Enhancing Profit and CO2 Mitigation: Commercial Direct Air Capture Design and Operation with Power Market Volatility

Zhiyuan Fan; Elizabeth Dentzer; James Glynn; David S. Goldberg; Julio Friedmann; Bolun Xu

doi:10.1016/j.eng.2026.02.025

Engineering ›› 2026, Vol. 62 ›› Issue (7) :283 -295. DOI: 10.1016/j.eng.2026.02.025

Research

research-article

Enhancing Profit and CO₂ Mitigation: Commercial Direct Air Capture Design and Operation with Power Market Volatility

Author information +

History +

PDF (4183KB)

Abstract

Current decarbonization efforts are falling short of meeting the net-zero greenhouse gas (GHG) emission target, highlighting the need for substantial carbon dioxide removal methods such as direct air capture (DAC). However, integrating DACs poses challenges due to their enormous power consumption. This study assesses the commercial operation of various DAC technologies that earn revenue using monetized carbon incentives while purchasing electricity from wholesale power markets. We model four commercial DAC technologies and examine their operation in three representative locations including California, Texas, and New York in the United States. Our findings reveal that commercial DAC operations can take financial advantage of the volatile power market to operate only during low-price periods strategically, offering a pathway to facilitate a cost-efficient decarbonization transition. The ambient operational environment such as temperature and relative humidity has non-trivial impact on abatement capacity. Profit-driven decisions introduce climate-economic trade-offs that might decrease the capacity factor of DAC and reduce total CO₂ removal. These implications extend throughout the entire lifecycle of DAC developments and influence power systems and policies related to full-scale DAC implementation. Our study shows that DAC technologies with shorter cycle spans and higher flexibility can better exploit the electricity price volatility, while power markets demonstrate persistent low-price windows that often synergize with low grid emission periods, like during the solar “duck curve” in California. An optimal incentive design exists for profit-driven operations while carbon-tax policy in electricity pricing is counterproductive for DAC systems.

Graphical abstract

Keywords

Direct air capture / Carbon dioxide removal / Profit-driven decarbonization / Climate-economic trade-off / Power market integration / Electricity price volatility / Climate change / Carbon incentive policy / Optimization / Carbon tax

Cite this article

Download citation ▾

Zhiyuan Fan, Elizabeth Dentzer, James Glynn, David S. Goldberg, Julio Friedmann, Bolun Xu. Enhancing Profit and CO₂ Mitigation: Commercial Direct Air Capture Design and Operation with Power Market Volatility. Engineering, 2026, 62 (7) : 283-295 DOI:10.1016/j.eng.2026.02.025

登录浏览全文

4963

注册一个新账户忘记密码

1. Introduction

1.1. Background and motivation

Current decarbonization efforts are failing short in meeting the net-zero greenhouse gas (GHG) emission target, and studies have shown that the most likely approach will be to overshoot 1.5 °C target followed by more intensive CO₂ removal from the atmosphere [1], [2], [3], [4]. Yet, the scale of potentially required carbon dioxide removal (CDR) deployment is enormous, ranging from 160 to 370 Gt-CO₂ net removal by 2100 for every 0.1 °C overshoot [1]. Depending on the chosen scenarios, up to 30 Gt-CO₂ of negative emissions per year will be required to balance the global carbon budget [5].

Direct air capture (DAC) is among the most scalable technologies for significant CO₂ removal [6], [7]. It extracts CO₂ from ambient air, while captured CO₂ can be stored geologically or mineralized for permanent removal (> 10 000 years) [8], or utilized as a carbon feedstock for synthetic fuels [9], chemicals [10], and building materials [11] to support a circular carbon economy. Among different DAC design concepts, adsorption-based design using solid sorbents offers the advantages of low regeneration temperature (< 100 °C) [7], [12], [13], flexibility with cyclic operation [14], [15], and modular design for scalability [16], [17]. During the adsorption and desorption cycles for DAC system regeneration, the rates of CO₂ capture and release exhibit nonlinearity depending on sorbent’s state of saturation and the cycle time is highly sensitive to different sorbent materials and operational decisions (Fig. 1 [18]). This allows for optimization opportunities in the DAC sorbent selection and operations and potential to achieve greater net-CO₂ removal. Also, compared to liquid solvents like potassium hydroxide (KOH), solid sorbents provide wider design flexibility, allowing for potential customization of sorbent material properties, such as cycle time and temperature preferences [19], [20].

While technological advancements and scalability of DAC systems are progressing rapidly, the challenges related to its economic efficiency and energy supply have become increasingly pronounced. Every gigatonne of CO₂ removal using DAC requires 167-305 TW·h of electricity (even after future technical improvement), potentially increase by 6-7 times with heating electrification, and costing hundreds of billion dollars [7], [12], [21]. In comparison, 1% of global total electricity generation in 2021 is 280 TW·h [22] and 1% of global gross domestic product (GDP) in 2021 is 968 billion dollars [23]. Such extensive power consumption is unlikely to be operated outside the supply-demand balance in power market, whose costs rely solely on subsidies funded by taxpayers. Due to its enormous potential scale, DAC’s energy demand could place a significant burden on the economy—particularly if it is planned or operated inefficiently, leading to both economic losses and additional GHG emissions from energy use. A recent study [24] calls for a more realistic analysis for commercial CDR deployment, considering scaling, energy, location, and cost in more practical ways.

