Enhanced carbonate dissolution:

Abstract The reaction of a mineral carbonate, such as limestone, with water and CO 2 to form bicarbonate in solution, is explored as a CO 2 mitigation strategy. Initial cost estimates for such a process range from $18 to $128 per tonne CO 2 sequestered, with an energy penalty of about 8% and with relatively low environmental impact. The regional availability and transport of water and mineral carbonate appear to be the primary determinants of the strategy’s cost and applicability. The bicarbonate-rich waste effluent would be released into rivers or coastal waters, ultimately adding a small amount to the existing, very large bicarbonate reservoir in the ocean. For many applications, this form of ‘marine’ carbon sequestration appears to be less costly, less affected by national and international regulations, more environmentally friendly and more effective over the long term than direct CO 2 injection into the ocean.


Introduction
Given the steadily increasing atmospheric CO 2 burden caused by human activities and the potential this poses for climate modi®cation [21,22], a variety of schemes to mitigate such increases have been proposed [17]. Among these have been chemical and biological methods of Energy Conversion & Management 40 (1999) 1803±1813 0196-8904/99/$ -see front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 1 9 6 -8 9 0 4 ( 9 9 ) 0 0 0 7 1 -0 www.elsevier.com/locate/enconman capturing CO 2 from waste gas streams and storing this carbon in forms or in locations isolated from the atmosphere. Examples include subterranean [14,19] and deep-ocean [16,29] injection of captured CO 2 , biomass formation/accumulation on land or in the sea [8,23] and use of CO 2 as feedstock for the production of carbonaceous compounds [3,43]. Proposed chemical methods of sequestering carbon include CO 2 reactions with naturally or arti®cially formed alkaline compounds, such as silicates, oxides and hydroxides, generally leading to the storage of carbon ultimately in the form of solid carbonate [12,27,28]. Reactions involving solid carbonate dissolution or`weathering' in the presence of water and CO 2 to form bicarbonate in solution are also thought to be capable of absorbing a signi®cant fraction of the anthropogenic CO 2 input. However, under natural conditions, this process is expected to take millennia [2,31]. Here, we wish to consider arti®cially enhancing mineral carbonate dissolution as a way of sequestering signi®cant quantities of CO 2 at a much accelerated pace. This would involve bringing mineral carbonate and water into direct contact with CO 2 -rich waste gas euent from centralized industrial or municipal sources; thus, at least partially, consuming CO 2 to form relatively innocuous and long lived cations and bicarbonate in solution which would directly or indirectly be added to the already large pool of these ions in the ocean.

Relevant chemistry
Carbonate weathering proceeds as follows: . gaseous CO 2 is dissolved in water; . and is subsequently hydrated to form carbonic acid; The hydrated CO 2 is then reacted with a mineral carbonate (in this example, CaCO 3 ) to form Ca 2+ and bicarbonate in solution; the net reaction being; While a wide range of metal carbonates can participate in such reactions, CaCO 3 and MgCO 3 and related complexes, such as CaMg(CO 3 ) 2 , are the most abundant (especially as contained in calcite, limestone and dolomite) and, therefore, the largest participants in natural carbonate weathering. However, rather than ®rst diluting waste gas CO 2 in the atmosphere and then allowing reaction 4 to proceed naturally (and slowly) as it currently does, it would be advantageous to place water and mineral carbonate in direct contact with waste gas streams whose pCO 2 is commonly several orders of magnitude higher than that of the atmosphere (e.g.,¯ue gas pCO 2 Q0.15 atms [39] vs atmospheric pCO 2 =3.7 Â 10 À4 atms [22]). This would allow CO 2 (aq) and carbonic acid (reactions 1 and 2) to form faster and in concentrations much higher than would naturally occur in water in contact with atmospheric CO 2 . This elevation in acid concentration would promote reaction 3 and the formation of bicarbonate in solution, half of which would be derived from the original waste gas CO 2 (reaction 4). By analogy, enhanced mineral weathering reactions occur naturally in elevated CO 2 environments, such as in decomposing, organic rich soils and in the deep ocean [4,30]. We suggest that such reactions be emplaced at sites of high anthropogenic CO 2 emissions so as to attenuate their release to the atmosphere, at least partially.

