Economics of carbon dioxide capture and utilization—a supply and demand perspective

Lately, the technical research on carbon dioxide capture and utilization (CCU) has achieved important breakthroughs. While single CO2-based innovations are entering the markets, the possible economic effects of a large-scale CO2 utilization still remain unclear to policy makers and the public. Hence, this paper reviews the literature on CCU and provides insights on the motivations and potential of making use of recovered CO2 emissions as a commodity in the industrial production of materials and fuels. By analyzing data on current global CO2 supply from industrial sources, best practice benchmark capture costs and the demand potential of CO2 utilization and storage scenarios with comparative statics, conclusions can be drawn on the role of different CO2 sources. For near-term scenarios the demand for the commodity CO2 can be covered from industrial processes, that emit CO2 at a high purity and low benchmark capture cost of approximately 33 €/t. In the long-term, with synthetic fuel production and large-scale CO2 utilization, CO2 is likely to be available from a variety of processes at benchmark costs of approx. 65 €/t. Even if fossil-fired power generation is phased out, the CO2 emissions of current industrial processes would suffice for ambitious CCU demand scenarios. At current economic conditions, the business case for CO2 utilization is technology specific and depends on whether efficiency gains or substitution of volatile priced raw materials can be achieved. Overall, it is argued that CCU should be advanced complementary to mitigation technologies and can unfold its potential in creating local circular economy solutions.

CO 2 problem are largely insufficient. The development of technically, environmentally, and economically viable ways of utilizing CO 2 as a feedstock for industrial production can imply a complementary route to existing mitigation strategies such as the deployment of renewable energy and other green technologies.
As almost all materials that surround us in our everyday lives are carbon-based, the option to use recovered CO 2 to substitute fossil carbon sources remains an attractive possibility that could be worth further investigation. In the last years, several public funding programs, e.g., by the US Department of Energy or the German Federal Ministry of Education and Research have encouraged research in this field and already the first technological breakthroughs and advances to a demonstration scale can be observed (Federal Ministry of Education and Research 2014;US DOE n.d.). Even though the first CO 2 -based products are just entering global markets in the near future their number and scale is expected to grow (Aresta et al. 2013). Consequently, based on a literature review, the potential supply and demand of the commodity CO 2 is presented in this paper to discuss the fundamentals of the commodity CO 2 from an economic perspective of comparative statics.
Since the utilized CO 2 in most cases is reemitted at a later point in time a simple aggregation of the used volumes of CO 2 is not an indicator of ecologic performance (von der Assen et al. 2013). Instead, a detailed environmental analysis is necessary to calculate the real carbon footprint of a certain CCU technology compared to a conventional technology (von der Assen et al. 2015). Indeed, the same principle applies to the business case of CO 2 utilization. In some cases, using comparatively cheap CO 2 as a feedstock and replacing more costly and volatile priced fossil-based raw materials can lead to a cost reduction which sets the business case for CCU. However, for those production processes that use CO 2 still inefficiently or are not competitive to conventional fossil-based production, there is no business case until further research and development or political incentives prove otherwise. While CO 2 can generally be used in many processes, this paper focuses on potential commodity CO 2 from industrial capture and does not include biological fixation and conversion via the cultivation of crops or algae for example for making biofuels.
Supply side: potential sources and cost of CO 2 The potential sources of waste CO 2 emissions are numerous. Industrial plants emit CO 2 in different quantities and at diverse qualities. Several capture technologies can be applied, for example adsorption, absorption, cryogenic separation, or membranes (de Coninck and Benson 2014). The costs of capturing CO 2 at a certain source depend on the technological efforts that must be undertaken to collect the CO 2 in the required quality from the industrial exhaust gas. Thus, the costs are largely influenced by the concentration of CO 2 in the exhaust gas. Moreover, the CO 2 needs to be purified and any toxic or hazardous chemicals removed (Aresta and Dibenedetto 2010). Furthermore, a larger plant size can lower the investment and operating costs per captured tonne of CO 2 through economies of scale (Faulstich et al. 2009;Möllersten et al. 2003). Consequently, despite technical feasibility, not all emitting sources represent economically viable options at current conditions. Table 1 summarizes current data on respective global emission volumes, concentrations, estimated capture rates, and benchmark capture costs per type of source for the largest point sources of CO 2 based on a literature review. The presented data are based on several selected sources with heterogeneous technological and financial assumptions, calculation methods, and reference years. Therefore, the collected data do not allow for a detailed cost comparison or technological discussion. Nevertheless, it is useful to provide general insights on potential large-scale supply of CO 2 as a commodity. The presented emission volumes represent recent global direct CO 2 emissions. The capturable emissions are calculated by applying the estimated lowerbound capture rate as a benchmark to the global emission volume per type of source. The benchmark costs represent minimum cost of CO 2 captured per tonne that are possible to achieve with a certain benchmark technology, a so-called best practice process. The purity of the recovered CO 2 depends on the respective process but can be assumed between 95 and 99.9 %. The detailed underlying assumptions are further explained in the following and in Table 2.