1.2. Literature review and research gap

Previous studies on the economic efficiency and GHG life cycle assessment (LCA) of DAC have shown that a grid-connected system is sensitive to the fossil electricity mixture, greatly affecting its capture efficiency and potentially lead to net CO₂ emissions rather than removal [25], [26]. These studies are generally based on average electricity price and emission intensity assumptions, seldom include renewable profiles [27], and not considering temporal volatility in electricity price and emission as shown in Table 1 [12], [13], [21], [22], [23], [24], [25]. These assumptions are based on constant DAC operations—approximately 100% capacity factor—and thus exclude potential operational optimization. Typically, techno-economic analyses focus on minimizing system costs to achieve specific CDR targets without accounting for the utility of individual DAC systems as autonomous agents. However, for commercial DAC projects, it is more rational to also consider the DAC system operator standpoints, who are profit-driven, especially when upfront investment is substantially made through private funding sources [28]. Commercial operations are also subjected to high price volatility introduced by large expansions in renewable power supplies [29]. A well-known example is the “duck curve” observed in California (CA) due to high solar generation during the daytime [30], [31]. The real-time price in power markets updates sub-hourly, CA and New York (NY) update the prices every 5 min [31], [32], and Texas (TX) every 15 min [33], and both price and emission intensity change rapidly. Intuitively, while exposed to market volatility without a minimum CO₂ removal constraint, DAC operations may be expected to accelerate when the power price is low and reduce or shut down when the price is high. Additionally, DAC operations are sensitive to ambient temperature and relative humidity while different types of DAC systems have different preferred environments [34], [35]. These characteristics point to the need for site-specific DAC operation modeling with higher resolution.

1.3. Contributions

Based on the review over existing literature and research gap, our contributions are presented:

Profit-driven optimization from DAC system perspective. Our model optimizes DAC system operations with a profit-maximization objective, thereby capturing the realistic behavior of these systems under conditions of price volatility and evolving policy environments/penalties (e.g., incentives and carbon taxes).

High and flexible temporal resolution. Our model employs a 5-min temporal resolution, matching the electricity market price data, which can be flexibly aggregated to lower resolutions with corresponding adjustments to the DAC parameters. Our analysis reveals that certain temporal dynamics are critical and can only be observed at high temporal resolutions.

Spatial and ambient environment sensitivity. We analyzed three representative US electricity markets—NY, CA, and TX—each characterized by distinct price profiles driven by different dominant renewable energy sources and unique ambient conditions (e.g., temperature and relative humidity) that significantly influence DAC operations. Our findings indicate that optimal DAC siting may be even more critical than engineering improvements.

Reverse engineering future DAC research and development. Our analysis indicates that under conditions of highly volatile electricity prices, carbon intensities, and ambient environments, flexible DAC operation/capability is not only strictly preferred but also offers significant utility—highlighting a previously underestimated research direction for future DAC system research and development.

2. Models

There are primarily two types of operations for different DAC systems: cyclic and continuous. In cyclic systems, typically low-regeneration temperature sorbents (100 °C), a stationary chamber holds the sorbent material, which undergoes adsorption-desorption cycles through variations in temperature, pressure, humidity, or a combination thereof (e.g., temperature-pressure swing design). In contrast, continuous systems typically employ a material loop for high regeneration temperature solvents (800 °C for CaCO₃ regeneration), where sorbent or solvent materials are conveyed between different pieces of equipment for each process step. Although the materials in continuous systems also experience absorption-regeneration cycles, key equipment and its energy consumption—such as adsorption platforms and regeneration units—operates continuously by constantly moving materials. This study models both processes, with a particular focus on cyclic operation due to its low-regeneration temperature and non-linear optimization characteristics. Both models are presented in rigorous mixed integer linear programming (MILP) optimization problems we propose additional strategic bidding algorithm that improves the computation.

2.1. Cyclic model

The objective function maximizes the total profit of the DAC operation at each time step within the horizon t∈T.

(1)

m a x \sum_{t} π d_{t} - λ_{t} C_{t} (P^{a} u_{t} + P^{d} v_{t}) - S z_{t}

where π is the CO₂ value constant, including selling price, subsidies, and carbon tax; d_t is the continuous non-negative desorption amount of the DAC system at time period t; πd_t is the total revenue by captured (desorbed) CO₂, minus the cost of power consumption corrected by ambient conditions at each time step $\lambda_{t} C_{t}\left(P^{\mathrm{a}} u_{t}+P^{\mathrm{d}} v_{t}\right)$ (where λ_t is the electricity price at time period t, potentially modified by CO₂ intensity; C_t is the energy consumption correction factor as a function of relative humidity and temperature; P^a and P^d are the electricity consumption per unit time period during the absorption phase and desorption phase, respectively; u_t and v_t are binary variables), minus the cost of material consumption (sorbent degradation and replacement and fixed operation and maintenance (O&M)) for switching cycles Sz_t (where S is the switching cycle cost associated with sorbent material consumption and z_t is binary variable). The fixed material replacement cost is effectively scaled proportionally with the operating capacity factor, reflecting a linear relationship of utilization on sorbent/solvent lifetime.

Binary constraints for absorption/desorption status:

(2)

u_{t} + v_{t} \leq 1

where the DAC system can only be either adsorption or desorption at one time period t, but it can be neither absorbing nor desorbing (i.e., staying idle). Absorption and desorption rate (using inequality here to avoid contradiction with binary constraints, which is relaxation of the constraints) are subjected to state-of-saturation $\bar{X}$ (the maximum available DAC capacity) correct to model the nonlinear behavior. Specifically, only adsorption (capture) process is corrected by ambient condition $\eta_{t}^{\mathrm{a}}$ of each time step, where desorption is a controlled high-temperature process that is not affected by ambient condition:

(3)

a_{t} \leq η_{t}^{a} (β_{1}^{a} + β_{2}^{a} X_{t})

(4)

d_{t} \leq β_{1}^{d} + β_{2}^{d} X_{t}

Additionally, the absorption rate and desorption rate shall be bounded by the binary variable for each as well.

(5)

a_{t} \leq M u_{t}

(6)

d_{t} \leq M v_{t}

where $\eta_{t}^{\mathrm{a}}$ is the adsorption rate correction factor as a function of relative humidity and temperature; $\beta_{1}^{\mathrm{a}}$ and $\beta_{2}^{\mathrm{a}}$ are the first-order and second-order coefficient for absorption, respectively; $\beta_{1}^{\mathrm{d}}$ and $\beta_{2}^{\mathrm{d}}$ are the first-order and second-order coefficient for desorption, respectively; at is the continuous non-negative absorption amount of the DAC system a_t time period t; X_t is the continuous non-negative state-of-saturation capacity of the DAC system at time period t; M is a sufficiently large number which does not bind the absorption and desorption rate if u_t and v_t are 1.

