Modeling Surfaces
Tutorial Created with Software Release: 2025-2
Topics: Catalysis & Reactivity , Energy Capture & Storage , Metals, Alloys & Ceramics , Thin Film Processing
Methodology: Periodic Quantum Mechanics
Products Used: MS Maestro , Quantum Espresso
|
80 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
Abstract:
In this tutorial, we will learn to model adsorption of small molecules on a surface via a multi-step workflow. From a bulk structure, we will build slabs, optimize those slabs towards convergence of surface energy, build adsorbate structures and calculate adsorption energy.
Tutorial Content
1. Introduction to Surface Modeling
Computing atomic-scale models of surfaces and interfaces with density functional theory (DFT) is a powerful way to address research questions in heterogeneous catalysis, thin film deposition (CVD, ALD, etc.), sensors, nanoparticles, photovoltaics and batteries.
In this tutorial, we will learn to model adsorption of small molecules on a surface using various tools available in Materials Science Maestro. The overall workflow entails building various slabs from an optimized bulk structure, optimizing the slabs to achieve convergence of a property of interest (e.g. surface energy), building adsorbate structures, and finally, computing adsorption energies. In doing so, we will use several panels in Materials Science Maestro, each linked here with the corresponding help documentation: Surface Energy, Enumerate Adsorbates, Adsorption Energy Calculations, Thermochemistry Viewer.
Here is a summary of the overall workflow:
workflow for modeling a surface and adsorbate to find adsorption energy
This procedure for calculation of adsorption energies can be used to inform novel materials design as well as help to gain mechanistic understanding of surface chemistry at the atomic level.
If you are unfamiliar with handling periodic structures or performing solid-state calculations with Materials Science Maestro, two foundational tutorials are available and recommended: Building and Manipulating Crystal Structures and Electronic Structure Calculations of Bulk Crystals Using Quantum ESPRESSO. In the current tutorial, we will begin with an optimized TiO2 bulk structure, as described in the latter foundational tutorial.
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.
- Double-click the Materials Science icon
- (No icon? See Starting Maestro)
- Go to File > Change Working Directory
- Find your directory, and click Choose
- 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/surfaces_slabs.zip
- After downloading the zip file, unzip the contents in your Working Directory for ease of access throughout the tutorial
- Go to File > Save Project As
- Change the File name to surface_tutorial, click Save
- The project is now named
surface_tutorial.prj
- The project is now named
We will produce a TiO2 slab from a previously generated bulk rutile structure. Information on the generation and optimization of bulk crystals is in the Electronic Structure Calculations of Bulk Crystals Using Quantum ESPRESSO tutorial. Let’s import the bulk structure now:
- Go to File > Import Structures
- Select
bulk_TiO2.mae - Click Open
- A new entry is added to the entry list titled bulk_rutile_TiO2
- It is both 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 and includedthe entry is represented in the Workspace, the circle in the In column is blue by default
3. Building a Slab Model of a Surface
The Surface Energy panel will be explored in two sections of this tutorial. This section will show how the Surface Energy panel constructs unoptimized slab models from a bulk structure by cleaving the surface and providing a vacuum gap in which to add an adsorbate. We wish to build a slab that best models the semi-infinite crystal while minimizing the system size for computational efficiency. Towards that end, we will build several slabs in this step, and in Section 4, we will continue using the panel to compute the surface energy and check its convergence, indicating which slab is a good minimal model of the bare surface.
In addition to the procedure described here, slabs can also be generated with the Build Slabs and Interfaces panel or can be imported from the Import Slabs panel. Alternatively, the Nanoparticle Builder can be used to generate a cluster model of a surface.
