Carbon dioxide direct air capture for effective climate change mitigation based on renewable electricity: a new type of energy system sector coupling

Recent research clearly indicates that 1.5–2 °C global temperature moderation pathways include fast and deep defossilisation of all energy sectors globally. However, past and present greenhouse gas (GHG) emissions are not compatible with directly achieving the target. Negative carbon fioxide (CO₂) emission technologies (NETs) are required to achieve the ambitious targets of the United Nations Framework Convention on Climate Change 21st Conference of the Parties, the Paris Agreement, (UNFCCC 2015) within this century. Kriegler et al. (2017) quantified the negative emission demand at around 10 Gt_CO2 in the 2050s, reaching about 20 Gt_CO2 later in the second half of the twenty-first century, which implies a need to start scaling up the NET capacities in the late 2030s and for massive investments in the 2040s.

Most integrated assessment models (IAMs) strongly rely on bioenergy carbon capture and storage (BECCS) while typically ignoring CO₂ direct air capture (DAC), as recently documented by Rogelj et al. (2018) and criticised by Williamson (2016) and further confirmed by the Intergovernmental Panel on Climate Change (IPCC) SR1.5 ( 2018). Nevertheless, the substantial obstacles for BECCS (Harper et al. 2018; Fajardy and Dowell 2018; Fridahl and Lehtveer 2018; King and Bergh van den 2018) may not allow a massive scale-up, since huge land areas are needed due to very low area efficiency, rise of water stress issues, low energy return on energy invested induce a burden for the entire energy system, the high cost of bioenergy-based solutions and last but not least societal barriers are expected to be substantial. And last but not least, BECCS as a base generation technology would have not much added value for an energy system mainly based on low-cost variable renewable energy (VRE) (Dowell and Fajardy 2017), i.e. solar photovoltaics (PV) and wind energy, which require flexibility and not an inflexible base generation. Recent research clearly indicates that future energy systems will be mainly based on very high shares of VRE, as pointed out by Jacobson et al. (2017), and can be even lower in cost than the present energy system as further emphasised by Breyer et al. (2018). The IPCC SR1.5 report (IPCC 2018) confirmed for the first time that 100% renewable energy (RE)-based systems shall be seriously taken into account. Brown et al. (2018) pointed out the technical feasibility and economic viability of 100% RE systems, in particular for the power sector. Sustainable bioenergy solutions are confirmed by most researchers in the field of high renewable shares (Breyer et al. 2018; Brown et al. 2018; Child et al. 2018), but not by all (Jacobson et al. 2017). A common view has been developed that full hourly resolution of energy system description is a required methodology of the best possible validation of results for highly renewable energy systems (Breyer et al. 2017a, 2018; Brown et al. 2018; Pursiheimo et al. 2018; Jacobson et al. 2018). The long underestimated role of VRE in climate change mitigation (Breyer et al. 2017a) may explain why electricity-based DAC solutions have not been much regarded in IAMs. However, recent progress in adopting real-world development in VRE, as initiated by Creutzig et al. (2017), may trigger more focus on DAC and also on IAMs for evaluating a more balanced solution space in the near future. This paper aims to examine the feasibility of renewable electricity-based DAC solutions.

DAC technology is on the rise in the debate on suitable NET options (Williamson 2016; Goeppert et al. 2012; Workman et al. 2011; Smith et al. 2016). DAC can combine several key features, in particular an excellent area footprint for large-scale deployment, no major conflicts with land use and an excellent match to the renewable electricity-based energy system of the future (Fasihi M, Efimova O, Breyer Ch. Techno-economic assessment of CO₂ direct air capture plants, submitted), which will be mainly based on solar PV and wind energy (Jacobson et al. 2017; Breyer et al. 2018). This implies additional key advantages, such as very low-cost energy supply, good energy system integration, access to areas of excellent energy resources and the potential to decouple the locations of DAC and electricity generation, if needed.

In the following sections, this research will focus on the Maghreb region with its excellent preconditions for low-cost RE and vast tracks of unused land. The preferred RE supply for DAC also matches the electricity supply in the region as supported by respective energy system transition analyses. The DAC cost is mainly presented for 2040 and 2050 considerations as the estimated scale-up years of the technology. It is emphasised that this analysis is the first to the knowledge of the authors based on full hourly and high spatial resolutions of the entire DAC system so that a deeper understanding of the required system design can be gathered.

