O​​​​​​​​​​​​​​​​​​​ur city CDR article from April (“Taking Out the Carbon Trash“) proposed an unconventional idea: Mine gigatons of solid carbon directly from the air via a global array of small, city-based capture units. We covered the deployment details in our April essay, but only hinted at the science.

In this essay, we’ll look more closely at how an emerging collection of scientific advances, each progressing along its own path, might eventually be combined to create an entirely new (and arguably, much needed) approach to carbon removal.

First, the need

At the rate the CDR industry is maturing—with many technologies still struggling to achieve commercial scale, and time running out to make climate-relevant reductions in atmospheric CO2—policy makers will almost certainly end up taking an all-of-the-above approach to carbon removal. No single technology will shoulder the entire burden.

Into this mix, solid-state carbon harvesting is worth exploring not only because its capture potential, speed, and efficiency are highly compelling, but also because it fundamentally reimagines downstream conversations on issues like removal verification and long-term monitoring, power consumption, permitting, and pipelines. Granted, with solid carbon harvesting we end up with a whole new set of concerns like how to move billions of tons of solid carbon and what to do with it. But even so, the idea is worth exploring because the enormous removal potential and relative policy simplicity are both compelling.

How solid carbon harvesting works

The second generation CDR panel we proposed using in our carbon trash article isn’t a sponge that needs to be coaxed into releasing trapped CO2 through heat or moisture. Rather, the most efficient panel and the one with the lowest operational costs is one that converts CO2 directly into a stable solid product, eliminating the need for regeneration cycles, compression systems, and long-term storage infrastructure.

While you can’t buy a panel like this yet, the individual parts already exist and are being fine-tuned. The basic architecture consists of two layers: the molecular sponge, and the electrochemical wedge.

Layer 1: The molecular sponge

The exterior skin of the collector faces the open air. Because carbon dioxide is highly dilute in our atmosphere—making up only about 420 parts per million—this layer requires an astronomical surface area to trap passing molecules. Researchers utilize materials called metal-organic frameworks (MOFs) and ultra-thin graphene membranes that act as atomic-scale molecular sieves. The interior of these materials is so incredibly porous that a single gram can have an internal surface area of an entire football field. As wind drifts through these channels (wind that exists abundantly in urban areas like freeway and skyscraper corridors), the internal chemical architecture selectively stalls CO2 molecules, packing them tightly together while letting the rest of the air flow through freely.

Layer 2: The electrochemical wedge

Once the CO2 is concentrated inside the molecular sponge, it needs to be broken down. Because carbon dioxide is incredibly comfortable in its chemical state, forcing it to let go of its oxygen atoms requires a jolt of electricity. There are several ways to provide this. The trick is to find a way that’s free and effective—as in not requiring a constant input of energy. There are at least four approaches that could work in this layer. The last two are the most likely candidates for our city-based CDR model because they don’t require heat or electricity to break apart CO2:

  • Gallium: In this approach, we push the concentrated CO2 up against a liquid-metal alloy substrate of gallium and indium. When a mild electrical current passes through this material at room temperature, it breaks the carbon-oxygen bonds apart. Oxygen gets released back into the air, while carbon falls out in various solid forms. There are at least two problems with this approach, however (at least as far as our large-scale deployment model is concerned): (1) gallium and indium aren’t particularly abundant metals, meaning that CDR panels built using them might be prohibitively expensive, and (2) the electrical input required to break apart CO2 will limit the versatility of these panels and also raise the operational cost.
  • Molten salts: Several research groups and startups are exploring high-temperature electrochemical pathways that dissolve carbon dioxide gas in a liquefied bath of lithium and sodium carbonate salts. Companies like SkyNano Technologies, for example, have successfully used this technology to split carbon dioxide (from concentrated streams), releasing pure oxygen and forcing the remaining carbon to assemble into high-value structures like carbon nanotubes and graphite. One attraction of this approach is that the immediate byproducts are valuable industrial materials. The problem with trying to scale this approach, though, is that—as with the liquid gallium approach—constant electrical input is required, meaning high operational costs.
  • Light-driven covalent organic frameworks (COFs): In this approach, our second layer isn’t liquid gallium or salts requiring a constant inflow of heat and electricity, but a robust manufactured sheet of highly ordered polymers made of common elements like carbon, hydrogen, nitrogen and oxygen (called covalent organic frameworks, or COFs). When sunlight hits one of these sheets, it excites electrons inside the pores where the CO2 is trapped, separating the carbon and oxygen bonds without any external power. Here again, this approach isn’t bulletproof: requiring sunlight will be a problem if these panels are installed in tunnels, for example. But then again, COFs don’t need electricity and liquefied minerals. Rather, organic photochemistry allows COFs to grow solid carbon (although not the perfectly regular nanotube variety—more like carbon “cotton”) that easily sloughs off into collection hoppers. Today’s photocatalytic frameworks typically yield products like carbon monoxide or syngas, but if we turn this same technology loose on CDR, future generations might be engineered to produce carbon.
  • Wind-driven COFs: This architecture replaces solar energy with wind energy, making it ideal for dark urban canyons, subways, or highway tunnels. The structural panel incorporates a flexible composite backing made of piezoelectric polymers—materials that generate a micro-electrical current whenever they bend or vibrate. As wind causes the panel’s exterior skins to flutter, this movement gets converted into voltage pulses, which provides the electrical nudge needed to split the CO2 and drop carbon cotton into the collector. While promising, the piezo-catalytic approach is still highly speculative since the high-frequency mechanical energy used to drive these reactions in a laboratory settings (like ultrasonic vibrations) are orders of magnitude more intense than the gentle, low-frequency vibrations that would be generated by wind. More research is needed.