Possible C 2 SEA reactor parameters
We envision applying the CO 2 SequEstration as bicarbonAte (C 2 SEA) process by allowing a CO 2 -rich euent gas stream to¯ow over or through a porous pile, bed or slurry of limestone (or other high CaCO 3 content) particles which are wetted by a continuous spray or¯ow of water. Referring to Fig. 1, a CO 2 -rich gas stream (1) enters the reactor vessel (5) by one or more entry ways (e.g., 2, 3 and/or 4). The gas stream then passes over or through a wetted, porous bed of calcium carbonate (e.g., limestone) particles within the reactor. This carbonate mass is sprayed (6) and wetted with and partially submerged in a water/carbonic acid solution which is unsaturated with respect to bicarbonate ion. This arrangement exposes the incoming gas to a large surface area of water/solution in the form of droplets and wetted carbonate particle surfaces in (5), facilitating hydration of the entering CO 2 to form a carbonic acid solution within the reactor. CO 2 -depleted gas then exits the reactor (7).
High-CO 2 waste gas is passed through such a reactor so as to contact the water and wetted surfaces, forming carbonic acid via reactions 1 and 2. The carbonic acid solution formed reacts with the carbonate to form calcium ions and bicarbonate in solution which is either recirculated or bled from the reactor and replaced with unreacted water within the reactor at a rate which maximizes bene®t/cost. Assuming a waste gas pCO 2 of 0.15 atms (e.g., for coal-derived¯ue gas; [39]), the carbonic acid formed would then react with the calcium carbonate to produce HCO À 3 at a rate of roughly 2 Â 10 À6 mmols HCO À 3 cm À2 carbonate surface area s À1 (2H 50%), if experiments with calcite in highly undersaturated solutions are representative [1,33]. Since half of this bicarbonate formation would be derived from waste gas CO 2 and the other half from the calcium carbonate, it would take 2.3 tonnes of calcium carbonate and 0.3 tonnes of water to react 1 tonne of CO 2 to form 2.8 tonnes of HCO À 3 in solution. Sequestering 1 tonne of CO 2 per day by this method would then require some 10 4 m 2 of carbonate surface area assuming the above bicarbonate formation rate.
If spherical carbonate particles are assumed, it can be shown for CaCO 3 [42] that A=k t/r where A = the carbonate surface area in m 2 , t = CaCO 3 mass in tonnes, r = particle radius in m and k = 1 m 3 tonne À1 . By specifying uniformly sized spherical particles of radius 10 À2 m, it would then be necessary to present 100 tonnes of carbonate to the incoming gas stream to eect a sequestration of 1 tonne CO 2 /day, aording a pile lifetime of roughly 40 d. Assuming a rhombohedral packing of spherical carbonate particles with a resulting total carbonate particle surface area/packed volume, AaV, of 222 m 2 /m 3 (=2.22 m 3 m À3 r À1 ), the minimum required pile volume is about (10 4 m 2 )/(222 m 2 m À3 )=45 m 3 per tonne CO 2 sequestered day À1 . This required pile volume will decrease as the speci®ed carbonate particle size decreases, the heterogeneity in particle size increases and as particle shapes deviate from a sphere. For example, Walter and Morse [42] found that natural calcium carbonate AaV can be several orders of magnitude higher than that predicted from spherical particles, thus for a given mean carbonate particle size, the preceding estimation method should yield an upper bound for the reactor bed volume.
Water in equilibrium with calcium carbonate and CO 2 (gas) at a pCO 2 of 0.15 atms (H258C) will contain 6.7 Â10 À3 molar bicarbonate [9] or 4 Â 10 À4 tonnes HCO À 3 /tonne H 2 O. At this concentration, a minimum discharge of 7 Â 10 3 tonnes H 2 O/day would be needed to accommodate the above production and discharge of 2.8 tonnes HCO À 3 day À1 as required to sequester 1 tonne CO 2 (gas) day À1 from the original waste gas stream. We stress that this is an absolute minimum H 2 O requirement per tonne CO 2 sequestered because the rate of bicarbonate formation will decline (nonlinearly) to zero as bicarbonate saturation is approached [32], making it unlikely under practical application that equilibrium conditions will be attained. This means that the reactor bicarbonate concentrations will be likely to be below bicarbonate saturation prior to discharge, requiring a higher water¯ow rate through the reactor in order to accommodate the discharge of 2.8 tonnes HCO À 3 day À1 . For example, attaining a half-saturated bicarbonate concentration (5 Â 10 À3 molar) would increase the above water requirement to 10 4 tonnes H 2 O tonne À1 CO 2 sequestered day À1 . By comparison, we note that coal ®red power plants consume roughly 800 tonnes H 2 O per tonne coal burned [34] or about 400 tonnes H 2 O per tonne CO 2 produced. However, reusing this`free' water as the sole source of C 2 SEA reactor water would allow a sequestration of about 4% of the CO 2 produced in coal ®red electricity generation. Such a quantity is not trivial, given policies, e.g., the Kyoto Protocol, calling for only small fractional reductions in CO 2 emissions.

