World Resource Review Vol. 16 No.2

CAPTURING DIRECTLY FROM THE ATMOSPHERE

Frank S. Zemanand Klaus S. Lackner Departmentof Earth and EnvironmentalEngineering Columbia University 918 Mudd, MC 4711 500 West 120dtStreet New York, NY, 10027 USA Email: [email protected]

Keywords: Air extraction,COb sequestration,climate change,kraft process

SUMMARY The increasingconcern over rising CO2levels in the atmosphereand their potentialclimate effectsis fuelling researchaimed at carbonmanagement. One areaof researchfocuses on capturingthe CO2after combustionand sequesteringit underground.Capture schemes would operateat the site of generationtaking advantageof the elevatedconcentrations of CO2in the effluent. Here, however,we proposean indirectmethod of carboncapture that removes CO2from the atmosphere.This processcombines existing technologies with recenttechnological innovations into a novel carboncapture concept termed air extraction. The processuses dissolved sodium hydroxide to removethe CO2from ambientair. The resultantsodium carbonate solution is causticizedusing calcium hydroxideto regeneratethe sodiumhydroxide solutionand precipitate calcite. The calcite is then thermallydecomposed to producelime and CO2.The lime is hydratedto completethe process.These latter stages are usedroutinely in the paperindustry and the calcinationof limestoneis centralto the cementindustry. This paperreviews the processand highlights pertinentresearch to developa likely cost-effectiveprocess. It is shownthat the proposedprocess is well defmed and technicallyfeasible. The wide parameterspace and potentialfor improvementssuggest that further efficiency improvementsare attainable. The most significantimprovements will be derived from efficient heatmanagement.

() 2004 World ResourceReview. All rightsreserved. 157 World Resource Review Vol. 16 No.2

1 INTRODUCTION

Recentyears have seenattention drawn to issuessun-ounding changes in the global climate inducedby humanactivities. Concernover these potential changeshas led to the formation of internationalagencies, such as the IntergovernmentalPanel on ClimateChange (IPCC), whose mission is to consider this problem. The IPCC hasstated that carbondioxide (COJ is the greenhouse gascontributing the largestportion of the anthropogenicincrease in global radiative forcing, which the IPCC defmesas an externallyimposed perturbation in the radiative energybudget of the Earth's climate system. AnthropogenicCO2 producesapproximately 62% of the increaseor 1.46 W1m2 out of a total increase of 2.43 W/m2(Houghton and Ding, 2001). The increasein atmosphericCO2 levels can be largely attributedto the combustionof fossil fuels, with a small contributionfrom cementproduction and land usechanges. The total global emissionsreached 6,611 million metric tons of carbonin 2000, which representsa 1.8%increase over 1999(Marland, Bodenand Andres,2003). Giventhe world's economicdependence on fossil fuels andthe expectedincrease in consumption, methodsfor mitigating the atmosphericrelease of CO2need to be developed. The primary targetsfor mitigationare power plantsas theyproduce large concentratedstreams of CO2(Herzog and Drake,1996). Thesesources account for aboutone third of the worldwide CO2emissions. Even aftereliminating all emissionsfrom power plants,the remainingtwo thirds would still be releasedto the atmosphere.Roughly half of all emissionsarise from smalldistributed and oftenmobile sources. Mitigating their CO2impact on the atmosphereis more difficult. One approachis to provide fuels that are carbonfree, which is the reasonmuch emphasisis being placedon hydrogenas a transportationfuel. Here,we proposean alternativemethod for mitigating CO2emissions from sourcesother than powerplants. We proposethat CO2 is removeddirectly from the atmospherein a costeffective, industrialprocess hereafter referred to as air extraction. Air extractioncan be viewed as a variationof flue gasscrubbing where the fluid is at atmospherictemperature and pressure with a CO2concentration of 0.037%. The implementationof the processdiscussed here uses a sodium hydroxide (NaOH)based, alkaline liquid sorbentto removethe CO2from the ambientair by producingdissolved carbonate ions. In orderto recoverthe sodiumhydroxide, the resultantsodium carbonate (NazCO3) solution is mixed with calciumhydroxide (Ca(OH)J to producesodium hydroxide and calcium