Collection and harvest

Once the carbon and oxygen have been separated, the next step is collection. An important prerequisite is to make sure the carbon being separated from CO2 molecules doesn’t coat the collection panel surface and thereby slow or stop the collection process. For this reason, all the Layer 2 approaches mentioned above are continuously self-cleaning. In the gallium approach, for example, carbon flakes don’t stick to gallium. And in the COF models, the continuous thermal expansion and contraction of day-night cycles, or the slight flutter caused by wind, subjects the panels to mechanical stresses that force the accumulating carbon layer to crack, curl, and flake away like dry bark, continuously exposing fresh reaction sites underneath.

In all these solutions, shed carbon simply drops into a collection trough at the base of the panel. The material can be collected and transported using standard vacuum systems and service vehicles, and stored at regional concrete plants, asphalt bays, farms, or other existing locations. There is no post-collection step that requires additional heat or electricity, and no transport and handling requirements that require pipelines, long-term monitoring and verification, hazardous waste concerns, and so on.

Technology readiness levels (TRLs)

A complete, self-sustaining, self-cleaning architectural panel operating reliably in an unmanaged outdoor environment isn’t nearly ready for prime time. Measured in terms of technology readiness level (TRL), these panels currently rate at around a 2 or 3, where 1 indicates the beginnings of basic research, 5 indicates successful lab and real-world validation, and 9 means ready for commercialization. However, the foundational building blocks this system relies on have already achieved validation in laboratory settings and have TRLs of 4 or 5. For example, the baseline chemistry of room-temperature solid carbon conversion was proven at RMIT University, where researchers successfully scaled benchtop units to a 2.5-liter dimension capable of continuously splitting CO2 gas for a month straight without any catalytic degradation. More recently, researchers scaled up the chemistry via a bubble-column reaction framework that accelerates bond-splitting to a fraction of a second. This work has been funded by grants from the Australian Research Council (ARC) alongside a $2.6 million commercialization agreement with ABR, an Australian environmental tech corporation.

SkyNano’s molten salt approach mentioned earlier is maturing as well, with their core intellectual property and heavy assets recently acquired by American Energy Technologies Company (AETC). AETC has integrated these assets into its commercial industrial graphite production facilities, marking the first true hand-off of solid-state capture technology from grant-funded startups to established materials manufacturing corporations.

Removal potential

In sterile, highly optimized laboratory settings, advanced nanomembranes and catalysts can capture roughly 1.5% of passing air, yielding an ideal removal potential of 500-1000 kilograms of CO2 per square meter annually (see footnote 3). In the messy reality of unmanaged outdoor deployments, this ceiling is much lower. Due to unpredictable wind fluctuations, microclimate stagnation, solar shadow-blocking, and dust or pollutant fouling on the active layers, actual operational efficiency will likely drop closer to the 0.1% to 0.3% range. This brings the expected field removal potential down to roughly 50-150 kilograms of CO2 per square meter, per year.

Even adjusting for these real-world conditions, however, solid carbon conversion panels might end up being incredibly effective, and could bring a 10 gigaton annual removal target easily within range. Specifically, removing 10 gigatons of CO2 per year—assuming real world efficiency— would require deploying around 100 million square meters (roughly 38 square miles) of harvesting surfaces, divided across whichever environments offer the simplest construction and highest return on investment. This removal target could be met, for example, by outfitting a tiny percentage of the world’s warehouse walls, commercial rooftops, and highway noise barriers with collection panels, or deploying these panels across a few thousand modestly-sized rural carbon farms.

Still, it’s important to emphasize again that this promised effectiveness is currently pie in the sky—possibly on the near-term horizon but still well behind existing approaches like DAC, enhanced weathering, or biomass. To catch up in time to make a climate-relevant contribution by mid-century, heavy funding for fundamental materials research will be needed over the next five years, followed by test deployments around 2035, and scaled deployment by 2050—a steep hill to climb.

