Carbon Dioxide Sequestration in Cement Kiln Dust through Mineral

Feb 5, 2009 - Department of Civil and Environmental Engineering, University of Michigan, 1351 Beal Avenue, Ann Arbor, Michigan 48109, Department of Ge...
17 downloads 10 Views 466KB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 1986–1992

Carbon Dioxide Sequestration in Cement Kiln Dust through Mineral Carbonation D E B O R A H N . H U N T Z I N G E R , * ,† JOHN S. GIERKE,‡ S. KOMAR KAWATRA,§ TIMOTHY C. EISELE,§ AND LAWRENCE L. SUTTER| Department of Civil and Environmental Engineering, University of Michigan, 1351 Beal Avenue, Ann Arbor, Michigan 48109, Department of Geological and Mining Engineering and Sciences, Michigan Technological University, 1400 Townsend Drive, Houghton, Michigan 49931, Department of Chemical Engineering, Michigan Technological University, 1400 Townsend Drive, Houghton, Michigan 49931, and Michigan Technological University Transportation Institute, 1400 Townsend Drive, Houghton, Michigan 49931

Received October 14, 2008. Revised manuscript received January 5, 2009. Accepted January 9, 2009.

Carbon sequestration through the formation of carbonates is a potential means to reduce CO2 emissions. Alkaline industrial solid wastes typically have high mass fractions of reactive oxides that may not require preprocessing, making them an attractive source material for mineral carbonation The degree of mineral carbonation achievable in cement kiln dust (CKD) under ambient temperatures and pressures was examined through a series of batch and column experiments. The overall extent and potential mechanisms and rate behavior of the carbonation process were assessed through a complementary set of analytical and empirical methods, including mass change, thermal analysis, and X-ray diffraction. The carbonation reactions were carried out primarily through the reaction of CO2 with Ca(OH)2, and CaCO3 was observed as the predominant carbonation product. A sequestration extent of over 60% was observed within 8 h of reaction without any modifications to the waste. Sequestration appears to follow unreacted core model theory where reaction kinetics are controlled by a firstorder rate constant at early times; however, as carbonation progresses, the kinetics of the reaction are attenuated by the extent of the reaction due to diffusion control, with the extent of conversion never reaching completion.

Introduction Mineral carbonation is a naturally occurring weathering process where alkaline earth metals (Ca, Mg) combine with carbon dioxide to form stable carbonates. Even though the reactions are highly exothermic, they occur slowly in nature during the gradual physiochemical weathering of silicate minerals, such as wollastonite: * Corresponding author phone: (734) 764-6350; fax: (734) 7632275; e-mail: [email protected]. † University of Michigan. ‡ Department of Geological and Mining Engineering and Sciences, Michigan Technological University. § Department of Chemical Engineering, Michigan Technological University. | Michigan Technological University Transportation Institute. 1986

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 6, 2009

+ 2CaSiO3 + 4H2O + 2CO2 f Ca2+ + 2HCO3 + 2H + SiO3 f CaCO3 + SiO2 · H2O + H2O + CO2 (1)

Because of the thermodynamic stability of the carbonate products, however, mineral CO2 sequestration is an attractive means for reducing carbon emissions (1, 2). To improve the viability of the process for application on an industrial scale, attempts have been made to accelerate the carbonation process in the laboratory by using chemically and mechanically pretreated silicate mineral feedstocks, such as olivine and serpentine (3-9). A majority of previous research has focused on the aqueous batch reaction of pretreated feedstocks and CO2 at elevated temperatures and/or pressures. However, the mining, processing (i.e., grinding to increase surface area, acid extraction of oxides), and reaction conditions (i.e., increased temperatures and pressures) required for mineral carbonation of raw feedstocks are energy intensive and costly (10). An alternative source of metal oxides is solid alkaline waste materials that are rich in calcium and/or magnesium. The oxide content and high surface area of alkaline wastes make them potentially suitable for carbonation reactions without the need for extensive preprocessing. Several studies have investigated the feasibility and/or reaction mechanisms of CO2 sequestration in industrial wastes such as steel slag and coal fly ash (5, 11, 12), municipal solid waste incinerator ashes (MSWI) (13, 14), and cement kiln dust (CKD) (15). In addition, the impacts of carbonation on the stability of the waste material have also been investigated, particularly for its reduction of pH and heavy metal mobility (e.g. 1-4, 6, 16). One advantage of the utilization of alkaline wastes for carbon sequestration is that many suitable solid residues are produced at sites that also produce significant quantities of CO2, such as steel and cement manufacturing facilities. The major disadvantage is that suitable waste products are not as abundant as raw feed stocks (e.g., olivine, wollastonite, serpentine). Therefore their overall potential for carbon sequestration is not as great (17). Nevertheless, alkaline solid wastes are worth considering for reduction of emissions at their source and for their potential to serve as a means to introduce mineral carbonation as a sequestration technology (17, 18). Of the industrial related carbon emissions, the cement industry generates the largest percentage of nonenergy related (i.e., process-based) CO2 emissions in the U.S (17, 19) during the calcination of CaCO3 and other raw materials to CaO-based products. For each ton of clinker or cement produced, the industry also generates approximate 0.15 to 0.20 tons of cement kiln dust (CKD) (20, 21). A few manufacturing facilities recycle all or a portion of their CKD into the raw material line entering the kiln. However the degree to which CKD can be recycled is often restricted by its composition (trace metal and contaminants) and regional alkali standards (i.e., potential for alkali silica reactions with aggregates), which varies widely within and between plants (20-22). This study aimed to investigate the potential of CKD to sequester CO2. A combination of batch and column experiments were conducted with three CKD types in order to, first, demonstrate the feasibility of mineral carbonation in waste products and show its viability as a potential sequestration technology, and, second, to improve the fundamental understanding of the mineral carbonation process within alkaline wastes, specifically to identify those mechanisms that may be rate or extent limiting at ambient temperature and pressure conditions and without any modifications to 10.1021/es802910z CCC: $40.75