0 2004 World ResourceReview. All rightsreserved. 158 World Resource Review Vol. 16 No.2 with (Ca(OH)2)to producesodium hydroxide and (CaCO3) in a reactionknown as causticizing. This reactiontransfers the carbonateanion from the sodiumto the calcium cation. The calcium carbonateprecipitates, leaving behind a regeneratedsodium hydroxide sorbent. The precipitate is dried, washedand thermally decomposedto producelime (CaO). This step. known as calcination, is followed by hydratingthe lime. known as slaking, which completesthe process.The processof recycling of sodiumhydroxide using calcium hydroxide hasbeen in usein the pulp and paper industry since 1884,where it is known as the Kraft Process(Miner and Upton, 2002). This processalso involves the calcinationof limestone,which is at the heart of the cementmanufacturing industry. In short. eachunit processrequired is known to be technicallyfeasible and the only remaining questionssurround their efficienciesand costs. This paperwill presenta re'tiew of the indi~idual components comprisingthe air extractionprocess with a view to highlighting the potential benefitsand concerns. The issuessurrounding their integrationinto a functional air captureS}.stem will also bediscussed. The layout of the paperwill follow the path of the C~ moleculethrough the system. A simplified schematicof the processis shownin Figure 1.

Figure 1 Overviewof Air ExtractionProcess. Note that two reactionsare not shown; !hying (d), and hydrating(h).

0 Z~ WorldResourre Review AD rights reserved 159 World ResourccRcyicw Vol. 16 No.2

2 ALKALINE SODIUM SORBENTS

The removal of a gaseouscomponent through contactwith a liquid is known as wet scrubbihg. Wet scrubbingcan be divided into processeswhere there is a chemicalreaction between the sorbateand the sorbentand wherethe sorbateis physically dissolvedinto the sorbentsolution. For the air extraction processwe proposean alkaline sodiumsolvent which reactschemically with the entrainedCO2. The chemicalreaction for this processis shownbelow as reaction (1). 2NaOH (aq) + CO2(g) -+ Na2CO3(aq)+ H2O; (1: The aqueouscarbonate reaction can be simplified by omitting the cation, resulting in the following ionic reaction.

20H-(aq) + CO2 (g)-+ CO32-(aq)+ H2O 0) (2) 6.GO= -56.1 kJ/mol (Lllio= -109.4 kJ/mol)

Note that the enthalpyand free energyof the reactionare for a nominal I molar solution. The thermodynamicdata, given at 298K and a pressureof 1 bar, was obtainedfrom the availableliterature (Lide, 2000). As a comparison,the free energyof mixing CO2with nitrogenand oxygento form ambientair is given by

~G = RT In (P atrnlPCoJ-20 kJ/mol. (3)

Clearly, sodiumhydroxide provides a sufficientdriving force to effectively collect CO2from ambientair. Eventhough a lower binding energy might be desirable,the high binding energyof chemicalsorbents proves useful in absorbingCO2 from streamswith low partialpressures of CO2(White, et al.,). As an alternativewith a weakerbinding energy,one may considersodium or potassiumcarbonate buffer solutionsas a sorbent. In this casethe absorptioncan be describedby:

CO2(g)+ CO]2- + H2OQ) -+ 2 HCO]' (4)

IlGO = -14.3 kJ/mol (ilHO=-27.6 klimat)

Even in this caseit is possibleto mustera sufficientthermodynamic driving force to removeCO2 from the air. For a two molar solutionof bicarbonate

2004World Resource Review. All rightsreserved. 1(;0 World Resource Review Vol. 16 No.2

ions, the free energyof the reactionfrom ambientair is negativeif the bicarbonate concentrationstays below 0.15 molar. A similar resultcan be obtainedby calculatingthe massaction equilibriumusing empiricalvalues for the equilibrium constants. Reaction(4) is effectivelytrimolecular and is the result of a sequence of reactionswhich have fastkinetics at high temperaturesor very high carbonate to bi-carbonateratios. Otherwisethe processoccurs in the diffusion regime, which is much slower(Astarita, 1967),making this reactionimpractical for air extractionas it is kinetically limited. In a recentreview focusedon flue gas,White et al. (2003)do not specificallydiscuss absorption on sodiumhydroxide (NaOH), ratherthey consider aqueoussolutions of sodiumand potassiumcarbonate. These sorbents were comparedto mono-ethanolamine(MEA), the industrystandard, and it was concludedthat MEA providesa substantiallymore cost-effectivesolution (Leci and Goldthorpe,1997). At thesehigher CO2 concentrations there would be no advantagein the higherbinding energyof the hydroxideand the heatof this reactionis in any caseirretrievably lost. An additionalcomplication arises becausethe productcarbonate cannot be decomposedwith eitherheat or pressure. The resultantNazCa] mustbe efficiently decomposedchemically. We proposeto accomplishthe chemicaldecomposition using calciumhydroxide as an intermediary. A sodiumhydroxide solutionprovides a liquid sorbentthat is more easily cycled through a piping systemthan a calciumhydroxide suspension.Its binding energyis strongenough and its reactionkinetics fast enoughto obviatethe need for heating,cooling or pressurizingthe air. BecauseCO2 is so dilute any such actionwould result in an excessiveenergy penalty. The hydroxide solutionavoids all suchcomplications. Sincesodium hydroxide is cheaperthan potassium hydroxide our starting point for the air extractiondesign will be basedon sodium hydroxide.

