Removing carbon dioxide (CO2) directly from the air, a process called direct air capture (or DAC), is one of several approaches being developed to help reduce the concentration of this greenhouse gas in the atmosphere. Among the methods being scaled up, one of the more established involves exposing air to a strongly alkaline liquid, typically a solution of potassium hydroxide (KOH), commonly known as lye. The liquid chemically binds the CO2, converting it into dissolved salts called carbonates and bicarbonates. Large facilities using this principle are already operating or under construction, with one plant in Texas that is currently under construction, designed to remove 500,000 tons of CO2 per year.
Despite the maturity of the underlying chemistry, there has been a fundamental limitation in how well researchers can study it. Until now, the process has been something of a black box. Scientists could measure what went into a capture system and what came out, but the detailed chemistry happening inside, specifically in the thin zone where the air and liquid meet, was very difficult to observe directly. This is a meaningful gap, because what happens in that zone determines how efficiently the system works, and how it should be designed, especially for novel DAC liquids. As Jason Pfeilsticker (a Graduate Student in the group of RASEI Fellow Wilson Smith , and lead researcher on this project), explains, "This really is a case of if you want to know about something, just look at, really carefully, and in this case there was some work to do before we could take a detailed look".
Think of it like medicine before medical imaging. For centuries, doctors understood that the body had internal structures and processes, but could only examine them indirectly, through symptoms, pulses, and what came out of the body. The development of X-rays and later MRI scanning did not change human biology, but it transformed what could be understood and acted upon. A diagnosis that once required guesswork could suddenly be made based on the information gained from mapping out the internal structures of the body. This study, just published in ACS Energy Letters , represents a similar shift for CO2 capture: rather than inferring what is happening at the gas-liquid interface from indirect measurements, researchers in the group led by Wilson Smith at the University of Colorado Boulder have built an instrument that lets them watch it directly.
The instrument at the center of this work is a custom-designed laboratory flow cell. This device was designed and built specifically for this purpose and, to the teams' knowledge, is the only one of its kind. "There were so many different variables that we wanted to explore, but in order to design a better process and or screen novel DAC solvents, we needed to have a better picture of what was going on" explains Pfeilsticker, "You can change the solvent, the pressures, the flow, the reactor design, all of which affect the microenvironment and thus the DAC performance ". To get a clearer picture they set out to build a flow cell with built in features that enabled accurate spatial mapping of the kinetics of the reaction, in real time. Designing and building it required solving a series of practical problems. The cell needed to bring CO2 gas into contact with flowing KOH liquid through a porous membrane, closely mimicking the interface in a real capture system. It needed to be optically clear and stable enough to allow laser-based measurements without bubbles, vibrations, or chemical interference disrupting the readings. The flow inside needed to be smooth and predictable, what scientists call laminar flow, so that the measurements could be interpreted meaningfully. Each of these requirements shaped the final design, from the choice of materials to the geometry of the flow channels. However, this oversimplifies the actual process, these lessons were learned as part of an extensive prototyping process.
"We made at least 60 or 70 iterations of this cell during the project" explains Jason. "I was drawn to this project because I really like to make things, and this looked like a challenge that would use a great combination of scientific investigation, detailed design and hands-on building". Jason, who spends much of his free time working on motorcycles, or building electronics and musical instruments, knew he was going to need to iterate on the cell design. Early on the team considered getting design iterations professionally machined. But each of these would cost thousands of dollars to produce, and when you are learning what is important as you are designing, a small tweak here and there can become very expensive. A typical filament-based 3D printer would not be suitable for working with the chemicals involved in DAC. "We identified a resin that was chemically compatible with the base reagents we were using, and we found a cheap resin 3D printer online, that let us do some initial proof-of-principle work, then we upgraded to a better 3D printer for the project, and now we could print iterations for less than a dollar," said Jason. This not only made the process cheaper but sped-up design development as well. The team identified three big challenges as they worked through the designs: good seals, bubbles, and smooth flow of the liquid. The solutions for these came from a number of inspirations, including sealing mechanisms borrowed from drumheads, reactor geometry angles to reduce bubble formation to enable effective laser probing, shaping of the flow inlets and outlets to ensure laminar flow, and flow dampener design.
To explore the reaction and map out the kinetics of the process the team used a technique called confocal Raman spectroscopy to make their measurements. This works by shining a laser at a point in the liquid and reading the light that scatters back; different chemical species produce distinct signatures, making it possible to identify and quantify them. By scanning the laser across the cell in a grid pattern while the process was running, the team built up two-dimensional chemical maps, essentially pictures showing where carbonates and bicarbonates were forming and accumulating across the contact zone, at the scale of fractions of a millimeter, in real time.
What those maps revealed was not what simple intuition would predict. "We saw that the equilibrium reaction is in effect going backwards near the surface" explained Pfeilsticker. When fresh KOH first contacts CO2, the highly reactive hydroxide ions in the liquid rapidly consume the incoming CO2, converting it to carbonate near the membrane. But this rapid reaction locally depletes the hydroxide supply right at the interface. As the liquid flows further through the channel and more CO2 is absorbed, there are fewer hydroxide ions available near the membrane to drive the reaction forward. "Because it is laminar flow, there is no turbulent mixing" said Jason. The result is that a thin layer of bicarbonate, an intermediate chemical species in the conversion process, forms immediately next to the membrane, nestled between the membrane surface and the main hydroxide and carbonate-rich zone further into the liquid. This pattern becomes more pronounced further along the flow channel and represents a direct, spatial record of the chemistry unfolding in real time.
The team also found that operating conditions matter. Higher flow rates altered the shape and extent of the reactive zone, and doubling the concentration of KOH shifted the balance of products and appeared to reduce the hydroxide depletion effect near the membrane, potentially useful information for future system designs.
A key part of this work was the development of a computational model mirroring, and interpreting, what is going on inside the cell. Using the experimental observations to provide a framework to build the theoretical model allowed the team to effectively bound the scope and validate the model, in ways that would have been essentially impossible without the experimental data. The hope is that this model, which has now been validated with experimental data, in conjunction with flow cell maps can be used by future researchers as an initial screening tool in designing new DAC systems.
This work has the potential for significant impact. DAC Facilities using alkaline liquids are being built at the industrial scale. Researchers are actively developing new and improved capture liquids to make the process more efficient, cheaper, and use less energy. With a cell design that enables accurate mapping, and a computational model that enables faster screening, the process of optimizing the carbon capture reactions can be accelerated. On an industrial scale even small improvements in reaction efficiency and cost can have huge savings on the system scale. Current approaches just look at the input and corresponding output of the cell, like judging a medical treatment by whether the patient recovered, without being able to examine what really happened inside the body.
This research describes a detailed, data-driven approach to answering the questions about what is really happening at the reactive center of DAC: how does a given liquid behave, what is happening at the interface where the chemistry is happening, how does varying the conditions impact the reaction? The combination of the experimental and theoretical tools disclosed by this work provides insight into how these processes work, and the key variables that can be used to optimize it.
The application of these tools can potentially extend beyond DAC. Wherever chemistry and transport interact at an interface, such as electrochemical systems that convert CO2 into fuels or commodity chemicals, or in the separation of critical minerals. The design of this device was around one specific challenge, but has the potential for broad utility.
The transition from black box to observable system does not, by itself, solve the engineering challenges ahead. Models still need refinement, and scaling to industrial practice requires substantial research. But the ability to directly observe what is happening is a critical step in that process. What was previously assumed can now be tested. The reaction black box now has a window, that enables researchers to gain valuable insights into the inner workings of this critical process.
