Science Success Story

« Back

Comet Supercomputer Used to Illustrate Novel Methane Storage Applications

Research Findings Could Lead to Higher Density Carbon-Neutral Chemical Fuels

By Kimberly Mann Bruch, SDSC Communications



This animation illustrates how an arriving methane molecule (adsorbate) and carbon surface model (adsorbent) mutually modify each other's electron density. Here, the methane molecule "lander" touches down at the carbon "moon surface," "kicking up dust" (i.e., electron density, shown in pink and green) in the process. Credit: Robert K. Szilagyi, Montana State University


Porous carbon is a well-established class of materials with significant potential in wide-ranging applications due to its structural tunability, robust mechanical properties, and unique electrical, chemical and even magnetic properties. Thanks to XSEDE allocations, an international team recently used Comet at the San Diego Supercomputer Center at UC San Diego to study a particular sub-group of porous carbon called zeolite-templated carbon (ZTC) as a gas storage material. 

Using computational modeling of ZTC structural units (referred to as maquettes, or molecular models of surface sites of interest), the research team evaluated methane gas binding energy as a function of chemical composition. Led by Nicholas Stadie, an assistant professor of physical chemistry and materials science at Montana State University (MSU), and Robert K. Szilagyi, an MSU associate chemistry and biochemistry professor, the team's study results were published last month in the Journal of Physical Chemistry A.

"From a comprehensive set of calculations, we clearly identified the preference of methane toward nitrogen-substituted adsorption sites," said Szilagyi. "The theoretically estimated binding energies and the experimentally measured heat of adsorption values are in good agreement, validating the use of computational chemistry as a tool to design new porous carbon materials for methane storage applications – a key bridging technology to reduced-carbon or carbon-neutral chemical fuels for vehicles." 


Why It's Important

"Instead of every researcher needing to set up their own computational cluster, it is important to have a centralized high-performance computational facility that concentrates knowledge and expertise, ensures sustainability and energy utilization, and (with some compromises) serves the standard needs of an average user." – Robert K. Szilagyi, associate chemistry and biochemistry professor at Montana State University

Computer-aided rationalized design of new materials provides an efficient way to develop new or optimize existing energy storage and conversion technologies. Instead of a traditional trial-and-error approach, computer modeling of chemical processes reduces product-to-market times, minimizes chemical waste, and maximizes the use of human resources and brainpower. 

Specifically, methane storage and conversion to energy or chemical feedstocks without greenhouse gas emissions is a long overdue target for petrochemical industries. Computer modeling-based materials design allows for simulation of every step of methane storage or conversion at the atomic scale, which provides a clear understanding of potential inhibiting steps and provides insights into how to mitigate harmful emissions and toxic chemical waste.


Showcasing Nitrogen-Doping of Carbon Materials

While the experimental results utilized in the computational work were obtained in Stadie's laboratory, Stadie and Szilagyi next demonstrated that experimentalists could use simulation results to guide their strategies in synthesizing nitrogen-doped carbon materials for methane storage and delivery. The computational results clearly documented the benefit of focusing on nitrogen-doping of carbon materials when it comes to methane storage. Additionally, Szilagyi said this work serves as a map of the chemical accuracy of simplified but realistic models (maquettes) of porous carbon surfaces for many other problems. 

"We were really surprised by the clear and consistent preference for methane toward nitrogen-substituted porous carbon models, over both the unsubstituted and the boron-substituted models," said Szilagyi. "All of the reliable levels of theory explored showed a 3-5 kJ/mol energetic preference, per methane molecule." 

"This may not seem like a large effect to many, but this was just the size of effect we were expecting," added Stadie. "This effect has very important implications for methane storage at ambient temperature."

"An additional surprise came when we visualized the electron density differences – how playful the π-electron cloud is!" exclaimed Szilagyi (π-electrons are involved in lateral covalent interactions between atoms, as opposed to σ-bonds that lie along the axis between atoms). "Even a methane molecule with relatively low polarizability can cause significant perturbations to the π-electron cloud of the surface, which induces electrostatic interactions between the adsorbent and adsorbate." 