The LUT Energy System model of LUT University is used to analyse RE-based DAC systems in the Maghreb region. The LUT model combines full hourly resolution and coverage of the world structured into 145 regions (Breyer et al. 2017a, 2018; Ram et al. 2017a), whereof the Maghreb region is highlighted in this research. This section is divided into the methodology for all relevant DAC aspects and a brief overview on the LUT Energy System Transition modelling.

This section is divided into results for the CO₂ DAC system modelling and the energy system transition towards full sustainability in the power sector for the Maghreb region.

The results reveal the outstanding potential not only for 100% renewable energy supply in the Maghreb region, but also for huge CO₂ removal in the region. The total LCOE for 100% RE supply of the power sector by 2050 is around 50 €/MWh for the historic cost of the energy transition, but would be around 40 €/MWh if the structure of the energy system achieved in the year 2050 would have been built with an overnight approach in the same year. The power supply for near baseload conditions of the new CO₂ removal sector is lower, at around 30–33 €/MWh in the Maghreb region at the best sites. The reasons for the cost difference are mainly three aspects: (i) the energy supply for an entire country is based on the usage of different sites and on a weighted mix of the 50% best sites in the country (Bogdanov and Breyer 2016); (ii) the load curve of a country has to be matched for all 8760 h of a year, whereas the DAC system does not require this level of energy supply security and is finally run for about 8300 h a year; and (iii) for a national energy system, there are also some higher cost must-run units included, such as waste-to-energy plants for efficiency and additional sustainability reasons (Breyer et al. 2017a, 2018; Ram et al. 2017a; Bogdanov and Breyer 2016).

The new CO₂ removal sector shows several characteristics already known from SWRO desalination, in particular the tendency for high FLh in the least cost operation mode. The FLh for DAC units are around 8300 in the cost-optimised operation, whereas SWRO desalination plants show even a tendency towards almost full baseload operation (Caldera and Breyer 2018). This indicates that not much efficiency gain due to sector coupling, but an almost parallel operation of these energy sectors, can be expected,.

The potential energy system impact of the CO₂ removal sector can be very dominant. The electricity generation for the power sector is expected to be around 440 TWh for the Maghreb region in 2050. This demand may be increased by SWRO desalination and non-energetic industrial gas demand by about 115% (Breyer et al. 2017a; Aghahosseini et al. 2016). The global CO₂ removal demand may be around 10 Gt_CO2 in 2050, which may lead to a contribution of the Maghreb region of 5–10% of this demand due to the excellent resource conditions. The CO₂ capture of 1 Gt_CO2 by 2050 would require at a top site in the Maghreb region the following DAC system:

These results reveal that the CO₂ removal sector would achieve the same size as the total energy system by 2050 in the Maghreb region for 10% of global CO₂ removal and about the same size as the demand in the power sector for 5% of global CO₂ removal demand. The electricity supply would be almost fully solar PV based, due to cost reasons.

The area demand for 1 Gt_CO2 annual capture is estimated to be about 5244 km², thereof 83% for the PV power plants, 6% for the wind power plants, 3% for the HP units and 8% for the DAC units. The specific area demand is assumed to be 75 MW/km² for solar PV (Bogdanov and Breyer 2016), 8.4 MW/km² for wind energy (Bogdanov and Breyer 2016), 1 GW_th/km² for heat pumps and 2.5 Mt_CO2,cap/a/km² for DAC units (Fasihi M, Efimova O, Breyer Ch. Techno-economic assessment of CO₂ direct air capture plants, submitted). The area demand is equal to 0.08% of the total area of 6.31 million km² of the Maghreb region; hence, even a 10% global CO₂ removal contribution is negligible in area demand for the Maghreb region. It needs to be mentioned that the DAC systems even produce some water as a by-product of CO₂ capture at least for moderate climate conditions, and the DAC system can be operated in a mode close to being fully automatized.

The CO₂ capturing cost shows a steep decline to about 105 €/t_CO2, 70 €/t_CO2 and 55 €/t_CO2 for the years 2030, 2040 and 2050, respectively. It can be assumed that some further cost reduction after 2050 can be realised due to further effects of economies of scale and learning curve aspects. It should be also mentioned that the underlying learning curve for the DAC units has to be called rather conservative, since a rather low DAC capacity demand of less than 11 Gt_CO2 by 2050 for all CO₂ removal and all carbon capture and utilisation (CCU) activities for power-to-fuels and power-to-chemicals globally is assumed (e.g. a substantial DAC-based CCU potential for power-to-fuels had been identified earlier for the Maghreb region (Fasihi et al. 2017)), as well as a rather moderate learning rate of only 10% (Fasihi M, Efimova O, Breyer Ch. Techno-economic assessment of CO₂ direct air capture plants, submitted). Fasihi et al. (Fasihi M, Efimova O, Breyer Ch. Techno-economic assessment of CO₂ direct air capture plants, submitted) clearly point out that more realistic assumptions, i.e. about 22 Gt_CO2 global capturing capacity by 2050 and a 15% learning rate, would lead to an additional DAC capex reduction from 196 €/(t_CO2·a) to 81 €/(t_CO2·a), or an additional reduction of close to 60%. This effect alone would translate to a LCOD reduction from 55.3 €/t_CO2 captured to 45.6 €/t_CO2 or about 18% cost reduction.

