To offset the cost associated with CO2 capture and storage (CCS), there is growing interest in finding commercially viable end-use opportunities for the captured CO2. In this Perspective, we discuss the potential contribution of carbon capture and utilization (CCU). Owing to the scale and rate of CO2 production compared to that of utilization allowing long-term sequestration, it is highly improbable the chemical conversion of CO2 will account for more than 1% of the mitigation challenge, and even a scaled-up enhanced oil recovery (EOR)-CCS industry will likely only account for 4–8%. Therefore, whilst CO2-EOR may be an important economic incentive for some early CCS projects, CCU may prove to be a costly distraction, financially and politically, from the real task of mitigation.
The continued growth in anthropogenic CO2 emissions would appear to be characterized by one word—inexorable. Despite a growing number of climate change mitigation policies, anthropogenic CO2 emissions in the period 2000–2014 grew at an average rate of 2.6% per year, in contrast with an average rate of 1.72% per year in the period 1970–20001,2. Indeed, in the period 2010–2014, emissions increased from approximately 31.9 to 35.5 GtCO2 per year; an average rate of 2.75% per year2. With the exception of a one-year reduction from 2008 to 2009, every year of this century has seen a year-on-year increase in anthropogenic CO2 emissions.
It has become commonplace to discuss future emission trajectories in terms of scenarios from, for example, the International Energy Agency (IEA) or the IPCC. Both the IEA and IPCC project that a world commensurate with no more than 2 °C of warming above pre-industrial levels is one in which total anthropogenic CO2 emissions are reduced to something less than 20 GtCO2 per year by 2050, with further reductions to near-zero or even net-negative emissions by the end of the century. This is typically referred to as the two-degree scenario or 2DS. At the other end of the spectrum, allowing anthropogenic emissions to increase to 60 GtCO2 per year by 2050 is commensurate with warming of approximately 6 °C above pre-industrial levels—this is the six degree scenario, 6DS1,3.
The conclusion one can draw from the foregoing data is that if anthropogenic emissions of CO2 continue along any of the recent growth trends, we are poised to very significantly overshoot the 6DS. To even meet the 6DS, we would need to reduce the annual rate of growth of emissions to 1.4% and to meet the 2DS, the rate of growth needs to be −1.5% if global emissions peak in the 2020s. If emissions peak later, the required rate of reduction similarly increases. For the remainder of this analysis, we hypothesize a world, inspired by recent success in Paris, that reduces emissions to a level commensurate with the 6DS by 2020 and aims thereafter to transition to a world commensurate with the 2DS, focusing on the period to 2050. This allows us to introduce the quantity mitigation challenge (MC), the amount of avoided CO2 emissions (against a reference case) by a given date, tf, in order to reduce emissions to a level commensurate with meeting the 2DS, E2DS. E2DS is a function of the year in which emissions peak, tp, the emission rate in that year, Etp, and lastly the rate at which CO2 would be emitted in tf according to a low mitigation scenario (LMS) reference scenario, ELMS. Therefore, MC can be expressed as equation (1):
In addition to being a function of tf, ELMS is also a function of tp, and the average rate of growth of anthropogenic CO2 associated with the LMS scenario in the period (tf–tp). Therefore, ELMStp = Etp(1+r)(tf–tp). Thus, in order to meet the IEA's 2DS with the 6DS as a baseline, it is necessary to avoid the cumulative emission of approximately 800 GtCO2 in the period to 2050 (Fig. 1).
Globally, despite an increasing emphasis on renewable energy, annual investment in fossil energy has more than doubled in real terms in the period 2000–2013, totalling more than US$950 billion at the end of this period4. It is therefore not unreasonable to suggest that fossil fuels will continue to be important to, if not dominate, the world's energy landscape for some time to come, with some estimates indicating that fossil fuels will still account for over 65% of the total energy mix in 21005, despite increasing penetration of renewable electricity generation6. For this energy mix to be coherent with the long-term ambition of substantially mitigating anthropogenic CO2 emissions, the widespread deployment of CCS technology7,8,9 will most likely be a vital part of the least-cost energy system of the future, working in conjunction with renewable energy to deliver energy which is low carbon, available, and affordable.