Costs of CO 2 capture
Today, CO 2 capture is technologically feasible and industrial practice on a small scale around the world. However, due to a lack of incentives, large-scale capture is currently not economically viable. Hence, the costs of capture are essential when considering potential sources and technologies for recovering CO 2 emissions. Capture costs are generally defined as the costs of CO 2 separation and compression at a single facility (e.g., an industrial plant), disregarding any costs of transport, storage, or further conversion steps (Metz et al. 2005). They are usually derived from comparing a system with CO 2 capture to a reference system without capture. In the literature, two main measures for CO 2 capture costs exist: costs of capture and costs of avoidance of CO 2 .   According to the IPCC (Metz et al. 2005), the two measures are clearly defined as follows: (a) Cost of CO 2 captured represents the rather straightforward cost of capturing per amount of captured CO 2 . They can be determined through formula (1): These capture cost can reveal Bthe viability of a CO 2 capture system given a market price for CO 2 (as an industrial commodity)^ (Bogner et al. 2007). Consequently, if these costs can be reimbursed, e.g., through CO 2 utilization options or political incentives such as a carbon tax, then carbon capture could make economic sense. Due to this information value, costs of capture are the preferred measure in this study.
(b) Cost of CO 2 avoided gauges the effect of the overall emission reduction by calculating the cost of capturing per amount of CO 2 reduced compared to a reference process. The respective formula is (2): The cost of CO 2 avoided considers actual emission reductions and thus contains more ecologically relevant information. As the process of capturing usually requires additional energy and decreases the plant's efficiency, the capturing often produces additional CO 2 emissions. Therefore, the amount of CO 2 reduced or avoided compared to a reference system will largely be smaller than the amount of CO 2 captured. Consequently, the cost of CO 2 avoided will generally be higher than the cost of capture (Metz et al. 2005). In an optimal capturing case, where additional energy and emissions can be avoided, both measures will be equal. 1 The IPCC recommends the cost of avoidance especially for complete carbon capture and storage (CCS) systems and less for capture only analysis (Metz et al. 2005). For the market perspective of this paper, the value of the commodity CO 2 is however more important. Moreover, it is recommended, that any utilization technology should be accompanied by an LCA that measures the full ecologic impact of all production steps.
Unfortunately, the two measures are often commingled in the literature, so that a clear differentiation is not always possible. The preferred cost estimate of this study is the cost of CO 2 captured. Deviating cost measures were included where necessary. As described above, the cost of CO 2 avoided are potentially higher than the capture costs since the reduced emissions in the denominator are smaller than the captured emissions for most capture systems. The difference depends on the additional amounts of CO 2 emissions caused by the capturing efforts and how they are accounted for. This cost difference is expected to increase more or less proportionally with the efforts that are necessary for capture at the respective sources. Thus, for sources with higher capture costs, the difference in costs of capture and avoidance should be larger. Moreover, even if a measure is clearly and consistently selected in an assessment, varying assumptions and system boundaries limit the possibility of comparing cost data across studies (Metz et al. 2005). Instead, a comparison would only make sense, if the calculation methods and underlying assumptions were fully transparent and either consistent or could be aligned. To improve the measurement quality and comparability of capture costs-within a single study as well as across studies-the reference systems with and without capture should ideally be based on the same assumptions. Firstly, general conditions such as reference year, region, and type of data (real or hypothetical) should be identical at best. Secondly, the technical and economical parameters such as production process, plant size, fuel type, energy cost, and efficiency standards should be comparable. Especially the measuring of capital costs which can strongly influence the capture costs should be consistent.
In a cross-technology comparison, meeting these standards is often not possible. Especially, when emerging technologies and future scenarios are evaluated, reliable and consistent data can be scarce. Instead, a more heuristic approach must be adopted and the best available data analyzed. Thus, this paper summarizes the recent techno-economic literature on carbon capture in order to establish a large-scale picture of CO 2 supply in the near-term. A secondary database for the largest industrial CO 2 emitting sources is established in Table 2 that gathers the most recent and reliable cost data available. The presented measures and assumptions are heterogeneous and the data should be considered as estimates and benchmark values for best practice processes. To maximize cost data quality, recent peer-reviewed as well as broader government studies were preferably selected. Other studies were included to fill data gaps. The origin and relative assumptions of the capture cost studies are detailed in Table 2 as far as they were disclosed. Since capture from coal-and natural gas-fired power plants has been discussed the most extensively in the literature, the summarized average costs from the IEA study seem a reliable data source. Moreover, the capture costs for the higher concentrated sources of ammonia, hydrogen, and natural gas as well as cement production derived from the annually updated assumptions of US Energy Information Agency (EIA) seem a reliable data source for the purpose of analysis even though the data regionally cover only the USA. For the other potential sources of CO 2 , less research has been performed and average cost data are not available. Thus, recent peer-reviewed techno-economic studies have been included for capture from iron and steel, refineries, bioenergy, fermentation, and aluminum production. For capture from ethylene production a non-governmental, non-peer reviewed data source was included.
The original data have been adjusted slightly to allow for a common depiction in € per tonne on a 2014 basis. Firstly, cost data in US$ have been converted to € based on Oanda exchange rates 2 from the end of the respective year of data reference. Then, data that were older than 2014 were adjusted with the annual Chemical Engineering Plant Cost Index (CEPCI). The CEPCI is a composite index that reflects the development of equipment, construction labor, buildings and engineering and supervision costs over time since the 1960s (Chem. Eng. 2008;Chem. Eng. 2015a;Jenkins 2015;Mignard 2014). Thus, it helps to correct for changing economic conditions for chemical plants over time while slightly adjusting the original cost data from older reference years.