The cost assumptions for the two main electricity supply technologies may be also too high for the conditions of the year 2050, since the PV LCOE reaches about 12 €/MWh in 2050. This is almost the level which has been offered in the most competitive PV tender in Saudi Arabia in the year 2017, when 17.86 USD/MWh (14.3 €/MWh at an exchange rate of 1.25 USD/€) was offered by a consortium of internationally leading PV and power companies (PV-Tech 2017). The potential for solar PV in 2050 should be well below 10 €/MWh, which may reduce the LCOD by a further 5% or more.

The battery cost may be also further reduced by 2050, since the learning curve cost target would be only one-third compared to the assumed cost, which is equal to an additional cost reduction of about 10%. A power-to-gas-to-power (PtGtP) system was also included in the model as an additional option for seasonal electricity storage. However, for the conditions of the Maghreb region, no considerable capacities of this option were installed in the cost-optimised solution.

In the absence of cost numbers for heat storage at around 100 °C, the cost assumptions for high-temperature heat storage are used, which are most likely too high by at least a factor of 2, and represent a LCOD reduction of about 5%. The balancing role of heat storage could be different depending on the availability of electricity and heat and the demand period. For a periodical heat demand and a near to baseload electricity availability, heat storage makes it possible to have a smaller HP capacity. On the other hand, for a baseload heat demand, a heat buffer helps generate the heat when electricity is available. Therefore, by letting the HP run for lower FLh, its capacity would be more than that for the baseload case. In this scenario, the benefit comes indirectly from avoiding DAC shutdowns or scaling up the electricity generation or storage system. The systems in the Maghreb region operate in a condition closer to the second case, whereby most of the storage and balancing is provided by low-cost battery storage.

However, there is also a downside potential for the development of DAC units and respective LCOD. The major risk may be a too slow market deployment of DAC technology, not only for carbon direct removal, as investigated in this research, but also in particular for power-to-fuel and power-to-chemical options needed for defossilising the transport sector and chemical industry. Since DAC units can be used for all these markets, a substantial delay in reducing GHG emissions in these sectors would lead to a delay in capacity increase of substitution solutions, and this would automatically lead to a reduced and delayed DAC capacity demand, and thus a delay in achieving the mentioned DAC capex targets outlined in this research.

Summing up, the possible further cost reductions of the DAC units, PV systems, battery storage and TES account for about 40% of further cost reductions, which would translate from 55.3 €/t_CO2 captured to about 32 €/t_CO2 captured in a most cost-optimised but possible case in 2050. The applied WACC is 7%, which may be reduced to 5% for low-risk international cooperation activities, and would further reduce the LCOD by about 15% to levels well below 30 €/t_CO2 for captured CO₂ by 2050. Substantial ambitionless defossilisation targets for major GHG emitting sectors, in particular the marine- and aviation-based transport sectors and chemical industry, would be a burden for achieving the assumed cost targets. Significant slower market development of DAC systems would delay the expected cost targets, which in most unfavourable conditions may lead to cost levels in 2050, which would be achieved otherwise around 2030.

The results for the Maghreb region also reveal that the CO₂ removal industry may grow to be one of the largest industries on the planet, since the annual turnover may reach 300 to 550 b€ per year, based on the results for the Maghreb region, for 10 Gt_CO2 of removal per year. This turnover would be roughly equivalent to 0.27–0.50% of the global GDP, assuming a continued real GDP growth of about 2% per year, hence well affordable for the global society if a climate change survival strategy is the clear target of the people of the world.