State-of-saturation capacity of DAC system updates with absorption/desorption rates:

(7)

X_{t} - X_{t - 1} = a_{t} - d_{t}

where the change of state-of-saturation between time periods equals: adding absorption or subtracting desorption.

(8)

0 \leq X_{t} \leq \bar{X}

where the state-of-saturation is always lower bounded by 0 (non-negative) and upper bounded by declared maximum capacity $\bar{X}$.

Switching cycle constraints model the cost of each cycle, this includes cost of degradation of sorbent materials and fixed cost for regeneration heat supply. The initial state of the sign-variable equals indicating cycle switching is zero.

(9)

k_{0} = 0

The following constraints allocate idle states as continuing of the previous adsorption/desorption process, thereby ensuring that such states do not trigger cycle counting:

(10)

- M (1 - k_{t}) \leq k_{t - 1} + (u_{t} - v_{t}) - 0.5

(11)

M k_{t} \geq k_{t - 1} + (u_{t} - v_{t}) - 0.5

where k_t is the binary sign variable used to facilitate calculation of z_t by detecting status changes at time period t.

The solution matrix for k_t cross all possible combinations of sign-function is presented in Eq. (12). It logically interprets the idle status as continuation of the ongoing adsorption/desorption process, rather than triggering a new cycle that consume material lifespan and incur cost:

(12)

k_{t} = \{\begin{matrix} 0 & ({i f u}_{t} - v_{t} = - 1, o r u_{t} - v_{t} = 0 a n d k_{t - 1} = 0) \\ 1 & ({i f u}_{t} - v_{t} = + 1, o r u_{t} - v_{t} = 0 a n d k_{t - 1} = 1) \end{matrix}

Then the sign binary variable k_t can be used to determine the binary cycle counting variable z_t:

(13)

z_{t} \geq k_{t} - k_{t - 1}

where z_t will be minimized in the objective function that only when k_t=1 and k_t_-1=0, z_t=1.

The above formulation summarizes the optimization framework of the DAC system using the temperature-pressure-swing DAC technology (all three cyclic DAC technologies tested in this paper). In practice, the DAC input parameters (β values) including the piecewise linear approximation for absorption and desorption rate using quadratic coefficients will be determined by different DAC specifications. For cyclic DAC systems, which naturally alternate between adsorption and desorption modes, minimum stable load, ramping rates, and on/off cycle limits are not applicable.

2.2. Continuous model

The variables definition is consistent with the cyclic DAC MILP settings.

(14)

m a x \sum_{t} π d_{t} η_{t}^{a} - λ_{t} P_{t}^{d} - S_{c} d_{t}

where πd_t is the total revenue corrected by ambient environment conditions $\eta_{t}^{\mathrm{a}}$, minus the cost of power consumption $\lambda_{t} P_{t}^{\mathrm{d}}$ and the non-electric operational cost S_cd_t (S_c is the continuous operation cost per unit desorption, which include both thermal energy cost and solvent material consumption). Ambient condition correction is applied directly to desorption, because a continuous process is assumed in which CO₂ is seamlessly desorbed instantly after adsorption.

For continuous DAC, the operational flexibility is determined by two constraints. The first one is on/off limit constraints:

(15)

y_{t} - z_{t} = w_{t} - w_{t - 1}

(16)

y_{t} - z_{t} \leq 1

where w_t is the binary on/off status of DAC, y_t and z_t indicating switching on/off status of DAC.

(17)

\sum_{i = 1}^{I} (y_{i} + z_{i}) \leq K

Eq. (17) means the total times of switching on or off during the horizon (i∈1, 2,..., I) is limited by maximum flexibility index K. In practice, I=288=24×(60/5) 5-min time-steps leading to 1-day horizon, meaning the continuous DAC can maximum turn on/off by K times in one day.

The second one is min/max rate constraints:

(18)

\underline{d} w_{t} \leq d_{t} \leq \bar{d} w_{t}

the capture rate of CO₂ d_t is limited between the min/max capture rate, and we use 80%-100% as representative knowing this might vary for different technologies. Accompanied with the power consumption constraints:

(19)

P_{t}^{d} = \frac{d_{t}}{\bar{d}} \bar{P^{t}}

where the actual power consumption $P_{t}^{\mathrm{d}}$ is the fraction of nominal power consumption for capture $\overline{P^{t}} $.

2.3. Abatement accounting

Eventually, the net-CO₂ removal is calculated by summing the total desorption and subtracting the emission from power consumption: emission intensity et times power consumption.

(20)

C O_{2}^{n e t} = \sum_{t} d_{t} - e_{t} (P^{a} u_{t} + P^{d} v_{t})

where e_t is the electricity CO₂ intensity at time period t.

CO₂ capture efficiency is defined by net-CO₂ removal over total captured:

(21)

η_{C O_{2}} = \frac{C O_{2}^{n e t}}{\sum_{t} d_{t}}

2.4. Strategic bidding algorithm

The computational burden of both DAC MILP models is expensive, where a 1-year rolling-horizon optimization generally takes 4-8 h for cyclic model and 1.5-2.0 h for continuous model. We propose a strategic bidding algorithm for cyclic DAC operation that improves the computational efficiency and simulate the actual DAC system behavior. Based on the MILP benchmark using deterministic price scenarios, we empirically extract operational insights such as bidding thresholds based on market clearing horizon (typically 1 day) and operate with the following logic:

(22)

\{\begin{matrix} Activate & if λ_{t} (t) \leq λ_{opt} \\ Idle & otherwise \end{matrix}

when activeted during low-price period, the DAC system will periodically adopt adsorption-desorption cycles between the pre-determined depth:

c_{1} \bar{X} \leq X_{t} \leq c_{2} \bar{X}

where c₁ and c₂ can be learned statistically from observing rigorous MILP results (Fig. 2 for the workflow).

The proposed strategic bidding algorithm offers several advantages (Table 2):

Realistic bidding strategy. It simulates a scenario in which a DAC system submits a demand bid to a power system operator, expressed as a price threshold: If the actual electricity price exceeds the bid, the system remains idle; if the price falls below the threshold, the DAC system operates. This is a more realistic market representation than the theoretical MILP optimum.

