solution Phase Energy (au)(2)
    • The Figure shows the settings for the first term in the formula
    • Add determines whether the results from stage 1 or 2 are used, this is denoted by either (1) or (2) in the lines that are generated
    • The Property option menu will populate with options once a selection is made from the Add option menu
    • The prefactor determines the multiplicative factor for that term, for terms 1 and 3 it should be 1, for term 2 it should be -1

This final property will be the free energy determined by summing up the solution-phase electronic energy calculated with the extended basis set and the ZPVE and thermal corrections calculated with the smaller basis set.

  1. Click Save
    • The QM Multistage panel closes

Figure 4-10. Naming and running the job.

Return to the Reaction Network Profiler panel to name and run the job.

  1. Change the Job name to reference
  2. Adjust the job settings () as needed
    • This job requires a CPU host and can take several days.
    • If you would like to run the job, proceed to click Run. If you would prefer to proceed with imported files, please proceed to the next steps.
  3. Close the Reaction Network Profiler panel

Figure 4-11. Importing provided files.

We will assume that you have not run the calculation, and instruct for importing:

  1. Go to File > Import Structures
  2. Locate the ref_workflow-out_rxnwf.mae file in the Section_04 > reference directory and click Open

The entry list is updated with a new entry group, containing 18 entries in 2 subgroups.

Figure 4-12. Visualizing the imaginary frequency for one of the conformers from one of the sibling groups.

From the output, we have a conformational ensemble for our two transition states: there are 13 different conformers for TS-S and 5 different conformers for TS-R. If you performed the calculations yourself, the number of conformers and the conformer structure might differ, so expect slight variance in the results.

 

The vibrational frequency can be analyzed for each TS conformer. We can confirm that each output structure has one negative (imaginary) frequency.

We can also confirm that the frequency corresponds to the reactive mode by clicking the animate button shown in the image on the left.

You should see the bond form/break between the B,C,H atoms (and given indices) input earlier in the create stage of the tutorial.

Figure 4-13. Viewing the energy difference between the transition states in the output PDF.

We are mainly interested in the relative Boltzmann-averaged free energies of the two transition states, which can be used to compute the enantiomeric excess of the reaction.

Conveniently, the output files include a .pdf file, which contains the various useful energy diagrams. On your computer (outside of MS Maestro), open ref_workflow_e_diagrams.pdf in any PDF viewer and locate G_(kcal/mol)_298.15K_1.00E+00atm_stage_4 (which is G_(au) specified in the “Analysis” part of the QM multistage setup). The energy difference on the plot (1.7 kcal/mol) represents the Boltzmann-average free energy difference between two transition states with five (TS-R) and thirteen (TS-S) TS conformers, respectively.  

Figure 4-14. A plot of the relative energies of the conformers per transition state.

Another approach is to use the Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data to view the ensemble free energy profile for our two transition states: there are 13 different conformers for TS-S and 5 different conformers for TS-R. Be aware that the printed energies are in Hartrees (conversion to kcal/mol is possible in the Project Table).

Using the data from the Project Table, it can be useful to create a plot of the relative energies of the conformers, as shown in the Figure.

We can also use the Reaction Profile Viewer or the Scatter Plot tools to generate useful diagrams as needed.

The energy difference on the plot in the PDF (1.7 kcal/mol) represents the Boltzmann-averaged free energy difference between two transition states with 13 and 5 conformers, respectively. Import the value to a separate .xls file where you can calculate the enantiomeric excess (% ee) using the formula:

 

For ΔG○ = 1.70 kcal/mol, T = 298 K, and R = 1.9872 x 10⁻³ kcal/mol·K,  we get 89% ee.

5. Running RxnEnumProfiler for % ee prediction and analyzing results

Using the reference reaction network generated in Section 3 and the provided virtual library file, the next step is to proceed with the Reaction Network Enumeration Profiler panel. The provided virtual library file contains the components necessary to systematically modify and enumerate our catalyst. If you need further information about generating and working with R-group libraries, see the R-group enumeration with the Materials Science Suite tutorial.

Figure 5-1. Opening the Reaction Network Enumeration Profiler panel.