From one perspective, CCS is a readily deployable technology solution, relying on well-understood components7,8,9. Two leading options for decarbonizing both the power and industrial sectors are the oxy-combustion of fuel or post-combustion scrubbing of the exhaust gas arising from a conventional combustion process. Both of these technologies are highly mature. Alkanolamine gas scrubbing was first patented in the 1930s and has since been widely used for natural gas sweetening10. Oxy-combustion, which relies on the cryogenic separation of air, was developed by Linde in 1902 and was operating at 30,000 toxygen per day at the Shell Pearl gas to liquids project in Qatar in 2006. This is sufficient oxygen to supply a 2 GW oxy-combustion power plant. Similarly, CO2 transport and injection has been practiced at scale for EOR since the 1950s. As of 2014, there are over 3,000 miles of high-pressure pipeline which transport over 60 million tonnes of CO2 per year for EOR in 113 projects in the US alone, with approximately 120 projects worldwide11,12. Similarly, the distribution and capacity of CO2 storage locations are also reasonably well-characterized, with first order estimates of theoretical global CO2 storage capacity of approximately 11,000 GtCO2 (ref. 13). Of this, approximately 1,000 GtCO2 capacity is provided by oil and gas reservoirs with approximately 9,000–10,000 GtCO2 capacity provided by deep saline aquifers14,15,16. Furthermore, there is also significant potential capacity in unmineable coal seams, with the additional economic benefit that this is may be accompanied by the recovery of coal-bed methane.
In order to stabilize atmospheric CO2 concentrations at a level of 450 ppm, that is, a concentration consistent with a world with a high likelihood of not exceeding 2 °C of warming, it is expected that it will be necessary to store 120–160 GtCO2 via CCS in the period to 205017, with similar trends expected to the end of the century. Therefore we have more than enough CO2 storage capacity to meet this target and, even without identification of further storage sinks, sufficient to meet even ambitious CO2 sequestration needs for well beyond the next century, giving ample time for the likely lengthy transition from fossil fuels. Finally, the world's first commercial CCS-equipped power station has started operation at the Boundary Dam facility in Saskatchewan, Canada, with a second project also in operation in Alberta, where Shell are capturing the CO2 arising from H2 production18. CCS is inarguably a well-understood, mature technology that is deployable at commercial scale today.
However, despite CCS relying on well-known and well-understood technology components, the transition to its widespread deployment continues to be an uphill battle. The financing of this transition is a particular challenge, one which requires the combination of strong policy and price signals to ensure that low-carbon and energy efficiency investments offer a sufficiently attractive risk-adjusted return.
It is in this context that CCU is often mentioned. As a relatively benign material, it is possible to convert CO2 into a wide variety of end products, in addition to its potential for enhanced hydrocarbon recovery. In this context, therefore, why should we not actively and favourably consider the reuse of captured CO2?
Certainly it represents a beguiling opportunity—convert a waste product into high-value end products and kick-start a highly skilled regional manufacturing industry. Moreover, global demand for the potential products, such as methanol, appears healthy19.
Therefore, it is easy to see why the prospect of CO2 utilization is an attractive one for a wide variety of academic, industrial, and political stakeholders. However, serious questions arise when the narrative around CO2 utilization becomes one of utilization in parallel with storage or utilization instead of storage. As will be discussed subsequently in this paper, from the perspective of mitigating anthropogenic climate change, CO2 utilization is highly unlikely to ever be a realistic alternative to long-term, secure, geological sequestration.
The remainder of this paper is laid out as follows; we first discuss the scale at which various CCU options could be deployed, we then go on to discuss the rate at which they could be deployed before finally discussing how much of the CO2 used in the various options corresponds to permanent storage. In all cases, this is contextualized with reference to the aforementioned mitigation challenge.
It's a matter of scale
To put this in some perspective, current total global anthropogenic emissions are about 35.5 GtCO2 per year. Typical CO2 injection and storage conditions are approximately 10 MPa and 40 °C, corresponding to a CO2 density of approximately 600 kg m−3. This corresponds to approximately 1.64 × 108 m3 per day, or more than 1,033 million barrels (MMbbl) of CO2 per day. This is in contrast to current global oil production rates of approximately 87–91 MMbbl per day20,21. This means that global CO2 production today is approximately a factor of 10 greater than global oil production today, and, at current rates of growth, may be as much as a factor of 20 greater in 205022.