Near-optimal performance. It exhibits near-optimal performance—approximately 90% of rigorous MILP results in most cases—and maintains stable solutions even under negative or marginal profit scenarios.

Interpretability. It provides clear and simple operational guidelines for the DAC system by defining electricity price cut-offs and identifying the optimal state-of-saturation depth for the specified technologies.

Computational efficiency. It enhances computational speed by a factor of 1500–3000, reducing the optimization process to approximately 10 s for a one-year optimization. Essentially, the optimization problem is reduced to a single variable, $\lambda_{\mathrm{opt}}$, as opposed to managing hundreds of variables, including binary ones.

To address potential feedback effects of DAC operations on market prices, we provide a comparison between the price taker assumption presented here and price influencer cases using integrated DAC + Western Electricity Coordinating Council (WECC) power market simulation results. We find that from the kiloton to hundred-kiloton scale, these feedback effects remain manageable, while strong feedback appears when DAC deployment reaches the scale with gigawatt level peak load (Mt-CO₂·a⁻¹ scale or higher).

3. Trading-off CO₂ removal with economic efficiency

Our objective is to compare commercially deployable DAC technologies utilizing electricity as the primary energy source. We have chosen three sorbent DAC technologies and KOH liquid solvent system for comparison for their technical maturity and data availability (especially in cycle time and operational flexibility).

Metal-organic framework (MOF): a family of novel solid porous adsorbents with fast cyclic operation and potentially low CAPEX. Energy consumption is relatively high but rapidly improving through innovation. Challenges include sorbent material costs and unknown lifespan for synthesis on a large scale [36], [37].

APDES-NFC-FD (AN): a type of commercialized amine-functionalized sorbent with cyclic operation. It is very efficient in energy and material cost with much less sorbent consumption per ton CO₂ captured and has higher. Higher CAPEX, moderate flexibility in cycle time [38], [39].

SI-AEATPMS (SA): a type of early commercialized amine-functionalized sorbent, used in cyclic operation with very high sorbent consumption per ton CO₂ capture. These operations are expensive and inflexible, with long cycle times, and are used as a comparison benchmark. [38], [40].

KOH liquid solvent: a commercialized continuous operation DAC by looping KOH liquid solvent. The regeneration temperature is high at about 800 °C. It is expensive in CAPEX and energy intensive for high temperature solvent regeneration [13], [41].

All cyclic solid sorbent DAC technologies require nearly the same desorption-regeneration temperature (100 °C), but larger differences in costs, cycle times, and energy consumption: Table 3 shows the technical parameters of the four DAC technologies and Table 4 shows full-year baseline simulation results for three selected power markets. In this power system-centric study, we assume that, aside from cost considerations, no operational constraints or geographic differentiation on thermal energy consumptions are applied. We simulate each technology using 2022 electricity prices from the NY, CA, and TX power markets (NYISO, CAISO, and ERCOT, respectively), representing the three grid interconnection areas in the United States: Eastern, Western, and TX. We use average emission signal published by the Independent System Operator (ISO) for optimization as these are used as the primary indicator for measuring carbon emission intensity [42].

The comparison between profit-driven and full-capacity models for different DAC technologies unveils a non-intuitive outcome: Increasing CO₂ removal does not always translate into higher profits for DAC operators. MOF technology offers more CO₂ removal for larger plant capacities but is less profitable than AN technology. A profit-driven MOF technology offers the flexibility to avoid high-price periods, reversing non-profitable NY and CA cases profitable, and enhances economic efficiency in TX by 10 times. Profitability enables earlier deployment within each power market and policy framework. DAC operations may benefit by remaining idle and reducing their CO₂ removal efforts during certain periods. At full capacity, DAC systems could be unprofitable in the same market and under the same incentives, making such projects unviable. Investors may opt for more efficient and profitable technologies with smaller plant capacity and operate it profit-driven which reduces CO₂ removal. The outcome points to an economic-climate trade-off: From DAC system perspective, increasing economic efficiency may reduce net-CO₂ removal.

4. Power market and environmental implications

4.1. Annual temporal behaviors

A more detailed look at the annual temporal behavior in different locations explains the economic-climate trade-offs. By using the MOF and KOH technologies as an example, the distributions of net-CO₂ removal (Fig. 3) is negatively correlated with electricity price across different markets and consistent across DAC technologies. Results show a strong seasonal pattern. NY performs best during spring and fall when prices are low due to low heating and cooling demands [32]; CA has its worst period in November and December due to the price surge caused by unexpected cold waves [31]; while TX shows summer is the worst season due to its high cooling demand and low wind profile [33]. The seasonal price pattern is driven by supply-demand balance of electricity, fundamentally determined by climate patterns such as temperature and relative humidity. The impact of climate on DAC technology selection is non-trivial: ① Different DAC technologies have different preferred climate conditions; ② the climate impact can be both positive and negative, causing gain or loss in abatement capacity.

4.2. Daily temporal behaviors

We examine the hourly resolution profile from daily average price (aggregated from 5-min resolution) for DAC operations and sheds light on the divergence between monthly profits and CO₂ removal in different states. CA shows a distinct duck-curve daily pattern that prices are significantly lower during the day (Fig. 4), while in NY and TX, the wholesale electricity prices are relatively stable within a day, showing a typical 2-peak profile [43]. As a result, DAC systems in CA would operate at a high-capacity factor during the daytime while remaining largely idle in other periods. Notably, the daytime price in CA is often near zero or even negative during spring months (April-July) when the cooling demand is also low, yielding large profit potential despite the net-CO₂ removal being lower than in NY or TX. This example shows how DAC operations could benefit from additional renewable deployments which increase grid volatility but contribute to more low-price intervals. Regional climate conditions cause significant capture reduction for NY, while TX favoring KOH than MOF. DAC operations are relatively insensitive to CA’s climate conditions for both technologies. Different DAC technologies exhibit varying sensitivities to ambient temperature and humidity. Sorbent-based systems (e.g., MOF) perform best in temperate, low-humidity environments, as water vapor interferes with CO₂ adsorption and high temperatures weaken binding strength. In contrast, KOH-based liquid solvent systems benefit from hot and humid climates, where water availability and elevated temperatures enhance CO₂ absorption efficiency by forming Ca(OH)₂. Ambient temperature-humidity corrections generally act as efficiency adjustments, affecting about 5%-20% of net-CO₂ removal. However, when conditions reach tipping points (e.g., KOH technology in NY during nighttime), they can flip profitable operations into losses, leading to large uncertainty in optimization outcomes.