Evaluation of potential CO 2 supply
Based on the presented data in Tables 1 and 2, a benchmark CO 2 supply base can be established as a merit order and insights on the feasibility of selected utilization scenarios can be derived. The benchmark CO 2 supply is based on the best practice minimum capture costs. Upper limit or maximum costs are not presented since it is assumed that a variety of processes exist that could lead to even higher capture costs than those presented in Table 2. The supply cost function thus can be described by formula (3) that sorts and aggregates the capturable quantities of the potential sources (q i ) according to their benchmark capture costs (p i ): For the space U as the union of all intervals ⋃ n i¼1 q i−1 ; q i ð with i = rank of CO 2 emitting source n = number of ranked sources p i = benchmark capture cost of source ranked in ith position; in € per tonne; p 1 ≤ p 2 ≤ … ≤ p n q i = aggregated capturable emissions of sources ranked ≤i; in Mt/year For the following considerations, the scenarios will be split in a near-term view of up to 10 years and a long-term horizon of more than 10 years. This differentiation is considered useful for technology development since it is a common assumption that new industrial technologies can grow from lab to commercial scale within 10 years. For near-term scenarios, the presented current volumes and costs of CO 2 capture seem an adequate estimate. For future volumes of CO 2 emissions, numerous scenarios exist for different policy scenarios and time horizons, most notably those of the IPCC and the IEA. Future capture costs will vary depending on changing overall economic conditions and energy prices. Improved technological efficiency and performance usually decreases costs over time when the technologies are deployed (Finkenrath 2011). Assuming that capture technologies are advanced further in the future, at stable economic conditions current benchmark costs can likely be lowered or at least maintained.
As assigned in Table 1, the CO 2 point sources can be divided into four major groups of emitters:

I. High purity sources
For certain industrial processes such as ammonia production, the CO 2 emitted is very pure and capture requires only small additional efforts (IEA 2011 andUNIDO 2011). Therefore, these processes yield relatively cheap CO 2 as an output. These high-concentration sources represent only approx. 2 % of the 12.7 Gt capturable point source emissions (see Fig. 1). Today, capture of CO 2 is an established process predominantly in hydrogen, ammonia, and natural gas purification plants as they allow for comparatively cost efficient CO 2 separation (Wilcox 2012). While raw natural gas can contain CO 2 in different concentrations depending on the respective source, the processing of the gas to achieve pipeline quality often includes carbon dioxide separation (Baker and Lokhandwala 2008).