 2009 American Chemical Society

Published on Web 02/05/2009

TABLE 1. Elemental Composition of the Precarbonated Cement Kiln Dust (CKD) Samples, along with Estimated Mass Fractions of Initial Calcium Carbonate, Unbound CaO, and the Theoretical Capacity of the Material for CO2 Sequstration oxide Na2O MgO Al2O3 SiO2 P2O5 K 2O CaO TiO2 MnO Mn2O3 Fe2O3 SO3 Cl– Br– LOIe total initial CaCO3f Unbound CaOg %ThCO2h

percent of dry mass Typical Range Chanuteb Midlothianc Continentald for CKDa 0–2 0–2 3–6 11–16 – 3–13 38–50 – – – 1–4 4–18 0–5 – 5–25 – – – –

0.65 1.39 3.45 13.31 0.06 7.04 48.03 0.23 0.06 – 2.14 2.73 5.22 0.21 14.88 99.39 14.5 37 31.9

0.49 0.55 4.09 12.61 0.11 4.03 47.14 0.22 0.13 – 1.79 7.66 0.18 0.03 20.37 99.40 46.4 8 15.2

0.26 2.22 4.02 15.53 0.08 2.01 46.41 0.25 – 0.10 1.28 2.04 – – 25.80 100.00 57.6 – 13.7

a Values reported by Corish and Coleman (25). b From alkali bypass system of a dry kiln, Ash Grove Cement Company, Chanute Manufacturing Facility, Chanute, KS. c From a baghouse or electrostatic precipitator associated with a wet kiln, Ash Grove Cement Company, Midlothian Manufacturing Facility, Midlothian, TX. d From a baghouse or electrostatic precipitator associated with a wet kiln, Continental Cement Company, Hannibal, MO. e LOI: Loss on ignition. f Calculated (within (5%) from thermal mass-loss curves obtained by thermal gravimetric analysis of precarbonated samples. Calcite is assumed to be the predominant carbonate species formed during the carbonation reactions. g Obtained from Rietveld refinement of X-ray diffraction patterns: supplied by Ash Grove Cement Manufacturing Co. (Overland Park, KS) for precarbonated samples. No information was available for the sample from Continental Cement Manufacturing Co. h Based on eq 9; refers to the mass of CO2 that can be sequestered per initial dry mass of waste. The error associated with the calculated theoretical extents is (5% based on the uncertainty of CaCO3 estimations from TGA (refer to 6 above).

the waste. The work presented here was part of a larger study examining the carbonation potential of CKD, both fresh and from aged waste piles (15).

Materials and Methods Fresh CKD from three different cement manufacturing processes were used in this study, including two separate wet kiln systems and the alkali bypass system of a recently renovated dry kiln (Table 1). The CKD samples were received dry and stored at room temperature in airtight containers. Representative subsets were taken from each homogeneous mix for use in the carbonation experiments. Batch Experiments. The batch experiments were performed in a stainless steel 288 L (60 by 60 by 88-cm long) chamber, with ∼100% relative humidity and ∼ 80% CO2 atmosphere. One- to five-gram samples of CKD were placed in aluminum weighing tins, oven-dried at 80 °C for 24 h, and combined with nanopure, deionized water at desired waterto-solids ratios (0 to 1.25) (Supporting InformationFigure S1 and Table S1). Prepared samples were weighed and placed in the chamber for reaction with CO2 under ambient temperature and pressure conditions. Each experiment was carried out in triplicate. Control samples of raw kiln feed (i.e., no reactive oxide content) from the Ash Grove, Chanute,