Cumulatively, CA surpasses TX as the most profitable location for DAC deployment, followed by NY, despite its highest average electricity price and lower net-CO₂ removal. A strong positive correlation between monthly profit and CO₂ removal is observed in NY and TX which grow proportionally by using 50-70 USD·(MW·h)⁻¹ fixed electricity price (Fig. 5). However, the most profitable season in CA (April-July) does not provide the highest CO₂ removal, and it uses essentially free electricity. For January-March, profits and CO₂ removals in CA are similar to NY. In summary, local demands such as heating/cooling and load-following power services strongly affect the power system pricing and resilience. These factors need careful consideration if DAC systems are to be deployed at a large scale.

4.3. Importance of temporal resolution

Figs. 3(a-iii) and (b-iii) show that aggregating daily results to monthly averages can mask important details. For TX, July’s average price is inflated by some expensive days, even though 8 days had prices below 60 USD·(MW·h)⁻¹, keeping DAC marginally profitable for operation. In contrast, all 31 days in August had prices above 60 USD·(MW·h)⁻¹ but don’t have any extreme high price outliers, making DAC operations largely unprofitable for all days and reducing its capacity factor to near 0 (Fig. S3 in Appendix A). This mismatch underscores that high temporal resolution analysis is essential for accurately capturing power market volatility and guiding DAC operations.

Flexible DAC operations for profit maximization are rational for investments compared to full-capacity operations under grid-average properties when incentives are not sufficiently high. It can significantly improve profitability, turning some otherwise unprofitable projects into viable ones and accelerating investment payback periods—thereby encouraging new DAC deployments. It is also necessary to avoid possible misleading differences from using grid-average assumptions. Marginal pricing in power markets automatically synchronizes DAC operation with renewable profiles. The simultaneity can boost profit and carbon removal efficiency, absorb additional/curtailed renewable, and stabilize power prices, but limit the DAC operation with renewable capacity factor. This approach may extend to other flexible demands, such as green hydrogen production [44], and could motivate mobile DAC systems, like rail-based solutions leveraging different low-price periods in different markets [45]. Analyzing power market temporal behavior further helps optimize maintenance and labor, reducing costs.

5. Policy implications

While DAC profit maximization is rational from an economic perspective, it’s essential to recognize that CO₂ differs significantly from conventional commodities. Unlike most product economics driven by supply and demand, the value of CO₂ is primarily rooted in the urgent need to mitigate its adverse impact on global warming [46]. Although commercial opportunities for CO₂ utilization such as cement [47], synthetic fuels, and chemicals [48] are growing, the need to scale CO₂ removal have been shaped by government regulations through national policies, rather than market forces. For DAC operations, the profit arises only after the retirement of credits in a voluntary or compliance CO₂ market. Therefore, a careful policy design by optimizing the incentives for DAC operations or imposing carbon taxes on their energy consumption is pivotal. Better incentive policy design can address the economic-climate trade-off. However, applying carbon taxes to energy consumption for DAC plants may be counterproductive.

5.1. CO₂ incentives

Using the CA as an example, we evaluate the four DAC technologies under different CO₂ incentive policies (i.e., revenues received by selling captured CO₂ credits (Fig. 6)). Although very different in assumptions, all DAC technologies exhibit a qualitatively consistent trend.

All DAC technologies are assumed to have the same initial investments of total 1.05 × 10⁷ USD, but different CAPEX assumptions for each different plant (Table 3). MOF technology plant delivers the highest net-CO₂ removal if the incentive is sufficiently high. However, its relatively high OPEX makes it less competitive than AN technology when incentives are low. Both exceed the net-CO₂ removals using SA and KOH technologies, which have high CAPEX or OPEX. The net-CO₂ removal grows quickly beyond the anchoring points, where the incentive equals the non-electricity OPEX, referred as cycle cost for sorbent DAC systems. With electricity price volatility, the cycle cost sets a lower threshold for initiating CO₂ removal, while CAPEX defines the upper limit for the total CO₂ removal capacity.

Implementing DAC operation to maximize profit is not a simple binary decision. The amount of total CO₂ removal and total profit grow gradually with the increasing CO₂ incentives. CO₂ removal plateaus when it approaches a DAC plant’s maximum capacity, but profit keeps growing as the incentive increases, leading to shorter capital payback years. When incentive is sufficiently large, low CAPEX technology pays back earlier even with lower energy efficiency. Net-CO₂ removal volume is quite sensitive to the range of incentives applied where a small change can impact CO₂ removal significantly. In the United States for instance, the current 45Q legislation [49] provides a 180 USD·t⁻¹ tax credit for DAC with geological storage. This falls within the sensitive price range for MOF technology (Fig. 6), suggesting a large opportunity cost: Increasing the 45Q incentive from 180 to 250 USD·t⁻¹ would mean that using MOF is economically more favorable than using AN, and approximately doubles the net-CO₂ removal volume.

Fig. 7 compares CO₂ removal and profit for each DAC technology under different incentive assumptions. CO₂ removal initially increases with greater incentives for all DAC systems. Beyond a sharp corner point, CO₂ removal plateaus and additional incentive value translates into additional profit. This corner point corresponds to the optimal incentive policy design for the CO₂ selling price and differs for each DAC technology. Using this information as a tool will be helpful in designing an improved DAC incentive policy and optimizing the economic-climate trade-offs, as well as tracking the state-of-the-art technology.

5.2. Carbon-tax

With point-source carbon capture and storage systems (CCS), such as power plants, applying carbon tax on energy consumption may be a positive driving factor that encourages deployment [50], [51]. Emission baselines are positive and will be reduced for such systems, lowering commercial carbon tax expenses and making CCS an economically rational decision. Therefore, a carbon tax applied in these cases can lead to net CO₂ reductions. For DAC CO₂ removal, however, low-emission electricity consumption driven by a carbon tax optimizes capture efficiency while sacrificing capacity factor, and results in effectively higher electricity costs. We therefore consider an analysis of the effect of a carbon tax applied to electricity consumption in DAC operations. Since a carbon tax alters the electricity price signal but does not eliminate associated emissions, these emissions are still accounted for in incentive payments, thereby shifting the optimization outcome; the coexistence of both policies may reflect different valuation of CO₂ (in USD·t⁻¹) in practice and does not introduce double counting or conflict.