II. Fossil-based power generation
The largest CO 2 emitting group-the combustion of coal and gas for power generation-currently is responsible for approx. 76 % of the 12.7 Gt capturable emissions from point sources (see Fig. 1). However, CO 2 capture at power plants is often connected to significant efficiency losses of approx. 10-30 % of the output energy (de Coninck and Benson 2014; Finkenrath 2011). Consequently, coal and natural gas power plants currently lack business incentives for large-scale capture. If CCS was incentivized by political regulation (e.g. via emission performance standards) power companies could start to implement capture technologies for new plants and possibly retrofits. For power plants, economies of scale can play an important role for lowering capital and operating capture costs per tonne.

III. Large industrial emitters
Large industrial CO 2 emitting processes together currently make up for approx. 22 % of the 12.7 Gt capturable emissions from point sources (see Fig. 1). They include the production of industrial materials such as iron and steel, cement, aluminum as well as refineries. As these processes emit CO 2 in different quantities and qualities, CO 2 capture at such plants is also connected to varying efficiency penalties and benchmark costs. Moreover, a large number of other industrial manufacturing plants are potential candidates for CO 2 capture. Often, they are comparatively smaller than power plants (Bennaceur et al. 2008;Faulstich et al. 2009;Weikl and Schmidt 2010). Thus, economies of scale can be more difficult to achieve. For example, waste incineration so far has barely been analyzed in regard to CO 2 capture although reusing such CO 2 would conceptually close resource cycles. The comparatively small size of the incinerators however entails higher capture costs per tonne than those of other CO 2 sources (Faulstich et al. 2009).

IV. Natural wells
It must be noted, that part of the current market is covered by CO 2 from natural wells instead of recovered CO 2 emissions (Aresta and Dibenedetto 2010). For example, in the USA, approx. 45 Mt of CO 2 from natural reservoirs are currently used in enhanced oil or gas recovery (EOR/EGR) (Wilcox 2012). The cost of natural CO 2 is connected to the oil price (US EIA 2014) and relatively low at 15-20 €/t due to its often rather high purity (Aresta and Dibenedetto 2010). However, assuming that CCU technologies must be measured in regard to their environmental performance, the use of natural CO 2 carries certain disadvantages. Extracting CO 2 that is naturally stored underground for the purpose of using it in the production of fuels and materials will result in higher total emissions than when using CO 2 that is emitted anyway, e.g., by an industrial plant. Thus, it is recommended to replace CO 2 from natural wells currently in use with recovered CO 2 to achieve a net emission reduction (Aresta and Dibenedetto 2010;Metz et al. 2005). Since this paper focuses on recovered CO 2 emissions, CO 2 from natural sources is consequently not further included in the presented data.
Based on the presented emission and cost data a potential supply curve for the commodity CO 2 is established in Figs. 2 and 3. The low cost sources represent largely the high purity emitters (group I) and are detailed in Fig. 2. These can collectively provide approx. 300 Mt of CO 2 at a benchmark capture cost of approx. 33 €/t or less. The aggregated capturable CO 2 supply of 12.7 Gt CO 2 from all listed point sources is displayed in Fig. 3. Coal power plants can provide large amounts of CO 2 at relatively low benchmark capture costs of approx. 34 €/t. Nevertheless, even if coal power was phased out in the future, several other industrial processes would cause large amounts of CO 2 emissions that can be captured at benchmark costs below 100 €/t. Such a supply curve can also be designed for a single plant where different processes emit CO 2 at various capture costs, as exemplified in van Straelen et al. (2010).

