Energy-intensive carbon removal pathways must use low-carbon energy to actually remove carbon from the atmosphere and minimize net removal costs. Certifiable clean energy procurement is not a new issue. Various instruments ranging from renewable energy certificates (RECs) to power purchase agreements (PPAs) have been introduced and used for this purpose in the past, with 24/7 carbon-free energy (CFE) emerging as the gold standard. Although carbon removal companies have multiple procurement options, each comes with trade-offs. CDR companies and other ecosystem actors must be cognizant of these trade-offs and collaborate to adeptly manage them.

CDR and energy use

To compensate for hard-to-abate emissions and reverse—rather than just mitigate—climate change, we need to remove carbon from the atmosphere. This will be done using carbon dioxide removal (CDR), which is a suite of pathways that involve capture of atmospheric carbon dioxide and subsequent storage in a variety of places including soil, plants, minerals, the ocean, commercial products, and geologic formations.

Up to roughly 10 gigatons of annual CDR may be required by 2050 to meet temperature goals set out by the Paris Agreement. Around two gigatons of carbon dioxide are removed from the atmosphere today, mostly through conventional types of removal—such as planting forests—that make use of photosynthesis as an energy source. While valuable, these conventional methods can suffer from a low degree of permanence, low-quality verification, and excessive land use, which can exacerbate other societal issues such as rising food prices.

There are CDR methods such as biochar burial and bio-oil injection that use photosynthesis as a primary energy source while also addressing permanence and verification issues. However, to mitigate land-use issues and create a more diversified and sustainable portfolio of removal methods, pathways using artificial sources of energy are also required. These sources, such as solar panels and nuclear power, are much more efficient at generating energy per unit of land area and can be located on non-arable land.

CDR pathways that make use of artificial energy include direct air capture (DAC), direct ocean capture, CO₂ conversion into certain kinds of long-lived products, and enhanced weathering processes that are dependent on energy-intensive mineral treatment. The energy used by these removal processes must emit less greenhouse gas than the processes ultimately remove from the atmosphere, otherwise they would not actually be carbon negative. Costs per net ton of carbon dioxide removed from the atmosphere will also be higher when process emissions are higher.

For example, the theoretical minimum amount of energy required for direct air capture is around 22 kJ/mol, which corresponds to about 140 kilowatt-hours per gross metric ton (kWh/t) captured from the atmosphere. Current energy requirements for actual DAC systems are closer to 2,000 kWh/t, although this requirement will hopefully fall with continued innovation. Compression to 15 MPa for subsequent pipeline transportation and sequestration can require at least another 100 kWh/t. For the purposes of this estimate, I will assume a relatively more mature DAC system that uses around 1,200 kWh/t.

Idemat 2023, an emissions database, estimates the following emissions factors for the specified energy sources (note that the onshore wind and geothermal factors come from separate references):

• Coal: 1.077 kilograms of CO₂ equivalent per kWh (kg CO₂-eq/kWh)

• Natural gas: 0.544 kg CO₂-eq/kWh

• Oil/petroleum (for electrical generation): 0.873 kg CO₂-eq/kWh

• Solar PV: 0.093 kg CO₂-eq/kWh (some sources estimate a median value closer to 0.05 kg CO₂-eq/kWh)

• Onshore wind: 0.011 kg CO₂-eq/kWh

• Offshore wind: 0.008 kg CO₂-eq/kWh

• Nuclear: 0.008 kg CO₂-eq/kWh

• Hydropower: 0.006 kg CO₂-eq/kWh

• Geothermal: 0.122 kg CO₂-eq/kWh

At the time of writing, the Midcontinent Independent System Operator (MISO) balancing authority in the United States that covers a large portion of the Midwest is using about 45% natural gas, 32% coal, 11% nuclear, and 2% each of solar, wind, and hydro, with the remainder coming from imports from other areas. When paired with the above emissions factors and rounded up slightly to account for imports, this mix results in an average electricity emissions factor of roughly 0.6 kg CO₂-eq/kWh.