Results show that the trade-off between capture efficiency and net-CO₂ removal volume for DAC technologies is highly skewed, strongly favoring higher removal volume when costs (including taxes) are reduced. The imposition of a carbon tax on electricity use may harm both the net-CO₂ removal and the profitability of potential DAC operations. Fig. 7 shows that a carbon tax always falls within or below the raw wholesale price case for CO₂ removals for all DAC technologies. While the carbon tax marginally enhances CO₂ capture efficiency, it fails to compensate for the substantial reduction in net-CO₂ removal volume. For the MOF technology, which has high electricity consumption, applying a carbon tax significantly increases the needed to reach the incentive the same optimal corner point (i.e. 250 USD·t⁻¹ increases to 360 USD·t⁻¹). For AN technology, with lower energy requirements, a carbon tax still generates lower profit and no improvement in net-CO₂ removal volume. The added cost from the carbon tax outweighs the benefit of increased capture efficiency, rendering the system unprofitable to stop operation and resulting in less CO₂ removal instead of efficiency gain. The impact of a carbon tax on energy use for DAC operations is therefore negative, irrespective of the technology used or the CO₂ removal incentive offered. This means that a carbon tax on electricity use may be counterproductive for both profitability and net-CO₂ removal of DAC plant operations.

Eventually, a carbon tax policy on DAC electricity usage will not lead to more CO₂ removal. We suggest an exemption for DAC plants from such taxes or equivalent spending (e.g., require purchasing of renewable energy certificates (RECs)) on condition that their capture efficiency remains reasonable, perhaps at least 75%-80%, which should be set based on specific power market and technology requirements.

6. DAC design implications

Comparing different DAC technologies (Fig. 7 and Supplementary Information in Appendix A for more details) reveals that reducing system CAPEX, material costs, and cycle time has a significant impact on DAC operations.

Profit-driven operations are very sensitive to CAPEX and material costs, offering a substantial room for improvement. DAC adsorbents must be durable, withstanding repeated regeneration cycles while retaining the effective capacity over their material lifespans [52]. Orders-of-magnitude differences in costs per ton removal [38] and orders-of-magnitude differences uncertainty about material resilience [37] exist for each. Additionally, CAPEX estimation is highly sensitive to plant scale and the technology chosen, and thus technical improvements are reported to affect performance by a factor of 2-10 [12].

Shorter cycle times and faster adsorption and desorption rates also improve DAC performance. Accurate cycle time assumptions are often missing in techno-economic analysis for full-capacity operations, and when high-resolution temporal volatility is introduced, long-cycle operations may no longer remain profitable (Fig. 8). Flexibility in cycle times may be critical as price incentives change in different power markets. From our market integration analysis, an ideal target for DAC development may have cycle periods of 1-2 h and cycle costs below 80 USD·t⁻¹ of CO₂ captured (not accounting for electricity cost). With new research and development in this area, routine day-ahead power markets and hourly-resolution dispatch prices, DAC operations may eventually be able to work with hourly market clearing resolution with short cycle time. In summary, reducing system CAPEX, material costs, and cycle times has the potential to achieve the cost efficiencies necessary to ultimately accelerate DAC deployment in commercial markets [53].

7. Discussion

This paper considers temperature-pressure-swing adsorption/desorption solid sorbent and KOH liquid solvent for a total of four commercial DAC technologies. Our methodology is adaptable to other cyclic or continuous DAC technologies as well, including pressure-swing, and moisture-swing approaches, as well as membrane-based DAC systems that operates continuously. The economic model accommodates diverse heat/electrical assumptions with varying cycle times and operational flexibility, facilitating both profit-driven operations and techno-economic analysis. It can also be adapted to simulate heat electrification by aggregating thermal energy consumptions into the desorption/regeneration phase. Despite the significant differences in the DAC technology parameters tested, our results exhibit consistent and robust qualitative trends that are broadly applicable across larger contexts.

Our analysis demonstrates that the integration of DAC operations with power market volatility prompts new perspectives. For DAC system development and deployment, key implications are: ① DAC technology development should prioritize low CAPEX (< 100 USD·(tCO₂·a)⁻¹), low material costs (< 80 USD·(tCO₂)⁻¹), and high flexibility with shorter cycle time (1-2 h); ② DAC should connect to appropriate power markets with preferred environmental condition such as temperature and humidity; for example, a liquid solvent DAC system for TX. Profit driven operations can lead to earlier DAC deployment with higher economic efficiency, for example, MOF technology can increase profit > 10 times and shorten the investment payback period by 15 years. Optimal DAC designs should be tailored to specific climates and geographic areas, where siting of DAC plants is fundamentally dictated by local climate features, both the renewable profiles in the power market and the appropriate climate conditions (e.g., temperature and humidity). Different power markets favor different DAC technologies, and should be carefully considered in decision-making for plant design and deployment. While this study applied ISO-based average emission signal for optimization, recent work [54] shows that average and marginal signals can diverge substantially, with the marginal signal may provide a more accurate measure of incremental load impacts. Exploring this distinction represents an important avenue for future analysis.

For policymakers, consideration of these factors for profit-driven DAC operations may help to establish effective policies that maximize net-CO₂ removal volume and assure that efficiency does not take precedence over scale. High CO₂ capture efficiency and energy efficiency at the expense of plant capacity, material lifespan, operational flexibility, and economic viability are counterproductive. Good policy design prioritizes incentives that balance the economic-climate trade-offs, which can potentially double the MOF DAC removal if incentive is raised from 180 to 250 USD·t⁻¹. Carbon tax exemptions or establishing a minimum required capacity factor of DAC plants for eligibility of government subsidies may be needed. In regulated Cap-and-Trade markets, for example, profitability can be boosted during high CO₂ price opportunities, lowering CO₂ removal costs for other carbon market participants. Our findings apply to a carbon tax combined with net-removal incentives, and they generally extend to variants such as tax refunds or clean procurement requirements, provided that these policies effectively change the electricity price signals, and eventually lead to shift of profit-max optimization results.