Given that CCS is expected to account for the mitigation of approximately 14–20% of total anthropogenic CO2 emissions, in 2050 the CCS industry will need to be larger by a factor of 2–4 in volume terms than the current global oil industry. In other words, we have 35 years to deploy an industry that is substantially larger than one which has been developed over approximately the last century, resulting in the sequestration of 8–10 GtCO2 per annum by 205022 with a cumulative CO2 storage target of approximately 120–160 GtCO2 in the period to 205017 and between 1,200–3,300 GtCO2 over the course of the twenty-first century13. This is an exceptionally challenging task, similar in scale to wartime mobilization, but it is a task we should not be daunted by. Neither should we be distracted by focussing too much on the long-term solution without giving sufficient attention to the short-to-medium-term necessity of fossil-fuel decarbonization in a manner that allows them to operate in sympathy with intermittent generation from renewable sources23
It is important to note that when CO2 utilization has traditionally been discussed, this has been in the context of CO2-EOR in the United States. In this paper we include CO2-EOR within a definition that considers any use of CO2, physical or chemical, that prevents immediate release of CO2 to the atmosphere as part of CCU. EOR is already a very mature technology with a history reaching back several decades, having well-defined techno-economic parameters, and is often considered to be an important part of the CCU landscape. In the early years of its development, CO2-EOR faced the challenge of relatively low oil prices and relatively high CO2 prices. Reservoir management was therefore optimized to maximize profit, not CO2 sequestration. At the time of writing, CO2-EOR provides approximately 5% of the total US crude oil production24, and whilst it has the potential to be appreciably expanded25, it is important to note the relationship between CO2 price and oil price. At oil prices of approximately US$100 per bbl, CO2 needs to be available at less than US$45 per tonne (ref. 12) for CO2-EOR to be economically viable. This is the case in the US, where the business model is very mature and the CO2-EOR capacity exists onshore, but this may not hold for the rest of the world. Thus, current oil prices in the range of US$40–60 per bbl and CO2 costs of US$60–80 per tonne (refs 26,27) make CO2-EOR less viable as a means of balancing the costs of large scale CCS operations, and separate economic or policy incentives are likely to be required.
Nevertheless, there is little question that CO2-EOR offers a large, near-term option to store large quantities of CO2 at lower net cost, with more than 90% of the world's oil reservoirs seemingly suitable for CO2-EOR12, if treated early enough, before the reservoir pressure drops below the minimum miscibility pressure. Thus, there exists the theoretical potential to produce 470 billion bbl of additional oil, corresponding to a cumulative theoretical CO2 injection capacity in the range of 70–140 Gt (refs 12,28).
However, this may be a highly optimistic estimate of the total deployable CO2-EOR capacity. As illustrated in Fig. 2, the majority of this capacity exists in the Middle East and North Africa and in the US at 50% and 13% respectively, whereas the estimated CO2-EOR in South Asia is essentially zero and the Asia Pacific region accounts for only about 3%.
In other words, there appears to be an unfortunate disconnect between regions of substantial CO2-EOR potential and those regions with the largest anticipated population growth, dependence on fossil fuels, and hence requirement to sequester CO2 over the course of the next century. In fact, the only regions where it appears certain that there is sufficient CO2-EOR capacity to meet the CO2 storage requirements to 2050 are the Middle East and Africa—although the requirements are close in North America and the former Soviet Union. Given the size and rate of growth of the CO2-EOR industry in the US, it is likely that the US will be a leader in the deployment of CO2-EOR. If we accept the availability of a CCS-derived stream of CO2 as a prerequisite for CO2-EOR, it would make sense to estimate the scale of likely CO2-EOR activities as matching regional CCS targets. Thus, a more realistic estimate is likely to be on the order of 40 GtCO2 cumulatively injected for CO2-EOR. Thereafter, if we consider the average CO2 footprint of a barrel of oil consumed, 0.43 tCO2 per bbl (ref. 29), this results in revising the above estimate down to approximately 35 GtCO2, or something in the range of 4.5% of the total CO2 mitigation challenge.