Kansas, manufacturing facility were also prepared and reacted in the same manner to verify that observed dry mass change was due to reaction with CO2. During the experiments, the chamber was constantly supplied with humidified CO2 to replenish the carbon dioxide sequestered during the reactions and to create a slight positive pressure in the chamber, preventing the diffusion of laboratory air into the glovebox apparatus. The gas within the chamber was continuously circulated using a 12 V, 0.15 A fan (2000 rpm with air flow of 41.9 cfm), and CO2 concentrations were monitored with a MTI Quad Model Q30L gas chromatograph (Fremont, CA). Humidity and temperature within the chamber were monitored with a certified-traceable Humidity/Temperature Pen (Control Company, Friendswood, TX). After the specified reaction time, samples were removed from the chamber, weighed, and oven-dried at 80 °C for 24 h. The final dry mass of the samples was recorded, and the samples were then stored at room temperature in sealed 20-mL vials for material characterization. If needed, duplicate samples were combined to produce a bulk sample sufficient in size for material analysis. Material Characterization. The elemental composition of the precarbonated CKD samples was analyzed with a Philips PW 2404 X-ray spectrometer (Cambridge, UK). In addition, qualitative analysis of mineralogical composition in pre- and postcarbonated samples was assessed using powder X-ray diffraction (Siemens D 500 Diffractometer, Cherry Hill, NJ) equipped with a Cu X-ray tube operated at 50 kV and using 27-mA, medium-resolution slits. The carbonate content (wt% CO2) of pre- and postcarbonated samples during the batch experiments was determined by complementary analytical and empirical methods: (1) observed dry mass change of the samples, (2) thermal gravimetric analysis-differential temperature analysis (TGADTA), and (3) quantitative X-ray diffraction (QXRD) analysis using the relative intensity ratio (RIR) method. The gain in dry mass between the initial precarbonated samples and dried carbonated product was assumed to equal the mass of CO2 sequestered by the sample: MCO2(t) ) M(t) - M(t ) 0)

(2)

where MCO2(t) is the mass of CO2 sequestered after time, t, M(t) is the dry mass of the sample corrected for the mass gain due to hydration reactions, such as gypsum formation and/or the production of calcium-silicate gels, and M(t ) 0) is the initial dry mass of the sample. The change in carbonate content between pre- and postcarbonated samples is expressed in terms of %CO2 (dry mass): %CO2(t) )

MCO2(t) M(t)

× 100%

(3)

In order to verify that observed mass increases were due to CO2 sequestration, a subset of the samples were examined by a combination of thermal decomposition (TGA-DTA) and X-ray diffraction methods. The overall carbonate content of selected pre- and postcarbonated CKD samples was measured with a TAInstruments Simultaneous TGA-DTA (Model SDT 2960, New Castle, DE). Representative samples (∼22 mg) were heated in corundum crucibles in a nitrogen environment to 1100 at 20 °C/minute. Sample mass losses were monitored by TGA. Phase changes were simultaneously evaluated with DTA by measuring the temperature differences between the sample and inert, heat-treated corundum. Carbonate decomposition in both pre- and postcarbonated samples was observed in the temperature region of 500-850 °C. Other mass-loss regions associated with the loss of hydrated water (1000 °C), were also observed (Supporting Information, Figure S2). In addition, particle size and distribution analysis was completed on precarbonated samples using a Microtrac X100 laser particle size analyzer with an automated small-volume recirculation and a range of 0.04 to 700 µm. Degree of Carbonation. The degree of carbonation (ξ(t)) was estimated by comparing the observed change in dry mass of each sample (MCO2(t)) with the calculated theoretical extent of carbonation based on the composition of the precarbonated waste: ξ(t) )

MCO2(t) M(t ) 0) · ThCO2

(4)

where ThCO2 is the theoretical mass fraction of CO2 sequestration achievable based on stoichiometry and the reactive-oxide content of the waste. It was assumed that reactive calcium species in the waste (e.g., free CaO, Ca(OH)2, Ca2SiO3) were the major phases participating in the carbonation reactions. The carbonation of CaO (through Ca(OH)2) can be expressed form as (23): + 2CO2(g) + H2O(l) f 2H(aq) + CO3(aq)

(5)

CaO(s) + H2O(l) f Ca(OH)2(s)

(6)

2+ + 2OH(aq) Ca(OH)2(s) f Ca(aq)

(7)

2+ 2+ CO3(aq) f CaCO3(s) Ca(aq)

(8)

Because of the wide range in CKD composition, however, other oxides (e.g., MgO, K2O, and Na2O) may also contribute to the sequestration of CO2 through a variety of reaction pathways. Therefore, in order to estimate theoretical CO2 sequestration, consideration must be given to the waste composition and the extent to which the oxides are available for reaction. Similar to estimates of CO2 sequestration in mortars and concrete by Steinour (24), the theoretical extent (as a percentage of dry mass) of carbonation in CKD was calculated as follows: %ThCO2 ) 0.785(%CaO - 0.56 · %CaCO3 - 0.7 · %SO3) + 1.091 · %MgO + 0.71 · %Na2O + 0.468(%K2O - 0.632 · %KCl) (9) The mineral species in eq 9 are represented in percent dry mass, and the stoichiometric mass factors assume that all of the CaO (except that bound in CaSO4 and CaCO3) will form CaCO3, MgO will form MgCO3, and Na2O and K2O (less that bound in sylvite, KCl) will form Na2CO3 and K2CO3. Column Experiments. The importance of gas transport limitations on carbonation were assessed through a series of column experiments conducted under steady gas flow conditions with uniformly packed Chanute CKD and controlled influent mixtures of nitrogen, water vapor, and CO2. Operating conditions were systematically varied to examine the effects of changing column conditions on the extent of carbonation. Holding other variables constant, flow rate and influent CO2 concentration were varied in nine column experiments (Supporting InformationTable S2). Applied gas flow rates (20, 40, 80 mL/min) were selected such that the residence time of the influent gas ranged between 0.5 and 2 min, which would be comparable to residence times in a plant-scale system based on average CKD and flue gas generation rates. Influent CO2 concentrations (50 000, 100 000, 150 000 ppmv) comparable with CO2 composition in cement manufacturing off-gas were used (14 to 33% (26)) while maintaining a constant gas flow rate of 40 mL/min. In addition to the series of column experiments conducted on the Chanute CKD, the extent of carbonation at the column 1988

