Sorbents for CO2 capture from high carbon fly ashes

Sorbents for CO2 capture from high carbon fly ashes

Available online at www.sciencedirect.com Waste Management 28 (2008) 2320–2328 www.elsevier.com/locate/wasman Sorbents for CO2 capture from high car...

255KB Sizes 0 Downloads 74 Views

Available online at www.sciencedirect.com

Waste Management 28 (2008) 2320–2328 www.elsevier.com/locate/wasman

Sorbents for CO2 capture from high carbon fly ashes M. Mercedes Maroto-Valer a,*, Zhe Lu b, Yinzhi Zhang b,1, Zhong Tang b,1 a

School of Chemical and Environmental Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, United Kingdom b Department of Energy and Geo-Environmental Engineering and The Energy Institute, The Pennsylvania State University, 405 Academic Activities Building, University Park, PA 16802, USA Accepted 31 October 2007 Available online 21 February 2008

Abstract Fly ashes with high-unburned-carbon content, referred to as fly ash carbons, are an increasing problem for the utility industry, since they cannot be marketed as a cement extender and, therefore, have to be disposed. Previous work has explored the potential development of amine-enriched fly ash carbons for CO2 capture. However, their performance was lower than that of commercially available sorbents, probably because the samples investigated were not activated prior to impregnation and, therefore, had a very low surface area. Accordingly, the work described here focuses on the development of activated fly ash derived sorbents for CO2 capture. The samples were steam activated at 850 °C, resulting in a significant increase of the surface area (1075 m2/g). The activated samples were impregnated with different amine compounds, and the resultant samples were tested for CO2 capture at different temperatures. The CO2 adsorption of the parent and activated samples is typical of a physical adsorption process. The impregnation process results in a decrease of the surface areas, indicating a blocking of the porosity. The highest adsorption capacity at 30 and 70 °C for the amine impregnated activated carbons was probably due to a combination of physical adsorption inherent from the parent sample and chemical adsorption of the loaded amine groups. The CO2 adsorption capacities for the activated amine impregnated samples are higher than those previously published for fly ash carbons without activation (68.6 vs. 45 mg CO2/g sorbent). Ó 2007 Elsevier Ltd. All rights reserved.

1. Introduction In 2005 around 70 million tons of fly ash were generated by electric utilities in US, and around 60% was disposed (ACAA, 2005). Similarly, in Western Europe (15 EU countries), the total production of fly ash was around 44 million tons in 2003, with only 47% being used in the construction industry (ECOBA, 2003). Moreover, the concentration of unburned carbon present in fly ash has risen drastically over the last few years, due to the implementation of increasingly stringent Clean Air Act Regulations regarding NOx emissions, which are mainly addressed in coal com*

Corresponding author. Tel.: +44 1158466893; fax: +44 1159514115. E-mail address: [email protected] (M. Mercedes Maroto-Valer). 1 Present address: Sorbent Technologies Corp., 1664 East Highland Road, Twinsburg, OH 44087, USA. 0956-053X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2007.10.012

bustion furnaces by a combination of low-NOx burners and catalytic reduction systems. Although low-NOx burner technologies efficiently decrease the emissions level by lowering the temperature of combustion, they also reduce the combustion efficiency with a corresponding increase in the concentration of unburned carbon in the fly ash (Maroto-Valer et al., 1998). This has restricted the principal use of ash in the cement industry, since the unburned carbon tends to adsorb the air-entrainment agents, which are added to the cement to prevent crack formation and propagation (Hill et al., 1997). Consequently, fly ashes of highunburned-carbon content, also referred here to as fly ash carbons, derived from coal–fired combustors are an increasing problem for the utility industry, since they cannot be marketed as a cement extender and, therefore, have to be disposed. However, due to the increasingly restricted landfill use, the coal industry needs to find uses for these chars. Following this demand, the authors have previously