Demand side: carbon dioxide capture and utilization
The idea of using CO 2 as feedstock is as old as the chemical industry (Aresta and Dibenedetto 2010), but so far very few applications have been realized. The conversion of CO 2 with a catalyst evolved in the 1970s, when chemical engineers first succeeded in developing catalysis processes inspired by nature's CO 2 conversion cycles (Aresta and Dibenedetto 2010; Aresta et al. 2013). Due to the oil crises at that time, the discovery of alternative feedstock to lower the dependency of fossil resources was economically very attractive. With rising political and public awareness on climate change a large field of research has developed around possibilities to reduce industrial CO 2 emissions. Accordingly, technological research on CCU technologies slowly but surely has gained momentum in the last decades.
CO 2 can either be used directly or as feedstock for a variety of products. Overall, approx. 222 Mt of the commodity are used in industrial applications worldwide (see current est. volumes in Table 3). Firstly, direct utilization of liquid or gaseous carbon dioxide usually requires a very high purity especially in the food and beverage industry which currently consumes approx. 11 Mt CO 2 per year. Furthermore, around 6 Mt CO 2 are used as process gas in various industrial applications (IHS 2013). The largest direct use of 25 Mt of CO 2 can be found in EOR/EGR which represent a borderline case, as they combine a utilization and storage function (Global CCS Institute 2014). Largely, they are attributed to CCS rather than CCU since after the extraction of additional fuels through CO 2 , the CO 2 can potentially be stored permanently in the depleted oil and gas fields. As EOR/EGR is a potential market for recovered CO 2 , it needs to be included when analyzing market volumes of CO 2 (see Table 3). Secondly, the conversion of CO 2 to materials still is limited to few applications at a smaller scale, except for urea synthesis which globally currently consumes approx. 130 Mt CO 2 per year. Indeed, urea and ammonia production are often combined, so that an estimated half of the high purity CO 2 from ammonia production is used for urea synthesis while the rest is often vented (IEA 2013; Metz et al. 2005). Apart from that, a marginal amount of CO 2 is used for the production of several specialty chemicals, e.g., of salicylic acid used for making aspirin pills. Commercial plants producing CO 2 -based fuels currently can be found only at demonstration scale of several thousand tonnes, e.g., by the companies Carbon Recycling International (CRI) in Iceland and Audi and Sunfire in Germany (CRI 2016;Strohbach 2013;Sunfire 2014). As R&D on CCU technologies continues and some important breakthroughs have been observed further CO 2 -based products are expected to enter global markets soon as depicted in the near-term (up to 10 years) estimates in Table 3. Thus, the demand for CO 2 as a commodity might increase in the future.

CO 2 utilization and emission reductions
CCS aims to store large amounts of CO 2 underground for long periods of time-approx. 1000 years (Metz et al. 2005). By contrast, when CO 2 is used directly or as feedstock for materials and fuels it will be reemitted to the atmosphere depending on the durability of the product, ranging from days to several years. As described earlier, the amounts of CO 2 used thus do not correspond to the amount of CO 2 avoided. Each CO 2 utilization process has a different environmental impact which needs to be determined in a life cycle assessment (von der Assen et al. 2015). The crux lies in the efficiency gains connected to the process: If the fossil raw material consumption of a production process can be reduced by the introduction of a CO 2 -based process the environmental balance can be positive. A recent example illustrates how for polyols used for the production of foams up to 3 t of CO 2 emissions can be avoided per tonne of CO 2 used compared to a conventional production process (von der Assen and Bardow 2014). Hence, despite the short durability of CO 2 utilization compared to storage there is an unknown overall mitigation potential that can possibly be significantly larger than the volumes of CO 2 utilized. However, for a market perspective, the volumes of the commodity CO 2 that can be captured and used need to be matched. A judgment in regard to mitigation potential is not possible on that basis. While CCS is a recognized emission reduction instrument and commonly accounted for in existing carbon management schemes CO 2 utilization per se is not accounted as direct emission reduction. If CCU can lead to reductions in fossil raw material use it is possible that CCU indirectly affects emission accounting just as other efficiency measures.