2.1 Experimental Investigations into Sodium Hydroxide Sorbents The absorptionof CO2by NaOH solutionwas studiedin 1943by Tepe and Dodge (1943). Experimentswere performedusing a packedtower arrangementwith the inlet gashaving a CO2concentration of -2%. Tepeand Dodge investigatedthe effectsofNaOH concentration,NazCO) concentration, gas flow rate, solventflow rate, and solventtemperature on the CO2uptake rate. The effectswere quantified in terms of an over-all absorptionrate coefficient. The resultsshowed that the rate coefficientincreased with increasingNaOH

I) 2004 World ResourceReview. All rightsreserved. 161 World Resource Reyiew Vol. 16 No.2 concentrationto a maximumnormality of -1.8 and thendecreased. The rate coefficientdecreased with increasingN~CO3 concentration.The gas flow rate had no effect, while the rate coefficientincreased with increasingsolvent flow rates. The temperatureof the solventaffected the rate coefficient significantly with the rate coefficientfollowing the temperatureof the solventto the 6thpower. The absenceof sensitivityto gas flow rate and high sensitivityto temperature suggeststhat for theseCO2 concentrations, the reactionis limited by transport resistancein the liquid phase.More recentresults on MEA solutionsalso show that at elevatedCO2 concentrations absorption is limited by resistancein the liquid phase(Aroonwilas et al.,2001). Generallyin wet scrubbing,one can breakthe transportresistance into two distinct components,the air side and liquid sideresistance. The air side resistanceis dominatedby the diffusion barrier in the laminarboundary layer. Typically, sucha boundarylayer also exists on the liquid side. In the bulk fluid, dissolvedCO2 reacts with water,or hydroxide ions to form carbonateor bicarbonateions. In contrastto the reactionsof CO2with water,the reactions with hydroxidereactions are very fastand their reactiontime can be ignored (Astarita,1967). However,since diffusion coefficientsof CO2in air are roughly four orders of magnitudelarger than ionic diffusion coefficientsin water, it is easy to becomerate limited on the liquid side. The experimentsof Tepeand Dodge suggestthat the transportresistance is dominatedby the liquid phase. For a onemolar carbonateion concentrationin the liquid, the concentrationratio betweencarbonate ions in the fluid and CO2molecules in the gasis 66,000: I. It thus will take time to fill up a boundarylayer on the liquid side. This suggeststhat at sufficiently low partialpressures of CO2the extraction processwill be limited by air~sideresistance. The absorptibnof CO2from atmosphericair using an apparatussimilar to that of Tepeand Dodgewas studiedby Spectorand Dodge (Spectorand Dodge, 1946). For ambientair, CO2absorption rates were proportionalto G", whereG is the air flow rate and the coefficient a.varies from 0.35 at low flow ratesto 0.15 at high flow rates. This suggeststhat at low CO2concentrations, 0.031 %, the liquid side resistanceto transportceases to be dominant. It is likely that in theseexperiments, fluid surfaceregeneration was sufficiently fastto preventa built up of liquid-side flow resistance.The experimentsdid showthat wet scrubbingusing NaOH can effectively removeCO2 from atmosphericair at a rate wherethe unavoidableair

c 2004World Resource Review. All rightsreserved. 162 World Resource Review Vol. 16 No.2