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 6, 2009

scale was also assessed in the Midlothian and Continental CKDs at a gas flow rate of 40 mL/min and an influent CO2 concentration of 100 000 ppmv. In all column experiments, an initial gravimetric water content of 20% was used to ensure both sufficient gas-filled porosity for well distributed vapor flow across the column and availability of water so that it was not a limiting factor in the reaction. Column experiments were conducted in a 6-cm long, 5-cm diameter glass column (Ace Glass, Inc., Vineland, NJ), fitted with Teflon end caps (No. 50 thread) and viton O-ring face seals. The inlet cap was equipped with 10.9 mL (1.2-cm deep by 3.4-cm diameter) of 2.5-mm glass beads packed between two 150-mesh stainless steel screens. For each experiment, 120 g of dry CKD was mixed with 40 g of nanopure water. The moist CKD was packed into the column in 1-cm lifts with a 4.5-cm diameter tamping device (Supporting Information Table S2). The remaining CKD was divided into three samples for measurement of initial gravimetric water content. Prior to the study, porosity and bulk density of the CKD types were gravimetrically determined. High-purity carbon dioxide (CO2) and nitrogen (N2) (Airgas, Marquette, MI) were mixed and regulated with Mass Precision Gas Flow controllers (MC Series, Alicat Scientific, Tucson, AZ) to achieve the desired flow rate and input CO2 concentrations (Supporting Information Figure S3). The influent gas was humidified to near 100%. Effluent gas samples were analyzed with a gas chromatograph (MTI Analytical Instruments Quad 4 Model Q30L, Fremont, CA) until full breakthrough of the input CO2 concentration was observed. At the end of each experiment, the column was flushed with N2 gas and CO2 effluent concentrations were monitored to determine the amount of unsequestered CO2 within the column tubing, end caps, pore spaces, and dissolved in the aqueous phase. After nitrogen flushing, the columns were weighed, dismantled, and the gravimetric water content of the carbonated CKD was determined. Degree of Carbonation. The degree of carbonation achieved in each column was determined through frontal analysis of effluent CO2 concentrations. Based on preliminary column experiments (15), effluent analysis is an appropriate method for assessing the extent of sequestration. The overall degree of carbonation achieved in each column was determined, similar to the batch experiments, where the observed mass of CO2 sequestered from frontal analysis was compared to the amount theoretically possible for each CKD type (Table 1).

Results and Discussion The elemental compositions of the different CKD types used in this study are presented in Table 1 along with the typical range in oxide composition reported for CKD. The wastes were all rich in CaO and contain moderate to high amounts of K, Na, Fe, Si, and S oxides. The theoretical extent of carbonation calculated for each of the waste types based on their reactive-oxide content is shown at the bottom of Table 1. Inherent in this theoretical calculation is the assumption that the oxides in certain calcium and potassium phases (i.e., calcite, anhydrite, and sylvite) are unlikely to react with CO2 at ambient temperatures and pressures. The abundance of unreactive phases in the precarbonated wastes was deduced normatively, using material analysis and stoichiometry, along with simplifying assumptions regarding material composition. The fraction of calcite in the unreacted material has a significant impact on calculated theoretical capacities. On the basis of the theoretical calculations, the Chanute bypass dust has the greatest potential for sequestration, followed by the CKD collected from the Midlothian wet kiln. The smaller calculated theoretical extent in the CKD from Continental wet kiln is attributed to its higher initial mass fraction of calcite.

