Back to top
Ethylene is the most common carbon-based commodity chemical in the world and the key component in numerous household and construction products—from water bottles, carpeting, and toys to pipes, insulation, and much more. Each year, 300 million tons of ethylene are produced from fossil fuels, most commonly through a process called steam cracking. Steam cracking relies on hydrocarbons such as ethane, propane, butane, naphtha, or gas oil as the feedstock, along with abundant energy to combine the feedstock with steam and expose the mixture to the extremely high temperatures needed to break down the hydrocarbons into ethylene and other valuable chemicals. One of the byproducts of this process is carbon dioxide (CO2), with nearly 4 tons produced for every ton of ethylene.
Livermore researchers and partners are seeking alternative pathways to produce commodity chemicals—one that, as opposed to generating CO2 as a byproduct, uses CO2 as the feedstock, operating at lower temperatures and requiring much less energy. Chief scientist of Lawrence Livermore’s Energy Program Roger Aines explains, “Carbon dioxide is the lowest energy state of carbon. Therefore, we must put energy back into it to produce carbon-based elements that we need.”
To achieve this, researchers from Lawrence Livermore National Laboratory, TotalEnergies, and the SUNCAT Center for Interface Science and Catalysis—a partnership between Stanford University and SLAC National Accelerator Laboratory—have entered into a cooperative research and development agreement (CRADA) to design next-generation electrochemical reactors for industrial CO2 conversion using computationally driven algorithms.
An electrochemical reactor can transform CO2 into other substances at room temperature. In an electrochemical reactor, a purified stream of CO2 gas interacts with a copper-coated surface. The gas binds to the copper, which acts as a catalyst, uniting the chemical ingredients in a way that encourages them to react. Then electrical energy sent through the reactor causes the CO2 to break down into carbon monoxide (CO). CO molecules zooming around the surface find and bind to other CO molecules before reacting with water to make ethylene, which exits the reactor as a gas. “Achieving results at room temperature is appealing,” says Livermore staff scientist Victoria Ehlinger, “but the reaction still requires a great deal of energy. We’re trying to understand the fundamental physics of the process to lower the energy needs as much as possible.”
A Strong Alliance
The Lawrence Livermore−Stanford−TotalEnergies CRADA officially launched in 2019, building on decades of collaboration. Livermore and TotalEnergies first partnered to examine underground coal gasification, an alternative to coal mining, in the 1980s. In the late 2010s, TotalEnergies sought to make products with lower CO2 emissions and reached out to Livermore for assistance.
The latest CRADA brings together the unique skill sets of a national laboratory, university, and industry partner, enabling a holistic approach to the research process that determines what is experimentally and industrially relevant. For example, Stanford works to conduct the fundamental science research, TotalEnergies provides insight from a user perspective into the technology’s practical applications, and Lawrence Livermore bridges the gap between science and practical applications, lending its multiscale modeling capabilities and additive manufacturing knowledge to develop and test the technology.
“While the team at Lawrence Livermore focuses on the development of multiscale and multiphysics models, our colleagues at Stanford primarily focus on experiments for CO2 conversion using different reactor configurations and materials,” explains project co-lead Tiras Lin. The Livermore modeling aims to explain the results observed in the Stanford experiments, and the experiments inform the phenomena that should be included in a model. These two steps are frequently iterated between the teams. TotalEnergies researchers advise on both tasks, translating how the lessons learned could be applied to scaled-up industrial reactor design.
Beyond innovating technologies, CRADAs help develop a strong workforce pipeline as well as mentorship opportunities for graduate students and postdoctoral researchers. This particular CRADA connected the Laboratory with experimentalist Christopher Hahn, who originally worked on the Stanford side before joining Livermore as a staff scientist. Says Hahn, “Cooperative agreements like this one allow us to address—in real time—potential research and development challenges that may prevent a technology from succeeding in the commercial world. This interaction helps students and newer employees understand what it takes to create a useful product.”
Model, Experiment, Repeat
Before work to build an industrial-scale electrochemical reactor commenced, the team used multiscale modeling to define the relationship between the individual process steps at the atomistic scale and in the phenomena occurring at the device scale—providing both a zoomed in (more detail) and a zoomed out (less detail) view of the reactions taking place. “Modeling helps us see the bigger picture without having to physically perform the experiments every time,” says project co-lead Sarah Baker. “Modeling is a cyclical process. Our experts construct detailed models, which are then validated by our experimentalists. The experimentalists provide feedback to the modelers, the model is refined, and the cycle continues.”