side resistanceto flow hasbecome important. Spectorand Dodgeobtained approximately90% removal for atmosphericconcentration levels. The advantageof using a stronghydroxide for CO2capture is a high load capacityand a fast reactiontime. The resultsby Spectorand Dodge suggestthat a systemcan be built that will be limited by transportresistance in the air side of the air-liquid contactsurface. In a regimewhere the dominanttransport resistance is on the air side, it is possibleto estimatethe sizeof the CO2extractor by the air dragthe extractor causeson the flow. Apart from the pressuregradient driven momentumflow, momentumtransfer to the wettedsurface follows a similar transportequation as the CO2diffusion. As a consequence,a systemthat incurs a pressuredrop roughly equalto pV2,which extractedvirtually all of the initial momentum,will be ableto extracta substantialfraction of the CO2from the flow. To setthe scaleof the operation,at 10m/sthe air flow throughan openingof 1m2carries a CO2load that equalsthe CO2produced by generating70 kW of heatfrom coal(Lackner, Grimes and Ziock, 1999). A 100MW powerplant operatingat 33% efficiencywould require 9,000m2of wind crosssection, if CO2collection efficiencywould be about 50%.

3 CAUSTICIZATION Causticizationrefers to the transfonnationof sodiumcarbonate into sodiumhydroxide. It is generallyperfonned by addingsolid calciumhydroxide to the sodiumcarbonate solution. The solubility of calciumhydroxide is suchthat this fonDSan emulsionaccording to reaction(5).

(5)

This reactioncan alsobe written in its ionic form as follows.

(6)

dGO = -18.2 kJ/mol (L\HO= -5.3 kJ/mol)

This process step regeneratesthe sodium sorbent. The CO2 is removed as a solid through a filtration process. Zsako (1998) noted that one could also start with lime, which would slake immediately in the aqueous solution, but in air

c 2004World Resource Review. All rightsreserved. 163 World Resource Review Vol. 16 No.2 extraction, it is importantrecover the heatof the slaking reactionat elevated temperatures.Thus we have separatedthe slaking stepfrom the causticization. Experimentshave shownthat the causiticizationrate increaseswith temperature(Dotson and Krishnagopalan, 1990). The initial sodiumcarbonate concentrationfor those experimentswas -2.0 mol/l andthe sampleswere subjectedto constantstirring. Reaction(5) eventuallyapproaches equilibrium and causticizingefficiency is generallyin the rangeof 80 to 90%. Causticizing efficiency refersto the amountof sodiumcarbonate converted to sodium hydroxide. Dotson (1990)determined that the rate constantfor the reaction(5) increasedby a factor 3 asthe operatingtemperature was raisedfrom 353 to 393K. The rate constantdropped when the feed solutioncontained sodium hydroxide. The experimentalcausticizing efficiency for pure sodiumcarbonate and a mixture of sodiumhydroxide and sodiumcarbonate were -94% and -85% respectively. The rate constantfor the causticizationis driven by the concentrationof free Ca ions in solution. Highly alkaline solutionswill limit the availability of dissolvedCa++ at anytime and consequentlyreduce the rate of conversion. Elevatedtemperatures and active stirring reducediffusional resistance and thus will increasethe rate of reactions. It was also noted that the concentrationsof all the calciumspecies remain essentiallyconstant throughout the reactionsdue to their low solubility. This suggeststhat the efficiency and rate constantsmay changeif insufficient calcium is present. Dotsonprevented this occurrenceby using a 10%stoichiometric excessof lime. However,this excessresults in solid calciumhydroxide being entrainedwith the filtrate. This would producehigher energyconsumption in the lime kiln dueto the dehydrationreaction mentioned by Zsako(1998). The concentrationsof the variousspecies, both sodiumand calcium,have a profound effect on the quality of the resultantfiltrate. Konno et al. observed that the solid phaseof calcite is unstablein pureNaOH solutiongreater than 2 moVl and easily convertsto Ca(OH)2(Konno, Yasunoriand Kitamura,2002). Thesesolids becomestable in a I moVl NaOH solutioncontaining at least0.02 mol/l N~CO3. This latter solutionmixture suggeststhat for an initial sorbent concentrationof I mol/L, 96% of the hydroxideions were convertedto carbonate accordingto reaction(2). The presenceofN~CO3 alsoreduces the solubility of calcite. The solubility ofCa(OH)2 is stronglydependent on the NaOH concentrationand drops by a factor4, to 5 x 10-4moVl, as the NaOH increases from 0 to 0.5 moVl. Konno also observedthe concentrationsofCa2+ and NaOH

c 2004World Resource Review. All rightsreserved. 164 WorJdResourceReview Vol. 16 No.2

during the reaction. As expectedthe Ca2+concentration dropped and the NaOH concentrationincreased as the reactionprogressed. The initial Ca2+concentration was -1 x 10-3moM. Konno alsonoted that Ca(OH)2super saturation ratio is the driving force for nucleation. In effect, theseprocesses balance the solubility of calciumhydroxide againstthe solubility of calciumcarbonate. The values for the dissociationconstants are availablein the literature (Snoeyinkand Jenkins,1980).