FIGURE 1. Degree of carbonation as a function of both reaction time and initial water-to-solids ratio: (A) Chanute, (B) Midlothian, and (C) Continental. Samples were reacted at 25 °C and >98% relative humidity. Error bars express range in calculated degree as a function of the uncertainty associated with both the mass change measurements and the theoretical calculation. Batch Experiments. Reaction Mechanisms and Products. Calcite and trace amounts of ankerite (Ca(Fe2+,Mg)(CO3)2) and dolomite (CaMg(CO3)2) were the only carbonate species detected by XRD. Calcite precipitation was also confirmed with TGA-DTA, QXRD (15), and scanning electron microscopy (SEM) microanalysis and element mapping (27, 28). From XRD the major and minor mineralogical phases identified in representative samples of the pre- and postcarbonated CKDs are summarized in Supporting Information Table S3. The major reactive calcium species identified in the precarbonated CKD samples were CaO and Ca(OH)2. In all cases, no free CaO was observed in the reacted samples. Similar to CaO, in most cases Ca(OH)2 is absent in carbonated samples (Supporting Information Table S3). However, small fractions of Ca(OH)2 were observed in reacted samples of the Chanute CKD, suggesting that the carbonation reactions had not yet reached completion. Selected samples were analyzed by thermal decomposition (TGA-DTA) and QXRD to confirm carbonate formation and to verify that observed mass change was an appropriate measure of sequestration extent (Supporting Information Table S4, Figures S2 and S4). The results of TGA-DTA for selected samples, expressed as %CO2, are presented in Supporting Information Table S2, along with the observed change in dry mass measured during the experiments. Gravimetric determinations of sequestration extent agree reasonably well with the other characterization techniques and appear an appropriate method for assessing the degree of carbonation achieved in the various batch experiments. Degree of Carbonation. The degree of carbonation or the ratio of observed to theoretical extent for the different CKD types was determined both as a function of water-to-solids ratio and time (Figure 1). The maximum degree of carbonation achieved was similar (∼80%) in each of the CKD types. However, on a mass basis (i.e., mass of CKD required to sequester a given mass of CO2), the Chanute sample exhibits the greatest overall potential for mineral carbonation, due mostly to its high free CaO content and relatively low initial mass fraction of CaCO3 (Tables 1, and Supporting Information Table S2). In general, the degree of carbonation increases with increased initial water:solids ratio (Midlothian and Continental; Figure 1). There is a limit, however, in the ability of water to promote the carbonation reactions and appears to be related to the material composition of the waste. The Chanute dust contains measureable amounts of the mineral sylvite, which readily absorbs water from the atmosphere. The combined impacts of higher initial water:solids ratios and adsorption of water from the gas likely caused diffusion limitations in the transport of CO2 and Ca2+ to and from

reaction sites, lowering the degree of carbonation with higher initial water-to-solids ratios in the Chanute CKD. Based on regression analysis of degree of carbonation with increasing water:solids ratio (Figure 1), a significant linear relationship (confidence level ) 95%) exists between degree of carbonation and water content in all CKD types. The lowering of carbonation extent with increased water content in the Chanute waste, however, is very small (35%) in carbonation with water content in the Midlothian and Continental CKDs. In all CKD types, the degree of carbonation increases with time, with greater than 75-80% of the carbonation occurring in less than 2 days and more gradual conversion of oxides to carbonates as the reactions progressed (Figure 1). The exponential shape of the time-dependent curve indicates a slowing of the reaction rate as extent of carbonation proceeds. Investigation of the use of CaO sorbents in the removal of CO2 from flue gas (29-31) have observed CaCO3 coatings and pore blockages in carbonated CaO samples. The observed result was a decrease in carbonation efficiency of the CaO sorbents with time and with repeated calcination/decalcination cycles. Similar limits to the decreases in the rate of carbonation extent with time were observed by Huijgen et al. (11, 16) in their study of the carbonation of steel slag. Although the exact reaction mechanisms may differ between carbonation in sorbents (direct gas-solid reaction) and those occurring in aqueous reaction systems, a general decrease in reaction rate with extent of carbonation was observed in both systems. Analogous precipitation patterns and limits were observed in this study. SEM EDX microanalysis investigations indicate micropore blocking due to the CaCO3 precipitation, along with the propagation of a carbonate shell in Ca(OH)2 particles that grew inward and became denser with time ((28), Supporting Information Figure S5). The skin or shell development on reactive particles appears to limit CO2 and/or Ca2+ diffusion to reaction sites, as well as the overall degree of carbonation in the waste, particularly for the Chanute CKD. Reaction Mechanisms. The specific carbonation pathways in wastes such as CKD are not fully known. Transportcontrolled mechanisms such as CO2 and Ca2+ diffusion to/ from reaction sites, boundary layer effects (diffusion across precipitate coatings on particles), dissolution of Ca(OH)2 at the particle surface, and pore blockage or precipitate coating development can all affect the rate and extent of carbonation and their effects can change with time. The results of a kinetic study examining the reactivity of fly ash mixtures doped with calcium hydroxide (Ca(OH)2) suggest that the reaction rates are a function of not only the hydroxide to fly ash ratio but also the extent to which calcium hydroxides have been consumed in the reactions (32). Similarly, mineral carbonation studies in waste steel slag hypothesize that the ratelimiting mechanism during carbonation is the diffusion of Ca2+ to reaction sites (11, 16), influencing the extent of reaction with time. Therefore, the kinetics of carbonation appear to be influenced by a combination of reaction mechanisms (e.g., surface-based and transport controlled), which individually can be difficult to measure. The work by Shih et al. (33) and Lee (34) were applied here to describe the extent of carbonation with time. Both studies were based on the classical unreacted or shrinking core-type model (e.g. (35)), where the time required to completely convert an unreacted particle into product is controlled by two reaction regimes: one driven by chemical reaction and the other by diffusion. Lee (34) modified the classical rate equations (which assume complete conversion) to one that more accurately describes the kinetic behavior in the diffusion-controlled regime of CaOcarbonation, where the ultimate conversion ranges from 70-80% and does not proceed to complete conversion VOL. 43, NO. 6, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1989

FIGURE 2. Observed and predicted degree of CO2 consumption in cement kiln dust (CKD) based on eq 11 in text.

