Batteries or Silos: Optimizing Storage Capacity in Direct Air Capture Plants to Maximize Renewable Energy Use

The Paris agreement of 2015 aimed to constrain global warming to under 2 °C by 2100. Several mitigation options to cut down on emissions have been explored including the following: replacement of fossil fuel energy with renewable energy (RE) sources, deployment of post-combustion carbon capture and storage (PCCS) technologies and adoption of emission-free transport alternatives such as electric vehicles. However, there is a growing consensus among researchers that cutting down on emissions alone will not achieve this vision [1], [2]. This has necessitated research into technologies that remove already emitted CO2 from the atmosphere, also called Negative Emission Technologies (NETs) [3]. One of the innovative solutions being proposed and piloted is direct air capture (DAC) of carbon dioxide, which involves the capture of existing CO2 from the air independent of the production source.

DAC is preferred over most NETs because of its modularity, scalability, controllability, and its independence from emission sources [4], [5]. From a policy perspective, DAC plants offer the needful traceability of the captured mass of CO2 and introduce no known disruption to industrial processes; thus, they are seen as a feasible option for global emergency-type deployment to avert a climate crisis [6]. Furthermore, unlike PCCS that is deployed at emission sources, DAC offers partial decoupling of the capture process from the main grid thus relaxing the technical limitations associated with deployment of new technologies into the grid [4]. Among the DAC technologies being proposed in the literature, two have been piloted at industrial scale, i.e, the solid sorbent and liquid solvent technologies. The solid sorbent process has favorable energetics but suffers from the difficulty to produce large low-cost sorbents which must periodically be closed from the surrounding air during regeneration [7]. Aqueous solvent DAC process has the potential for operating at high capacity factors, but has an energy-intensive, high-temperature regeneration process [5]. The focus of this research is on the liquid solvent DAC because it is reported that it has a lower levelized cost of capturing CO2, it could achieve higher capacity factors and it could be easier to operate at large scale compared to solid sorbent DAC [8].

Despite its great potential to aid the net-zero emission mission, liquid solvent DAC is an energy intensive process that requires large amounts of stable power making it incompatible with intermittent renewable energy (RE) sources. A commercial-scale (1M ton-CO2) liquid solvent DAC plant requires 200–300 MW of steady electricity to meet its capture targets [7], [9], [10]. Large-scale deployment of DAC plants into the grid would require expensive power network upgrading, which may not favor investment in RE. An et al. [11] studied DAC energy consumption under different climatic conditions and showed that supplying DAC plants using natural gas isolated from the grid is close to 10% cheaper than connecting the plants to the grid. This is because the energy-intensive regeneration process could be achieved by burning natural gas directly without the inefficiencies associated with first converting the gas energy into electricity [10]. Most of the studies on DAC assume 90% availability of the DAC plant which is technically infeasible if the plant were to be powered using low-capacity factor RE such as solar and wind without deploying an expensive energy storage technology [7], [11]. This study delves into making the liquid solvent DAC technology adaptable to intermittent RE by investigating two most plausible storage options, i.e, storage of energy in batteries and storage of solids in silos. While solvent-based DAC employs calcium looping (Ca-L) to regenerate the solvent, the aspect of integration of storage of solids in silos – which is already being tested in flexibilization of PCCS – is yet to be explored.

DAC is different from PCCS in that the technology is used to trap CO2 from ambient air where its concentration is just about 0.04% [4] while PCCS captures emitted CO2 from the source of the emission like a fossil fuel power plant (FFPP) where its concentration is above 15% [12]. Recently, researchers have been exploring the use of flexible Ca-L in PCCS [12]. This makes PCCS to have some structural similarities with the liquid solvent DAC, which also employs Ca-L [7], [11]. Ca-L is a cyclic and reversible process of using a calcium compound to react with another chemical in a mixture to separate it from the mixture and later regenerating the separated substance by chemical, electrochemical or thermal means. Ca-L is used in PCCS, DAC and thermochemical energy storage (TCES). In the case of PCCS, CaO reacts with CO2 found in flue gas from a fossil fuel power plant (FFPP) to form CaCO3 in a carbonator, which is later thermally decomposed at high temperature (about 900 °C) in a calciner to liberate the CO2 and free the oxide for further use in the cycle [13]. In liquid solvent DAC, the solvent first reacts with CO2 in a contactor forming a carbonate in solution. The carbonate is then reacted with a solution of CaOH2 to precipitate CaCO3, which is heated in a calciner producing CO2 and CaO [7]. Ca-L is preferred in these processes because it uses chemicals that are readily available in nature, nontoxic and have high energy density [14]. Recently, making the high-energy Ca-L applications flexible with the prevailing energy environment has attracted a lot of research whose output inform the proposals in this paper.