[Ca~[OH-r < KoH=10-'49mor/l (8) [Ca++][CO32-] < Kco3=10-3.22mor/l (9) Given that the calcium concentrationis the samein both (8) and (9), we can solve for the carbonateconcentration as follows.

[CO]20]=

Assuminga 1 molar sodiumsolution we can neglectthe effect of calcium on the chargebalance and the sodiumconcentration must therefore balance all the negativeions. Ifwe furtherdefme the causticizingefficiency (E)as the ratio of hydroxide ions over sodiumions we obtainthe following relationship.

-1 &= (2~[OH-]+ KOH

The stablesolution suggested by Konno would containapproximately 1 mol per liter hydroxide ions. This would suggesta theoreticalcausticizing efficiency of 96%, which is slightly higherthan the experimentalvalue of Dotson et al. (1990). The differenceis likely dueto the omissionof ionic activity in the calculations. The experimentalwork discussedabove provides a pathway for recoveringsodium hydroxide from sodiumcarbonate. In the process,the carbon dioxide has beentransferred into a solid form of calciumcarbonate, which can be readily removed from the liquid and afterwashing and drying it can be thermally decomposed.The causticizationtakes place in an emulsionof calciumhydroxide. There are a numberof options for the implementation,but this unit processis well establishedin the pulp and paperindustry.

c 2004World Resource Review. All rightsreserved. 165 World Rcsourcc Rcvicw Vol. 16 No.2

4 CALCINATION OF LIMESTONE

The fmal stageof the air extractionprocess is the recycling of the calcite precipitate. This is accomplishedthrough thermalregeneration or calcination. Lime and limestoneare amongthe oldestmaterials used by humanity with the fIrst recordeduse in the Egyptianpyramids. The fIrst soundtechnical explanation of lime calcinationcame in the 18thcentury from the British chemist,Joseph Black (Boynton, 1966). Thereare three essentialfactors in the kinetics of dissociation; the dissociationtemperature, the durationof calcination,and the CO2in the surroundingatmosphere. The reactionis shownbelow.

CaCO3(s) -4 CaO(s) + CO2(g); LiliO = + 179.2 kJ/mot 12)

The fIrSt quantificationof the thennal decompositionwas perfonned in 1910and the resultwas a decompositiontemperature of 1171K in a 100%CO2 atmosphereat atmosphericpressure (Johnston, 1910). Currentpractices use lime kilns to dissociatethe calcite which vary greatlyin their perfonnance. The most importantperfonnance metric for air extractionis the thennal efficiency, which is the product of the theoreticalheat requirement and the availableoxide content divided by the total heatrequirement. Boyntoncompares three hypotheticalkilns and basesone on "the lowestfuel efficiency on record, that of the mostadvanced Gennanmixed-feed kiln." This advancedkiln is reportedto obtaina thennal efficiency of 85% for 93%available lime. The thennal efficiencyrefers to the proximity to the theoreticalminimum heatrequirement as defmedby reaction(12) while the availablelime refersto the amountof inert materialpresent, in this case 7%. This translatesinto a total heatrequirement of 3 .03 MMBtu per ton of lime or 4.5 GJ per tonne of CO2,This latterfigure is lower than the 4.8 GJ pertonne of CO2used in a previousair extractionfeasibility study(Zeman, 2003). The thennodynamicminimum heatrequirement of 4.1 GJ/tonneCO2 can be calculated from the enthalpyvalue in reaction(12). The potentialcost of air extractionwill be dominatedby this reactionand any improvementregarding the threekinetic factorslisted abovewill directly affectthe cost of the project. A lower dissociationtemperature will require less heat input as will a shorterduration of calcinationand a lower CO2content in the surroundings.Garcia.,Labiano et al. found that calcinationrate increasedwith temperaturein a neutralenvironment and decreased with increasingCO2 partial pressureand total pressure(Garcia-Libiano et al., 2002). Khinast et a1.echoed