  1. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries and includethe entry is represented in the Workspace, the circle in the In column is blue the rxnwf_input group in the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and workspacethe 3D display area in the center of the main window, where molecular structures are displayed
  2. Go to Tasks > Materials > Quantum Mechanics > Reaction Network > Reaction Network Enumeration Profiler Calculations
  3. Click Import Selected Entries

Figure 5-2. Setting up the panel for TS_R.

  1. Click Import Selected Entries
  2. Click Add Row
  3. Check Pick pairs
  4. Select the boron and carbon atoms shown in the Figure
  5. Repeat for TS_S, click Add Row
  6. Check Pick pairs
  7. Select the boron and carbon bonds shown in the Figure

Figure 5-3. Setting up the panel for TS_S.

  1. Repeat for TS_S, click Add Row
  2. Check Pick pairs
  3. Select the same boron and carbon illustrated in Figure
  4. Check Ensure mass is conserved
  5. Click Add Row
  6. Check Pick pairs
  7. Select the equivalent boron and carbon atoms as in the last figure
  8. Click Import File and navigate to Section_05 > R_group_ee.maegz
  9. Check Only enumerate reaction network files
    • It is highly recommended to first perform the enumeration to visually inspect the structures before submitting the calculation (i.e. run the calculation as it currently is before moving on to the next step)
  10. Go to the Reaction Network tab

Figure 5-4. Running the job.

  1. Set all of the parameters to match those that were used in Section 4 for the reference reaction, including the additional jobs using QM Multistage Workflow
  2. Change the Job name to rxn_net_enum_prof_ee

Recall that this first run will be only to enumerate the structures.

Note: An option for a machine learning force field (MLFF) is available as an alternative in the Jaguar Options. See the Reaction Network Enumeration Profiler panel documentation to learn more.

  1. Adjust the job settings () as needed
    • This job can be performed with minimal resources (e.g. 1 CPU on a localhost) in just a couple of minutes
  2. Click Run
  3. Close the Reaction Network Enumeration Profiler panel

Figure 5-5. The ten novel catalysis.

Once the job has completed or after importing you should see a new group in the entry list entitled rxn_net_enum_prof_ee with 20 entries. Feel free to visualize the structures in the workspace.

 

By checking the only enumerate reaction network files option, we only produce several RxnProfiler input files. Each of these files can then be separately used for regular RxnProfiler calculations. The purpose of using this option is to check if everything is in place and/or run each enumerated network individually, if needed. Alternatively, and this is the goal of the panel, we can run all these RxnProfiler workflows in parallel by unchecking “Only enumerate reaction network files” then clicking “Run”.

Upon running the RxnProfiler workflow on all ten catalysis, various output files are returned in the same directory. The information to calculate the % ee for each catalyst can be extracted from each pdf-file exactly the same way shown earlier in this tutorial using the Reference Network Profiler panel. The ten output folders are not provided in the provided files due to their gigantic file size. However, we provide % ee based on analysis of these files shown below:

 

We are only interested in those catalysts that provide better % ee with respect to the reference catalyst (computed to be 89 % ee). These four catalysts are further studied in Section 8.

6. Preparing the Reference Reaction Network for TOF prediction

This section prepares the reference reaction network for the TOF predictions. The steps follow similar steps to Section 3. If preferred, feel free to import in the results file from Section_06 > cat0_TOF_rxnwf.mae and proceed with the next section.

Figure 6-1. Selecting and including cat structure and opening the Create Reaction Network Profiler Input Structures panel.

  1. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries and includethe entry is represented in the Workspace, the circle in the In column is blue the cat entry in the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and workspacethe 3D display area in the center of the main window, where molecular structures are displayed
  2. Go to Tasks > Materials > Quantum Mechanics > Reaction Network > Create Reaction Network Profiler Input
  3. Reset the panel

Figure 6-2. Loading structures into the panel for steps 1 and 2.

  1. Click Import Selected Entries
  2. Change the Conformer group name to cat
  3. Change the Sibling group name to state1
  4. Click Add Group
  5. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries im
  6. Click Import Selected Entries
  7. Change the Conformer group name to im
  8. Change the Sibling group name to state1
  9. Click Add Group

Figure 6-3. Loading structures into the Create Reaction Network Profiler Input Structures panel for step 3.

  1. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries H2
  2. Click Import Selected Entries
  3. Change the Conformer group name to H2
  4. Change the Sibling group name to state1
  5. Click Add Group

Figure 6-4. Loading structures into the panel for steps 4 and 5.