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

developed a one-step activation protocol to produce activated carbons from high carbon fly ashes (Zhang et al., 2003; Maroto-Valer et al., 2002). The authors’ previous studies have shown that fly ash carbons only require a one-step activation process, since they have already gone through a devolatilization process while in the combustor (Zhang et al., 2003; Maroto-Valer et al., 2002). This is an important advantage compared to the conventional twostep activation process that includes a devolatilization of the raw materials, followed by an activation step. CO2 capture technologies aim to isolate the CO2 from the flue gas into a form suitable for transport and subsequent storage. Existing chemical absorption processes using monoethanolamine (MEA) were developed for enhanced oil recovery and will impose an energy penalty of about 30% and an increase of the price of electricity of at least 60% when applied to power plants (Yeh and Pennline, 2001). Adsorption processes for CO2 capture using high surface area solids have recently been proposed (Siriwardane et al., 2001). Adsorption can be classified as physical and chemical adsorption (chemisorption) based on the nature of the bonding between the adsorbate molecule and the solid surface. Chemisorption involves electron transfer, whereas the bonds formed in physical adsorption are held by van der Waals and coulombic (electrostatic) forces. The later are much weaker, generally below 10–15 kcal/ mole, and hence the process is easily reversed. (Yang, 1987). Materials like zeolites and activated carbons have high surface areas (>1500 m2/g) and adsorb selectively different gases depending on their surface area, pore size, pore volume and surface chemistry. They operate in pressureswing-adsorption (PSA) or temperature-swing adsorption (TSA) modes to desorb the adsorbed gases either by reducing the pressure or increasing the temperature, respectively. However, physical adsorption on zeolites systems may not be attractive for gas- and coal-fired power plants because these adsorption processes are energy intensive and expensive, particularly the PSA and TSA processes. Accordingly, new solid-based sorbents are being investigated, where amine groups are bonded to a solid surface, resulting in an easier regeneration step (Siriwardane et al., 2001; Zinnen et al., 1989; Soong et al., 2001). The supports used thus far include commercial molecular sieves and activated carbons. Previous work has explored the potential development of amine-enriched fly ash carbons (Gray et al., 2004; Arenillas et al., 2005). Although the amine treatment process used improved the CO2 capture capacities of the fly ash carbon samples, their performance was lower than that of commercially available carbon based sorbents (7 vs. 88 mg CO2/g sorbent) (Gray et al., 2004). However, none of the previously published work activated the fly ash carbons prior to amine impregnation, and, therefore, they have a very low surface area (<100 m2/g) compared to commercial activated carbon sorbents (>800 m2/g). Accordingly, the work reported here focuses on the development of activated fly ash derived sorbents for CO2 capture. In this work, two fly ash carbon samples were

2321

collected. Previous work has shown that the sorbent properties of the inorganic fraction is very low compared to the carbon present in ash (Serre and Silcox, 2000; Hassett and Eylands, 1999; Maroto-Valer et al., 2005a). Therefore, the samples were subjected to an acid digestion step to further reduce their ash concentrations. The samples were then activated using the protocols previously developed by the authors (Zhang et al., 2003), and the resultant activated samples were amine impregnated. The parent, activated and treated samples were tested for CO2 capture at different temperatures. 2. Experimental 2.1. Study samples Two high carbon fly ash samples, named FA1 and FA2, were procured and characterized for this study. FA1 was collected from a pulverized-coal-fired suspension-firing research boiler equipped with a low-NOx burner and burning high volatile bituminous coal. The FA2 sample was procured from a gasifier that uses a subbituminous coal. These samples were selected due to their high loss-on-ignition (LOI) values (Table 1), as this work focuses on the study of high carbon fly ashes as feedstock for activated carbons. A conventional acid (HCl/HNO3/HF) digestion step was conducted to remove the ash from the samples and concentrate the carbon; 50 g of sample was treated with the above acids at 65 °C for 4 h, as previously described (Maroto-Valer et al., 2005a). For the acid digestion, a mixture of HF (48–49%), HCl (6 N) and HNO3 (3 N) was added to the sample. The solution was stirred, and then filtered and rinsed with distilled water until the filtrate was neutral. The resultant carbon concentrated samples were dried overnight at 105 °C. The de-ashed samples were labeled as FA1–DEM and FA2–DEM and were used for the subsequent CO2 capture studies. 2.2. Loss-on-ignition analysis (LOI) The loss-on-ignition (LOI) content of the samples was determined according to the ASTM C311 procedure. Around 1 g of sample was oxidized in air for 3 h at 800 °C to constant weight in a muffle furnace. The LOI content was then calculated from the weight loss of the sample after oxidation. The LOI analyses were conducted in duplicate. Table 1 LOI and pore structure parameters of the studied samples Sample