- Ensure that bulk_rutile_TiO2 is 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 and includedthe entry is represented in the Workspace, the circle in the In column is blue in the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion
- Go to Tasks > Materials > Quantum Mechanics > Workflows > Surface Energy
- The Surface Energy Calculations panel opens
- Click Load Bulk from Workspace
- bulk_rutile_TiO2 appears next to the button in the panel
- Input (1 1 0) for Single plane (hkl)
- Input 6.000 Å for Vacuum
- This generates a reasonable vacuum buffer for a small adsorbate; buffers 10 Å or even 15 Å thick may be needed for larger adsorbates
- Click Add
- In the table, an entry appears for this (1 1 0) surface orientation with slab settings that can be adjusted
- Click Show Plane
- Two red planes in the workspace illustrate where the bulk structure will be cleaved in order to generate the slab for the yellow-highlighted entry
- Returning to the (1 1 0) entry in the table, increase the Thickness to 3.00 layers
- An extended bulk structure is created in the workspace to show the new cleaving planes
Integer numbers of layers are generally best, because they preserve the stoichiometry of the bulk. A slab whose stoichiometry is different from that of the bulk will likely have an electronic or redox state that is also different from the bulk, which is usually undesirable. In any case, an integer multiple of the bulk is mandatory for computing the surface energy in this panel.
It can be seen that this first proposal intersects atoms in the structure, making it a poor choice of cleaving planes. Stable surfaces are generally those that cut the minimum number of bonds. We can check other (1 1 0)-oriented cleavage planes by shifting the planes in the C direction.
Interactively adjust the Bottom, n parameter in the (1 1 0) entry in the table and see what happens to the planes in the workspace:
- Finally, set Bottom, n to 0.50
- At this value the planes cut symmetrically between oxygen atoms
- Click the Preview Slab button to generate the unoptimized slab structure in a new entry
- Two new entries are generated, an expanded bulk cell and the slab of interest
- Includethe entry is represented in the Workspace, the circle in the In column is blue the new bulk_rutile_TiO2_110_surface_bottom=0.50_thickness=3.00_vacuum=6.00A entry and check that it looks correct
- This symmetric choice of cleavage planes has produced a slab with two identical, non-polar slab faces, with overall stoichiometry (TiO2)6
If there is more than one element in the slab (as in this case of a binary oxide), cutting planes should be chosen so that the slab faces are non-polar, either consisting of a single layer (Tasker Type 1) or symmetric stack of sublayers (Tasker Type 2) that are uncharged overall - this is to avoid a dipole moment in the C direction. In the current example, each of the three layers in the slab are symmetric stacks of O2-..[Ti2O2]4+..O2-.
If possible, the slab should have identical top and bottom faces, again to minimize any dipole moment along the C direction. Clearly this is no longer possible when an adsorbate or other functional group is added.
If desired, the previewed slab can be used as an input structure for your own surface calculations with Quantum ESPRESSO (QE). In this case, make sure that you have optimized the cell parameters and atomic structure of the bulk before loading into this panel. The cell parameters of a slab should be kept fixed at optimum bulk values throughout the surface calculations.
It is of course possible to now run the Surface Energy panel to optimize the bulk and a single slab with QE (see Section 4). However for this tutorial, we would like to construct several additional slabs of varying thicknesses and will use them to test for convergence in the next section:
- Back in the Surface Energy Calculations panel, Add another (1 1 0) entry with the same Vacuum
- Set Bottom, n to 0.50
- Increase the Thickness to 4.00 layers
- Repeat steps 12-14 to add four further (1 1 0) entries with Thicknesses of 5.00, 6.00, 7.00, and 8.00 layers
- Ensure your panel matches what is shown in the Figure
In addition to evaluating several slab thicknesses, we may also want to evaluate slabs generated from other cutting planes
- Input (1 0 1) for Single plane (hkl)
- Maintain 6.000 Å for Vacuum
- Click Add
- In the table, an entry appears for this (1 0 1) surface orientation, with slab settings that can be adjusted
- Increase the Thickness to 3.00 layers
- Click Show Plane and confirm that setting Bottom, n to 0.50 gives good cutting planes
- Click the Preview Slab button to generate the unoptimized slab structure in a new entry
- Includethe entry is represented in the Workspace, the circle in the In column is blue the new bulk_rutile_TiO2_101_surface_bottom=0.50_thickness=3.00_vacuum=6.00A entry and check that it looks correct
- This symmetric choice of cleavage planes has produced a slab with two identical, non-polar slab faces
- Repeat these steps to generate additional entries for 4-8 layer thick (1 0 1) slabs
The table in the Surface Energy panel should now match that which is shown in the figure, which contains six (1 1 0) slabs, 3-8 layers thick, and six (1 0 1) slabs, also 3-8 layers thick.