Water considerations
With the relatively high water demand of the C 2 SEA process, it is likely that water availability and cost could, in many instances, determine the practical amount of CO 2 mitigation that can be eected for a given point source. No-cost water sources, such as power plant cooling water or other sources of recycled or reclaimed water, together with river or seawaterÐif availableÐshould, therefore, be considered ®rst. The use of seawater is particularly attractive for coastal power plants because: (i) there is virtually a limitless source of such water; (ii) carbonate dissolution is enhanced over that in freshwater due to increased ionic strength and (iii) bicarbonate waste can be directly dumped and diluted in the ocean.
Regarding (iii), we point out that, barring drainage into a landlocked basin, the ocean would be the ultimate repository of the bicarbonate and Ca 2+ generated by C 2 SEA, where 42 Â 10 12 tonnes of C are already present as dissolved bicarbonate plus a much smaller quantity of dissolved carbonate ion [5]. By comparison, if it were somehow possible to`bicarbonate' the entire yearly global fossil fuel CO 2 emission (22 Â 10 9 tonnes CO 2 /y [24]), this would only annually add about 0.1% to the ocean's bicarbonate burden, while processing an amount of water equivalent to less than 0.02% of the ocean's volume. The availability of calcium carbonate for C 2 SEA also does not appear to be an issue. The carbon mass of sedimentary carbonates is estimated to be 6 Â 10 16 tonnes [5] with a current mining output of roughly 10 9 tonnes crushed limestone per year for the US alone [41], enough to react about 4 Â 10 8 tonnes CO 2 /y using C 2 SEA. We also note that the majority of anthropogenic CO 2 emissions in the US occur in the eastern half of the country [39], fortuitously coinciding with the majority of the nation's limestone reserves and production [40], and thus minimizing the limestone (or CO 2 ) transportation distances to C 2 SEA potential facilities in this region.