FA1 FA1–DEM FA2 FA2–DEM

LOI (%)

59 97 38 97

Surface area (m2/g)

Pore volume (ml/g)

Smi

Sme

Vmi

Vme

57 32 48 99

18 21 236 632

0.027 0.014 0.021 0.038

0.020 0.026 0.256 0.702

2322

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

2.3. Porosity analysis The porous texture of the two samples was characterized by conducting N2 adsorption isotherms at 77 K using a Quantachrome adsorption apparatus, Autosorb-1 Model ASIT. The pore volume was calculated as the volume measured in the nitrogen adsorption isotherm at a relative pressure of 0.95 (V0.95). The total specific surface area, St, was calculated using the multi-point BET equation in the relative pressure range 0.05–0.35, as previously described (Zhang et al., 2003). The pore sizes of 2 and 50 nm were taken as the limits between micro- and mesopores and meso- and macropores, respectively, following the IUPAC nomenclature (Sing et al., 1985). From the adsorption isotherm, the micropore (<2 nm) volume, Vmi, was calculated using the aS-method, where non-graphitized non-porous carbon black Cabot BP 280 (SBET = 40.2 m2/g) was used as a reference adsorbent (Kruk et al., 1997). The mesopore (2–50 nm) volume (Vme) was calculated by subtracting the volume of Vmi from Vt. 2.4. One-step activation Sample FA1–DEM was steam activated in a horizontal tubular furnace. Around 3 g of sample were placed into a reactor and a flow of steam passed through. The reactor was then heated under nitrogen flow to the desired temperature before steam was introduced in the reactor, while the reactor was kept under isothermal conditions for a desired period of time as previously reported (Zhang et al., 2003; Maroto-Valer et al., 2002, 1998). The activation processes were conducted at 850 °C for 30, 60, 90, and 120 min, and the corresponding activated samples were named as FA1–AC-a, FA1–AC-b, FA1–AC-c, and FA1–AC-d, respectively. The 850 °C activation temperature was selected because previous work has shown that high surface area materials can be generated at this temperature. The porous properties of the activated carbons were systematically characterized using N2 adsorption isotherms, as previously described. 2.5. Impregnation of sorbents Alcohol amines have previously been used to impart an amine functionality to activated carbons and, therefore, to enhance their CO2 adsorption capacity (Xu et al., 2002). In this work, a primary (monoethanolamine, MEA), a secondary (diethanolamine, DEA), and a tertiary (methyldiethanolamine, MDEA) were used to modify the activated fly ash carbon FA1–AC-c (activated FA1 at 850 °C for 90 min). The activated fly ash derived sorbents were impregnated by immersing them in an amine solution of the desired compound, MEA, MDEA or DEA (Xu et al., 2002; Zinnen et al., 1989; Smith, 1999; Yeh and Pennline, 2001). The amine solutions used in this study (MEA, DEA, MDEA) have concentrations over 98%. In addition, a mixture of MEA and MDEA was loaded onto the sorbent. The resul-