4. Optimizing the Slab and Computing Surface Energy
Having used the Surface Energy panel to define and preview slabs in Section 3, we now continue using the panel to relax the atomic structure of the (1 1 0) and (1 0 1) slabs and compute their surface energies at the level of density functional theory with Quantum ESPRESSO.
We have chosen the following DFT settings for the slab calculations, balancing computational time against accuracy. For instance, we use the D3 correction to account for possible adsorbate-adsorbate interactions.
- PBE functional with D3 dispersion correction
- 40 Ry wavefunction cutoff and 200 Ry charge density cutoff for the plane wave basis
- Ultrasoft PBE GBRV pseudopotentials
- Grid plane distance of 0.1/Å for k-points (including Γ)
We specify these settings via the usual QE panels.
To ensure consistency, the atomic positions and cell of the bulk structure are also optimized with these settings and the slabs are built from the optimized bulk lattice parameters.
- Back in the Surface Energy Calculations panel, click Pseudopotentials
- In Path click Browse
- Navigate to where you saved the tutorial file and choose the
pseudopotentials > all_pbe_UPF_v1.5directory- The Pseudopotentials panel is automatically filled with the paths to the directory and to the Ti and O specific pseudopotential files
- Click OK to close the Pseudopotentials panel
- A message appears that the energy cutoffs were not defined. Click OK to close the message as we will define the cutoffs in the next step
Note: Pseudopotentials are not included with the Quantum ESPRESSO installation and must be downloaded separately. There are several libraries available (see this link for more information). The pseudopotentials used in this tutorial can be downloaded from GBRV pseudopotential site.
- Click Advanced Options
- In the Theory tab, change Spin treatment to Spin-polarized
- Change the Density functional type to GGA
- This should automatically change the Density functional to PBE
- Change Dispersion correction to DFT-D3
- Set Grid plane distance to 0.1/Å
- Check Include Γ-point
- When computing slabs, this grid plane distance will be applied in the x and y directions, but only one k-point will be used in the z direction
- Click Update K-point mesh
- Confirm that the rest of the Theory settings match those shown in the Figure
- Go to the SCF tab
- Click Update from pseudopotentials
- The energy cutoff for wavefunctions is set to 40.0 Ry and the cutoff for charge density to 200.0 Ry
- Click OK on the Info dialogue
- Click Save
- Change the Job name to surface_energy_all_slabs
- Adjust the job settings (
) as needed
- This job requires a CPU host and must be run on a Linux machine. The job takes approximately 10 hours on a 12 CPU host
- If you would like to run the job yourself, click Run. Otherwise, go to File > Import Structures, navigate to the provided tutorial files and Open
Section_04 > surface_energy_all_slabs > surface_energy_all_slabs-out.maegz. - Close the Surface Energy Calculations panel
- When the job completes or after importing, a new entry group is added to the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion entitled surface_energy_all_slabs-out containing the 12 DFT-optimized slabs. Feel free to visualize and stylize the outputs as you wish.
To choose which slab to use for the adsorbate models, we must examine convergence. We wish to choose a minimal model to balance computational expense with having an approximation that best mimics the semi-infinite crystal surface.
One approach could be to assess the convergence of a geometric parameter, e.g. Ti-O bond length for surface 6-coordinate Ti to bridging O. Here are the results from DFT optimization:
Unrelaxed bulk = 1.96 Å, relaxed 3-layer = 1.84 Å, 4-layer = 1.83 Å, 5-layer = 1.83 Å, 6-layer = 1.83 Å, 7-layer = 1.83 Å, 8-layer = 1.83 Å
On this surface, the geometric parameter is evidently quite insensitive to slab thickness, converging to <0.01 Å at 4 layers thick. Therefore, if studying a surface property that is controlled by this parameter, then a 4-layer slab model seems to be adequate.