Reactor euent
While no US federal regulations currently exist regarding bicarbonate in wastewater, thè hardness' and total dissolved solids load of C 2 SEA euent could limit the disposal of such waste into inland waterways which would ultimately deliver the bicarbonate to the ocean. Coastal power plants would again be at an advantage, where direct dumping of`hard' C 2 SEA euent to the ocean would likely be unrestricted. This could pose a distinct advantage over proposed open-ocean CO 2 injection which would not be immune to international law, e.g., the London Dumping Convention [25]. Also, the ecological impacts of C 2 SEA euent are likely to Table 1 Estimated cost per tonne CO 2 sequestered using a C 2 SEA reactor (e.g., Fig. 1 [41]. Crushing costs derived by specifying a particle radius reduction from 10 À1 to 10 À2 meters and using the energy estimation equation 10-49 of Boyce [6], with an assumed electrical energy use eciency of 40%. Energy cost here assumes osite electrical energy consumption [18]. Per km CaCO 3 transportation costs: rail [20]; ship/barge=20% of rail cost [26]. Required water volume assumes reactor euent bicarbonate concentration of 5 Â 10 À3 molar (50% saturation). Per tonne water cost typical of agricultural water.`Free' water refers to sources, such as seawater, power plant waste cooling water, river water, or other types of no-cost, untreated water. Water pumping cost assumes water must be pumped a total of 1.2 vertical meters (with or without water recycling through reactor) at 60% eciency using equation 20-13 of Snow et al. [36]. Per kWh e cost here assumes onsite power usage and, therefore, excludes $0.02 power distribution cost added to kWh e energy cost for crushing (above). Capital costs are calculated by assuming a capital outlay of $50 M for an onsite power plant C 2 SEA reactor and a payback rate of $0.04 $M À1 tonne À1 CO 2 sequestered [18]. Operating and maintenance costs are assumed to be 25% of capital costs [18]. Costs of other large scale CO 2 sequestration technologies from Herzog et al. [18] and Fujioka et al. [15]. See text for further details. be far smaller than those predicted/observed for direct marine CO 2 dumping and resulting ocean acidi®cation [10,38] because a substantial amount of waste CO 2 (and acidity) will have been neutralized to bicarbonate in the C 2 SEA process.
The residual CO 2 concentration in C 2 SEA euent would nevertheless pose a problem. Returning to the above solution resulting from the equilibrium among water, calcium carbonate and CO 2 (pCO 2 =0.15 atms, T = 258C), such C 2 SEA euent would have a pCO 2 about 400 times above ambient air. This pCO 2 dichotomy between the euent and air would lead to a rapid degassing of CO 2 from the former solution, causing carbonate ion oversaturation and likely precipitation of solid carbonate from the solution. Such eects could be minimized by direct dumping of the euent into the ocean, allowing copious subsurface mixing and dilution of the euent. By our calculation, a 100-fold dilution of the above reactor euent, followed by complete CO 2 equilibration with the atmosphere, would lead to only a 13% increase in carbonate ion concentration in the resulting seawater. Such an increase would be unlikely to cause a carbonate precipitation event, given the already supersaturated state of the ocean as apparently maintained by certain naturally occurring ions in seawater which inhibit carbonate precipitation [30]. Depending on the relative ionic burden in C 2 SEA euent, it could be denser than the receiving water, causing the euent to sink, further minimizing contact with and CO 2 loss to the atmosphere. The possibility for dilution/sinking would be more limited in inland waterways where undesirable euent CO 2 degassing prior to sucient wastewater dilution and/or prior to entry to the sea could result in in-stream carbonate formations analogous to travertines [13]. It may be possible to reduce such eects in freshwater by combining C 2 SEA euent with other wastewater streams whose ion and organic content inhibit carbonate precipitation from the euent in the face of inadequate dilution and signi®cant CO 2 degassing.

Cost and energy penalty estimates
With the preceding considerations and limitations in mind, our preliminary calculation of cost per tonne CO 2 sequestered using C 2 SEA technology ranges from about $18 to $128 (Table  1). By comparison, recent cost estimates of CO 2 capture, transport, and open ocean injection range from $90 to $180/tonne CO 2 [15]. Signi®cant cost savings for the C 2 SEA scheme can be achieved through the use of`free' sources of water, such as seawater, river water, or power plant cooling water, as well as the use of ship or barge transport of carbonate. The`free' water scenario costs are signi®cantly below the $50 tonne CO 2 tax currently levied in Norway ( Table  1), suggesting that such a C 2 SEA scheme might be an economically competitive CO 2 mitigation strategy.
In our estimates, we have included the cost ($1.45) and energy penalty (21 kWh e ) of crushing 2.3 tonnes of limestone (particle radius reduction from 10 À1 to 10 À2 m), as well as pumping the required water the equivalent of 1.2 vertical meters at 60% pump eciency ($2.85 and 57 kWh e , respectively). This amount of energy is envisioned for pumping the water to and from the site as well as internal water recycling. It could also include short distance pumping of the gas stream to/from/within the reactor if required. We, therefore, calculate an energy penalty of roughly 77 kWh e /tonne CO 2 sequestered (not considering the energy used in carbonate transport). If the combustion of 1 tonne of coal produces 2 tonnes of CO 2 and 2 Â 10 3 kWh e [39], then this C 2 SEA energy penalty amounts to about 8% of the energy produced. Herzog et al. [18] list energy penalties for CO 2 capture technologies alone (without disposal or sequestration), ranging from 13% to 37%. However, all of our cost and energy penalty estimates for C 2 SEA should be viewed as preliminary and will vary widely depending on the desired reactor size, con®guration and performance, the proximity to and cost of carbonate and water sources and the realized carbonate dissolution and CO 2 sequestration rates.

