I’m currently working on a project that requires explicit solvation effects in the aqueous phase, and I’d like to incorporate explicit water molecules in my Q-Chem calculations. I understand that Q-Chem is primarily designed for quantum chemistry calculations, but I’d like advice on the best approach for setting up explicit solvation.
Specifically, I’d appreciate guidance on:
Adding Explicit Water Molecules: Is there a recommended way to position and number water molecules around my solute? Should I consider any particular geometry or symmetry to best mimic solvation?
Combining with Implicit Models: If combining explicit water with an implicit solvation model (such as IEFPCM), are there best practices for balancing the interactions? How do I ensure that the combined approach accurately represents aqueous solvation?
Automating Water Placement: Are there tools or methods within Q-Chem, or external programs compatible with Q-Chem, that can help in positioning water molecules around the solute?
Benchmarking: For those who have used explicit solvation in Q-Chem, any advice on benchmarking to determine the optimal number of water molecules?
Any insights, example inputs, or references would be greatly appreciated!
To answer the more general aspects of your quite loaded query, it is hard to comprehensively answer here. A better, more thorough approach to learn relevant protocol would be to read papers related to your specific research problem.
You need to generate a set of configurations (maybe through MD or AIMD simulations), then identify the solvent molecules you want to be treated quantum mechanically for each of the configurations, which is problem specific. A often-used protocol is to have MD or AIMD trajectories to run for some time before handpicking configurations obtained at regular intervals. For solute–solvent systems, individual molecules in these configurations are more or less randomly oriented, so exploiting spatial symmetry is mostly not possible.
While trying to determine the extent of explicit solvation, you need to ensure that the artifacts introduced by dismantling the condensed-phase system into different subsystems treated with different levels of theory does not impact the energetics and properties you are studying. In short, the property you are studying should be converged with respect to the number of explicit solvent molecules around the solute. This can be quite problem specific.
Regarding automating water placement, there are many MD tools out there that can directly create a water box around the solute, which you can then use to run MD simulations.
This is a question for the literature. You can find a review on implicit solvation models here, https://doi.org/10.1002/wcms.1519
and one on mixed implicit/explicit solvation (often called the “semicontinuum” or “cluster-continuum” approach) here, https://doi.org/10.1002/wcms.1440
Many MD packages will place solvent molecules around a solute, e.g., Tinker is a free one that will do this.