(34). Although their study examined direct gas-solid reaction of CaO with CO2, a very similar rate expression was developed by Shih et al. (33) to describe the reaction kinetics of Ca(OH)2 with CO2 in a humid, low temperature environment. The rate expression used by Shih et al. (33) assumes that the rate of carbonation is controlled by surface reaction on the Ca(OH)2 particles, along with the coverage of reaction product on the particle surface. The assumption here, as with these other studies (33, 34) is that during the initial stages of reaction, the rate of carbonation is dependent on the rate constant, k, where k has the form time-1 (or first-order) and is likely influenced by a number of factors including surface area, humidity, temperature, and CO2 concentration. As conversion increases, the rate of reaction is attenuated by the extent of the reaction due to the precipitate of a carbonate coating on reacting particle surfaces (refer to Supporting Information Figure S5). Thus, the overall rate of conversion can be expressed as (34):

(

ε(t) dε(t) )k 1dt εu

)

n

(10)

Integrating eq 10 with an n ) 1 yields:

[

( )]

ε(t) ) εu 1 - exp -

kt εu

(11)

where ε(t) is the extent of conversion at time (t) and εu is the ultimate conversion possible. Equation 11 is essentially the same model proposed by Shih et al. (33). A similar rate expression was used by Lee et al. (34) to define rate constants for both the first and second phases of the reaction: surface reaction control regime and diffusion control regime, respectively. Although this could be done with our experiments, the limited number of data points (six measurements in time) makes the proper distinction and quantification of these two regimes dubious. Therefore, eq 11 was used to express the extent of carbonation as a function of time, where an overall or lumped εu and k were fit to the observed data rather than the two regimes individually (Figure 2). Thus, the rate constant, k, is assumed to be predominantly controlled by the chemical reaction regime (i.e., initial stages of reaction), and the ultimate conversion (εu) is predominantly controlled by the diffusion regime. The values for the lumped rate constant (k) and ultimate conversion (εu) were optimized for each CKD and for the CKDs overall by minimizing the root-mean-square of the 1990

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 6, 2009

sum of the normalized residuals between predicted and experimental degrees of carbonation (Table S5; Figure 2). From the results of these experiments, no clear relationship can be drawn between reaction rate and the initial mean particle size, specific surface area, or particle size distribution. Instead, the rate constants, k, appear to be inversely proportional to the amount of free CaO or Ca(OH)2 in the unreacted material; meaning that the greater the mass fraction of free oxides, the longer the first phase of the conversion will take. This assumption is supported by the results of the column experiments discussed later. Column Experiments. In order to assess the influence of changes in gas flow rate and concentration on sequestration performance, a series of seven column experiments were conducted using the Chanute CKD. Other operational parameters such as initial moisture content and gas conditions (humidified) were held constant. Effects of Flow and Concentration on Carbonation. The total amount of sequestration in each column was determined from both observed mass change and frontal analysis of column effluent (see Supporting InformationTable S6, Figures S6-S8). The overall sequestration performance in the Chanute columns (Supporting InformationTable S6 and S7) was comparable to that observed under the batch experiments. The average degree of carbonation achieved in the columns was 76.9% (standard deviation ) 2.7%, n ) 7), and under comparable conditions, the carbonation obtained during batch studies was 77.2%. This equivalent sequestration performance, both between the column and batch studies and among the columns, indicates a lack of macroscale effects (e.g., preferential flow paths) on overall carbonation at the column scale. Even though the complexity of the reaction system has increased, the degree of carbonation achieved was comparable to the batch experiments. In addition, no discernible trend in the degree of carbonation was observed with changes in influent CO2 concentration, indicating that physical (e.g., mass transport limitations) rather than reactive mechanisms may be more limiting under the conditions examined, as was indicated in the rate analysis conducted on the batch experiments. Carbonation performance by the columns was also compared to ideal behavior, in which all of the CO2 introduced to the system is consumed by the waste. For each column, the mass of CO2 injected and the corresponding mass sequestered with time were normalized by the theoretical extent of sequestration and plotted together in conjunction with the trend for ideal sequestration (Figure 3; Figure S6). On average, 85% (standard deviation ) 5.6%, n ) 7) of the sequestration is achieved while following ideal behavior, with the remaining carbonation taking place more slowly. Similar trends were observed in the batch experiments, where approximately 90% of the observed CO2 sequestration was achieved in less than2 days and additional carbonation occurred gradually over the remainder of the experiment (total time ) 8 days). Effect of CKD Composition. Two additional column experiments were also conducted with the Midlothian and Continental CKDs to assess the impact of material composition (e.g., reactive oxide content: free CaO and Ca(OH)2) on carbonation performance at the column scale. As with the Chanute CKD, the Midlothian and Continental CKDs achieved degrees of carbonation comparable to that obtained in the batch experiments (Supporting Information Table S7). The degree of carbonation achieved follows ideal behavior (all CO2 injected was consumed) and appears correlated to the free oxide content of the wastes. In other words, the greater the mass fraction of free oxide content (CaO and Ca(OH)2), the greater overall mass of CO2 sequestered by the material.