The scope of this research has been the CO₂ capture by DAC units and is comparable to other research in the applied metric and system boundaries. However, it needs to be mentioned that gaseous CO₂ is not an intrinsic safe material to be stored over the period of many thousands of years. Leakage from geological storage has been studied and is summarised by Leung et al. (2014). They separate the cases for caprock leakage which is in the order of tens of thousands of years and leakage through permeable pathways which causes more concerns. In addition, the injection wells are identified as the most likely leakage pathway. The CO₂ injection also can cause fractures in the geologic formation which happened for caprock in a project in Algeria, and also in the world’s biggest storage site in Norway (Montastersky 2013). A leakage rate of already 0.1% per year has been identified as critical (Enting et al. 2008). Hazeldine (2009) describes conditions for leakage of CO₂ injections in geologic formations. Fogarty and McCally (2010) investigate CCS from a medical perspective and raise concerns about risks of CCS for human health. As long as remaining risks and uncertainties can be removed, it may be of interest for all to imply at least one further conversion step from gaseous CO₂ to a compound, i.e. solid material, which shall be chemically inert and which should have a high combustion point, comparable to SiC or carbon fibres. Such a conversion will require further energy and conversion techniques, but is indispensable for responsible long-term carbon storage. This conversion would be also required for all other CO₂ capture technologies, whether BECCS, point source carbon capture or other options, and therefore no differentiator in the CO₂ capturing industry. Research on CO₂ conversion to stable compounds for long-term storage is currently in its infancy.

All CO₂ capture technologies would need, to some extent, a collection system from distributed units and a buffer unit before utilisation, conversion, transportation or permanent storage of CO₂. Such a system would also have some additional costs and energy demand. It is important to note that the presented cost levels in 2040 and 2050 could be achieved only if the implementation of CO₂ capturing technologies starts in rather small scales in the 2020s in order to stimulate a learning curve. Otherwise, without an early start and learning all the options, there might be more economic barriers to achieve the outlined cost levels.

The targets of the Paris Agreement to limit the anthropogenic temperature rise to a 1.5–2 °C increase above pre-industrial levels imply not only net zero GHG emissions and thus a zero GHG emission energy system by the mid-twenty-first century, but also substantial permanent CO₂ removal. This has to be started in the 2030s and massively scaled up in the 2040s and 2050s. CO₂ DAC is among the most promising CO₂ removal technologies, also called negative emission technologies. IAM-based scenarios have neglected this option partly due to limited understanding of the real cost dynamics of RE technologies, and also a high focus on the BECCS option in the past, which raises increasing debate on a more balanced discussion on the various NET options. The water demand of DAC is in practical terms zero and the area demand is very low. Even more importantly, DAC systems can be built on barren land, as shown for the Maghreb region, which is dominated by desert environments. The area demand for 10% of global CO₂ removal demand in 2050 would require less than 0.08% of the Maghreb region for the full DAC system including full energy supply.

The costs for the captured CO₂ have been calculated as 105 €/t_CO2, 70 €/t_CO2 and 55 €/t_CO2 for the years 2030, 2040 and 2050, respectively, for areas of several tens of Gt_CO2 annual capturing potential. All calculations are based on full hourly modelling for all included sub-systems, which are mainly solar PV power plants, wind power plants, battery storage, heat pumps, thermal energy storage and DAC units. The DAC units are operated at near baseload conditions for the least cost of captured CO₂. The derived CO₂ capturing cost may be further reduced by about 50%, thereof by 40% pts due to reductions in sub-system costs, mainly DAC units, battery systems, PV systems and TES, and further by 10% pts due to lower than assumed weighted average cost of capital. This is because, in an international cooperation programme, the business risk can be reduced if guarantees are issued by countries.

The results show that CO₂ captured cost could reach levels of about 55 €/t_CO2 and even lower than 30 €/t_CO2 for the best conditions in the Maghreb region. Almost all required energy would be solar PV based, which is the least cost source of energy in the applied scenario and abundantly available in the Maghreb region. The additional energy demand of the CO₂ removal energy sector may grow to levels comparable to the entire energy throughput for the energy system. Not much sector coupling benefits are expected due to similar operation modes as found earlier for SWRO desalination plants which are also operated in a least cost mode at near baseload operation. Concerns about risks of geological storage of gaseous CO₂ may require further conversion to solid compounds with additional energy and area demand, which may not be so costly, as already documented for the sustainable CO₂ removal based on renewable electricity and DAC units.

Key global climate change mitigation strategy options can be derived from the structural insights for the Maghreb region. An energy transition in the power sector leads to very high shares of RE, as a consequence of environmental sustainability, technical feasibility and economic viability. For achieving the targets of the Paris Agreement, NETs are required and CO₂ DACCS represents a very attractive NET option, which can be implemented globally and scaled to very large volumes, if needed.

CO₂ removal may be an outstanding economic opportunity for the Maghreb region due to excellent resource conditions, strong CO₂ removal demand and a strong competitive edge of the Maghreb region for CO₂ removal.