c 2004World Resource Review. All rightsreserved. 166 World Resource Review Vol. 16 No.2

similar fmdingsregarding the effects of CO2partial pressure but also found a decreasein reactionrates with increasingparticle size (Khinast et al., 1996). The effect of stearnon calcite decompositionwas also investigated. It was found that stearnenhanced the calcinationrate and evencompensated for the presenceof CO2over the solid (Thompsonand Wang,1995). Thomson'sexperiments producedgreater than 95% conversionafter 40 minutes in an environmentof 21 % stearnand 79% helium at a temperatureof 753 K. The positive effects of stearn on calcinationmay be synergisticwith other advancementsin improving the efficiency of calcination.In the caseof lime mud calcination,Theliander proposed a novel stearndrying systemthat reducesthe overall systementhalpy by 33% (Thelianderand Hanson,1993). Thereare somenovel designsfor calcination units as well. The decompositionof limestonewas testedin a solar reactorand achievedan averageof -50% conversion(Imhof, 1997).

5 AIR EXTRACTION AS CARBON CAPTURE We areproposing to captureCO2 directly from the atmospherein a cost effective manner. As such,a brief comparisonwith the industrystandards will provide a benchmarkfor future work. Stericallyhindered amines (SHA) and MEA are consideredpotentially successfulCO2 capture technologies. They are regeneratedusing steamand their thermalenergy requirements are 700 and 900 kcal/kg CO2for KS-2 andMEA, respectively(Mimura et al., 1997). Thesevalues canbe convertedto 2.9 and 3.8 GJ/tonne CO2. It shouldbe noted that Mimura et al. captured90% of the CO2generated by the powerplants; the remainderwould haveto be mitigated by othermeans. An economicanalysis of CO2capture using MEA obtaineda cost of $50 per tonne of CO2avoided (Zappelli et al., 2003). Calciumbased sorbents have also beeninvestigated for use in CO2capture. The thermalrequirements will follow the limits outlined above;however, the durability of the sorbentis also an importantcost factor. This has beendiscussed for dry CO2cycles. Abanades summarized several studies of the carbonation/calcinationcycle and found very good agreement(Abanades, 1998). Starting at around80% conversion,performance rapidly droppedto lessthan 20% conversionat 14cycles where it stabilized. Thesetests were conductedunder different conditionsthan proposed for air extraction.Most importantlythe air extractionprocess would hydrate or slakethe resultinglime. Generallylime is regeneratedin the hydrationprocess, thus we do not expecta greatreduction in captureefficiency from one cycle to the next. The processof hydrationis

0 2004 World ResourceReview. All rightsreserved. 167 CaO(s)+ H2O (I) -+ Ca(OH)2(s) LiliO = -64.5 kllmot (13) The hydrationcauses both expansionand the liberation of heatwhich causesthe particle to split, exposingfresh surfaces and therebyreducing the effectsof sintering. The inclusion of this reactionin the carbonationprocess will likely alterthe perfonnanceand durability of the lime cycle, as will the lower temperaturesof reaction. We believethat thesedifferences preclude the useof durability datapresented by Abanadesas a deterrentfrom further study. Zsako (1998)stated that slakedlime undergoesdehydration at temperaturesabove ~700K under ambientconditions. This defmesthe rangeof possibleoperating temperaturesfor the hydrationprocess. Hydration is highly exothennicand can provide usefulheat energyif it is perfonnedefficiently at high temperatures. Thereare othermethods for removingCO2 from the atmosphere including physicalabsorption and refrigerationamong others. Steinberg comparedeight differentmethods of CO2removal and found alkaline wet scrubbingto be the leastenergy intensive (Steinberg and Dang,1977). The estimatedenergy required to strip air of CO2was 0.4 kWh(e/lbmethanol or 4.4 GJ/tonne CO2. The CO2was strippedusing a 0.0 I N K2CO3solution and was desorbedby ,from ,and low pressuresteam. It shouldbe noted that this calculationdid not include lossesduring the conversionof primary fuel to electricity. The feasibility of air extractionwill dependon the overall cost in comparisonto alternateremoval technologies. The costper unit removalwill further dependon the energyrequirements, the durability of the sorbents,and costsexternal to the process.These external costs could include excessivewater lossesfrom the wet scrubbing.As this part of the processwill be in contactwith the openatmosphere, evaporation can be expected.This canbe minimized by adjustingthe sorbentconcentration which variesthe vaporpressure until it matchesthat of the ambientair. The thennophysicalproperties of sodium hydroxide solutionare known for a wide range of concentrations(Olsson, Jernqvistand Aly, 1997). Other externalcosts will be delineatedduring bench and pilot scaledemonstrations. The calcinationreaction is likely the most energy intensivefor the statedprocess. This highlightsthe need for efficient heat managementwithin the system. Additionally, any significantlime degradation~

