DFT and Ab Initio Calculations for Water Splitting Including Real-Time Time-Dependent Density Functional Theory


Lawrence Berkeley National Laboratory (LBNL)

Capability Expert

Lin-Wang Wang


Computational Tools and Modeling

Node Readiness Category

2: High-Temperature Electrolysis (HTE)
1: Low-Temperature Electrolysis (LTE)
1: Photoelectrochemical (PEC)
3: Solar Thermochemical (STCH)


We have several capabilities and expertise using ab initio simulation to study low-temperature electrolysis (LTE) and photoelectrochemical (PEC) related phenomena:

  1. Band alignment crossing realistic interface (e.g., crystal-Si/amorphous-SiO2) can be calculated using HSE functional and realistic structure models with several hundred atoms.
  2. Carrier dynamics, charge transport and conductivity crossing interface, from band edge state to defect state, hot carrier cooling, can be investigated using a few techniques (real-time time dependent density functional theory; non-adiabatic molecular dynamics, Marcus theory, and multiphonon process through electron-phonon coupling).
  3. Oxygen evolution reaction (OER) and hydrogen evolution reaction (HER) reaction steps, surface stability, and Pourbaix diagram can be calculated using electrochemistry calculation method with implicit solvent model and Poisson Boltzmann equation. In particular, we can carry a fix electrode potential grand canonical calculation for electrochemical reaction at the solid/liquid interface.
  4. For large systems with >10,000 atoms, we can carry out self-consistent field density functional theory (DFT) calculations for the charge density and potential using our divide-and-conquer linear scaling LS3DF code, followed with band edge state and density of state calculations.
  5. Genetic algorithm surface structure search can be used to explore the possible surface structure (e.g., reconstruction, or molecule/surface attachment configuration), or cluster atomic configuration. Besides, defect non-radiative decay rate (Shockley-Read-Hall, SRH recombination) can be calculated using a special electron-phonon coupling algorithm. Defect level can be investigated using HSE calculations, and electronic structure of >1,000 atom nanosystems can be studied with charge patching method.

Capability Bounds‎

For many of the ab initio theoretical studies, we need to have a good idea for the atomic structures, or some atomic structure candidates from which we can investigate their energies. At this stage, we cannot simulate the material synthesis process, although limited ab initio molecular dynamics can be carried out for tens of picoseconds with hundreds of atoms. For some of the strongly correlated systems, the DFT (and DFT+U) method has its limitation and uncertainty. Cautions are needed in studying such systems.

Unique Aspects‎

Many of our methods are developed in-house and can be adapted according to the problem at hand. We are not just using the commercial DFT codes. The LS3DF method is a linear scaling method for large system (>10,000 atoms) calculations. Rt-TDDFT uses a new algorithm that speeds up the calculation by factor of 10, and it includes detailed balance and decoherence for carrier cooling. We have the unique algorithm to calculate the electron-phonon coupling constant within a single SCF calculation, which enable us to study SRH recombination rate and other charge transfer processes. Our in-house genetic algorithm code is capable of searching surface and deposition structures. The fixed potential grand canonical calculation can be used to study many electrochemistry processes for a given electrode voltage. We have tools and experience for carrier dynamics, transport, and cooling processes.


Most of the codes are ready to use. They can be run on our own computers within the group. We also have a computer center account at NERSC, although the largest calculation (e.g., >10,000-atom LS3DF calculation) might depend on the availability of a larger supercomputer resource. We have previous experience working with other groups in the HydroGEN project, and our calculations help to explain the experimental results.


Ab initio simulation can provide atomistic insights to understand the experimental phenomena and reaction mechanism, thus help to improve the design of the system. Although there are always uncertainties related to ab initio simulations, such atomistic insights usually cannot be provided by other direct experimental measurements. The ab initio calculation will be most useful when it can work in sync with experimental characterizations to provide cross checking.


Figure 1. The calculated OER reaction intermediate steps, and the resulting OER overpotential for Co-doped graphene systems with N bonding

Figure 2. The calculated volcano plot for the relationship between the OER over potential, and the O*, HO* free energy difference for different transition metals on graphene substrate

Figure 3. The charge transfer calculation between the crystal-Si and amorphous SiO2, calculated with hybrid HSE functional and Marcus theory. The HSE functional gives the correct band alignment between different semiconductor systems.

Figure 4. Large scale density functional theory electronic structure calculations using the linear scaling LS3DF code for hybrid perovskite MAPbI3. The random orientation of the MA molecule causes an electrostatic potential fluctuation, which localized the hole wave function as shown in the figure. The system contains 20,000 atoms.


  1. L.W. Wang, "Divide and conquer quantum mechanical material simulations with exascale supercomputers," National Science Review 1 (2014): 604.
  2. Z. Wang, S.S. Li, L.W. Wang, "An efficient real-time time-dependent DFT method and its applications to ion-2D material collision," Physical Review Letters 114 (2015): 063004.
  3. H.H. Pham, M.-J. Chen, H. Frei, L.W. Wang, "Surface proton hopping and fast-kinetics pathway of water oxidation on Co3O4 (001) surface," ACS Catalysis 6 (2016): 5610.
  4. F. Zheng, H.H. Pham, L.W. Wang, "Effects of the c-Si/a-SiO2 interface atomic structure on its band alignment: an ab initio study," Physical Chemistry Chemical Physics 19 (2017): 32617.
  5. G. Gao, F. Pan, L.W. Wang, "Theoretical investigation of 2D hexaaminobenzene coordination polymers as Li-S battery," Advanced Energy Materials 8 (2018): 1801823.
  6. Y. Zhou, G. Gao, Y. Li, W. Chu, L.W. Wang, "Transition-metal single atoms in nitrogen-doped graphenes as efficient active centers for water splitting: a theoretical study," Physical Chemistry Chemical Physics 21 (2019): 3024.
  7. Y. Zhou, G. Gao, J. Kang, W. Chu, L.W. Wang, "Transition metal-embedded two-dimensional C3N as a highly active electrocatalyst for oxygen evolution and reduction reactions," Journal of Materials Chemistry A 2019, DOI:10.1039/c9ta01389j.