Exploration of the reaction pathways of (1) oxygen evolution reactions and (2) electrolyte decomposition using the SC-AFIR method

Exploration of the reaction pathways of oxygen evolution reactions using the SC-AFIR method

Introduction

Understanding reaction pathways is crucial for improving the activity and durability of electrocatalysts. However, exploring reaction pathways is usually cumbersome because candidate pathways must be hypothesized manually, and verifying these assumptions with first-principles calculations is computationally expensive. We focused on GRRM20 with Matlantis as a solution to this problem. While this method has been applied to thermal reactions, there are very few examples of its application to electrochemical reactions [1]. Therefore, we investigated whether this method could reproduce known electrochemical reaction pathways without prior knowledge. We targeted RuO₂, a catalyst for the oxygen evolution reaction (OER). In this system, the adsorbate evolution mechanism (AEM), in which adsorbed species participate in the reaction, occurs. Previous studies have reported that the rate-determining step is *O → *OOH [2]. We aimed to evaluate the usefulness of this method and demonstrate its potential application to other electrochemical reactions and catalyst design in the future.

Computational Details

A 3×2 supercell of the rutile-type RuO2(110) surface was constructed. The RuO₂ slab consisted of four atomic layers, with a vacuum layer of 20 Å. The surface was oxygen-terminated [3]. Two water molecules were placed on the RuO2(110) surface, and an artificial force was applied to these water molecules and five oxygen atoms on the catalyst surface. The artificial force was set to 500 kJ/mol. The artificial force-induced reaction method was performed using the single-component algorithm (SC-AFIR) implemented in the GRRM20 program. After identifying the reaction pathways that produced O2 from the reaction pathways determined by SC-AFIR, an energy diagram was calculated using a computational hydrogen electrode (CHE) model.

Results and Discussion

SC-AFIR allowed us to obtain more than 600 locally stable structures. By tracing the reaction pathways starting from the most stable structure, we found two main pathways for the OER. The first pathway involved adsorbed OH reacting sequentially to form OOH, which then produced OO, corresponding to the AEM (Pathway 1 in Figure 3). The second pathway involved oxygen atoms present on the catalyst surface spilling over and combining with adjacent oxygen atoms to form *OO (Pathway 2 in Figure 3). During the exploration of the reaction process, a side reaction was also observed in which hydrogen dissociated from H₂O was adsorbed onto the RuO₂ surface. This was attributed to an artifact of the vacuum slab model, in which hydrogen species that would realistically dissolve into the electrolyte were excessively adsorbed on the surface. Here, the side reaction was removed from the computational model and an energy diagram was constructed. Figure 3 shows that the free energy change for the rate-determining step was 1.74 eV for *O → *OOH in Pathway 1 and 3.51 eV for *O → *OO in Pathway 2. The lower free energy change along Pathway 1 indicates that this path is more favorable, which is consistent with previously reported first-principles calculations [2].

As described above, by using GRRM20 with Matlantis, we succeeded in exploring the oxygen evolution pathways on RuO₂(110) simply by providing the reactants, without any prior knowledge of the reaction.

Computational Details

Items Details
Number of atoms 161 atoms (Ru48, O109, H4)
Model (110) plane, 3×2 supercell, 4 RuO2 layers, O termination, 20 Å vacuum layer, 298.15 K
PFP v8.0.0, R2SCAN_PLUS_D3
GRRM20 SC-AFIR, gamma = 500 kJ/mol, all but 11 atoms are fixed in position as Frozen Atoms (see Figure 2).
Options
NoFC
DownDC=15
EQOnly
KeepLUPPath
Nrun=1
SaveDataIndirectory
SC=FindUDC
RandBox-X=2.56 8.80
RandBox-Y=2.22 10.49
RandBox-Z=11.15 18.03
AddInteraction
gamma=500
Target=1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
END
ReadBareEnergy
TrafficVolCheck
TimeScale=1e3
Siml_mLatest=370
Siml_tLargest=1110
SimlTemperature=300 500 1000
RTemperature=10000
ScaleStructCheckThreshold=3
MatchDecScale=7
MatchDecTarget=1-4
sublink = ./kickpfp.py

References

[1] Ichino, T. et al. ChemRxiv. 2024, 10.26434/chemrxiv-2024-9h20v.