In marginal-price-based markets, profit driven DAC operation will automatically avoid peak hours with high wholesale electricity prices and absorb additional/curtailed renewables during off-peak hours. Power system operators will benefit from such flexible DAC operations and reduce system cost, emissions, and instability with renewable-following DAC operations. Specifically, solar simultaneity suggests that the prices can be much lower than grid averages, enabling profitability and encouraging DAC expansion. Certain market strategies and policies might also help to increase the DAC system capacity factors and maximize CO₂ removal, such as firm power purchase agreements.

8. Conclusions

Our global energy system is currently undergoing a fundamental transformation towards a net-zero emission target. Both CO₂ removal using DAC and power grid decarbonization play pivotal roles in this transition. Given that this energy transition is anticipated to be lengthy and gradual, demanding an unprecedented level of investment, it is crucial to maintain the stability of the energy system throughout the transition period and ensure economic viability. With profit-driven investments in technically mature, large-scale DAC plants, operations under power market volatility can be financially advantageous, climatically beneficial, and supportive a smooth energy transition. Taking the DAC system perspective for profit driven operation introduces an economic-climate trade-off with implications extending throughout the entire lifecycle of DAC system operations, reflecting the efficient use of current technologies, future technology research and development, power markets integration, and decarbonization policy designs. Our insights and methods open future work for potential DAC system design and sorbent materials customization for specific power market and climate zones, and provide power market services such as demand response. These are important topics to a successful low-carbon integrated power-energy system.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	Calvin K, Dasgupta D, Krinner G, Mukherji A, Thorne PW, Trisos C, et al. Climate change 2023:synthesis report. Report. Geneva: IPCC; 2023.

[2]	Lamboll RD, Nicholls ZRJ, Smith CJ, Kikstra JS, Byers E, Rogelj J. Assessing the size and uncertainty of remaining carbon budgets. Nat Clim Chang 2023; 13 (12):1360-7.

[3]	Tollefson J. Is it too late to keep global warming below 1.5 °C? The challenge in 7 charts [Internet]. London: Nature; 2023 Nov 21 [cited 2025 May 25]. Available from: https://www.nature.com/immersive/d41586-023-03601-6/index.html

[4]	Buck HJ, Carton W, Lund JF, Markusson N. Why residual emissions matter right now. Nat Clim Chang 2023; 13(4):351-8.

[5]	Marcucci A, Kypreos S, Panos E. The road to achieving the long-term Paris targets: energy transition and the role of direct air capture. Clim Change 2017; 144(2):181-93.

[6]	Goldberg DS, Lackner KS, Han P, Slagle AL, Wang T. Co-location of air capture, subseafloor CO₂ sequestration, and energy production on the Kerguelen plateau. Environ Sci Technol 2013; 47(13):7521-9.

[7]	McQueen N, Gomes KV, McCormick C, Blumanthal K, Pisciotta M, Wilcox J. A review of direct air capture (DAC): scaling up commercial technologies and innovating for the future. Prog Energy 2021; 3(3):032001.

[8]	Alcalde J, Flude S, Wilkinson M, Johnson G, Edlmann K, Bond CE, et al. Estimating geological CO₂ storage security to deliver on climate mitigation. Nat Commun 2018; 9(1):2201.

[9]	Parigi D, Giglio E, Soto A, Santarelli M. Power-to-fuels through carbon dioxide re-utilization and high-temperature electrolysis: a technical and economical comparison between synthetic methanol and methane. J Clean Prod 2019; 226:679-91.

[10]	Gulzar A, Gulzar A, Ansari MB, He F, Gai S, Yang P. Carbon dioxide utilization: a paradigm shift with CO₂ economy. Chem Eng J Adv 2020; 3:100013.

[11]	Liu B, Qin J, Shi J, Jiang J, Wu X, He Z. New perspectives on utilization of CO₂ sequestration technologies in cement-based materials. Constr Build Mater 2021; 272:121660.

[12]	Fasihi M, Efimova O, Breyer C. Techno-economic assessment of CO₂ direct air capture plants. J Clean Prod 2019; 224:957-80.

[13]	Sabatino F, Grimm A, Gallucci F, Van Sint AM, Kramer GJ, Gazzani M. A comparative energy and costs assessment and optimization for direct air capture technologies. Joule 2021; 5(8):2047-76.

[14]	Wiegner JF, Grimm A, Weimann L, Gazzani M. Optimal design and operation of solid sorbent direct air capture processes at varying ambient conditions. Ind Eng Chem Res 2022; 61(34):12649-67.

[15]	Jiang L, Liu W, Wang RQ, Gonzalez-Diaz A, Rojas-Michaga MF, Michailos S, et al. Sorption direct air capture with CO₂ utilization. Pror Energy Combust Sci 2023; 95:101069.

[16]	Ozkan M, Nayak SP, Ruiz AD, Jiang W. Current status and pillars of direct air capture technologies. iScience 2022; 25(4):103990.

[17]	Smith P, Davis SJ, Creutzig F, Fuss S, Minx J, Gabrielle B, et al. Biophysical and economic limits to negative CO₂ emissions. Nat Clim Chang 2016; 6 (1):42-50.

[18]	Young J, Mcilwaine F, Smit B, Garcia S, Van Der Spek M. Process-informed adsorbent design guidelines for direct air capture. Chem Eng J 2023; 456:141035.

[19]	Beuttler C, Charles L, Wurzbacher J. The role of direct air capture in mitigation of anthropogenic greenhouse gas emissions. Front Clim 2019; 1:10.

[20]	Bose S, Sengupta D, Rayder TM, Wang X, Kirlikovali KO, Sekizkardes AK, et al. Challenges and opportunities: metal-organic frameworks for direct air capture. Adv Funct Mater 2024; 34(43):2307478.