Evaluation of potential CO 2 demand
The presented status quo of CO 2 utilization has demonstrated the limited demand for the commodity CO 2 . However, ongoing worldwide CCU-related research covers a diverse array of utilization options. In the best case Brecycling of CO 2 from anthropogenic sources provides a renewable, inexhaustible carbon source and could allow the continued use of derived carbon fuels in an environmentally friendly, carbon neutral way^ (Mikkelsen et al. 2010). Even in direct utilization innovations are possible that go beyond the substitution of fossil feedstock. As currently shown by the CO 2 -based dry cleaning innovation of the US company CO 2 Nexus large-scale dry cleaning with CO 2 could potentially lead to economic and environmental benefits by replacing and reducing the a Current data is based on the 2013 estimates from Aresta et al. (2013) b Near-term data is based on the former 2016 estimates from Aresta et al. (2013) and includes CCU technologies that could be implemented within the next 10 years c Data from IHS (2013), worldwide data without Latin America and Asia except Japan d Estimated as constant by the author, not included in the cited original literature e Data from Global CCS Institute (2014) f Estimate from (Metz et al. 2005) consumption of a combination of valuable raw materials such as water, natural gas, and energy (Madsen et al. 2014). At the same time, replacing existing refrigerants with CO 2 as a standard coolant in automobile air conditioning systems is currently considered at the European Commission (EurActiv.com 2013; Malvicino 2011). Consequently, even when the CO 2 is directly used it can potentially substitute various substances that are hazardous or have a higher climate impact (Aresta and Dibenedetto 2010).
Overall, future estimates for the potential of CO 2 utilization in the literature vary, but all range around the same maximum potential. For chemical materials, the estimated large-scale potential is around 200 Mt CO 2 p.a. (212 Mt est. by Aresta et al. (2013), 200 Mt by Mikkelsen et al. (2010), 180 Mt est. by VCI (2009), and 115 Mt est. by Metz et al. (2005)). By contrast, the estimated large-scale potential for fuel production with CO 2 is much bigger with approx. 2 Gt CO 2 p.a. (VCI 2009). Altogether, large-scale CCU can hence potentially require a maximum of 5-6 % of the estimated 37 Gt of anthropogenic CO 2 emissions (Le Quéré et al. 2014). In order to build a potential demand for the commodity CO 2 , next to the utilization potential the storage potential must also be taken into account. Thus, CCS targets of the IEA will also be included in the scenarios (IEA 2013). Since EOR/EGR is usually considered as storage, it will be attributed to the CCS and not the CCU scenarios in this study. For direct utilization no long-term estimates exist, but due to the limited volumes currently required even significant demand changes are expected to have a minor impact on overall CO 2 demand. Consequently, the following five demand scenarios were identified:

A. CCU current
This scenario represents the current (2013 est.) CO 2 demand presented in Table 3. The aggregate CO 2 demand amounts to approx. 200 Mt CO 2 excluding EOR/EGR.

B. CCU near-term
This scenario includes the expected near-term development of CO 2 utilization over the next 10 years based on the near-term scenario of Table 3. Next to a continued direct utilization of approx. 17 Mt p.a. (excluding EOR/EGR) the production of some CCU-based fuels consumes an estimated 20 Mt of CO 2 and the aggregated demand for CO 2 amounts to 250 Mt p.a. Similar demand scenarios of mixed CCU material and fuel activities have also been projected in the range of 250-350 Mt by Quadrelli and Centi (2011) and 300-400 Mt by Aresta and Dibenedetto (2010) for the medium-term.

C. CCU and CCS near-term
This scenario combines the projected shorter-term development of 250 Mt CO 2 used with 40 Mt CO 2 stored (Carbon Visuals 2014). A combined 290 Mt CO 2 thus depict a realistic demand for CO 2 in the near-term of up to 10 years.

D. CCU fuels long-term
This scenario illustrates a large-scale potential of CO 2 utilization including a transformation of the fuels sector to synthetic CO 2 -based fuels. Due to the relatively low current prices of fossil energy, the transition away from fossil fuels is a larger endeavor however that will take much more than 10 years to implement. Implementing this scenario would be connected to severe infrastructure investments in particular in regard to renewable energy and fuel refineries and thus requires strong political efforts. Consequently, this scenario serves rather as an optimistic long-term vision of CCU with 2300 Mt of CO 2 (250 Mt for materials and direct use and 2050 Mt for fuel production (VCI 2009).

E. CCU fuels and CCS long-term
The combination of the 2300 Mt CO 2 of scenario D and the 2050 target for CCS deployment according to the IEA (2013) of 7870 Mt CO 2 provides a long-term overall potential demand of combined use and storage of more than 10 Gt CO 2 .
In summary, scenarios A, B, and C represent realistic, nearterm scenarios that show a potential demand of 200-300 Mt CO 2 and represent less than 1 % of anthropogenic emissions while scenarios D and E with 2-10 Gt show an optimistic, long-term potential of materials, fuels, and storage and consume potentially 5-27 % of the current estimated anthropogenic emissions of 37 Gt (Le Quéré et al. 2014).

Supply and demand scenarios
As a next step, the merit order of supply of CO 2 from industrial point sources will be matched with the described demand developments for CCU and CCS. For this, the discussed current and potential demand volumes (d) for scenarios A to E are combined with the supply of CO 2 in order to determine the equilibrium best practice benchmark capture prices (p * ) for each scenario. These are determined by applying formula (4) to the presented data.
with d = current and potential demand volumes for scenarios A to E p * = equilibrium best practice benchmark capture prices For simplification purposes, it is assumed that the cost of capture correspond to the price of carbon dioxide paid by the consumer which implies that no profits are made on the commodity CO 2 . The presented scenarios are useful to understand the potential allocation of captured emissions in global supply and demand of CO 2 . The analysis of comparative statics is again grouped into short-term and long-term visions.