Another common approach is to assess the convergence of the surface energy, especially if interested in a related quantity like the adsorption energy. While various formulae exist for computing surface energy, we follow Sun and Ceder (Efficient creation and convergence of surface slabs. DOI:10.1016/j.susc.2013.05.016) in using the following:
Esurf (hkl) = [Etot - n*Ebulk (hkl)] / 2A
where Etot is the DFT total energy of the slab containing n bulk TiO2 units, A = a*b*sin(γ) is the cross-sectional area of the slab, calculated from the lattice parameters, and Ebulk is the DFT energy of the crystalline solid oriented along (h k l) so that its Brillouin zone in x & y is identical to that of the slab. It is shown by Sun and Ceder (Efficient creation and convergence of surface slabs. DOI:10.1016/j.susc.2013.05.016) that this minimizes errors in calculation of the surface energy. The panel therefore carries out QE calculations on (h k l)-oriented bulk structures, as well as on the slabs.
The surface energies computed in this way are included in the Project Table as the property 'QE Surface Energy (J/m^2)'. If needed, Go to the Property Tree (
), expand All > Materials Science > Secondary > QE Surface Energy (J/m^2). Check that your values match those in the Figure below.
For the 3-8 layer (1 1 0) TiO2 rutile slabs, the sequence shows odd-even oscillating convergence towards Esurf = 0.62 J/m2. The 4-layer slab is within 5% of convergence, which is adequate for the adsorption calculations in this example, bearing in mind the high computational cost of DFT slab calculations. However, if computing resources allow, results could be verified later on a 6-layer slab (with a surface energy converged to within 1.5%) or on an even thicker slab.
The thickness needed for convergence depends on the material and crystallographic orientation. The surface energies of the (1 0 1) TiO2 slabs show fairly smooth convergence to within ±0.08 J/m2 (6%) already at 4-5 layers. The 4-layer slab is therefore an adequate model to proceed with for this surface.
Note that the bare (1 0 1) surface evidently has a higher Esurf than bare (1 1 0), meaning that (1 0 1) is less stable. This agrees with both experiment and prior theory work on TiO2 (see the References below). We would typically proceed to use the most stable surface, i.e. (1 1 0), as the model of the bare surface. In this tutorial, however, we will demonstrate adsorption enumeration for both (1 1 0) and (1 0 1) slabs.
5. Building Adsorbate Structures
This section will demonstrate use of the Enumerate Adsorbates panel to prepare slab models (referred to as ‘adsorbates’) where a gas molecule has adsorbed onto a substrate slab. These systems can then be used to compute adsorption energies. We will enumerate several structures by combining gas molecules trimethylphosphine, P(CH3)3 and hydroxylamine, NH2OH with the bare TiO2 substrates from the Surface Energy panel (Sections 3 & 4) and with a hydroxylated substrate from a separate QE calculation. The tool described herein can be used to build a single or many adsorption structures from any combination of substrate(s) and gas molecule(s).
Note that instead of periodic cells, nanoclusters can also be used as substrates for Adsorption Enumeration.
Additionally, the Custom R-Group Enumeration tool (help documentation and tutorial) can be used for enumerating groups adsorbed at multiple sites and various coverages within the surface cell.
The Enumerate Adsorbates panel uses rules to specify substrate sites and molecular orientation, as will be shown in this section. Complementing this, the Adsorption Site Finder panel uses force fields in Monte Carlo simulated annealing to generate random orientations and locations for an adsorbate (Locating Adsorption Sites on Surfaces tutorial). Structures from the Enumerate Adsorbates panel can therefore be used as input for the Adsorption Site Finder panel.
