Researchers have created the first complete quantum mechanical water model.
They used machine learning to develop the model, which provides a detailed and accurate description of how large groups of water molecules interact with each other.
“We think we’ve found the missing piece to a complete, microscopic understanding of water,” said Joel Bowman, professor of theoretical chemistry at Emory University and senior author of the study. “It appears that we now have everything we need to know to describe water molecules under any conditions, including ice, liquid or vapor, over a range of temperatures and pressures.”
The researchers developed free, open-source software for the model, which they dubbed “q-AQUA.”
The q-AQUA software offers a universal tool for analyzing water. “We expect researchers will use it for everything from predicting whether an exoplanet might have water to deepening our understanding of water’s role in cellular function,” says Bowman.
Bowman is one of the founders of the field of Theoretical Reaction Dynamics and a leader in exploring the mysteries underlying questions such as why we need water to live.
The study’s first author is Qi Yu, a former Emory graduate student at Bowman Lab who is now a postdoctoral fellow at Yale. Co-authors include Emory graduate student Apurba Nandi, a graduate student in Bowman Lab; Riccardo Cone, a former Emory postdoctoral fellow at the Bowman Lab who is now at the University of Milan; and Paul Houston, former Dean of Science at Georgia Institute of Technology and now Professor Emeritus at Cornell University.
Water covers most of the earth’s surface and is essential for all living organisms. It consists of simple molecules, each composed of two hydrogen atoms and one oxygen atom bonded by hydrogen.
Despite the simplicity and ubiquity of water, describing the interactions of clusters of H2O molecules under arbitrary conditions poses great challenges.
Newton’s law governs the behavior of heavy objects in the so-called classical world, including the motion of planets. However, extremely light objects at the level of atoms and electrons are part of the quantum world governed by the Schrödinger equation of quantum mechanical systems.
“The hydrogen atom is the lightest atom of all and therefore the most quantum mechanical one,” explains Bowman. “It has the quantum weirdness of being a particle and a wave at the same time.”
Although large, complex problems can be broken down into parts to be solved in the classical world, objects in the quantum world are too “fuzzy” to be broken down into discrete parts.
Researchers have attempted to create a quantum model of water by decomposing it into the interactions of clusters of water molecules. Bowman likens it to people at a party grouped into talk groups of two, three, or four.
“Imagine trying to come up with a model to describe the conversations in each of these groups of people that can be extended to the whole party,” he says. “First you collect data from two interlocutors and determine what they are saying, who is saying what, and what the conversation means. It gets more difficult when you try to model the conversations between three people. And when you have up to four people, it becomes almost impossible with so much data coming at you.”
For the current paper, the researchers used powerful machine learning techniques that allowed computers to capture the interactions of groups of two, three and four molecules.
“It was very difficult to get it down to the four-body level, and something that nobody had done and released before,” says Bowman. “We knew if we could achieve that, we would be a long way from having a near-complete solution. In a way it was the capstone of the whole process.”
Instead of words coming out of people’s mouths, the analyzes involved thousands of numbers coming out of computers. Unlike humans, however, the individual water molecules are all identical. This symmetry allowed the researchers to build on the model for interactions between sets of two, three and four water molecules, making it applicable to even larger sets of molecules.
“The four-body interaction of water molecules appears to be the final one that governs all interactions of water molecules,” says Bowman.
To test their model, the researchers ran computer simulations over a temperature range for up to 256 water molecules interacting simultaneously in groups of two, three and four molecules. The results showed that the model was very accurate even at this scale.
“We think we can get our model up to 3,000 or 4,000 interacting water molecules,” says Bowman. “The computational effort will increase greatly, but these are simulations that we intend to run next after we have demonstrated the proof of concept for our model.”
The model could also serve as a springboard for developing similar, simplified models that require less computational power but are still accurate enough to make useful predictions about the quantum mechanics of water, Bowman says.
In the meantime, the authors hope other researchers will download the free q-AQUA software and use it to delve deeper into unanswered questions about water.
“We’re about 70 percent water by weight,” says Bowman, “yet we don’t really understand chemically how water molecules interact with biological systems. Now that we have a good template to understand how water molecules interact with each other, we have a basis to deepen our understanding of the role of water in vital biochemical processes.”
The research appears in The Journal of Physical Chemistry Letters.
Source: Emory University