As I am writing these lines, the supposed climate leaders of the world meet at the COP 26 in Glasgow to find ways to ramp up their national commitments towards the well below 2°C climate target.
Since most of our energy consumption and industrial production still hinges on fossil fuels, the structural changes that lie ahead of us are huge and make the efforts so far (if any) look tiny.
Yet the solutions are there. The basket of climate change mitigation options, strategies to reduce greenhouse gas emissions from burning fossil fuels, industrial processes, and agriculture, is full. On the supply side, options include renewable low carbon energy in all its forms, such as wind energy, solar photovoltaics, or hydropower. On the demand side, there is a wide range of measures, mainly energy efficiency, electrification (switch from other fuels to using electricity), new services or shifts of transport modes. Many technological solutions have matured, supporting policies been developed, and market volumes for wind, PV, and electric vehicles are on the rise.
So what is the problem? Well, most countries have delayed ambitious climate change mitigation for so long that we have now run out of time. The scale-up of the different mitigation strategies requires huge economic and societal transformations, but too many fossil fuel assets still have some operating time left before they will retire, creating a huge lock-in of already committed emissions [1]. To break this lock-in, entire industries (like coal-fired power generation) need to disappear and new ones (like new vehicle use and appliance sharing solutions) need to be tested and scaled up quickly. That transformation requires tremendous efforts and meets massive resistance by incumbent players in many places.
But what if there was an easy way of doing it? Some techno-fix that would help us continue business as usual? This is where ‘negative emissions technologies’ come into place. The idea is simple. A chemical plant captures CO2 from the exhaust of a power station or even from ambient air (where the concentration of CO2 is now around 413 ppm), produces a stream of pure carbon dioxide, which is then compressed, transported and stored (‘sequestered’) in some rock formation where it hopefully will stay forever. This technology is called carbon capture and storage (CCS) in the case of the power station and direct air carbon capture and storage (DACCS) for the case of ambient air. It actually works already beyond the lab scale and there are DACCS demonstrators with a capture capacity of up to 4000 tons per year [2].
For me as a system scientist, the question is now: With DACCS being feasible from an engineering point of view, what are its system-wide implications and is the technology fit for scale-up?
Together with my team member Kavya [3], Sumukha, a former master student, Felix Creutzig [4], and supported financially by the Eva Mayr-Stihl Foundation [5], we set out to address these questions. We combined process modelling with state-of-the-art life cycle assessment (LCA), conducted a detailed sensitivity analysis, and compared a hypothetical 1 gigaton DACCS scale-up with other, more established uses of low carbon energy. The result of this effort, that started with Sumukha’s master thesis that she handed in in early 2019 and that stretched out over several years, is now published in compact form in Nature Energy [6]. It is the first detailed comparison of major DACCS technologies and of DACCS with other mitigation options, and directly builds upon a number of fantastic assessments that were published before [some core citations: 7,8,9,10].
Two direct air capture (DAC) technologies, one based on absorption of CO2 by KOH, transfer to CaCO3, and then release by calcination at high temperatures (beyond 900°C), and one using amines (-NH2) to capture CO2 and release it between 100°C and 120°C, were assessed and compared. A lot of work went into compiling data on the size of the different components, like fans, reactors, pumps, and pipes, their material composition, their energy consumption, and – crucial – the material and energy efficiency of the different process steps. Using our best knowledge, we established a material and energy consumption inventory of the different technologies, and we performed a sensitivity analysis on those crucial parameters whose exact value is unknown or can be flexibly be set. An example of an unknown parameter is the exact sorbent recovery rate for the high temperature process, and an example of a flexible parameter is the carbon intensity of the electricity and heat supply of the DAC plant.
Importantly, we not only asses the energy and climate implications of these technologies but also look into material consumption (both for the assets and for sorbents), particulate matter emissions, water use, and land occupation.
The central numerical results and all the data we used can be found in our paper [6]. Here, I want to summarize our main findings, focusing on qualitative findings and make an attempt to place DACCS into the wider spectrum of climate change mitigation options.
Carbon capture efficiency: The main question is of course: How does the overall carbon balance of the technology look like? By how much do the GHG emissions from manufacturing and installing the assets, energy supply and sorbent production diminish the gross emissions reduction by the DAC step? Or, in different words, what is the life cycle GHG emissions of the technology? It should be clearly negative, as the functional unit (LCA term the useful stuff a technology does) is 1 ton of CO2 removed from the atmosphere. Here, our work confirms others and shows that with a low-carbon energy supply, the high temperature and the low temperature DAC processes have a net carbon removal of up to 73% and 86% per ton of CO2 captured and stored. That means that the net climate impact for 1 ton of CO2 captured and stored is down to -730 kg (high temperature) and -860 kg (low temperature). These technologies are clearly life cycle carbon negative, and the low temperature process systematically outperforms the high temperature process by a factor of 1.3–10 in all environmental impact categories studied. This makes sense because the low temperature process not only requires less operational energy, which dominates life cycle impacts, but also requires low temperature heat only, which is easier to obtain from sunlight and for which even heat pumps can be used. To be fair towards the high temperature process one must say that this is a design based on the premise of using off-the shelf chemical engineering solutions to guarantee a rapid scale-up.