tant slurry was continuously stirred for about 20 min, and then dried in air at 120 °C until constant weight. The samples before and after chemical loading were characterized by conducting N2 adsorption isotherms at 77 K, as described previously. The BET surface area, total pore volume and mean pore size of the sorbents were obtained by analyzing the N2 isotherms. 2.6. CO2 adsorption studies The CO2 adsorption properties of the samples were characterized using a Perkin Elmer thermogravimetric analyzer (model PE-TGA 7). Around 10 mg of sample were placed in a platinum crucible, heated up to desired temperature (30, 70, 100, and 120 °C) in 100 ml/min pure N2 flow, and held at this temperature until the weight of the sample was stable (10–20 min). The gas flow was switched from N2 to 99.8% bone-dry CO2 at a flow rate of 100 ml/min to measure the CO2 adsorption performance, and was then changed back to a 100 ml/min N2 flow at the same temperature for the desorption test. The flow rate used (100 ml/ min) was set to exclude the effect of interparticle diffusion. The weight change of the sorbents was recorded and used to determine the adsorption capacities of the samples. The adsorption capacity was reported in mg CO2/g-sorbent and was used to evaluate the performance of the samples prepared (Maroto-Valer et al., 2005b). 3. Results and discussion 3.1. Sample de-ashing and one-step activation The LOI values of the studied samples are listed in Table 1. The LOI of the parent samples, FA1 and FA2, are 62% and 38%, respectively, which are higher than those reported in previous studies that are typically <15% (Zhang et al., 2003; Hill et al., 1997). The de-ashing step used in the present paper can successfully concentrate the unburned carbon, where the resultant samples, FA1–DEM and FA2–DEM, have LOI values as high as 97%. The acid digestion step was used to reclaim the carbon from fly ash for its subsequent use as feedstock for activated carbons, and a commercial plant may use a different separation technology. However, it was not the purpose of this work here to develop a beneficiation process. It should be noted, however, that there are developed processes that use an aqueous hydrofluoric acid (HF) process for producing ultra-clean coal without requiring the on-going purchase of chemicals (Steel and Patrick, 2004). The pore structure parameters of the studied samples are listed in Table 1. The total surface area (Smi + Sme) and total pore volume (Vmi + Vme) of sample FA2 are significantly higher than the values reported for FA1 (284 m2/ g and 0.277 ml/g vs. 75 m2/g and 0.047 ml/g). Previous studies have shown that some porosity is developed during the combustion process, with surface areas typically below 100 m2/g (Zhang et al., 2003). The higher surface area

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

Table 2 LOI and pore structure parameters of the activated samples and their CO2 adsorption capacity at 30 °C Sample

FA1– AC-a

FA1– AC-b

FA1– AC-c

FA1– AC-d

Burn-off (%) BET S.A (m2/g) Vtotal (ml/g) Dn (nm) Vmi (ml/g) Vme (ml/g) Vmi/Vtotal (%) CO2 uptake (%) CO2 uptake (mg CO2/g sorbent)

11 372 0.234 1.24 0.177 0.057 75.6 3.12 31.2

14 538 0.329 1.21 0.254 0.075 77.2 3.66 36.6

48 818 0.665 1.26 0.400 0.265 60.1 4.03 40.3

67 1075 0.774 1.34 0.463 0.311 59.8 6.95 69.5

described earlier (Zhang et al., 2003), where the increase in pore volume is associated with the development of both micropores and mesopores. Extending the activation time, and correspondingly increasing the burn-off, seems to promote the formation of mesopores. For instance, the sample FA1–AC-c has 75% micropores, while the sample FA1– AC-d has only 60% micropores. This has been previously observed for different carbons (McEnaney and Mays, 1989), anthracites (Maroto-Valer et al., 2005b; MarotoValer and Schobert, 1998) and also unburned carbon (Zhang et al., 2003) and it is related to the removal of pore walls and enlargement of micropores. 3.2. Effect of temperature on CO2 adsorption The CO2 adsorption studies of the FA2 and FA2–DEM samples were conducted at both 30 °C and 75 °C, as described before, and the capacity results are shown in Fig. 1. At 30 °C, the parent sample FA2 adsorbs around 17.5 mg CO2/g, while FA2–DEM adsorbs 43.5 mg CO2/ g. This is consistent with the much higher total surface area of the latter, 284 m2/g vs. 731 m2/g for FA2 and FA2– DEM, respectively (Table 1). As expected from a physical adsorption process, the CO2 adsorption capacities of both samples decreased to 10.2 and 22.0 mg CO2/g for FA2 and FA2–DEM, respectively, when the temperature was raised from 30 to 75 °C. The CO2 adsorption/desorption profiles of the activated sample FA1–AC-c are presented in Fig. 2, where both the adsorption and desorption time was 150 min. The sample exhibited a CO2 adsorption capacity of 40.3 mg/g (4.03%) at 30 °C. With increasing temperature, the CO2 uptake on the activated carbon decreased to 18.5 and 7.7 mg/g at 70 and 120 °C, respectively. This behavior is typical of a physical adsorption process and has previously been reported for zeolite-based adsorbents and anthracites (Xu et al., 2002; Maroto-Valer et al., 2005b). However, typical flue gas has temperatures typically up to 150 °C; therefore, it is necessary to develop sorbents that can operate at these higher temperatures (Xu et al., 2002; Maroto-Valer et al., 2005b). It should also 50 CO2 Adsorption, mg-CO2/g