  1. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries TSHHim/2(a)
  2. Click Import Selected Entries
  3. Change the Conformer group name to TS1
  4. Change the Sibling group name to TS1
  5. For Add parent group name, be sure that state1 is shown in the dropdown, and click Add
  6. For Indices of atoms to restrain, input 12,56,63,77
  7. For Atom index pairs of distances to restrain, input (12,56);(56,77);(77,63)
  8. For Atom index triples of angles to restrain, input (12,56,77);(56,77,63)
  9. For Atom index quadruples of dihedrals to restrain, input (12,56,77,63)
  10. Check Transition state
  11. Click Add Group
  12. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries c1-HH-a
  13. Click Import Selected Entries
  14. Change the Conformer group name to state2
  15. Change the Sibling group name to state2
  16. For Add parent group name, be sure that TS1 is shown in the dropdown, and click Add
  17. Click Add Group

Figure 6-5. Loading structures into the panel for steps 6 and 7.

  1. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries TS-2-R1
  2. Click Import Selected Entries
  3. Change the Conformer group name to TS2
  4. Change the Sibling group name to TS2
  5. For Add parent group name, be sure that TS1 is shown in the dropdown, and click Add
  6. Change the Parent group name to state2
  7. For Indices of atoms to restrain, input 1,61,62
  8. For Atom index pairs of distances to restrain, input (1,61);(61,62)
  9. For Atom index triples of angles to restrain, input (1,61,62)
  10. Check Transition state
  11. Click Add Group
  12. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries am
  13. Click Import Selected Entries
  14. Change the Conformer group name to state3
  15. Change the Sibling group name to TS1
  16. For Add parent group name, be sure that TS1 is shown in the dropdown, and click Add
  17. Change the Parent group name to TS2
  18. Click Add Group

Figure 6-6. Loading structures into the panel for step 8 and running the panel.

  1. Return to the entry list, and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries cat
    • Duplicate the cat entry and rename the new entry to capr
  2. Click Import Selected Entries
  3. Change the Conformer group name to capr
  4. Change the Sibling group name to state3
  5. For Add parent group name, be sure that TS1 is shown in the dropdown, and click Add
  6. Change the Parent group name to TS2
  7. Check Ensure mass is conserved
  8. Change the Reaction network name to cat0_TOF
  9. Click Run
    • The panel does not require any host. The job will finish immediately.
  10. Close the Create Reaction Network Profiler Input Structures panel

Feel free to view the Energy level Diagram once again before closing the panel.

Figure 6-7. Viewing the Create Reaction Network Profiler Input Structures panel.

When the job finishes, a new entry group is added to the entry list, as shown in the Figure. The structures themselves are the same as the initial structures, but now they are now with the necessary information for running the reaction network in the next section.

Alternatively, the file can be imported from Section_06 > cat0_TOF_rxnwf.mae

7. Running and Analyzing the Reference Reaction Network for TOF prediction

In this section, we will execute and analyze a reaction network calculation on the reference reaction for a TOF prediction using the reaction network input prepared in the previous section. These steps follow a similar fashion as Section 4.

Figure 7-1. Selecting the reaction network input group and opening the Reaction Network Profiler panel.

  1. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries the output group, cat0_TOF (8)
  2. Go to Tasks > Materials > Quantum Mechanics > Reaction Network > Reaction Network Profiler Calculations
  3. Check Ensure mass is conserved on import
  4. Click Import
    • (8 structures imported) appears to indicate the input has been loaded into the panel

Figure 7-2. Setting parameters in the Reaction Network Profiler panel.

  1. Check Deduplicate structures using this RMSD 0.25 Å
  2. Check Return conformational search job files
    • Setting the maximum number of conformers to 1 will only return the lowest energy conformer
  3. Click Jaguar Options

Figure 7-3. Jaguar options.

We will perform the quantum mechanical calculations at the wB97X-D/6-31G* level of theory. Of course, other choices of functional and basis set will be appropriate depending on the study.

  1. Set the Theory to wB97X-D
  2. Set the Basis Set to 6-31G*
  3. For SCF, change Maximum iterations to 150
  4. Check Switch to analytic integrals near convergence
  5. Set the Convergence criteria to Accurate
  6. Uncheck Use symmetry
  7. Set Additional keywords: isolv=7 solvent=toluene
  8. Ensure the remaining settings match the Figure and then click OK

Figure 7-4. Additional parameters in the panel. 