0 2004World Resource Review. All rightsreserved. 168 .. 6 CONCLUSION Our previous feasibility study mentioned earlier obtained a price range of ~$25- 75 per tonne of CO2extracted (Zeman, 2003). The large range is due to the fluctuations in the price of natural gas and the uncertainty regarding the cost of solid membranes. This range does match up with the quoted cost of MEA capture and does suggest further studies may provide a viable option for extracting CO2 from the atmosphere. Air extraction is not necessarily more expensive than MEA capture in a flue stack. Even though the air contactors must be much larger than the equivalent contact surfaces in the flue stack, their contribution to the total cost may be very small. The recovery MEA sorbents requires similar amounts of energy, and becausethe air is clean and hydroxides are not subject to oxidative losses, the make-up costs are low in a hydroxide system. It is however clear, that the biggest challenge for air extraction will lie in the efficient management of heat and not in the design of the physical reactions. The information highlighted in this paper maps out a large parameter space. We believe that such a broad discussion will prove useful in designing new implementations of processes that have been optimized in the past for entirely different goals. Our objective is quite different from that of the individual studies cited. The overall goal of the process is the maximize CO2 capture while minimizing energy consumption. Variations in the operating conditions of each reaction should be investigated in order to quantify their effect on the entire process. For example, heating the fluid in the causticization reactor representsa compromise between increased energy consumption and longer reaction time. Sophisticated designs may maximize the amount of heat recaptured at elevated temperatures. For example, if the lime is hydrated using steam then the energy output of this reaction increases by the heat of condensation. This would be offset by steam generation elsewhere, but the higher quality heat obtained may be easier to harness. In the end, the minimum theoretical penalty for calcination is 4.1 GJ per tonne of CO2 and the maximum energy recoverable from hydration is 1.5 GJ per tonne of CO2. The resulting minimum net energy penalty is 2.6 GJ per tonne of CO2, which is already lower than the practical energy cost of amine solutions.~

0 2004 World ResourceReview. All rightsreserved. 169 World Resource Review Vol. 16 No.2

Theoretical efficiencies may not be achievable in practice but they are an indication of the potential of the method.

REFERENCES Abanades,J.C., The maximumcapture efficiency of CO2using a carbonation! calcinationcycle ofCaO/CaCO3,Chemical Engineering Journal, 90, 303-306 (2002). Aroonwilas, A., P. Tontiwachwuthikul., A. Chakma, Effects of operating and design parameters on CO2 absorption in columns with structured packings, Separationand Purification Technology,24,403-411 (2001). Astarita, G., Mass Transfer with , pp. 131-152, Elsevier Publishing Company, New York (1967). Boynton, R.S., Chemistry and Technology of Lime and Limestone, Interscience Publishers, New York pp. 3,255,258 (1966). Dotson,B.E., A. Krishnagopalan,Causticizing Reaction Kinetics, Tappi Proceedings1990 Pulping Conference,234-244 (1990). Garcia-Labiano,F., A. Abad, L.F. de Diego, P. Gayan, J. Adanez,Calcination of calciumbased sorbents at pressurein a broadrange of CO2concentrations, ChemicalEngineering Science, 57, 2381-2393(2002). Herzog,H.J., E.M. Drake, Carbondioxide recoveryand disposalfrom large energysystems, Annu. Rev. Energy Environ.,21, 145-66(1996). Houghton,J. T., Y. Ding, co-chairs,Climate Change 2001: the Scientific Basis, IntergovernmentalPanel on Climate Change(2001). Imhof, A., Decompositionof limestonein a solar reactor,Renewable Energy, 10 No. 2/3, 239-246(1997). Johnston,J.,.l Am. Chem.Soc., 32, p. 938 (1910). Khinast, J., G.F. Krammer, Ch. Brunner, G. Staudinger,Decomposition of limestone:the influence of CO2and particle size on reactionrate, Chemical EngineeringScience, 51(4), 623-634(1996). Konno, H., N. Yasunori, M. Kitamura, Crystallization of aragonite in the causticizing reaction, Powder Technology, 123, 33-39 (2002). Lackner,K.S., P. Grimes, H.J. Ziock, Carbondioxide extractionfrom air: Is it an option, Proceedingsof the 24th Annual TechnicalConference on Coal Utilization and Fuel Systems(1999). Leci, C.L., S.H. Goldthorpe,Assessment of CO2Removal from PowerStation Flue Gas,Energy ConversoMgmt., 33 (5-8), 477 (1997). Lide, D.R., editor in chief, CRC Handbook of Chemistryand Physics81st Ed., CRC PressLLC, Boca RatonFlorida (2000). c 2004World Resource Review. All rightsreserved. 170 World Resource Review Vol. 16 No.2