reported here for FA2 could be due to the gasification conditions under which this sample was generated. In the case of FA2, the de-ashed sample (FA2–DEM) has a total surface area of 731 m2/g, compared to only 284 m2/g for the parent sample (FA2). This increase in surface area may partly come from the removal of the fly ash, confirming that the fly ash is virtually non-porous, and only contributes to the sample weight. Demineralization of low rank coals and heavy oil fly ash have shown to increase the surface area due to removal of mineral matter and opening of existing pores (Gomez-Serrano et al., 2004; Seggiani et al., 2005). However, the same de-ashing step resulted in a decrease of the surface area for sample FA1, from 75 to 53 m2/g for FA1 and FA1–DEM, respectively. This surface area decrease was accompanied with a pore size enlargement, where FA1–DEM has more mesopores and less micropores than the parent FA1. Leaching of heavy oil fly ash with H2SO4 has shown to decrease the surface area due to sulfate precipitation and occlusion of preexisting pores (Seggiani et al., 2003). Similar precipitation and occlusion phenomena may be responsible here for the surface area reduction observed, but further studies are needed to understand the effect of de-ashing on the porous structure of fly ash carbons. Sample FA1–DEM was steam activated and the porous properties of the activated samples are presented in Table 2. For the purpose of comparison, the surface areas and pore volumes reported here are corrected to ash-free basis by subtracting the measured isotherm for a pure ash sample after complete carbon burn-off in air at 850 °C, as previously described for a series of density fractions and fly ash samples (Zhang et al., 2003; Maroto-Valer et al., 2001). The one-step activation process significantly increased the surface area and total pore volume with surface area and total pore volumes up to 1075 m2/g and 0.774 ml/g, respectively. As expected, the longer the activation time, the higher the burn-off levels, where after 30 min activation, the burn-off was 11% compared to 67% after 120 min activation. Furthermore, within the activation temperature studied here (850 °C), both the surface area and pore volume increase with the rise in the activation time. This is associated with the increase in burn-off as

2323

40

30 ºC 75 ºC

30 20 10 0 FA2

FA2-DEM

Fig. 1. CO2 adsorption capacities of samples FA2 and FA2–DEM at 30 °C and 75 °C.

2324

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

Fig. 2. CO2 adsorption/desorption profiles of the activated sample FA1– AC-c at 30, 70 and 120 °C.

be noted that for the sorbents developed, no CO2 remains after desorption and the process is reversible. Therefore, the samples can be regenerated after desorbing the CO2 and used in cyclical operation. 3.3. Effect of porous properties on CO2 adsorption Table 2 lists the CO2 adsorption capacities at 30 °C of the activated carbons with different porous properties. The sorbent with the highest surface area and larger pore volume (FA1–AC-d) presented the highest CO2 adsorption capacity (69.5 mg CO2/g sorbent). Previous studies conducted by the authors on CO2 adsorption on activated anthracites showed that the activated anthracites with the highest microporosity presented the highest CO2 adsorption capacity (Maroto-Valer et al., 2005b). The results reported here on activated fly ash carbons show the inverse trend: highest microporosity is related with the lowest CO2 adsorption. However, the micropore volume of the activated anthracites varied between 35% and 70%, while the microporosity of the activated fly ash carbons studied here only varied between 60% and 77%. It is not known whether a series of activated fly ash carbons with a wider range of microporosity would have presented the same trend as that observed for the activated anthracites. 3.4. Amine impregnation of sorbents FA1–AC-a and FA1–AC-b were not considered for amine impregnation as they present relatively low surface