Barking up the wrong tree?

Our search for better CDR solutions shouldn’t stop with carbon panels. Other solutions might be waiting as well. Take lasers, for example. Researchers at the University of California, Davis have demonstrated that by zapping carbon dioxide gas with high-energy vacuum ultraviolet (VUV) lasers, CO2 molecules can be temporarily forced into a triangular shape. When the molecule snaps out of this shape, it shears into a pure oxygen molecule and a lone, solid carbon atom, which falls to the ground. Like our carbon collection panels, this idea is still mostly sci-fi but only in terms of finished product—the science and technology exist in pieces and are just waiting to be assembled together into a solution.

So would this solution mean a future where CDR lighthouses shoot giant laser beams into the sky? Not quite. Put into action into one single area, this idea might look more like Christmas tree lights strung across vacant public land or farmland (although this need not be the case; we can also think in terms of incorporating this tech into the home-based carbon collection cans discussed in the city CDR article). This is because VUV light at the wavelength needed to split CO2 can’t travel very far because the air absorbs and scatters 90% of its power every few centimeters. Therefore, strings of tiny LED lasers—maybe in the neighborhood of 300 million for a two square mile carbon farm (which may seem like a lot but a typical television has millions of these; see footnote 1)—might be what a deployed solution looks like, all prewired into rolls and then unfurled about five meters above ground level. This height allows the system’s radiant heat—a few tenths of a degree C—to safely diffuse so it won’t bake the dirt below. At this height, to prevent strong winds from scattering the lightweight carbon dust across the landscape, the overhead wires can be backed by a static collection mesh that catches the soot as it forms. Every few hours, the lasers turn off and the mesh drops the accumulated carbon down as heavy, dust-free pellets into the soil. Electric drone tractors then move underneath this array to continuously plow and churn freshly settled carbon residue into the earth.

How effective would this process be? A two-square-mile field with a gentle breeze will have roughly 32.4 million metric tons of CO2 drift through its lower airspace every year. Assuming a cumulative 32.2% multi-pass efficiency (i.e., in the lab, lasers have a single pass efficiency of 5%, but the air moving through this grid will be zapped multiple times before it exits), this laser grid can theoretically capture and deposit around 2.8 million tons of carbon soot annually (see footnote 2)—roughly five times the amount of carbon per square meter as even our highly effective passive panels. After a period, this array and operation can be moved to a new location and improve the soil in this area—a service that farmers everywhere would gladly pay if the costs pencil out to be lower than fertilizers.

Like our panels, though, this idea also isn’t ready for prime time, but is just an example of one of other CDR approaches we might want to evaluate. And all of these approaches will likely have tradeoffs, which need to be weighed against carbon removal effectiveness. In this case, the biggest downside of the laser approach is power—likely requiring a dedicated small modular nuclear reactor (SMR) capable of generating roughly 380 megawatts (unless more efficient lasers are invented). While this is well within the capability range of SMRs being discussed to power datacenters, the idea of building a thousand new nuclear plants to capture CDR somewhat boggles the mind. The approach might be intriguing as a one-off demonstration project, but scaling it will require more thinking.

Questions

Some of the questions that might come up as we evaluate various approaches for collecting solid carbon include:

  • What would the world possibly do with 10 gigatons of solid carbon per year, short of building giant piles? The first part of this answer is that scaling CDR to 10 Gt/yr of CO2 doesn’t result in 10 Gt/yr of solid carbon. Rather, by stripping away the heavy oxygen atoms, the physical harvest weight collapses down to just 0.27 tons of solid carbon per ton of CO2 captured (since carbon dioxide gas is composed of roughly 27% carbon and 73% oxygen by weight). So removing 10 Gt/yr of CO2 will generate only 2.7 Gt/yr of solid carbon—still a lot, but more akin to a “smaller” commodity market for which ample downstream use will surely develop, from stronger concrete to pavement to plastics, to soil additives, and beyond. Truly extra supplies could even be blown back into abandoned mines and wells.
  • What happens if this method becomes so successful that we start drawing down CO2 to dangerously low levels? There is zero likelihood of this happening anytime soon, but it’s a genuine issue if mining the skies is free and profitable and if the price of elemental carbon stays high. We’ll have many decades of fantastically successful extraction before we get to this decision point, so there will be ample time to work out the kinks in the supply and demand systems, systems for monitoring carbon harvesting, and for thinking about how to rebalance CO2 levels (fossil fuels anyone?) should this need ever come to pass.
  • Won’t the extra oxygen be a hazard? The Earth’s atmosphere contains roughly 1.2 million gigatons of pure oxygen, making up about 21% of the air. When we capture 10 gigatons of CO2 and split it, we release 7.27 gigatons of pure oxygen back into the air annually. Over a 50-year deployment timeline, this means releasing roughly 363 gigatons of new oxygen, which represents a roughly 0.0003% increase in O2 levels (363 Gt of newly-released O2 divided by 1,200,000 Gt of existing O2), which will have no discernible impact. To the contrary, it’s repaying a centuries-old environmental deficit. Every time a carbon atom from a fossil fuel is burned, it steals two oxygen atoms from the open air to form CO2. Because humanity has burned over 1.5 trillion tons of fossil fuels, atmospheric oxygen concentrations have been slowly declining for more than 150 years. For every 1.0 ppm increase in atmospheric CO2, the sky loses roughly 2.15 ppm of pure O2. Solid-state carbon removal reverses this dynamic, functioning as a synthetic plant that returns oxygen to the global pool.
  • Why cities?: This is just a starting point for discussion given that cities already have centuries of experience with financing and managing waste infrastructure. But if we think of carbon not as a waste product but a harvestable crop or a value added for cropland, then one might make an equally strong case for carbon farms instead of carbon garbage trucks. Using passive arrays arranged at varying heights and angles, a few thousand concentrated regional carbon farms around the world could easily reach multi-gigaton scale without massive land footprints, and without using arable land, power, or water.