If we were to use this average emissions factor in an emissions assessment of a DAC system located in MISO, the system would have energy emissions of around 1,200 kWh/t * 0.6 kg CO₂-eq/kWh = 720 kg CO₂-eq/t. This means that for each ton (1,000 kg) of CO₂ the system removes from the atmosphere, it would emit at least 720 kg. If it costs $500 to capture and sequester the gross ton, then the net removal cost, which represents the total cost to remove one ton from the atmosphere when considering process emissions, would be around $1,800/t, as around 3.6 gross tons would need to be captured and stored in this case to remove one net ton. This cost level would be a serious hurdle for this system. With an even dirtier mix or the use of more energy, the system would emit more than it removes.

For real systems, emissions margins are even tighter due to emissions from other system components, such as sorbents and plant construction. The negative effects on process emissions and net removal costs do not even begin to factor in the negative societal consequences of continued and increased fossil fuel extraction and combustion if the processes use fossil fuels as their source of energy.

This problem could also apply if a new CDR project were to displace existing buyers of clean energy. While this would allow the CDR project to claim the use of the clean energy, it might cause the previous buyers to begin consuming dirtier energy. In this case, the CDR project would in reality be responsible for the deployment of dirtier sources. There is a risk of this scenario in grid regions that have fully developed low-carbon resources, such as hydropower plants, but where new demand is met by spinning up resources such as gas peaker plants.

CDR projects must ensure clean power they consume is additional, similar to additionality claims for carbon offset projects in general. Additionality attempts to account for whether the renewable energy or carbon project would have proceeded without the customer’s purchase; if it would have, the project is not additional.

Additionality of low-carbon energy is especially important for carbon removal projects to minimize the opportunity cost of using these resources for CDR instead of decarbonizing other parts of society. Using CDR to hasten the deployment of more clean energy can help contribute to further economies of scale and learning-by-deployment and even support clean energy development for regions that might not have had it otherwise.

Ultimately, energy-intensive CDR projects need to use newly built, low-carbon energy. Not using clean energy to power these systems would be akin to simply venting a huge portion or even all the captured CO₂. Fortunately, procuring low-carbon power is not an entirely new challenge.

Types of clean energy procurement

In any given region of the electrical grid, electricity generally comes from many power plants using different sources of energy. It can also be imported from or exported to other regions. This mixed nature of the supply poses an issue for incentivizing clean energy generation, as at any time this generation is technically spread across all users in a grid region but some are more willing to pay for it than others. Additionally, many states in the United States implemented renewable portfolio standards (RPS) to increase the generation of renewable electricity by setting renewable generation targets for electricity suppliers, but grid regions often transcend state boundaries.

To allow for better market signals and more liquidity, renewable energy certificates (RECs), called guarantees of origin (GOs) in Europe or energy attribute certificates (EACs) more generally, were introduced. One REC is created for each renewable megawatt-hour generated, after which it can be traded across companies and regions and finally retired by the entity that has used the credit. Various registries track REC creation and trading.

There is also a voluntary market for RECs, where energy customers purchase and retire RECs in addition to paying their regular utility bills. Many corporate “100% renewable energy” claims are based on a company retiring a number of RECs that is commensurate with its non-renewable electricity consumption. Such retirements are the basis of market-based scope 2 emissions claims, where companies calculate their emissions based on their purchase of RECs rather than the actual electricity mix at their operating locations.

Anyone can buy RECs just like they can buy carbon offsets. Similar to carbon offsets, additionality is a key concern for voluntary purchases. Many clean energy projects now proceed on economic grounds, and they would occur regardless of whether others bought the projects’ RECs or not. Already-existing renewable projects also generate RECs, and many may continue to operate regardless of how many companies buy their RECs. Purchases of RECs from such projects have little if any additional impact on the climate other than creating a weak market signal in favor of renewables. This desire to move away from less additional RECs toward ones bundled with new renewable energy projects has driven significant interest in another instrument known as a power purchase agreement (PPA).

PPAs originally enabled power project developers to lock in selling prices for generated energy to utility companies. In the past couple of decades, they have emerged as a powerful tool to enable higher-quality and more additional renewable energy procurement. Google was one of the leaders in using PPAs for clean energy procurement, and many other companies followed suit.