How XSEDE Helped

While the team had access to a modest laboratory workstation cluster and MSU's Hyalite HPC system, accessing Comet allowed for the speedy completion of especially intense computational experiments that would otherwise have been very challenging. The team was able to run potential energy surface mapping calculations where each grid point was calculated at the appropriate level of theory without truncations and simplifications.

"Access to a high-performance computational facility, where users do not need to worry about how to maintain hardware, compile software, and tune executions, allows for experimentalists to harness the power of computational chemistry as an independent and routine research tool in their laboratory." – Nicholas P. Stadie, assistant professor of physical chemistry and materials science at Montana State University.

"The additional computational resources provided by the XSEDE allocation allowed us to complete a more thorough study, including computational control and blank simulations, which are mandatory for experimental work but are often omitted in theoretical studies," said Szilagyi. "A ‘computational control' can be evaluating various levels of theories or alternative structural arrangements of adsorbate molecule on the adsorbent model that do not correspond to what we anticipated based on literature examples, while a ‘computational blank' simulation is where we limit the emergence of a specific interaction whose importance is being probed by constraining molecular geometry and thus indirectly showing its energetic and structural significance. Thus, rather than saving time, this allocation rather created time by tying up all of the loose ends and confirming the results from various initial starting points and with different preconceptions."

"The experimentalist's approach to computational chemistry has great potential to elevate theoretical models to work directly alongside observations from our laboratory," continued Stadie. "The successful combination of theory and experiment elevates the impact of computations in rationalizing experiments and generating experimentally testable hypotheses."


What's Next?

The approach developed in the first two years of this work has direct implications for the ongoing experimental research directions in the Stadie laboratory, from assisting with screening synthetic conditions to down-selecting gas adsorption studies and predicting new technological applications beyond gas storage. The research team has already started evaluating other molecular substitutions, exploring larger molecular models with oxygen-containing functional groups, considering other adsorbates, and scaling computational methods up to larger, periodic systems, just to name a few. 

In addition to undergraduate student Rylan Rowsey (now a graduate student at UC San Diego) and MSU graduate student Erin Taylor, Stadie and Szilagyi collaborated with computational soft-materials scientist Stephan Irle from Oak Ridge National Laboratory. Additional collaborators included Prof. Hirotomo Nishihara from Tohoku University in Japan and scholars Tamas Szabo (Hungary), Eva Scholtzova (Slovakia), Amrita Jain (Poland), and Monica Michalska (Czech Republic).

This research was supported by the U.S. Department of Energy's Office of Energy Efficiency and Renewable Energy (EERE) under the Hydrogen and Fuel Cell Technologies and Vehicle Technologies Offices (DE-EE0008815). Support for the analysis of theoretical results was provided by the U.S. Department of Energy Fossil Energy and Carbon Management

Program, Advanced Coal Processing Program, C4WARD project. Additional support was provided by the American Chemical Society Petroleum Research Fund and the MSU Undergraduate Scholars Program. Computation on Comet was supported via XSEDE, which is supported by National Science Foundation grant ACI-1548562. Additional computations were carried out using the Hyalite HPC System, operated and supported by the University Information Technology Research Cyberinfrastructure at MSU.


About SDSC 

The San Diego Supercomputer Center (SDSC) is a leader and pioneer in high-performance and data-intensive computing, providing cyberinfrastructure resources, services and expertise to the national research community, academia and industry. Located on the UC San Diego campus, SDSC supports hundreds of multidisciplinary programs spanning a wide variety of domains, from astrophysics and earth sciences to disease research and drug discovery. SDSC's newest National Science Foundation-funded supercomputer, Expanse, supports SDSC's theme of "Computing without Boundaries" with a data-centric architecture, public cloud integration and state-of-the art GPUs for incorporating experimental facilities and edge computing.


Media Contact: 

Kimberly Mann Bruch, SDSC Communications,


Related Links:

San Diego Supercomputer Center:

Montana State University:

UC San Diego:

National Science Foundation:


At a Glance:

  • XSEDE allocations were used by an international team of researchers to study a type of carbon called zeolite-templated carbon as a material for gas storage

  • The team used San Diego Supercomputer Center's Comet for creating detailed simulations.

  • Results were published in Journal of Physical Chemistry A.