Addressing Dynamics at Catalytic Heterogeneous Interfaces with DFT-MD: Anomalous Temperature Distributions from Commonly Used Thermostats

Density functional theory-based molecular dynamics (DFT-MD) has been widely used for studying the chemistry of heterogeneous interfacial systems under operational conditions. We report frequently overlooked errors in thermostated or constant-temperature DFT-MD simulations applied to study (electro)catalytic chemistry. Our results demonstrate that commonly used thermostats such as Nosé–Hoover, Berendsen, and simple velocity-rescaling methods fail to provide a reliable temperature description for systems considered. Instead, nonconstant temperatures and large temperature gradients within the different parts of the system are observed. The errors are not a “feature” of any particular code but are present in several ab initio molecular dynamics implementations. This uneven temperature distribution, due to inadequate thermostatting, is well-known in the classical MD community, where it is ascribed to the failure in kinetic energy equipartition among different degrees of freedom in heterogeneous systems (Harvey et al. J. Comput. Chem.1998, 726−740) and termed the flying ice cube effect. We provide tantamount evidence that interfacial systems are susceptible to substantial flying ice cube effects and demonstrate that the traditional Nosé–Hoover and Berendsen thermostats should be applied with care when simulating, for example, catalytic properties or structures of solvated interfaces and supported clusters. We conclude that the flying ice cube effect in these systems can be conveniently avoided using Langevin dynamics.

Dear Editor, Thank you for your email regarding our manuscript. We wish to thank the reviewers for their valuable work and hope that the updated manuscript will be found suitable for publication in The Journal of Physical Chemistry Letters.
We acknowledge the first r eviewer's n ote r egarding t he p ublication v enue b ut f ully agree with their assessment that "in JPCLett the paper probably has a greater chance to be seen by those who need to read it".
Please note that the TOC figure o f t he m anuscript h as b een updated.
Below we respond to the second reviewer's comments in detail. These comments have been included in black, our responses in blue, and modifications t o t he t ext i n italics.
Sincerely, on behalf of all the authors,

Karoliina Honkala
Reviewer 2 a) Did the authors noticed a similar paper about the temperature distributions among the different degrees of freedom using classic MD simulation (Yan LM, et al. Adv. Manuf. 2013, 1(2) 160)? b) What were the differences between these two papers: the simulated systems, the MD method, and the thermostats? c) What is the progress of the present manuscript compared with the pervious paper? a) We thank the referee for pointing out this paper, which focuses on classical molecular dynamics simulations for bulk water. We would like to point out that the flying ice cube effect is well known in classical MD literature, which was already extensively cited in our original manuscript. We now cite the Yan et al. paper in the main manuscript (see citation 68) and briefly d iscuss i t i n s upporting i nformation, w here w e p resent our results for bulk water.
The following text has been added to chapter 3B in SI: This short evaluation already demonstrates the unpredictability of the kinetic energy distribution when using the Nosé-Hoover thermostat. The effect is not in the strict flying i ce c ube c ategory as the lowest-frequency mode is under-represented. Similar variability in the kinetic energy drainage has been observed in a classical MD evaluation of a bulk water system [22] where the effect depended on the time step and exhibited the opposite direction from low to high frequencies.
b) As an answer to referee's question how our work differs from the Yan et al. 2013 work, we point out the following: • Yan et al. focused on classical MD whereas we discuss DFT-MD, which has its own considerations, especially convergence issues.
• We discuss multiple thermostats and demonstrate the flying ice cube effect over multiple DFT-MD implementations, while Yan et al. only used one software and the Nosé-Hoover thermostat.
• We extend the discussion from homogeneous fluids to heterogeneous or twocomponent systems, demonstrating the possible issues for several systems of current interest instead of only looking at simple model systems like bulk water.
• Finally and most importantly, we highlight that the flying ice cube effect is not just a minor theoretical detail for MD experts to nitpick about, but a very concrete concern, and that some of the DFT-MD results presented in literature might suffer from this problem. c) Please see our response b). 1