M~l~d, G., T.A. Boden,and R.J. Andres,Global, Regional,and National CO2 Emissions.In Trends: A Compendiumof Data on Global Change,Carbon Dioxide Information Analysis Center,Oak RidgeNational Laboratory,U.S. Departmentof Energy, Oak Ridge, Tenn.,U.S.A. (2003). Mimura, T., H. Simayoshi, T. Suda, M. Iijima, S. Mituoka, Developmentof energysaving technologyfor flue gas carbondioxide recoveryin power plant by chemicalabsorption method and steamsystem, Energy ConversoMgmt., 38, S57- 62 (1997). Miner, R., B. Upton, Methods for estimatinggreenhouse gas emissions from lime kilns at kraft pulp mills, Energy,27, 729-738(2002). Oates,J.A.H., Lime andLimestone: chemistry and technology,production and uses,p. 213, Weinheim:Wiley-VCH, New York (1998). Olsson,J., A. Jemqvist, G. Aly, Thermophysicalproperties of aqueousNaOH- H2Osolutions at high concentrations,International Journal of Thermophysics, 18(3),779-793 (1997). Snoeyink,V.L., D. Jenkins,Water Chemistry, p. 295, JohnWiley and Sons,New York (1980). Spector,N .A., B.F. Dodge,Removal of carbondioxide from atmosphericair, Trans.Am. Inst. Chem.Engrs., 42,827-48 (1946). Steinberg,M., V-D. Dang, Productionof syntheticmethanol from air and water using controlledthermonuclear reactor power -I. Technologyand energy requirement,Energy ConversoMgmt., 17, 97-112(1977). Tepe,J.B., B.F. Dodge,Absorption of CarbonDioxide by SodiumHydroxide Solutionsin a PackedColumn, Trans.Am. Inst. Chem.Engrs., 39, 255 (1943). Theliander,H., C. Hanson,Steam drying and fluidized bed calcinationof lime mud, Tappi Journal,76(11),181-188 (1993). Thompson,W.J., Y. Wang,The effects of steamand carbondioxide on calcite decompositionusing dynamic x-ray diffraction, ChemicalEngineering Science, 50(9), 1373-1382(1995). White, C.M., B.R. Strazisar,E.J. Granite, J.S. Hoffman, H.W. Pennline, Separationand Captureof CO2from Large StationarySources and Sequestration in GeologicFormations -Coalbeds and DeepSaline Aquifers,.J: ofthe Air & WasteManagement Association, 53, 645-715(2003). Zappelli, P., E. Gambarotta,V. Meda, A. Tortorella,CO2 separation from flue gas: Evaluationof a solid sorbentbased process, Proceedings of the 2nd Annual Conferenceon Carbon Sequestration,Exchange Monitor (2003). Zeman,F .S., An investigationinto the feasibility of capturingcarbon dioxide directly from the atmosphere,Proceedings of the 2nd Annual Conferenceon Carbon Sequestration,Exchange Monitor (2003). Zsako,J., M. Hints, Use of thermalanalysis in the study of sodiumcarbonate

c 2004World Resource Review. All rightsreserved. 171 World RcsourccRcyicw Vol. 16 No.2 causticization by means of dolomitic lime, Journal of Thermal Analysis, 53, 323- 331 (1998).

c 2004World Resource Review. All rightsreserved. 172