Supporting Information Available Schematics of laboratory setup, as well as environmental scanning electron microscope images, X-ray diffraction, and thermal gravimetric analysis of pre- and postcarbonated samples of CKD along with additional supporting tables and figures. This information is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited

FIGURE 3. Cumulative mass of CO2 sequestered with influent concentration: (a) Chanute CKD. Concentrations are normalized by influent concentration; and (b) Chanute, Midlothian, and Continental CKDs. Concentrations are normalized by theoretical mass of sequestration possible for each CKD type and influent concentration. Columns were operated at 40 mL/min with an influent CO2 concentration of 100 000 ppmv. On the basis of stoichiometry, variations in material composition, theoretical determinations of sequestration capacity, and assuming a sequestration efficiency of 80%, CKD generated within the U.S. has the potential to recapture between 0.74 and 5.12 Tg CO2 per year. From estimates of combustion and process-related CO2 emissions (21), this equates to up to 10-13% of the CO2 emitted from the calcination process (∼6.5% reduction in overall U.S. cement related CO2 emissions, or a 0.33% reduction in global CO2 emissions). Thus, the reuse of CKD for CO2 sequestration has the potential to be a valuable means for meeting current voluntary or future mandatory emission reduction goals. While the impacts of mineral carbonation in alkaline wastes may not compare in scale to those of other sequestration technologies, it is worthy of investigation because of the stability of the end products, beneficial use of waste materials, and favorable thermodynamics of the carbonation reactions.

Acknowledgments This work was funded by the Institute of Hazardous Materials Management (IHMM) and the Michigan Academy of Certified Hazard Materials Managers, along with support from the National Science Foundation (NSF) through a Sustainable Futures Integrated Graduate Education and Research Traineeship (IGERT). We thank Scott Schlorholtz at Iowa State University’s Materials Analysis and Research Laboratory for his assistance in material analysis.

(1) Huijgen, W. J. J.; Comans, R. N. J. Carbon Dioxide Sequestration by Mineral Carbonation Literature Review Update 2003-2004; Energy Research Centre of the Netherlands, 2005. (2) Huijgen, W. J. J.; Comans, R. N. J. Carbon Dioxide Sequestration by Mineral Carbonation Literature Review; Energy Research Centre of the Netherlands, 2003. (3) Lackner, K. S.; Wendt, C. H.; Butt, D. P.; Joyce, E. L.; Sharp, D. H. Carbon dioxide disposal in carbonate minerals. Energy 1995, 20 (11), 1153–1170. (4) Fauth, D. J.; Goldberg, P. M.; Knoer, J. P.; Soong, Y. Carbon dioxide storage as mineral carbonates. Abstr. Pap. Am. Chem. Soc. 2000, 220, U395–U395. (5) Fauth, D. J.; Soong, Y.; White, C. M. Carbon sequestration utilizing industrial solid residues. Abstr. Pap. Am. Chem. Soc. 2002, 223, U565–U565. (6) Fauth, D. J.; Baltrus, J. P.; Soong, Y.; Knoer, J. P.; Howard, B. H.; Graham, W. J.; Maroto-Valer, M. M.; Anderson, J. M. Carbon Storage and Sequestration as Mineral Carbonates. In Environmental Challenges and Greenhouse Gas Control for Fossil Fuel Utilization in the 21st Century; Kluwer Academic/Plenum: Boston, 2002. (7) Gerdemann, S. J.; O’Connor, W. K.; Dahlin, D. C.; Penner, L. R.; Rush, H. Ex situ aqueous mineral carbonation. Environ. Sci. Technol. 2007, 41 (7), 2587–2593. (8) Chen, Z. Y.; O’Connor, W. K.; Gerdemann, S. J. Chemistry of aqueous mineral carbonation for carbon sequestration and explanation of experimental results. Environ. Prog. 2006, 25 (2), 161–166. (9) O’Connor, W. K.; Dahlin, D. C.; Rush, G. E.; Dahlin, C. L.; Collins, W. K. Carbon dioxide sequestration by direct mineral carbonation: process mineralogy of feed and products. Miner. Metallurg. Process. 2002, 19 (2), 95–101. (10) Huijgen, W. J. J.; Comans, R. N. J.; Witkamp, G. J. Cost evaluation of CO2 sequestration by aqueous mineral carbonation. Energy Convers. Manage. 2007, 48 (7), 1923–1935. (11) Huijgen, W. J. J.; Witkamp, G. J.; Comans, R. N. J. Mineral CO2 sequestration by steel slag carbonation. Environ. Sci. Technol. 2005, 39 (24), 9676–9682. (12) Soong, Y.; Fauth, D. L.; Howard, B. H.; Jones, J. R.; Harrison, D. K.; Goodman, A. L.; Gray, M. L.; Frommell, E. A. CO2 sequestration with brine solution and fly ashes. Energy Convers. Manage. 2006, 47 (13-14), 1676–1685. (13) Bertos, M. F.; Li, X.; Simons, S. J. R.; Hills, C. D.; Carey, P. J. Investigation of accelerated carbonation for the stabilisation of MSW incinerator ashes and the sequestration of CO2. Green Chem. 2004, 6 (8), 428–436. (14) Li, X. M.; Bertos, M. F.; Hills, C. D.; Carey, P. J.; Simon, S. Accelerated carbonation of municipal solid waste incineration fly ashes. Waste Manage. 2007, 27 (9), 1200–1206. (15) Huntzinger, D. N. Carbon Dioxide Sequestration in Cement Kiln Dust through Mineral Carbonation. Ph.D. Dissertation, Michigan Technological University, Houghton, 2006. http:// services.lib.mtu.edu/etd/DISS/2006/Geological&MiningEng&Sci/ huntzingerd/diss.pdf. (16) Huijgen, W. J. J.; Comans, R. N. J. Carbonation of steel slag for CO2 sequestration: Leaching of products and reaction mechanisms. Environ. Sci. Technol. 2006, 40 (8), 2790–2796. (17) Huntzinger, D. N.; Eatmon, T. D. A life-cycle assessment of Portland cement manufacturing: comparing the traditional process with alternative technologies. J. Cleaner Prod. 2009, doi: 10.1016/j.clepro.2008.04.007. (18) Mazzotti, M.; Abanades, J. C.; Allam, R.; Lackner, K. S.; Meunier, F.; Rubin, E.; Sanchez, J. C.; Yogo, K.; Zevenhoven, R. Mineral carbonation and industrial uses of carbon dioxide; Intergovernmental Panel on Climate Change; Cambridge University Press: New York, 2005. (19) U.S. EPA. Inventory of U.S. Greenhouse Gas Emissions and Sinks; United States Environmental Protection Agency: Washington, D.C., 2005. VOL. 43, NO. 6, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1991

