The CP-PAW code package for first-principles calculations from a user's perspective

TL;DR

CP-PAW unifies the projector augmented-wave method with Car-Parrinello dynamics to enable robust first-principles simulations of condensed-phase systems, including solids, liquids, and correlated oxides. The paper outlines the theoretical foundations (PAW, Car-Parrinello, k-point treatments, and local hybrids), prerequisites, and a hands-on tutorial suite covering wave-function optimization, structure optimization, MD, and solid-state scenarios. It also details practical aspects such as on-the-fly augmentation setup, thermostats, mass renormalization, and electrostatic decoupling, with guided exercises on Malonaldehyde, ferrite/hexaferrum, and PrMnO$_3$. The work demonstrates CP-PAW’s capability to handle complex electronic structures, magnetic order, and phase transitions, emphasizing its workflow for end users and reproducible tutorials through public data and documentation.

Abstract

CP-PAW is a combined electronic structure and ab-initio molecular dynamics code to perform mixed quantum and classical simulations of atomistic condensed phase systems, such as solids, liquids, and molecular systems. As the name suggests, the CP-PAW code unifies the all-electron projector augmented-wave method with the Car-Parrinello approach to determine not only the electronic and nuclear ground state of condensed matter, but also to study their properties and dynamics. In addition to briefly outlining the underlying theory, the focus will be on unique aspects of CP-PAW and how to correctly employ them as a user. How to install CP-PAW using the new build system will also be briefly mentioned.

Figures (9)

• Figure 1: Trajectory $x(t)$ of a one-dimensional harmonic oscillator obtained with the Verlet algorithm for different time steps $\Delta$ with $\Delta/T_0=0,\frac{1}{15},\frac{1}{5},0.99/\pi,1.01/\pi$ from top to bottom. $T_0$ is the exact oscillation period. The dots represent the calculated points, while the line is a plane-wave passing through the calculated points. The top row is the exact result. In the bottom row, the time step exceeds the stability limit, resulting in an exponential divergence. Within the stability limit, i.e. rows 2,3 and 4, the trajectory maintains the qualitative behavior with a frequency increasing with the time step.
• Figure 2: Frequency of the discretized one-dimensional harmonic oscillator, as a function of the time step $\Delta$ in terms of the exact oscillation period $T_0=2\pi/\omega_0$.
• Figure 3: Decay time $T$ of the total energy as a function of vibrational frequency $\omega_0$. The decay time describes the convergence of the total energy as $E(t)\sim\exp(-\frac{t}{T})$. The friction parameters $a=\alpha\Delta/2$ are red for $a=1$, blue for $a=0.5$, green for $a=0.25$ and orange for $a=0.125$, respectively. The stability limit of the Verlet algorithm is $\omega_0\Delta=2$, whereas the grey region is not accessible with any friction parameter. The dashed lines are the results from the discretized equations, while the full lines are from the continuous open trajectories.
• Figure 4: Two isomers of malonaldehyde before and after an intramolecular proton transfer.
• Figure 5: Long and short OH bond distances (red and black) in $a_{B}$ as a function of time in ps for malonaldehyde. For comparison, the double-bond network (blue) and intermolecular O-O distance (green) are also shown.