Conclusion: Why do we need this approach?

It’s still early, but a future where carbon can be extracted directly from the air and converted into a stable, usable material is scientifically conceivable. If the underlying research can be advanced, the engineering challenges solved, and the manufacturing processes affordably scaled—all significant ifs—then solid-state carbon harvesting could become an important addition to our current portfolio of carbon dioxide removal approaches.

Granted, many existing forms of CDR are already more mature, better researched, and further along the path toward commercialization. Yet solid-state carbon harvesting offers a distinct combination of potential advantages that make it worth exploring. By converting atmospheric CO2 directly into a stable solid product, it has the potential to simplify measurement, reporting, and verification while reducing the need for complex downstream handling. Just as importantly, it reframes carbon dioxide removal from the management of an unwanted waste product to the harvesting of a potentially valuable industrial material.

Whether this particular architecture ultimately succeeds is, however, only part of the story. The larger lesson is that carbon dioxide removal should not be viewed as a finished engineering field whose fundamental design choices have already been made. Throughout history, transformative technologies have often emerged by combining existing discoveries in unexpected ways rather than waiting for a single revolutionary breakthrough. The materials science described throughout this essay—molecular sieves, advanced catalysts, photocatalytic frameworks, nanostructured surfaces, electrochemical conversion, and self-cleaning materials—is already advancing rapidly, albeit along separate tracks. The question is whether these advances can eventually be integrated into systems that make carbon removal simpler, more efficient, more affordable, and ultimately more sustainable than the CDR approaches available today. Continuing to pursue research along these lines broadens the range of options available to a world that cannot afford to assume today’s technologies represent the limits of tomorrow’s possibilities.

Technical footnotes

  1. Because vacuum ultraviolet (VUV) light at 157 nanometers is quickly absorbed by oxygen in the air, the laser beam cannot shoot long distances. The path length must be short—around 10 centimeters. If you manufacture tightly-spaced rolls of LED lasers and hang these about 15 feet above the ground (high enough so electric tractors can pass underneath), then you can fit roughly 300 million of these lasers over a single 2 square mile tract (roughly 5.18 million square meters). This may seem like an astronomical number, but for comparison sake, a single commercial 4K television screen contains roughly 24 million micro-LEDs. Stamping out 300 million diodes onto flexible, rolling wire might use the exact same (or similar) high-speed automated assembly lines that manufacture consumer electronics today.
  2. Again, a two-square-mile farm is roughly 5.18 million square meters. If you assume a light ground-level breeze, the atmosphere delivers 1,287,200 cubic meters of air per second through this field. At 420 ppm, this volume translates to 1.03 metric tons of CO2 gas passing through the farm’s airspace every single second. Over a year, that’s 32.4 million metric tons of CO2. When you apply a cumulative 32.2% multi-pass efficiency to that volume, and multiply it by the strict atomic weight of carbon (27% of a CO2 molecule), you get roughly 2.8 million tons of solid carbon soot removed per year.
  3. The density of CO2 in ambient air is roughly 0.8 grams per cubic meter (which translates to 420 ppm). A 1-square-meter vertical plane exposed to a continuous 2 meter-per-second breeze processes approximately 63 million cubic meters of air annually (1 sqm x 2 m/s x 31.5 million seconds/yr). Multiplying this flow by the CO2 gas density yields a total passing mass of roughly 50,450 kilograms of CO2 per year. A laboratory capture efficiency of 1.5% yields 500-1000 kg; a real-world field efficiency of 0.2% yields 50-150 kg.