The use of flexible Ca-L for PCCS is motivated by the fact that the current energy systems have more FFPP being used, not to power the base load as conventionally done, but as back-up to the intermittent, non-dispatchable RE sources [12]. The desire to make the Ca-L process adaptable to the new variable output power plants has attracted a lot of attention. For instance, in 2016 the European Union released close to 1.5 million Euros to finance the FlexiCal project intended to make calcium looping flexible to enable PCCS operate flexibly in backup FFPPs that operate at low-capacity factors [15]. Similarly, in 2020 the US Department of Energy through its Advanced Research Projects Agency – Energy, rolled out the FLExible CCS (FLECCS) project aimed at making PCCS process flexible with variable output FFPPs [16]. In both projects, the idea of introducing storage silos for CaO and CaCO3 in the Ca-L to transform it into a Ca-L with storage (Ca-LS) became dominant [13], [17], [18].

Typically, two fluidized bed reaction chambers are used in a Ca-L process, namely, a carbonator and a calciner. The carbonator uses solid CaO to capture CO2 from the flue gas (at about 15% concentration) from the FFPP. The calciner regenerates CO2 (at about 99% concentration) from the resultant CaCO3. Ca-LS introduces flexibility to Ca-L process flow by means of two storing silos; one that stores CaO and another that stores CaCO3 both maintained at selected temperatures [17]. Ca-LS modifies the reactant flow dynamics between the two main reaction chambers to enable the calciner to operate uninterrupted despite the intermittent supply of flue gas at the carbonator. A simplified Ca-LS flow diagram is shown in Fig. 1 [12].

In the case of TCES, Ca-LS has been proposed to enable large-scale energy storage to make concentrated solar power generators (CSPGs) dispatchable. This idea has been explored in over 63 research papers under the SOlar Calcium-looping integRAtion for ThermoChemical Energy Storage (SOCRATCES) project financed by the European Union [19]. The main solution proposed and piloted in this project is the use of a solar-powered Calix Flash Calciner (CFC) to facilitate the endothermic decomposition of CaCO3 during high solar irradiance hours, storing the resultant CaO and CO2, and later reacting the two in a carbonator to liberate heat in an exothermic process [20], [21], [22], [23]. The heat liberated in the process is used to generate electricity at night.

The aforementioned projects have given a fair groundwork for developing adaptable PCCS and TCES systems. This paper considers applying the Ca-LS concept to enable flexibilization of the energy-hungry liquid solvent DAC process. In such a DAC plant, the contactor which captures CO2 from air is supposed to operate without any interruption in order to achieve economically viable results. The calciner then regenerates the captured CO2 at high concentration (about 99%) for sequestration. The main problem faced in powering the DAC plant is that the regeneration is an energy-hungry process that consumes more than 90% of the total energy supplied to the plant [7]. Studies have estimated that sustaining the two process would require north of 200 MW of constant electric power supply for a 1M ton-CO2 DAC plant [7], [9], [10], [24]. It is currently technically infeasible to provide such kind of power from RE sources, which are intermittent by nature. Furthermore, connecting such a high-power load to the current grids would not only be technically challenging but also counterproductive because modern power grids are carbon-intensive.

This study builds on the Ca-LS idea to develop an energy optimization model for a liquid solvent DAC process with storage (LSDAC-S) with the aim of making the DAC plant operations adaptable to the current and future energy infrastructure dominated by RE. Other methods of optimizing Ca-LS such as energy saving methods, and optimization of process kinetics have been reported in [22], [25]. Astolfi et al. [26] designed a control model for flow rates of solid silos integrated into PCCS in FFPP using simple rule-based methods. More recent research have employed mathematical programming for optimization of design and operations of PCCS focusing on the process thermodynamics without the inclusion of storage of solids in silos [27], [28], [29]. Moreover, most studies that schedule DAC as part of energy systems assume flexible DAC operations without mentioning how this flexibility could be achieved [9], [30]. This work focuses on making DAC operations flexible to intermittent RE by introducing storage in the calcium loop and optimally scheduling the operations using a linear programming (LP) model. Unlike previous flexible models that concentrated either on PCCS for low-capacity factor backup FFPPs or TCES with concentrated solar power, this work delves into the operational flexibilization, and optimization of a standalone LSDAC-S supplied by a hybrid energy system with two different storage technologies. The contributions of this paper are as follows:

• •
  remodeling the chemical process flow network of a liquid solvent DAC plant to include storage silos in the calcium loop and allowing for flexible operations;
• •
  development of an LP optimization model that matches energy supply and material flow in the LSDAC-S through storage of both solids in silos and energy in a battery energy storage bank (BESB);
• •
  development of novel mass balance constraints for the material flow in the LSDAC-S plant to ensure it meets the target CO2 removal rate, making the LP model easy to incorporate into larger net zero expansion and planning models for power networks;