areas and CO2 uptakes (Table 2). The two potential samples for amine impregnation are FA1–AC-c and FA1–AC-d. There is a trade-off between surface area and burn-off, where both samples AC–FA1-c and FA1–AC-d have a high surface area, but FA1–AC-c was selected as it presents a lower burn-off (Table 2). The high burn-off of sample FA1–AC-d (67%) implies that after activation only small quantities are available for impregnation. Therefore, activated sample FA1–AC-c, which was generated from the steam activation of sample FA1 at 850 °C for 90 min, was selected for the amine impregnation studies. The activated sample FA1–AC-c was loaded with MDEA, DEA, and MEA, as previously described. The addition of MEA to an aqueous solution of MDEA has been shown to significantly increase the rate of absorption of CO2 (Hagewiesche et al., 1995), and accordingly, in this work MDEA and MEA were sequentially loaded onto the activated sample. The amine loaded samples were marked as AC-MDEA, AC-DEA, AC-MEA, and AC-MM. The amine loading onto the activated carbons is calculated by the weight difference of activated carbons before and after impregnation and is presented in Table 3. The weight differences for all of the amine loadings were between 29% and 46 wt%. Similar impregnation methods in carbon molecular sieves resulted in loadings between 20% and 65 wt% (Zinnen et al., 1989). Table 3 lists the surface areas and pore volumes (Vtotal, Vmi and Vme) for the parent activated carbon, FA1–AC-c, and its amine loaded counterparts. The surface area and the total pore volume of the sample FA1–AC-c are 818 m2/g and 0.665 ml/g, respectively. The impregnation process results in a decrease of the surface areas, indicating a blocking of the porosity. For example, after MDEA loading onto the activated carbon, the surface area and the total pore volume for sample AC-MDEA decreased to 204 m2/g and 0.203 ml/g, respectively. Table 3 also shows the decrease in micropore and mesopore volume due to the impregnation process. It can be seen that the amine loading decreases the micro- and mesopore volume, probably due to pore filling effects. However, there appears to be significant differences in the way that the amines used are filling up the pores of the activated sample. For example, MDEA had the most significant effect on micropore filling, with a micropore volume decrease of 73%. In contrast, loading of MDEA + MEA

Table 3 Amount of amine loaded, surface area and pore volumes (Vtotal, Vmi and Vme) for the parent activated sample and its amine loaded counterparts Amine loading (wt%) SBET (m2/g) Vtotal (cm3/g) Vmic (cm3/g) Vme (cm3/g) Decrease in Vmi after chemical loading (%) Decrease in Vme after chemical loading (%)

FA1–AC-c

AC-MDEA

AC-DEA

AC-MEA

AC-MM

818 0.665 0.400 0.265

46 204 0.203 0.110 0.092 73

34 265 0.288 0.126 0.162 69

39 241 0.397 0.143 0.254 64

20 MDEA +20 MEA 302 0.414 0.244 0.170 39

65

39

4

32

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

resulted in a decrease of micropore volume of only 39%. For the case of mesopore filling, MDEA also presented the highest ability to decrease the pore volume (65%), followed by DEA, MDEA + MEA, and MEA with volume decrease of 39%, 32%, and 4%, respectively. The mechanism of pore filling using different amines is not well understood, but it is assumed that the different pore filling effect was due to the difference in the molecular size and shape of the amines used. For instance, MEA that is a primary amine and has the lowest molecular weight (69 g/mol) of the amines studied here, is very specific towards filling and/or blocking the micropores and virtually does not accumulate in the mesopores. MDEA that is more branched and has a higher molecular weight (119 g/mol) appears to build up in the mesopores and may block off the micropores without actually filling them. It should also be considered that the loading amount may also affect the pore coverage and blockage, with AC-MEA having the lowest loading and MDEA presenting the highest loading. Finally, for the AC-MM sample (sequential loading of 20 wt% of MDEA followed by 20 wt% of MEA), the mesopore volume reduction seems to be a simple additive effect, since the predicted calculated value of 30% (100  [(65/46)  0.2 + (4/29)  0.2] = 30) is very close to the observed value of 32%. In the calculation above, 65 and 4 are the pore volume reductions observed for MDEA and MEA, respectively, and 46 and 39 are the amine loading wt% for MDEA and MEA, respectively (Table 3). However, the micropore volume reduction is significantly lower than that expected from an additive effect, where the predicted calculated value of 65% (100  [(73/ 46)  0.2 + (64/39)  0.2] = 65%) is much higher than the observed value of 39%. This can be explained as the sample was first loaded with MDEA that filled the mesopores and blocked the micropores, with the subsequent MEA loading mainly filling mesopores, as most of the micropore blockage has already taken place by MDEA. 3.5. CO2 adsorption studies of impregnated activated samples Fig. 3 shows the CO2 adsorption/desorption profile of the sample AC-MDEA (activated sample FA1–AC-c impregnated with MDEA). In contrast to the samples prior to impregnation (Fig. 2), with increasing temperature, the CO2 uptake increases and reaches a maximum at 100 °C, probably indicating a chemical adsorption process (Maroto-Valer et al., 2005b). As previously observed for the acti-