While modeling enables the researchers better understand how the device will function once scaled up, 3D-printed prototypes of the reactor and its components—made initially at the Livermore Valley Open Campus’ Advanced Manufacturing Laboratory (AML)—help them learn more about the design and overall framework. AML houses some of the Laboratory’s most sophisticated manufacturing equipment, allowing external partners in government, industry, and academia to utilize these resources to accelerate market-driven innovation and reduce production costs and time.
Additively manufactured reactors represent a significant step toward derisking the industrialization and optimization of this game-changing technology while developing design metrics that can promote commercial feasibility. The 3D-printed prototypes enable the team to quickly bring their designs to life and test them by running small-scale experiments and having researchers at Stanford measure the resulting chemical reactions.
The team’s modeling and experimental results are used to build an end-to-end model that simulates CO2 mass transfer across an industrialized reactor from start to finish. With this model, the researchers can optimize the reactor by simulating how different “ingredients” affect the conversion process. Lin explains, “Understanding how to develop a model for CO2 conversion that accurately captures the important phenomena across the reaction and is suitable for optimization is a key step. Using the model, we can then leverage design optimization to tell us which components of the reactor we need to tweak for improved performance.”
Optimizing for Scale-Up
Overall, design optimization is an automated, behind-the-scenes process that eliminates the tedious trial-and-error nature of human design. To computationally design the reactor, a set of governing equations are used to guide the model, helping it determine how close or how far the resulting simulations are from the desired design criteria. Using this information, the computer can adjust the design accordingly, rerun the simulation, and recalculate the design as many times as needed until the criteria are met.
Optimization comes with tradeoffs, however. Staff scientist Thomas Roy explains, “If working with two different metrics, we cannot always optimize both. Instead, we can see what happens when we switch the priority from one to the other or gradually shift the priority to generate a spectrum of results.” For example, choosing whether to optimize the CO2 conversion rate or the energy efficiency of the device may yield different results. Roy focuses on algorithm development as the team aims to improve parameters and develop better code, which the team uses to build the simulation.
Ehlinger adds, “We are working on methods to scale up computation so we can achieve both complexity of geometry and complexity of physics. We can solve the physics in our simplified, one-dimensional model. But, geometrically, we want to go to a 3D model, which increases the complexity of the physics equations. Our simulation must be in 3D given the more complex shapes possible with the Laboratory’s advanced manufacturing capabilities.” As models grow in size and geometric complexity, researchers solve for selected physics parameters such as electrochemistry, ion and gas transport, and side reactions to achieve convergence in the simulation.
Other factors that can be optimized include how the reactor will respond to catalysts other than copper. “Catalysts dictate the products,” says Ehlinger. While hundreds of possible catalysts exist, copper is ideal for ethylene given its industrial efficiency and low cost. The end-to-end model can also simulate the overall design of the catalyst, from the thickness of the copper layer to the texture of its surface. For example, 3D microstructures can be added to the catalyst’s surface and the model can run a series of simulations to determine the best shape and size for optimal performance. Experimentalist Jeremy Feaster says, “The reactor is like a car, but the catalyst is like the engine. If you put a good engine in a less-than-ideal car, then the engine cannot perform as it should.” The team’s modeling approach makes it possible to find the best reactor design that allows the catalyst to shine.
Once the team has optimized every detail within the model, they can use this knowledge to build the industrial-use reactor. To build a scaled-up version of the modeled reactor, Lawrence Livermore and TotalEnergies have begun a follow-on CRADA with industry partner Siemens Corporation, a leader in manufacturing with experience making similar devices for hydrogen production.
Ehlinger reiterates that lowering energy requirements remains the key to producing commodity chemicals from CO2 at less-than-extreme temperatures. “Our technoeconomic studies always place electricity as the top cost,” she says. Adds Feaster, “A huge part of this project has been asking ourselves, ‘How do we build a world that we will be proud of?’ Working alongside partners who see the value of this technology and have a shared vision for creating a sustainable world has been one of the most encouraging aspects of this work.”
—Shelby Conn
For further information contact Sarah Baker (925) 422-3811 (baker74 [at] llnl.gov (baker74[at]llnl[dot]gov)) or Tiras Lin (925) 424-4396 (lin46 [at] llnl.gov (lin46[at]llnl[dot]gov)).