Other issues and alternatives
Regarding the actual rates of C 2 SEA sequestration, experiments with impure calcium carbonate minerals (e.g., dolomite) yield dissolution rates lower than those of calcite [11]. However, we are unaware of any rate measurements conducted on more C 2 SEA-relevant limestone, or on carbonate dissolution in highly CO 2 -supersaturated seawater (pCO 2 >0.01 atm). Factors negatively aecting such rates could include the presence and interaction of ions, such as SO 2À 4 and PO 2À 4 and certain metals, which can impede or inhibit dissolution [30]. If such eects are present, it may be necessary to oset them by: (i) increasing the reactor's carbonate surface area by decreasing the carbonate particle size or increasing the carbonate bed volume; or (ii) reducing the concentration or reactivity of contaminants from the incoming gas, carbonate, and/or water streams. The possible need for¯ue gas desulfurization (FGD) prior to the C 2 SEA process could be one example. However, because FGD also commonly employs wet limestone (to convert SO 2 (gas) to CaSO 4 (solid) [37]), possible integration of such FGD and C 2 SEA processes might be considered. While wet limestone FGD diers from C 2 SEA in at least three important areas (it generates CO 2 , it consumes O 2 and it produces a solid byproduct), the practical design and implementation of C 2 SEA reactors could, nevertheless, bene®t from models developed for the former process [7].
It may also be advantageous to increase the carbonate dissolution rate by maximizing the production and concentration of carbonic acid in the reactor solution. This could be achieved by elevating the pCO 2 of the incoming waste gas stream, either by mechanically elevating the total gas pressure within the reactor or by pre-concentrating the CO 2 contained in the incoming gas stream using various existing technologies [17]. Under circumstances where CO 2 emissions are far from water and carbonate resources, the overland piping or shipping of concentrated gaseous or liquid CO 2 to a centralized C 2 SEA reactor (rather than the transport of carbonate and water to the site of CO 2 generation) might prove cost eective at rates below $0.01 tonne À1 CO 2 km À1 [35]. Also, to the extent that the hydration rate of CO 2 (reactions 1 and 2) limits the carbonate dissolution rate in the reactor, it may be appropriate to add the highly eective catalyst carbonic anhydrase to the reactor solution. The cost/bene®t of the above potential reactor modi®cations needs to be investigated.

Summary and conclusions
The C 2 SEA approach appears to have some advantages over other large scale CO 2 sequestration technologies proposed to date. First, unlike other chemical CO 2 capture methods (e.g., those using monoethanol amine or highly alkaline solutions), the process does not require exotic, expensive, or environmentally damaging reactants or procedures. While in some cases regionally scarce, the C 2 SEA reactants, water and calcium carbonate, are globally abundant, inexpensive and innocuous. The reactor would have a relatively low energy penalty, incurred primarily in reactant transport to and within the reactor. The reactor waste products, Ca 2+ and bicarbonate, are relatively benign and, once in the ocean, will add little to the burden of these ions already present. Coastal dumping of fossil carbon in the form of bicarbonate appears to have economic, ecological and legal advantages over direct, open ocean CO 2 injection. Furthermore, the lifetime of carbon stored as marine bicarbonate will be substantially longer than carbon injected as CO 2 into the subsurface ocean, Caldeira and Rau [44]. The C 2 SEA method would circumvent the risk of signi®cant CO 2 degassing inherent in direct subterranean and submarine CO 2 injection and storage. We believe that the ecacy and economics of the C 2 SEA process in partially sequestering CO 2 from the atmosphere on an industrial and possibly global scale deserves further study.