Short-term scenarios
The current and near-term potential is depicted in Fig. 4. It becomes evident that there is enough CO 2 from high purity sources at a comparatively low equilibrium capture cost of approx. 33 €/t to cover the CCU demand in the short-term. From a global perspective, the volumes captured from high purity sources such as fermentation, bioenergy, natural gas production, hydrogen, and ammonia would suffice to cover the current and upcoming CO 2 demand. Consequently, for the small total volumes required in each specific case the source is usually chosen based on local availability, respective quality, and cost of available CO 2 . A trend towards using a plant's own or a nearby partner's waste emissions can be observed at several demonstration plants (Chem. Eng. 2015b;German Embassy Pretoria 2013;Tieman 2013).

Long-term scenarios
Considering the long-term development of CCU and CCS technologies, an optimistic full implementation of described target scenarios D and E is illustrated in Fig. 5. Next to the high purity sources, other industrial plants will become relevant to capture the demanded CO 2 emissions. Currently, CO 2 from coal can be captured in large amounts (>7.5 Gt p.a.) at benchmark costs of approx. 33 €/t. Beyond coal and gas, other industrial processes such as iron and steel, ethylene, and cement production can provide CO 2 at benchmark capture costs of less than 70 €/t. Considering the laid out business as usual supply base the long-term scenario D for CCU including fuels could be achieved with CO 2 at an equilibrium CO 2 price of approx. 33 €/t. For a combined long-term target for CCU and CCS, the equilibrium price per tonne based on current

Scenarios without fossil-fired power generation
As stated earlier, the future development of CO 2 emission volumes and prices is unclear due to a variety of political scenarios. The static equilibrium of the current CO 2 supply with long-term demand scenarios illustrated in Fig. 5 can only provide insights for a business as usual scenario. If however major technological or political changes occur and significantly larger volumes of CO 2 will continually be mitigated or removed the investments into CCS and potentially also CCU can turn into a dead end and lead to sunk costs in the long run. For future scenarios assuming a larger share of renewable energy, remaining fossil power plants might have to work at lower load factors and efficiencies (Finkenrath 2011). Thus, capture costs at these plants could potentially be higher than the current data suggest. Moreover, for other industrial plants, new technologies can become available in the long run that emit significantly less CO 2 . Then, the relevancy and cost of carbon capture could change-in both directions. For example in the iron and steel production the recently demonstrated Hisarna process can on the one hand reduce approx. 20 % of the conventional CO 2 emissions. On the other hand, the process allows for a very efficient combination with CO 2 capture (Pfeifer 2015). Consequently, advancing green technologies across industries will impact available volumes and costs of recovering CO 2 emissions.
Since CCS technologies are largely considered as an instrument to improve the carbon footprint of continued fossilbased power generation (Metz et al. 2005), their implementation becomes less relevant once the energy sector would be based on renewables. To understand the effects on CCU, a modified merit order for CO 2 supply excluding power generation based on coal and natural gas has been developed in Fig. 6 and matched with a demand scenario excluding CCS. It becomes evident that in the near-term (scenario B) CO 2 supply for CCU would be provided from industrial high purity sources at an equilibrium price of approx. 33 €/t, in line with scenario B depicted in Fig. 4. Meanwhile the long-term, largescale potential of CCU including fuels (scenario D) would be served from recovered emissions of various industrial production processes such as iron and steel, ethylene, and cement at equilibrium cost of approx. 68 €/t. This modified supply scenario shows that the future development of CCU technologies is independent from the fossil power industry. By contrast, the amounts of CO 2 required even for the visionary potential of CCU can possibly be recovered from various industrial sources at estimated costs of less than 70 €/t. CCU technologies thus do not conceptually contradict renewable energy or energy efficiency technologies. Instead, they can be seen as a complementary route.