It is, however, important to further note that, given the appropriate incentives and regulatory environment, it is possible to operate a CO2-EOR operation so as to maximize the storage of CO2 per bbloil recovered30. This can have the effect of reducing the amount of oil recovered per tCO2 injected from approximately 3.33 bbloil per tCO2 to 1.11 bbloil per tCO2. At the lower end, once the CO2 emissions associated with the consumption of that oil are accounted for, this can result in the storage of up to 0.52 tCO2 stored per tCO2 injected, increasing the contribution of CO2-EOR to something in the range of 8% of the total CO2 mitigation challenge. A final point for consideration here is that oil derived from CO2-EOR could well displace oil that would otherwise be derived from unconventional sources which are known to have a CO2 intensity of 108–173% of conventional oil31. This displacement effect is estimated to be on the order of 80%, owing to market elasticities30. Therefore, assuming a constant demand, the deployment of CO2-EOR could lead to the avoidance of CO2 that would otherwise be emitted by the production of unconventional hydrocarbon resources, in addition to the reduced environmental and social risks of oil production via CO2-EOR in mature fields relative to unconventional hydrocarbon production.
Obviously, CO2-EOR is not the only route to CO2 utilization—there are also CO2 conversion options. There has been active interest in the chemical conversion of CO2 into platform chemicals, plastics, and other materials and fuels since the 1850s32,33,34,35 with the synthesis of salicylic acid, sodium carbonate via the Solvay process, and urea developed in 1869, 1882, and 1922 respectively36,37,38. It is therefore important to recognize that the focus on CO2 utilization is not a recent phenomenon. Overall, current annual global CO2 utilization is on the order of 200 Mt (ref. 35) and it has been suggested that this is likely capped at approximately 650–700 Mt in 2050 (ref. 33). Whilst this estimate was made in 2006, it is in line with current growth rates of the global chemical industry39. Further, of these conversion products, approximately 75% is accounted for by compounds which would not correspond to long-term sequestration of CO2 as the incorporated CO2 is released once the products are used. Therefore, given a 3% per year growth rate of CO2 utilization and a sequestration rate of 25%, this corresponds to a cumulative total of 15.42 GtCO2 utilized by 2050 and 3.86 GtCO2 sequestered—about 0.49% of the 800 GtCO2 mitigation challenge.
Mineral carbonation is another process that is under consideration40. Whilst this process does correspond to the effectively permanent sequestration of CO2 in a solid form, this is a reaction that happens naturally—albeit at an exceptionally slow rate. Accelerating the rate of these reactions requires mining (or other collection process), transporting, crushing, grinding and handling of vast quantities of material suitable for carbonation. This requires very large quantities of decarbonized electricity—which then begs the question: is there not a more profitable purpose to which we could put this decarbonized electricity—electrification of heating, or charging an electric vehicle, for example, and allow the carbonation of this material to take place naturally, noting that this may take an extremely long time?
Furthermore, whilst it is possible to convert CO2 into liquid fuels such as methanol for use in ground transport41, this would result in the near-immediate release of the CO2 to the atmosphere, and, although potentially reducing emissions relative to a baseline, cannot be considered to contribute directly and significantly to the CO2 mitigation challenge; capturing CO2 directly from a vehicle is unlikely to be feasible in the medium term.
Leaving the toxicity of methanol to one side, at 43–44 GJ per tmethanol (ref. 42), the energy required to convert CO2 into methanol is substantial relative to the energy density of methanol (19.7 GJ per tmethanol). This corresponds to an energy return on energy invested (EROEI)43 of approximately 0.45. More than 80% of this energy is associated with the generation of renewable electrolytic H2, with approximately 10% required for the capture of CO2 from a fossil-fired power station. If we were to consider the direct capture of CO2 from the air as the CO2 source, then one might expect the specific energy footprint of CO2-derived methanol to increase to the order of 60 GJ per tmethanol, or an EROEI of approximately 0.33. This represents a substantial quantity of renewable energy, which compares extremely poorly with the methanol's energy density (lower heating value basis), and could arguably be put to better use elsewhere.
By way of comparison, conventional coal and oil–gas production processes have an EROEI of approximately 46 and 20 respectively44,45, with wind, solar photovoltaic, geothermal, and biodiesel having an EROEI of approximately 18–20, 10, 9 and 2–5 respectively44,46.
Given that a fuel or energy needs an EROEI of at least 3 to be considered useful to society43,44, the energy required to produce methanol would have to be reduced by a factor of 6–10, depending on the source of the CO2, in order to become viable: this is a substantial challenge.