Finally, note that the Molecular Deposition tool (help documentation and tutorial) is a different but related methodology for studying physisorption of layers or multi-layers of molecules with force-field-based molecular dynamics.
Figure 5-1. The optimized slabs in the entry list with the imported hydroxylated slab included in the workspace.
First, let’s import a hydroxylated 4-layer TiO2-(1 1 0) slab, functionalized with OH groups, that has been optimized for you:
- Go to File > Import Structures, navigate to the tutorial files and import
Section_05 > optimized_hydroxylated_slab.mae
Your entry list now contains 13 optimized slabs:
- 3-8 layers for (1 1 0)
- 3-8 layers for (1 0 1)
- 1 hydroxylated slab
The newly-imported hydroxylated surface has a coverage of 2 hydroxyl (OH) groups per 1x1 surface cell. In experiment, the coverage of OH groups depends on the processing conditions, with higher coverages at lower temperatures and higher partial pressures of water vapor.
Compare the hydroxylated slab with the bare TiO2-(1 1 0) slab. You can see that, on each face, H has been added to surface-oxygen and OH has been added to surface-titanium, a net addition of H2O to each face. This should not cause any substantial change in the electronic structure of the TiO2 substrate.
Here we provide two gas molecules for importing (feel free to draw them yourself if you prefer):
- Go to File > Import Structures, navigate to the tutorial files and import
Section_05 > gas_molecules.mae- A new entry group is added to the entry list entitled gas_molecules (3). It contains P(CH3)3 and two identical entries of NH2OH
Depending on the size of the gas molecules, larger surface cells may be needed to provide space for them to adsorb. We will proceed with the (1 1 0) 4-layer cell, the (1 0 1) 4-layer cell and the hydroxylated 4-layer cell. Let’s make larger surface cells:
- Includethe entry is represented in the Workspace, the circle in the In column is blue surface_energy_all_slabs_110_bott_0.50_thick_4.00_vac_6.00A (the optimized (1 1 0) 4-layer slab)
- Click on the
button (bottom-right corner of the workspacethe 3D display area in the center of the main window, where molecular structures are displayed) to open the Periodic Structure Tool Window
- Go to Build Cell
- Click Extents
- Insert 1 for the +B direction
- Click Apply
- An extended view of the crystal is displayed in the workspacethe 3D display area in the center of the main window, where molecular structures are displayed
- Cell 1 x 2 x 1 is shown in green in the bottom right
- Click on the
button to open the Periodic Structure Tool Window
- Click Make P1 Cell
- A new entry is added to the bottom of the entry list
Note: If you are not familiar with these cell manipulations, revisit the Building and Manipulating Crystal Structures tutorial.
Note: Rotations and expansions of the cell are also possible via the Redefine Lattice panel.
- Rename the new entry bare_TiO2_110_4layer_1x2
Figure 5-7. The three expanded cells are renamed in the entry list. The (1 0 1) cell is shown in the workspace.
- Repeat Steps 3-11 for the (1 0 1) 4-layer cell and the hydroxylated 4-layer cell, ensuring that you are using the optimized structures. EXCEPT, make the following specifications:
- Make the (1 0 1) cell 2x2 by inserting 1 for extents in both the +A and +B directions
- Rename the (1 0 1) cell bare_TiO2_101_4layer_2x2 accordingly
- Rename the hydroxylated cell hydroxylated_TiO2_110_4layer_1x2
The Figure shows the named entries in the entry list. Feel free to also import Section_05 > expanded_slabs_for_enumeration.mae from the provided files to confirm that your slabs match.
- 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 three expanded substrates in the entry list
- Go to Tasks > Materials > Enumeration > Adsorption
- The Enumerate Adsorbates panel opens
- In the Substrates selection, with Load structures from Selected Entries chosen, click Load
- The three substrates are loaded into the table
- The first substrate is shown in the workspace with one atom shown as a ball proposed as the ‘most exposed’
The panel automatically proposes the most exposed atom in each structure as its Reactive Atom.