There are several types of PPAs, with each involving a combination of a newly built clean power project and the retirement of generated RECs by the buyer. Onsite PPAs involve the development and ownership of a clean energy project by a developer on a buyer’s property with a predetermined rate for the energy. Offsite physical PPAs involve a project developer building, owning, and operating a clean energy project at a site separate from the buyer’s site that is still in the same wholesale electricity market. The generated energy is delivered to the buyer via a power marketer, and there is a predetermined rate for the energy. This structure is only legal in states with competitive electricity markets.

Virtual PPAs (VPPAs) are financial arrangements where a developer builds, owns, and operates a new clean power project and sells the generated electricity into a wholesale power market that may or may not be the one where the buyer is. If the price the developer gets for the electricity is lower than the agreed-upon price in the VPPA, the VPPA buyer makes up the difference, and if the price the developer gets is higher, then the buyer gets the difference.

Figure 1. Flows in a VPPA from LevelTen Energy

PPAs that offer bundled RECs are superior to unbundled REC purchases, as PPAs are generally more additional. In each case, new clean power is being built. In the ideal scenario, these new builds would be displacing fossil-based electricity and moving us closer to a world powered with 100% low-carbon power. Occasionally, buyers will choose VPPAs in locations where they will offset the dirtiest generation, as Boston University did with their wind VPPA in South Dakota. Some call this concept “emissionality,” and benefits may be able to be accounted for as avoided emissions.

PPAs are not the end of the story. A new concept called 24/7 carbon-free energy (CFE) has emerged to address identified downsides with RECs and PPAs. The “carbon-free” label is a bit misleading, given that all sources of energy at least involve some level of supply chain emissions. Regardless, 24/7 CFE is generally understood to include renewable sources, and some also view it as inclusive of other low-carbon electricity sources such as nuclear energy and plants equipped with carbon capture and storage.

Google was again a leader in identifying the need for 24/7 CFE and crafting a strategy to pursue it. Even for energy-buying behemoths like Google, VPPAs are especially popular. This is likely due to (a) additionality issues with RECs; (b) a lack of space for onsite PPAs; and (c) regulatory restrictions and other complexities associated with offsite physical PPAs. However, the company realized that despite the benefits of VPPAs, its facilities were still actually drawing dirty power from the grid that may not have been perfectly offset by VPPAs. Even with a physical PPA that involves direct delivery of generated renewable electricity to Google’s sites, there would still likely be an imbalance of supply and demand due to intermittency. 24/7 CFE creates an even stronger claim of direct and complete use of clean electricity and supports the development of an electrical grid that is 100% clean all the time.

Figure 2. Hours of direct carbon-free electricity usage at a Google data center throughout the year, emphasizing heavy solar generation in the middle of the day

This procurement strategy involves having a supply of clean energy that is additional and temporally and geographically matched with energy demand from the buyer’s sites. It can be done through PPAs, but they must be in the same grid region as the energy buyer and temporally aligned with the buyer’s energy demands. Specifically, temporal alignment means that the generation of low-carbon power matches with when buyers are drawing power, generally on an hourly basis.

24/7 CFE helps avoid a potential mismatch between the increase of dirty power in a new energy buyer’s grid region and the offsetting of less dirty power in the grid region with newly contracted low-carbon power. This can occur as adding extra demand to the grid in one location can lead to more production of dirtier energy, such as that from gas- or petroleum-powered peaker plants, while new clean power added in another area may not offset electricity that is as dirty. This is particularly the case given that areas with favorable geographical and political conditions for renewables will likely already have other renewable installations and may even be curtailing excess renewable power when it exceeds demand.

If every energy buyer practiced 24/7 CFE, the entire grid would be low carbon. 24/7 CFE also helps create valuable market signals to developers incentivizing them to supply more clean power during periods of the day that currently do not have sufficient clean power generation. However, 24/7 CFE faces a number of challenges. Intermittency and the lack of energy storage is probably the most difficult aspect at this time. The infrastructure for better temporal tracking of generated renewable energy is still being built out by organizations like EnergyTag, although it has already been implemented on a small scale by M-RETS. From a corporate perspective, there are currently few direct rewards to putting in the extra work and resources to use 24/7 CFE instead of just a PPA other than a reputational boost and long-term strategic positioning. Regulators are only beginning to catch up. Finally, many stakeholders lack awareness about the concept, which is difficult to fully understand.