So, yes! From the life cycle GHG perspective, DAC has matured to a level where it can become part of the spectrum of mitigation solutions. But there are many other facets to consider when scaling up such a material and energy-intensive process! The following aspects need particular attention:
Impacts other than climate: Low-temperature DACCS needs 3-4 times more materials than the mitigation options, and it has a similar land occupation and water use (see Table from the paper below). This finding substantiates a trend: many climate strategies are very material-intensive, and thus also capital-intensive. But these numbers are not infeasible.
Table source: Madhu et al., 2021, [6].
Risk: Next to the sorbent cycling rate and the carbon intensity of the operating energy mix, the risks and the overall potential of geological storage is the largest remaining uncertainty for DACCS. A recent report on a large Australian carbon storage project explains the difficulties this technology faces. “Each project will need to be carefully tailored to the precise geologic circumstances of the reinjection site.” [11] and concludes that “rolling out CCS rapidly and at gigatonne scale in many hundreds of places around the world is not easy to envisage. We are still in the stage of CCS experimentation, and are well before a standardisable and inexpensive approach can be widely used.”
Performance compared to other mitigation options, costs: Two inputs/factors for DACCS are particularly scarce: (a) low carbon energy (everybody wants it!) and (b) ample low-risk storage locations. The figure from the paper below addresses point (a), where we show how much overall GHG savings (y axis) can be achieved by using a given amount of low-carbon energy (x-axis). The steeper the lines, the more ‘climate-friendly’ the energy use is.
To our surprise, the low temperature DACCS is roughly as good as the good vehicle fleet electrification and even slightly better than using heat pumps for buildings. That means it is not ruled out by efficiency per se, like the poor life cycle performance of synthetic liquid fuels compared to battery-electric propulsion.
But the required low-carbon energy input will continue to be scarce and the other solutions provide a direct service AND reduction of GHG emissions, rather than just being a waste removal, what CCS is. These emissions savings plus the service provided should clear the business case for the mitigation options rather than for the burn first and remove later-options.
Figure source: Madhu et al., 2021, [6]. Inspired by figure 4 in [12].
Summary:
Our findings show that DACCS energy use is not prohibitive, corroborating earlier findings, and that material consumption is high but not a real limit given current material production volumes. From the life cycle perspective on technology, we conclude that DACCS should be further developed so that technology learning will lead to increases in efficiency.
Should we now all go for/wait for DACCS? No! While the mitigation options like battery vehicles and heat pumps are proven and currently rolled out on the markets, CCS and DACCS are still in their infancy. Moreover, it is by no means clear how the required large-scale geological storage of the captured CO2 streams will work and when it will become available.
Thus, one must not fall prey to the fallacy that DACCS is an established competitor to the established mitigation options. It is not. Especially not in the coming years, which are crucial for climate change mitigation.
Moreover, as long as low-carbon renewable energy is a globally scarce resource and the many established emissions mitigation options meet a huge unseized deployment potential, mitigation can be expected to be more cost effective and create larger environmental co-benefits than a DAC roll-out.
This has profound implications for climate policy: During the coming two decades, we will have to do it the hard way and actually transform large parts of our energy infrastructure.
Meanwhile, rich people might use the DACCS prototypes to offset their emissions (while increasing their material, water and land footprint! So if you don’t like that, you’d better get a bicycle!) and that could provide the funding for the further development of the DACCS technology.
In several decades from now, DACCS may play a larger role in compensation emissions that are otherwise too expensive or even impossible to mitigate, like residual GHG emissions from agriculture, cement plants, and the chemical industry.
Summarizing the summary, we arrive at the title of this piece: We conclude from our research that DACCS is a potential building block, but not a wildcard or silver bullet to fighting climate change.
References:
[1] Tong et al. (2019), https://doi.org/10.1038/s41586-019-1364-3
[3] https://www.indecol.uni-freiburg.de/en/team-contact-network
[4] https://www.mcc-berlin.net/ueber-uns/team/creutzig-felix.html
[5] https://eva-mayr-stihl-stiftung.de/en/foundation
[6] https://doi.org/10.1038/s41560-021-00922-6
For a read-only version, check: https://rdcu.be/cAnay
For a copy of the work or additional material, please contact Kavya or me directly: https://www.indecol.uni-freiburg.de/en/team-contact-network
[7] Deutz and Bardow, 2021: https://www.nature.com/articles/s41560-020-00771-9
[8] Realmonte et al., 2019: http://dx.doi.org/10.1038/s41467-019-10842-5
[9] Keith et al., 2018: https://doi.org/10.1016/j.joule.2018.05.006
[10] Holmes and Keith, 2012: https://doi.org/10.1098/rsta.2012.0137
[11] https://www.carboncommentary.com/blog/2021/7/30/the-struggles-to-make-ccs-work
[12] Kätelhön et al., 2019: https://doi.org/10.1073/pnas.1821029116