- Change the Reactive atom filter to Ti
- The most exposed titanium atom is selected in the first two substrates
- Note that whether the atom is at the ‘top’ or ‘bottom’ of the slab is arbitrary and inconsequential, but can be changed manually if desired
The panel cannot identify an exposed titanium atom in the case of the hydroxylated surface. This is, of course, expected given that none of the titanium atoms are exposed.
- Select (click-to-highlight) the third row in the Substrates table
- The hydroxylated slab is shown in the workspace
- Click Pick in WS
- Click on an exposed H to manually select the hydrogen as the Reactive Atom
The three slabs are now all imported into the panel, and each is tagged with a reactive atom (an exposed Ti for each of the bare slabs and a hydrogen atom for the hydroxylated slab)
We can now proceed to load the gas molecules:
- Keep the Enumerate Adsorbates panel open, and return to the entry list to 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 three gas molecules
- In the Gas molecules selection, with Load structures from Selected Entries chosen, click Load
- The three gasses are loaded into the table
- The first molecule is shown in the workspace with one atom shown as a ball proposed as the ‘most exposed’
We have duplicated the hydroxylamine substrate because we could imagine being interested in analyzing adsorption via O-coordination or N-coordination:
- Select (click-to-highlight) the second row in the Gas molecules table
- The hydroxylamine_1 molecule is shown in the workspace
- Click Pick in WS
- Click on the O to manually select the nitrogen as the Reactive Atom
- Select (click-to-highlight) the third row in the Gas molecules table
- The hydroxylamine_2 molecule is shown in the workspace
- Click Pick in WS
- Click on the N to manually select the nitrogen as the Reactive Atom
- Keep the Adsorption direction set to Automatic
- The output with the fewest steric interactions will be chosen
- Check Enumerate and set to 3 rotations around the adsorption axis
- This will enumerate three rotational orientations of each adsorbate. For hydroxylamine in particular, this will result in a variety of useful starting models
- Change the Job name to enum_adsorbates_Ti_gas
- Confirm that your panel resembles that which is shown in the Figure
- Adjust the job settings (
) as needed
- This job requires a CPU host and takes approximately 2 minutes
- If you would prefer to proceed with pre-generated outputs, go to File > Import Structures, navigate to the tutorial files and import
Section_05 > enum_adsorbates_Ti_gas > enum_adsorbates_Ti_gas-out.maegz. Otherwise, click Run - Close the Enumerate Adsorbates panel
The output is an entry group entitled enum_adsorbates_Ti_gas (27) containing the 27 possible enumerated outputs from the input (matching each surface, with each gas, in each orientation 3 x 3 x 3 = 27)
- Includethe entry is represented in the Workspace, the circle in the In column is blue the first entry, entitled trimethylphosphine_on_bare_TiO2_110_4layer_1x2_top-28 which is associated with the adsorption of trimethylphosphine on the (1 1 0) 4-layer slab in one rotated orientation
The adsorbate atoms may appear to be distributed across the cell boundaries. This is purely visual, but we can make useful adjustments if we like:
- Right-click on the P atom in the workspace and click Center Cell on Atom
- The cell adjusts such that the P atom is at the center of the periodic cell
Note: The Manipulate Cell panel can also be used for related adjustments
Feel free to visualize any of the other enumerated adsorbates as you wish. It is recommended to extend to a 2x2x1 supercell using the periodic structure tools to check for overlap of the adsorbate with its image neighbors (rotation may relieve hindrance here). Always remember to revert back to the 1x1x1 cell afterwards before proceeding to any calculations.
Some of the enumerated structures may be unsatisfactory and can be discarded. Settings can be adjusted in the Enumerate Adsorbates panel and it can be re-run until you are satisfied with the adsorbate structures. Note that if you want to generate a single adsorbate structure, just import one substrate, one gas molecule and include no directional enumeration.
Each enumerated structure is valid input for the Adsorption Site Finder panel, where the adsorbate can be randomly translated and rotated so as to sample more configurations (Locating Adsorption Sites on Surfaces tutorial).