[2] Zagalskaya, A. et al. ACS Catal. 2020, 10(6), 3650-3657

[3] Fornaciari, J. C. et al. Electrochimica Acta, 2022, 405, 139810.


Exploration of the reaction pathways of electrolyte decomposition

Introduction

Precise control of the solid electrolyte interphase (SEI) at the electrode/electrolyte interface is crucial for the stable operation of lithium-ion batteries. Complex reactions, including electrolyte reduction, occur at the negative electrode interface, making comprehensive analysis with conventional computational methods difficult because of the high computational cost. We therefore applied GRRM20 with Matlantis, which enables systematic reaction pathway exploration at a realistic computational cost, to electrolyte components.

In this study, we modeled the reductive decomposition of vinylene carbonate (VC), a representative SEI-forming additive, near the negative electrode surface, and explored the decomposition pathways using the SC-AFIR method. In particular, we focused on CO₂ evolution, which is characteristic of the VC decomposition reaction, and analyzed the resulting reaction pathway.

Computational Details

We systematically explored the reaction pathways of VC using the SC-AFIR method implemented in GRRM20 with Matlantis. Following the previous study [1], we performed calculations on the Li-coordinated VC adduct (Figure 2).

Results and Discussion

Applying the SC-AFIR method to the VC + Li model yielded 188 equilibrium structures (EQs) and 2,343 path tops (PTs) within a few days of computation.

An example of a reaction pathway is shown in Figure 3. By comparing it with a previous study [1] that applied the ADDF method based on quantum chemical calculations, we confirmed that the pathway for the formation of CO₂ from VC + Li was successfully reproduced. Although the calculated energies deviate somewhat from those reported in the previous study, the high computational speed of GRRM20 with Matlantis enabled us to systematically screen reaction pathways. Promising candidate pathways can subsequently be refined with quantum chemical calculations and discussed in more detail.

These results demonstrate the effectiveness of utilizing GRRM20 with Matlantis to analyze extremely complex reactions in electrolytes. Because this approach is significantly faster than performing full quantum chemical calculations, we anticipate it will be useful for predicting and controlling diverse reactions at the electrode/electrolyte interface in complex battery systems.

Computational Details

Items Details
Number of atoms 9 atoms
Model Vinylene carbonate + Li
PFP PBE_U_PLUS_D3 v8.0.0
GRRM20 SC-AFIR method

# Options
SaveDataIndirectory
NoFC
DownDC=15
EQOnly
KeepLUPPath
Nrun=1
AddInteraction
gamma=200
Target=1, 2, 3, 4, 5, 6, 9 (Specify atoms other than H)
END
TrafficVolCheck
TimeScale=1.0
Siml_mLatest=70
Siml_tLargest=210
SimlTemperature=300 500 700
RTemperature=7000

References

[1] https://iqce.jp/SRPS/SRPS2012/Miyamoto.pdf

Profile of the case study provider

Asahi Kasei Corporation

Company logo

what's new

Large-Scale MD Simulation of GaN/AlN Heteroepitaxial Thin-Film Cracking Using LightPFP

ceramicsemiconductors

Analysis of the Ozone Oxidation Reaction Mechanism of Rubber Antidegradants

PolymerReaction pathway exploration

Interfacial Thermal Resistance Calculation at a Metal–Polymer Interface

Interfacial Thermal ResistanceSemiconductorAdhesive Materials

High-Accuracy, High-Speed Binding Energy Evaluation of Cyclin-Dependent Kinase 2 Inhibitors

Bio and Pharma

Data-Driven Method for Discovering Low Adsorption Energy Molecules on Si Surfaces

ScreeningMaterials InformaticsSemiconductors