Conclusions and perspectives
With carbon capture and utilization, recovered CO 2 emissions could be turned into a valuable feedstock for the production of consumer goods. CCS in contrast aims to sequester recovered CO 2 permanently underground. For both technology fields CO 2 represents a commodity good that is potentially supplied from capture at industrial plants. The comparative statics of CO 2 demand and supply show that depending on the targeted scenario for CCU and CCS different industrial sources of CO 2 emissions will play a role: For the near future, smaller scale CCU scenarios, industrial plants with higher CO 2 concentration and lower benchmark capture costs of approx. 33 €/t will be relevant. Meanwhile, for the long-term large-scale scenarios including the fuel sector and CCS coal-fired power plants or other larger emitters must also be included and CO 2 captured for benchmark costs of up to 65 €/t will be needed. Even if fossil-fired power generation is phased out, capture at industrial production processes can provide more than enough CO 2 for large-scale CCU visions. Consequently, reusing recovered industrial CO 2 emissions can unfold its environmental and economic potential in creating regional or local circular economy solutions. For example, half of the total CO 2 emissions from waste incineration in Germany would sufficiently supply CO 2 as a carbon source for the domestic polymer consumption (Bringezu 2014). Moreover, capture of CO 2 from ambient air remains a potential alternative for recovering CO 2 emissions. However, technological development is at present still at an early stage and the costs of air capture are highly uncertain and depend on a variety of factors (Lackner et al. 2012). In any case, further research on environmental and economic impacts should accompany the technological advancement of carbon capture and utilization technologies. Harmonizing methods and measures in environmental and techno-economic assessments and establishing best practices would improve the interpretability of the results enormously.
Since the expected near-term status of CO 2 utilization does not involve large volumes of CO 2 smaller regional solutions gain importance. When new plants are set up that reuse CO 2 emissions these can be planned next to a convenient source of CO 2 at sufficient quality and a competitive cost. Emissions from fossil-fired power plants are not required to meet the potential CO 2 utilization demand. Even large-scale visions for CCU can therefore not serve as an argument to prolong fossil-fired power generation. When implementing large-scale CO 2 utilization scenarios involving synthetic fuels based on power-to-liquid or -gas technologies a broader infrastructure especially for renewable energy but also for CO 2 supply will be needed. Until then, from a mitigation perspective, differentiating recovered CO 2 by source can even be misleading as in sum it does not play a role to the environment where the reused CO 2 comes from. Instead, market mechanisms will balance supply and demand. Nevertheless, sustainability aspects always need to be considered when further deploying CCU technologies for example by conducting lifecycle analysis and considering alternative technologies based on renewable energy and raw materials.
Moreover, if energy efficient CCU technologies can be developed, the presented CO 2 supply benchmark cost range of 10-100 €/t CO 2 can prove to be a relatively cheap alternative feedstock compared to more expensive or volatile priced chemicals based on fossil raw materials such as crude oil. The search for such technologies that use CO 2 to replace fossil raw materials and their derivatives consequently remains attractive even in times of relatively low or decreasing energy prices. Currently, in existing emission trading schemes, the CO 2 is largely underpriced. At current conditions, investments into CCU or CCS hence can only allow for future profits if substitution of expensive raw materials is possible, increased process and production efficiencies can be achieved or government subsidies compensate potential losses. In the future, however, the combined total of approx. 6 Gt of global CO 2 emissions that are currently regulated by some form of carbon pricing instrument (World Bank 2014) could increase in amount and price. A sufficiently high carbon tax or emission trading price could then incentivize further CCU and CCS activities in certain regions. If the combined costs of capture, transport, and storage of a certain emitting source are lower than the CO 2 tax or certificate price CCS will have a business case. For example, the relatively old as well as high Norwegian carbon tax, especially for offshore petroleum businesses of up to 69 €/t CO 2 in 2014 (World Bank 2014) has led to several investments by the affected players into CCS (de Coninck and Benson 2014) and energy efficiency (Bruvoll and Larsen 2004). In contrast, CCU technologies-as efficiency measures generally-are indirectly impacted by the carbon price. Depending on respective energy requirements and prices, certain technologies are profitable even at current conditions. A higher carbon price can be expected to stimulate the deployment of more CCU technologies. Best practice benchmark capture cost can give an indication but do not include potential substitution or efficiency effects connected to CCU processes. An equilibrium price of carbon dioxide for enabling the implementation of CCU from an economic perspective is thus technology specific.
Eventually, while policy makers and businesses must continue to work on mitigating global emissions, pathways for using waste emissions should be pursued complementary. For any desired future, reliable political targets and regulations will be important to permit optimal investment decisions when technologies are implemented and scaled up. Overall, moving ahead, CCU should be considered as a means for improving regional resource security and as enabler of smaller circular economy solutions. Making use of recovered CO 2 seems to be one option for imitating nature's no waste philosophy into industrial design and consumption.