It is important to note that these are just candidate, unoptimized adsorbate structures. The next step is to run DFT calculations (with Quantum ESPRESSO for periodic slabs or Jaguar for nanoparticles) to optimize the adsorbate structures and compute properties. For this tutorial, we will now proceed to compute adsorption energy for just the trimethylphosphine_on_bare_TiO2_110_4layer_1x2_top-28 system, but feel free to follow the steps in Section 6 for any of the systems.
6. Computing Adsorption Energies
This final section will introduce the Adsorption Energy and Thermochemistry Viewer panels which are used to optimize the adsorbate structure and compute thermodynamics (i.e. adsorption free energies) for the system. To run these jobs, all that is required is an adsorbate structure, for example, any of the 27 structures that we generated with the Enumerate Adsorbates panel in Section 5. It is also possible to use structures generated by the Adsorption Site Finder panel (Locating Adsorption Sites on Surfaces tutorial) as input here.
The Adsorption Energy and Thermochemistry Viewer panels support batch operation, inputting multiple structures at once. Here we will only look at one of these structures, trimethylphosphine on the (1 1 0) 4-layer bare TiO2.
As an alternative to the Adsorption Energy panel, one can optimize substrate and adsorbate slabs at DFT level using Quantum ESPRESSO (as described in this tutorial), use Quantum ESPRESSO to also compute the gas molecule in a periodic cell and then combine the DFT total energies manually to obtain the adsorption energy.
Although beyond the scope of the current tutorial, we note here that fragmentation or diffusion reactions of adsorbates on surfaces can likewise be studied via Quantum ESPRESSO slab models, including computation of activation energies with the Nudged Elastic Band approach.
- 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 trimethylphosphine_on_bare_TiO2_110_4layer_1x2_top-28 entry from the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion
Note: Alternatively, you can select a batch of enumerated entries from some or all of the output of the Enumerate Adsorbates panel.
- Go to Tasks > Materials > Quantum Mechanics > Workflows > Adsorption Energy
- The Adsorption Energy Calculations panel opens
Note: We do not need to select the substrate or gas molecule. The adsorbate entry alone includes the requisite information about its components for the calculation.
The Adsorption Energy Calculations panel allows you to Generate Preview Structures before running the calculation to allow for modifications to be made to the periodic and/ or molecular system. If a job fails, or completes without successfully calculating all of the adsorption energy parameters, we can use the Restart job option to troubleshoot.
For slabs such as this, we will use periodic DFT (Quantum ESPRESSO), but note that the panel can also compute adsorption onto non-periodic systems such as nanoparticles with Jaguar.
The panel also allows adsorption energies to be computed using machine learning force fields (MLFF) and more detail is given in the Atomic Layer Deposition tutorial.
- Ensure that Periodic with Quantum Espresso is chosen
- Click Advanced Options
-
Adjust the Advanced Options so that they match those used previously in the tutorial:
- On the Theory Tab: Spin-polarized, GGA, PBE, DFT-D3, Grid plane distance 0.1, include Γ point, click Update K-point mesh, ensure that there is only one k-point in the z-direction (i.e. 2 x 2 x 1), editing that value manually if necessary
- On the SCF Tab: custom energy cutoff for wavefunctions 40.0 Ry and custom energy cutoff for charge density 200.0 Ry. Change the Max steps in SCF to 150
- On the Optimization tab change the Number of steps to 150
- Click Save
- Check Adsorption free energy
- Electronic (internal) energy is always computed and output to 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 by default as well
- Set the Temperature to Start at 273.15 K, the Number of steps to 5 and the Step increment to 50.0 K
- Set the Pressure to Start at 0.001 atm, the Number of steps to 3 and the Step multiplier to 10.0
- These settings specify pressures of 0.001, 0.01 and 0.1 atm
- Retain the Gas phase entropy lost upon adsorption at 30%
- The entropy contribution is approximated with Jaguar. For an adsorbate that is mobile on the surface, 30% entropy loss is a good approximation. See the documentation for further discussion
- Retain the default Jaguar Options, these settings do not need to match the Quantum ESPRESSO settings:
- The Jaguar calculation is purely for calculating the entropy of the gas molecules (a measurable quantity) via vibrational analysis
- Quantum ESPRESSO calculations are calculating electronic energies (not a measurable quantity)
- Since Jaguar and Quantum ESPRESSO are calculating different, relatively independent quantities their settings do not need to match
- Change the Job name to adsorption_energy_Ti_PMe3
- Adjust the job settings (
) as needed
- This job requires a CPU host and must be run on a Linux machine. The job takes approximately 20 hours on a 12 CPU host
- If you would prefer to proceed with pre-generated outputs, go to File > Import Structures, navigate to the tutorial files and import
Section_06 > adsorption_energy_Ti_PMe3 > adsorption_energy_Ti_PMe3-out.maegz. Otherwise, click Run - Close the Adsorption Energy Calculations panel
When the job finishes or after importing, a new group of entries is created containing the optimized adsorbate and the individual gas and substrate molecules from which it was derived.