[21]	Realmonte G, Drouet L, Gambhir A, Glynn J, Hawkes A, Köberle AC, et al. An inter-model assessment of the role of direct air capture in deep mitigation pathways. Nat Commun 2019; 10(1):3277.

[22]	World energy outlook 2022. Report. Paris: IEA; 2022.

[23]	DataBank:world development indicators. Report. Washington, DC: World Bank Group; 2023.

[24]	Herzog H, Morris J, Gurgel A, Paltsev S. Getting real about capturing carbon from the air. One Earth 2024; 7(9):1477-80.

[25]	Terlouw T, Treyer K, Bauer C, Mazzotti M. Life cycle assessment of direct air carbon capture and storage with low-carbon energy sources. Environ Sci Technol 2021; 55(16):11397-411.

[26]	Deutz S, Bardow A. Life-cycle assessment of an industrial direct air capture process based on temperature-vacuum swing adsorption. Nat Energy 2021; 6 (2):203-13.

[27]	Breyer C, Fasihi M, Aghahosseini A. Carbon dioxide direct air capture for effective climate change mitigation based on renewable electricity: a new type of energy system sector coupling. Mitig Adapt Strateg Glob Change 2020; 25 (1):43-65.

[28]	Carter L. Private investment in climate action. Report. New York City: United Nations Development Programme; 2020.

[29]	Maniatis GI, Milonas NT. The impact of wind and solar power generation on the level and volatility of wholesale electricity prices in Greece. Energy Policy 2022; 170:113243.

[30]	Calero I, Canizares CA, Bhattacharya K, Baldick R. Duck-curve mitigation in power grids with high penetration of PV generation. IEEE Trans Smart Grid 2022; 13(1):314-29.

[31]	caiso.com [Internet]. Folsom: California Independent System Operator; c2025 [2025 May 25]. Available from: https://www.caiso.com.

[32]	nyiso.com [Internet]. Rensselaer: New York Independent System Operator; c2025 [2025 May 25]. Available from: https://www.nyiso.com.

[33]	ercot.com [Internet]. Austin: Electric Reliability Council of Texas; c2025 [2025 May 25]. Available from: https://www.ercot.com.

[34]	An K, Farooqui A, McCoy ST. The impact of climate on solvent-based direct air capture systems. Appl Energy 2022; 325:119895.

[35]	Sendi M, Bui M, Mac Dowell N, Fennell P. Geospatial analysis of regional climate impacts to accelerate cost-efficient direct air capture deployment. One Earth 2022; 5(10):1153-64.

[36]	Azarabadi H, Lackner KS. A sorbent-focused techno-economic analysis of direct air capture. Appl Energy 2019; 250:959-75.

[37]	Sinha A, Darunte LA, Jones CW, Realff MJ, Kawajiri Y. Systems design and economic analysis of direct air capture of CO₂. Ind Eng Chem Res 2017; 56 (3):750-64.

[38]	Leonzio G, Mwabonje O, Fennell PS, Shah N. Environmental performance of different sorbents used for direct air capture. Sustain Prod Consum 2022; 32:101-11.

[39]	Wurzbacher JA, Gebald C, Brunner S, Steinfeld A. Heat and mass transfer of temperature-vacuum swing desorption for CO₂ capture from air. Chem Eng J 2016; 283:1329-38.

[40]	Marinič D, Likozar B. Direct air capture multiscale modelling: from capture material optimization to process simulations. J Clean Prod 2023; 408:137185.

[41]	Keith DW, Holmes G, St Angelo D, Heidel K. A process for capturing CO₂ from the atmosphere. Joule 2018; 2:1573-94.

[42]	Schäfer M, Cerdas F, Herrmann C. Towards standardized grid emission factors: methodological insights and best practices. Energy Environ Sci 2024; 17 (8):2776-86.

[43]	Ruggles TH, Farnham DJ, Tong D, Caldeira K. Developing reliable hourly electricity demand data through screening and imputation. Sci Data 2020; 7 (1):155.

[44]	Ruhnau O, Schiele J. Flexible green hydrogen: the effect of relaxing simultaneity requirements on project design, economics, and power sector emissions. Energy Policy 2023; 182:113763.

[45]	Bachman E, Tavasoli A, Hatton TA, Maravelias CT, Haites E, Styring P, et al. Rail- based direct air carbon capture. Joule 2022; 6(7):1368-81.

[46]	Groom B, Venmans F. The social value of offsets. Nature 2023; 619 (7971):768-73.

[47]	Shah IH, Miller SA, Jiang D, Myers RJ. Cement substitution with secondary materials can reduce annual global CO₂ emissions by up to 1.3 gigatons. Nat Commun 2022; 13(1):5758.

[48]	Bhardwaj A, McCormick C, Friedmann J. Opportunities and limits of CO₂ recycling in a circular carbon economy:techno-economics, critical infrastructure needs, and policy priorities. Report. Columbia University Libraries; 2021.

[49]	Ramseur JL. Inflation reduction act of 2022 (IRA):provisions related to climate change. Report. Washington, DC: Congress.gov; 2022.

[50]	Sun L, Chen W. Impact of carbon tax on CCUS source-sink matching: finding from the improved China CCS DSS. J Clean Prod 2022; 333: 130027.

[51]	Victor N, Nichols C. CCUS deployment under the U.S. 45Q tax credit and adaptation by other North American Governments: MARKAL modeling results. Comput Ind Eng 2022; 169:108269.

[52]	Low MY, Barton LV, Pini R, Petit C. Analytical review of the current state of knowledge of adsorption materials and processes for direct air capture. Chem Eng Res Des 2023; 189:745-67.

[53]	Lackner KS, Azarabadi H. Buying down the cost of direct air capture. Ind Eng Chem Res 2021; 60(22):8196-208.

[54]

Sukprasert T, Bashir N, Souza A, Irwin D, Shenoy P. On the implications of choosing average versus marginal carbon intensity signals on carbon-aware optimizations. In: Proceedings of the 15th ACM international conference on future and sustainable energy systems; 2024 Jun 4-7; Singapore, Singapore. ACM; 2024. p. 422-7.

[55]	ERA5 hourly data on pressure levels from 1940 to present. Report. Reading: Climate Data Store; 2018.