The relatively low energy density of methanol also presents substantial challenges to its use as a fuel. Gasoline has an energy density of 46.4 MJ per kg and upon combustion produces 3.09 kgCO2 per kg, whereas methanol has an energy density of 19.7 MJ per kg and upon combustion produces 1.38 kgCO2 per kg.
As can be observed from Fig. 3, owing to the reduced energy density of methanol, its use as a fuel will result in the emission of approximately 5% more CO2 than would have otherwise been the case.
Moreover, the processes for converting CO2 to methanol do not have a perfect yield. There will be some fraction of CO2 purged from the process—typical numbers are 0.08 tCO2 purged and 0.67 tmethanol produced per tCO2 feedstock42. Consider, then, that 1 bbloil will yield 19 gallons of gasoline, and supply 2,469 MJ per bbloil, therefore emitting 164.46 kgCO2 per bbloil. To deliver the same amount of energy requires 125.36 kgmethanol per barrel of oil equivalent. When this methanol is combusted, and accounting for the CO2 that was emitted in the initial production of the methanol, this corresponds to approximately 188 kgCO2 per barrel of oil equivalent or approximately 14% more CO2 than would have been produced had conventionally-sourced crude oil been used. This demonstrates the difficulty in using methanol production as a carbon sequestration process.
In order to compare CO2-EOR and methanol production on the basis of energy service, we first recall that, depending on the version of EOR practiced30, between 1.1–3.3 bbloil per tCO2 are produced and that each bbl will produce 12 gallons of diesel and 19 gallons of gasoline, which delivers 4,284 MJ per bbloil. In the default CO2-EOR case, 3.3 bbloil per tCO2 are produced and where the EOR operation is optimized for storing CO2, this is reduced to 1.1 bbloil per tCO2.
This leads to the net emission of 0.43 and −52 tCO2 per tCO2 injected, respectively and delivering 4,760–14,279 MJ per tCO2 injected or between 0.03 and −0.11 kgCO2 per MJ (Table 1). Displacing this service with CO2-derived methanol would require the production of 242–725 kgmethanol, leading to the emission of approximately 0.08 kgCO2 per MJ. Thus, from the perspective of both EROEI and a carbon balance, the utilization of CO2 for EOR would appear to be preferable to the conversion of CO2 to methanol. In all cases, CO2-derived methanol would appear to increase the quantity of CO2 emitted whilst delivering the same service and, under some circumstances, CO2-EOR can result in the net sequestration of CO2, whereas it does not appear that this is feasible with methanol.
It's a matter of time
A further point which must be taken into account is the period for which each utilization option actually stores the CO2. It is well-accepted that in order to mitigate the effects arising from anthropogenic CO2 emissions, it is necessary to permanently sequester the CO2 that is excess to the earth's carbon cycle. Chemicals such as urea or methanol store CO2 only until they are used; once urea is applied as fertilizer or methanol is used as a fuel, the CO2 is immediately released to the atmosphere—corresponding to a storage duration of perhaps six months. The conversion of CO2 into polymers might store the CO2 for several decades, perhaps as much as 50 years. This is in contrast to geological sequestration, which can be considered permanent.
It's a matter of rate
In order to reduce global CO2 emissions to 80% of 1990 levels by 2050, it will be necessary to reduce anthropogenic emissions by approximately 42 GtCO2 per year by 2050 compared to a 1990 baseline in line with the IEA and IPCC scenarios. To achieve this, it is anticipated that, amongst other things, it will be necessary to sequester a cumulative 120–160 GtCO2 in the period to 20503,15,22, or 16–20% of the cumulative mitigation challenge. This corresponds to a rate of CO2 sequestration of approximately 2.5 GtCO2 per year by 2030, increasing to 8–10 GtCO2 per year by 20503,15,22, with further increases in the rate of sequestration in the period to 21001.