- Open 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 (
)
Associated with the adsorbate entry, the Adsorption Energy (as defined in the documentation) is listed as -33.36 kcal/mol, indicative of a strong Lewis acid-base interaction between Ti and P.
Adsorption free energies at the specific temperatures and pressures are also given with the same entry in the Project Table (and can be exported as .csv)
- Close 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
- Open the Thermochemistry Viewer panel via the WAM button (
) directly in the entry lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion
- Alternatively, access the panel via Tasks > Materials > Quantum Mechanics > Molecular Quantum Mechanics > Thermochemistry Viewer
- This panel can read any energy output and plot it as a function of pressure or temperature
- If the panel is not populated with data, select the output and click Load Data for Selected Entries
- Select Joules for the Base units
The axes can be fine-tuned for easier visualization.
- Click Axes…
- The Axes Parameters panel opens
- For Temperature set the Minimum to 0 and the Maximum to 600
- For Adsorption Free Energy set the Minimum to -150 and the Maximum to -75
- Click OK
Note that you can use the Copy button to copy and paste the graph image outside of MS Maestro.
The data shows that adsorption of trimethylphosphine onto bare TiO2-(1 1 0) at this coverage is thermodynamically favorable (ΔG<0) at all of these temperatures and pressures. In Section 5, we generated 27 adsorbate structures for various molecules on various TiO2 facets and sites. We can repeat the adsorption energy calculations from this section for any of those other adsorbate structures. Comparing the resulting thermochemistry curves would tell us whether adsorption of those molecules onto the other surfaces is more or less favorable than trimethylphosphine onto bare TiO2-(1 1 0).
7. Conclusion and References
In this tutorial, we learned how to generate atomistic models of surfaces and model the adsorption of small molecules onto the surfaces via a multi-step workflow.
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 related practice, proceed to explore other relevant tutorials:
- Introductory and advanced topics in solid-state modeling:
- Machine learning for solid-state materials:
For further reading:
- Help Documentation on: Surface Energy, Enumerate Adsorbates, Adsorption Energy Calculations, Thermochemistry Viewer
- Efficient creation and convergence of surface slabs. DOI:10.1016/j.susc.2013.05.016
- Optimisation of accurate rutile TiO2 (110), (100), (101) and (001) surface models from periodic DFT calculations. DOI:10.1007/s00214-006-0189-y
- The surface science of titanium dioxide. DOI:10.1016/S0167-5729(02)00100-0
- The stability of ionic crystal surfaces. DOI:10.1088/0022-3719/12/22/036
- Introduction to Computational Chemistry, 3rd Edition
- Essentials of Computational Chemistry: Theories and Models, 2nd Edition
- Molecular Modelling: Principles and Applications, 2nd Edition
8. 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
Recent actions - This is a list of your recent actions, which you can use to reopen a panel, displayed below the Browse row. (Right-click to delete.)
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