  1. Set Initial Hessian for transition state searches to GFN2-xTB
  2. Set Tolerate negative (imaginary) frequencies greater than to -30.00 wavenumbers
  3. Check Return Jaguar job files
  4. Click Open next to Specify additional jobs using a Jaguar multistage workflow

Figure 7-5. The Input tab in the QM Multistage Workflow panel. 

  1. Set the Basis Set to 6-311G++**

Figure 7-6. The Solvation tab in the QM Multistage Workflow panel. 

  1. Go to the Solvation tab
  2. Set the Solvent model as PCM
  3. Change the Solvent to toluene

 

Figure 7-7. Adding the convergence keyword. 

  1. Scroll down within the stage, and in the Additional keywords input box, add nops=1
    • This is a useful general Jaguar keyword which employs special methods to facilitate SCF and/or geometry convergence
  2. Click Append Stage
    • A 3rd stage is added to the multistage workflow

Figure 7-8. Adding an analysis stage. 

  1. Change the Stage type to Analysis
  2. For Property name, input G_(au)
  3. Next to Property terms, click Add

Note: G_(au) is the G_(kcal/mol)_298.15K_1.00E+00atm_stage_4 value in the results.pdf file. This energy represents the Boltzmann-average free energy.

Figure 7-9. Inputting the Property terms and saving the multistage workflow. 

  1. Use the Add and Property dropdowns to write the following formula:
    • 1.0*Total Free Energy (au) *K *atm(1) + -1.0*Solution Phase Energy (au)(1) + 1.0*Solution Phase Energy (au)(2)
    • The Figure shows the settings for the first term in the formula
    • Add determines whether the results from stage 1 or 2 are used, this is denoted by either (1) or (2) in the lines that are generated
    • The Property option menu will populate with options once a selection is made from the Add option menu
    • The prefactor determines the multiplicative factor for that term, for terms 1 and 3 it should be 1, for term 2 it should be -1

This final property will be the free energy determined by summing the energy calculated with the extended basis set and the thermal correction from the smaller basis set.

  1. Click Save
    • The QM Multistage panel closes

Figure 7-10. Naming and running the job.

Return to the Reaction Network Profiler panel to name and run the job.

  1. Change the Job name to TOF_ref
  2. Adjust the job settings () as needed
    • This job requires a CPU host. The job can be completed in about 4 days when parallelized for 10 simultaneous subjobs on 8 CPU each

If you would like to run the job, proceed to click Run. If you would prefer to proceed with imported files, please proceed to the next steps.

  1. Close the Reaction Network Profiler panel

Figure 7-11. Importing provided files.

We will assume that you have not run the calculation, and instruct for importing:

  1. Go to File > Import Structures
  2. Locate the cat0_TOF-out_rxnwf.mae file in the Section_07 > TOF_ref directory and click Open

The entry list is updated with a new entry group, containing several entry subgroups.

Feel free to analyze vibrational frequencies. We can confirm that each transition structure has a negative (imaginary) frequency and no negative frequencies are present for the equilibrium geometries.

Figure 7-12. Lowest Energy Conformer Free Energy Diagram for the catalytic cycle corresponding to im hydrogenation into am as printed in the output PDF.

We are mainly interested in the relative energies of the two transition states, which can be used to compute the enantiomeric excess of the reaction.

Most conveniently, the output files include a .pdf file, which contains the various useful energy diagrams. On your computer (outside of MS Maestro), open cat0_TOF_e_diagrams.pdf in any PDF viewer and locate G_(kcal/mol)_298.15K_1.00E+00atm_stage_4.

 

 

The RxnEnumProfiler panel automatically calculates TOF from, see the work by Kozuch et al. cited in the References for more details on the TOF calculation:

 

The approximate (energetic span model approximation) formulas are:

These metrics are available in the cat0_TOF_tof.csv file. Opening this csv file with any spreadsheet program that you have available:

In this example, we requested only the lowest energy conformer, that’s why the values are identical. The TOF is 11.63 s-1 based on the exact “Sum Over States” formula, and 23.41 s-1 based on the approximate "Energetic Span Approximation” formula. The file also contains information on determining states within energetic span model approximation, specifically those are state 1 and TS1.