Each strategy, from RECs to PPAs to 24/7 CFE, builds on the prior and gets closer to a more additional, impactful, scalable, and defensible model. 24/7 CFE is likely the ideal case, other than onsite/behind-the-meter clean generation, for energy-intensive clean tech pathways, ranging from water electrolysis to direct air capture. However, these more ideal options are more difficult to implement. CDR companies requiring large amounts of artificial energy need to understand and carefully navigate the options available to them and begin strategizing for future procurement.

Options for CDR companies

The current primary business model of carbon removal companies is to remove carbon dioxide from the atmosphere and sell associated offsets to emitting entities. The use of these offsets allows the emitters to “cancel out” a portion of their greenhouse emissions. Given that there is usually no delivery of a tangible product, a high degree of trust and verification is required. The verifiable use of clean energy for energy-intensive CDR pathways is paramount to the validity, and therefore development, of the industry.

CDR companies currently have multiple options for clean energy procurement. Below, I outline the options as well as trade-offs associated with each.

Behind-the-meter clean energy

The best option is to use behind-the-meter, low-carbon energy and retire any generated RECs (if relevant). In this scenario, a new clean power project, such as a solar or wind farm or a small modular nuclear reactor, would be built alongside the CDR facility. This could be financed as part of the overall project, although it could also take the form of an onsite PPA.

This power could be from solar and storage, wind and storage, a mix of solar and wind, onsite natural gas with integrated carbon capture and sequestration (only when upstream methane leaks are very low), a small modular nuclear reactor, sustainable biomass combustion, or some other low-carbon power source. The challenges with this approach include matching energy demand and supply, increased project complexity, financing issues for smaller projects and less mature companies, and potentially increased land-use requirements.

Deployment in a (marginally) clean grid

Another worthwhile and defensible option is to deploy in a grid region that either runs completely on clean energy or that supplies clean energy for increased or “marginal” demand. Iceland, for instance, has a grid that is powered by around 70% hydropower and 30% geothermal energy, and it is highly likely that any new demand added to their grid would be met with these clean resources. Buyers would have to ensure that the proper RECs/GOs are retired for their consumption, however, to avoid any double counting of renewable generation. This was actually an issue in Iceland.

Companies should note that this approach can be problematic in regions that have clean electricity on average but dirty marginal generation. In some areas, such as the U.S. Pacific Northwest, increased generation can be dirtier than average generation if clean resources are already being fully utilized. New buyers need to be exceedingly careful to not be responsible for increasing dirty energy usage on the margin in these cases, which is something that a focus on additionality and 24/7 CFE can help address. Asking the utility or other system operators in an area about how the local grid responds to increased demand can be informative here.

24/7 carbon-free energy

If behind-the-meter and clean grid approaches cannot work, then the next best option would be the use of 24/7 CFE. It may be possible to integrate a degree of behind-the-meter deployment with a PPA or a series of them in the same grid region to achieve 24/7 CFE. As noted above, there are several challenges to implementing 24/7 CFE ranging from intermittency to underdeveloped tracking infrastructure.

Regulators in the EU will be requiring geographic and hourly matching of clean energy for the production of green hydrogen, and it is likely that similar requirements will eventually apply to energy-intensive CDR pathways. Therefore, while difficult, CDR companies must begin creating an energy procurement strategy to get ahead of this. A potential stopgap measure could be the use of blended PPAs that involve both solar and wind, possibly linked with storage, in the same grid region as the project to help cover generation during more hours of the day.

Otherwise-curtailed renewable energy

An option pursued by some in the clean tech world is the use of renewable energy that would have otherwise gone to waste and been curtailed. Excess renewable generation during some parts of the day (e.g., oversupply of solar during the middle of the day in California) means that renewable plants sometimes have to be throttled to limit excess energy. This is a waste of renewable potential. While carbon removal systems could potentially make use of this energy, there are multiple barriers. The primary one relates to poor capital utilization: it is not cost effective to have a CDR facility that only runs for part of the day, as less carbon is removed per dollar spent on capital equipment. There may also be other technical and economic issues related to continuously ramping the systems up and down rather than running them more continuously.