(20) Hawkins, G. J.; Bhatty, J.; O’Hare, A. Cement kiln dust production, management, and disposal. Portland Cement Association R & D Report, Serial No. 2737, 2003. (21) Van Oss, H. G.; Padovani, A. C. Cement Manufacture and the Environment, Part II: Environmental Challenges and Opportunities. J. Ind. Ecol. 2003, 7 (1), 93–127. (22) Helmuth, R.; Stark, D.; Diamond, S.; Moranville-Regourd, M. Alkali-Silica Reactivity: An Overview of Reseach; Strategic Highway Research Program Report SHRP-C-342. National Research Council: Washington D.C., 1993. (23) Snoeyink, V. L.; Jenkins, D. Water Chemistry; John Wiley & Sons: New York, 1980; p 463. (24) Steinour, H. H. Some Effects of Carbon Dioxide on Mortars and Concrete - Discussion. J. Am. Concr. Inst. 1959, 905907. Concrete Briefs. (25) Corish, A.; Coleman, T. Cement Kiln Dust. Cement 1995, 40– 42. (26) Hendriks, C. A.; Worrell, E.; de Jager, D.; Block, K.; Riemer, P. Emission Reduction of Greenhouse Gases from the Cement Industry. In Greenhouse Gas Control Technologies; Eliasson, B.; Riemer, P., and Wokaun, A., Eds., Elsevier: Oxford, UK, 1999. (27) Anderson, C. P. Effects of Carbonation on the Mineral Composition of Cement Kiln Dust. M.S. Thesis, Michigan Technological University, Houghton, 2006. http://services.lib.mtu.edu/ etd/THESIS/2006/Geological&MiningEng&Sci/andersonc/ thesis.pdf. (28) Anderson, C. P.; Sutter, L. L.; Huntzinger, D. N.; Giekre, J. S. Effects of Carbonation on the Mineral Composition of Cement

1992

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 6, 2009

(29)

(30)

(31)

(32)

(33)

(34) (35)

Kiln Dust. Presented at the Twenty-Ninth Conference on Cement Microscopy, Quebec City, PQ. Canada, May 20-24, 2007; Quebec City, PQ. Canada, 2007. Wei, S. H.; Mahuli, S. K.; Agnihotri, R.; Fan, L. S. High surface area calcium carbonate: Pore structural properties and sulfation characteristics. Ind. Eng. Chem. Res. 1997, 36 (6), 2141–2148. Mess, D.; Sarofim, A. F.; Longwell, J. P. Product layer diffusion during the reaction of calcium oxide with carbon dioxide. Energy Fuels 1999, 13 (5), 999–1005. Agnihotri, R.; Mahuli, S. K.; Chauk, S. S.; Fan, L. S. Influence of surface modifiers on the structure of precipitated calcium carbonate. Ind. Eng. Chem. Res. 1999, 38 (6), 2283–2291. Biernacki, J. J.; Williams, P. J.; Stutzman, P. E. Kinetics of reaction of calcium hydroxide and fly ash. ACI Mater. J. 2001, 98 (4), 340–349. Shih, S. M.; Ho, C. S.; Song, Y. S.; Lin, J. P. Kinetics of the reaction of Ca(OH)2 with CO2 at low temperature. Ind. Eng. Chem. Res. 1999, 38 (4), 1316–1322. Lee, D. K. An apparent kinetic model for the carbonation of calcium oxide by carbon dioxide. Chem. Eng. J. 2004, 100 (1-3), 71–77. Levenspiel, O. Chemical Reaction Engineering: An Introduction to the Design of Chemical Reactors; John Wiley and Sons, Inc.: New York, 1962; p 501.

ES802910Z