8. Running and Analyzing the Reaction Network Enumeration Profiler for TOF Prediction

With our TOF reference reaction input and the various catalyst structures we are ready to utilize the Reaction Network Enumeration Profiler panel for TOF prediction.

Figure 8-1. Opening the panel.

  1. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries and includethe entry is represented in the Workspace, the circle in the In column is blue the cat0_TOF group in the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and workspacethe 3D display area in the center of the main window, where molecular structures are displayed
  2. Go to Tasks > Materials > Quantum Mechanics > Reaction Network > Reaction Network Enumeration Profiler Calculations
  3. Click Import Selected Entries

Figure 8-2. Setting up the panel.

  1. Click Import Selected Entries

Similar to Section 5, every catalyst needs to have the B-C atoms selected as the R-Group Site Indices.

  1. In the cat conformer group, click Add Row
  2. Check Pick pairs
  3. Select the boron and carbon bonds shown in the Figure
  4. Repeat for TS1, state2, TS2, capr

Figure 8-3. Adding the R-group file.

  1. Set the Reaction network name to Cat0_TOF
  2. Check Ensure mass is conserved
  3. Click Add Row
  4. Click Import File and navigate to Section_08 > R_tof_prediction.maegz
    • This field contains the four fragments of interest based on the results in Section 5
  5. Go to the Reaction Network tab

Figure 8-4. Running the calculation.

  1. Set all of the parameters to match those that were used in Section 7 for the reference reaction, including the additional jobs using QM Multistage Workflow
  2. Change the Job name to rxn_net_enum_prof_TOF
  3. Adjust the job settings () as needed
    • This job runs on CPUs
  4. Click Run
  5. Close the Reaction Network Enumeration Profiler panel

We will discuss the results below.

Once again since the results folders are so large they are not provided and instead will be discussed. From analyzing the TOF results we see that we have three suitable novel candidates with two of the options out ranking the other.

In computational chemistry, particularly with DFT, numbers are less important than a trend, see the work by Bursch et al. and cited in the References. In the context of catalyst design, the objective of computational studies—using DFT as the final level of refinement—is to identify catalyst candidates that offer improved enantioselectivity, while retaining or exceeding the TOF of the reference catalyst. In the first stage, we narrowed the pool of 10 catalysts down to 4 based on their relative enantioselectivity, prioritizing this metric over TOF. In the second stage, this subset was further reduced to 3 catalysts based on relative TOF. We note here, although two catalysts appear to be an order of magnitude slower based on TOF relevant to the reference catalyst (~2-3 s-1 vs 12 s-1), we still consider them possible candidates due to the small free energy difference—just ~1 kcal/mol—in the apparent activation energies of their full catalytic cycles, which falls within the typical uncertainty range of DFT calculations.

The summary of our design can be visualized as:

The experimental validation for one of the two catalysts, specifically, the variant bearing two tert-butyl groups, is described in the work by Hamza et al. and cited in the References. Hamza et al. experimentally obtained a 92% ee on one of the modeled novel catalysts from the original value of 78% ee.

Note that a final tab is available in the Reaction Network Enumeration Profiler panel which can be used to generate descriptors for machine learning model creation. Feel free to read about these capabilities in the help documentation.

9. Conclusion and References

In this tutorial, we learned to leverage the Reaction Network Enumeration Profiler towards the design of asymmetric catalysts.

For further learning:

For introductory content, focused on navigating the Schrödinger Materials Science interface, an Introduction to Materials Science Maestro tutorial is available. Please visit the materials science training website for access to 70+ tutorials. For scientific inquiries or technical troubleshooting, submit a ticket to our Technical Support Scientists at help@schrodinger.com. 

For self-paced, asynchronous, online courses in Materials Science modeling, including access to Schrödinger software, please visit the Schrödinger Online Learning portal on our website.

For some related practice, proceed to explore other relevant tutorials:

For further reading:

10. Glossary of Terms

Entry List - a simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion

Included - the entry is represented in the Workspace, the circle in the In column is blue

Project Table - displays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data

Scratch Project - a temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project

Selected - (1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries

Working Directory - the location where files are saved

Workspace - the 3D display area in the center of the main window, where molecular structures are displayed