As discussed previously, CO2-EOR is a potential sink for a substantial amount of CO2. One of the major barriers—if not the major barrier—to higher levels of CO2-EOR on a global basis is an insufficient supply of affordable CO2. In 2004, there was a supply shortfall of approximately 40 MtCO2 per year for CO2-EOR in the Permian Basin. Subsequently, between 2007 and 2010, an additional supply of approximately 5 MtCO2 per year was sourced in response to this demand28. This is very possibly the world's first example of a demand pull on anthropogenic CO2 capture. Recent years have seen a steadily increasing share of this CO2 supply being provided by anthropogenic sources; as of 2010 this was 12 Mt per year12. This represents a very significant rate of increase in the size of this industry, and we would cautiously suggest that a global rate of increase in CO2-EOR activity of 11% per year is feasible, given appropriate initial conditions such as secure supplies of CO2. From a baseline of approximately 0.06 GtCO2 per year used for CO2-EOR, this could grow to perhaps 26–27 GtCO2 per year in 2050. This could correspond to a cumulative total of approximately 40–60 GtCO2 injected, and 35–70 GtCO2 stored. As previously, this represents about 4–8% of the ∼800 GtCO2 mitigation challenge by 2050.
Concerning other options for CO2 conversion, data from some recent estimates of current and near-term market sizes is presented in Table 2. It should be noted that the two largest sinks for CO2—urea and methanol—do not correspond to storing CO2 for any significant period of time. Similarly, the technological category appears to be a catch-all for CO2 utilization in food and drink manufacture, fire suppression, as an inerting agent and dry ice, and other miscellaneous activities. Again, these options do not correspond to long-term sequestration of CO2.
It is worth considering for a moment the rates of growth implicit in the figures presented in Table 2. Given that the current rate of growth of the global chemical industry is approximately 3% per year39, it is difficult to accept that this could, in any way, be indicative of a long-term trend. Furthermore, there appear to be significant assumptions in these data35 surrounding the rate of displacement of CO2-derived products in the market. Other, more conservative estimates of CO2 utilization for the manufacture of chemicals place an upper limit of 650–700 MtCO2 per year on total global utilization33.This implies a growth rate of 3% year in the period 2010–2050, which is in line with the current rate of growth of the global chemical industry39. This would correspond to a cumulative total of 15.42 GtCO2 utilized in the period 2010–2050. As discussed previously, only about 25% of these products correspond to sequestering the CO2 for any significant duration: therefore this total is reduced to 3.86 GtCO2—or slightly less than 0.5% of the CO2 mitigation challenge of 800 GtCO2 by 2050.
Putting it in perspective
When we take these data and then compare them for the period to 2050, it becomes clear how negligible the contribution of CCU will be to the global CO2 mitigation challenge (Fig. 4).
This emphasizes the danger of reinforcing the narrative that CO2 utilization is key to making CCS profitable in a simplistic commercial sense. If this narrative continues, it introduces the very real risk that emission mitigation targets will not be met and that CCS through geological storage will not be deployed in any meaningful way. From a commercial and policy perspective, CCU should be encouraged when and only when CO2 is useful as a cheap feedstock, or when it can robustly and reliably shown that the CO2-derived product can reasonably displace the incumbent product, that is, deliver the same service at the same price, and also not result in an increase in the emission of CO2 associated with delivering that service. The driver should be feedstock substitution and the production of materials at a lower cost and with lower fossil carbon content. The primary driver should not be locking up CO2, as this can never happen at the required magnitude without geological storage.
Underpinning research into CO2 conversion should continue in order to expand options and reduce costs. CO2-EOR, whilst no panacea, can be deployed at a sufficient scale to facilitate the deployment of CO2 transport infrastructure and potentially stacked CO2 storage options. There is clearly a role for this technology to play in some early CCS demonstrations, as exemplified by the Sask Power Boundary Dam and the Air Products steam methane reformer projects in Canada and the United States, respectively. The key to climate change mitigation is scale, and it is generally accepted that the CCS cost reduction will be primarily achieved via deployment at scale47,48. Whilst CO2-EOR projects can be deployed at a sufficient scale to facilitate learning, leading to material cost-reduction, the same is not true for the majority of CCU technologies. Thus, from the perspective of mitigating climate change, CCU can, at most, be seen as supplementing CCS to a small extent. Any proposals for its large-scale deployment should be accompanied by a careful and thorough analysis of associated primary and associated opportunity costs.
About this article
Geoengineering: neither economical, nor ethical—a risk–reward nexus analysis of carbon dioxide removal
International Environmental Agreements: Politics, Law and Economics (2018)