Other clean technologies, including energy storage, demand management, and green hydrogen, are also competing for this cheap source of clean energy. Ultimately, it will probably be more efficient to address curtailment with such pathways rather than use it for carbon removal in most cases.

Traditional power purchase agreements

The next option involves the use of offsite physical or virtual PPAs. Offsite physical PPAs involve the actual delivery of generated power to a buyer through a power marketer, but they can be complex, are only available in certain states, and generally would not be used for 100% annual matching due to intermittency. Virtual PPAs can more easily allow for 100% annual matching and may offset a commensurate amount of fossil fuel use, but this is less certain and may not be the case when deployed in regions that already feature heavy renewable penetration.

VPPAs might also undermine the perceived validity of the offset being sold; a new DAC plant that increases natural gas use in West Virginia but offsets a similar amount of natural gas use in California may frustrate or harm some stakeholders, hurt the reputation of the project, and complicate the path toward net zero.

Emissionality

As mentioned above, there is also the emissionality approach that involves selecting a VPPA based on where it would have the most significant marginal climate impact. This option would involve selecting a VPPA location for a CDR project in a way that displaces dirty energy as much as possible, but it faces the same issues as VPPAs that are outlined above. The crediting of benefits from emissionality also wades into the territory of avoided emissions, which are important but can introduce complexities surrounding baseline estimations and counterfactual scenarios.

Traditional renewable energy certificates

One of the least ideal options, but one that some companies may have to pursue in the near future, is simply connecting to a mostly dirty grid and buying RECs or engaging in a related utility green tariff program. It may also be possible to pay more for higher-quality RECs, just as some pay more for higher-quality offsets. Depending on the emissions accounting standard used (e.g., see ISO 14064-1, Appendix E), this approach may even be allowable for emissions assessment, although this is likely to change as requirements for clean energy procurement become stricter. Regardless, RECs will always pose risks of non-additionality, double counting, and reputational damage.

Doing nothing and waiting

The worst option of all is to use dirty electricity, not buy RECs, and simply wait for the electricity grid to decarbonize on its own. This option would inflate net removal costs as discussed before and potentially cause a reputational hit not just to the company making this choice but the entire industry. Grid decarbonization may also take well over a decade or two depending on location. Except in experimental stages, no company should take this route, particularly when there are few true barriers other than minor costs to purchasing RECs.

Compromise

Clean energy procurement for carbon removal projects presents a stark trade-off between ease and integrity. Buying RECs is easy but generally not additional, PPAs are somewhere in the middle, and 24/7 CFE and behind-the-meter approaches are the most defensible but the hardest to implement.

This same issue is currently playing out with clean hydrogen subsidies, both in the U.S. and in the EU. If the requirements for low-carbon energy supply to water electrolyzers are too strict in the beginning, the industry will fail to take off. If the requirements are too lax, it defeats the purpose of producing green hydrogen at all. Environmental groups are pushing for stricter standards while profit-motivated entities are pushing for looser standards (although some blue hydrogen proponents may also be pushing for stricter standards to favor their own pathway). Some wealthy groups must be paying for the anti-additionality advertisements I have been receiving on YouTube! The IRS actually accepted public comment on the topic, and the responses are available online and are quite interesting.

While there is a significant amount of nuance here, the solution for both green hydrogen and CDR will likely involve some level of compromise and clever policymaking. The EU chose to phase in hourly matching requirements for green hydrogen to allow near-term scaling while preserving the ultimate integrity of the industry, and it would not be surprising to see a similar approach implemented in the U.S. for hydrogen and eventually for CDR. A compromise for any kind of energy-intensive clean technology helps balance the immediate need for scaling and deployment with the ultimate need for true climate impact.

CDR companies should strive for low-carbon, onsite generation and 24/7 CFE, and they have a chance to lead the market and help build out the necessary infrastructure. If this is not possible, high-quality PPAs should be used in the interim. Combining approaches and changing the procurement strategy with time and as regulations come into effect may be necessary. CDR developers, offset buyers, standards authors, registries, investors, and regulators must all be aware of what constitutes high-quality clean energy procurement to build a more genuine and robust CDR ecosystem.