2325

Fig. 3. CO2 adsorption/desorption profiles of the activated and MDEA impregnated sample (AC-MDEA) at 30, 70, 100 and 120 °C.

vated sorbents, after the desorption cycle no CO2 remains, indicating the reversibility of the process. Therefore, the samples can be regenerated and used in cyclical operation. The CO2 adsorption capacities of the sample FA1–AC-c and its amine loaded counterparts are presented in Table 4. As previously described for activated fly ash carbons (Fig. 2), the CO2 uptake decreases with increasing temperature, as expected from a physical adsorption process. Previous work on CO2 adsorption using fly ash carbons has reported CO2 capacities at 120 °C of around 1.1–3.2 mg CO2/g sorbent (24.4.–72.9 lmol CO2/g sample), probably due to the low surface area of the samples (27 m2/g) (Gray et al., 2004). The work reported here has shown that the activation of fly ash carbons can improve their adsorption capacities to values up to 7.7 mg CO2/g sorbent at 120 °C. As shown in Fig. 2, the activated carbon exhibited a CO2 adsorption capacity of 40.3 mg CO2/g sorbent (4.03%) at 30 °C and with increasing adsorption temperature, the CO2 uptake on the activated carbon decreased to 18.5 and 7.7 mg CO2/g sorbent at 70 and 120 °C, respectively. This is probably because the CO2 adsorption on activated fly ash carbon prior to impregnation is a physical process, where both the surface adsorption energy and the molecule diffusion rate increase with increasing temperature, as previously reported for zeolite-based adsorbents and anthracites (Xu et al., 2002; McEnaney and Mays, 1989). The loading of MDEA, DEA, and MDEA + MEA resulted in a decrease of the CO2 adsorption capacity as measured at 30 °C, where the CO2 adsorbed decreased to 17.1 mg CO2/g sorbent for AC-MDEA, 21.1 mg/g for

Table 4 CO2 adsorption capacities mgCO2/g sorbent of the activated and impregnated samples at 30, 70, 100 and 120 °C Temperature, °C

FA1–AC-c

AC-MDEA

AC-DEA

AC-MEA

AC-MM

30 70 100 120

41.8 18.5 – 7.7

17.1 30.4 40.6 16.1

21.1 37.1 16.3 4.2

68.6 49.8 25.3 5.5

32.6 35.8 43.2 23.7

2326

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

AC-DEA, and 32.6 mg CO2/g sorbent for AC-MM compared to the parent sample (41.8 mg CO2/g sorbent). This decrease in the CO2 adsorption capacity is probably related to the fact that the loading of amines significantly decreased the surface area and pore volume of the activated carbon due to pore filling effects (Table 3), which probably leads to the low CO2 adsorption capacity compared with the parent sample. Xu et al. (2002) and Arenillas et al. (2005) also reported low CO2 adsorption capacities at ambient temperature for a series of samples impregnated with polyethylenimine, and they concluded that the CO2 adsorption is a kinetically diffusion-controlled process, so at low temperatures it takes longer for the CO2 to diffuse into the particles. In contrast to the loading of MDEA, DEA and MDEA+MEA, the loading of MEA led to an increase in the CO2 uptake compared to that of the parent activated carbon (68.6 vs. 41.8 mg CO2/g sorbent). This is probably as a result of the chemical reaction between MEA and CO2 under the test conditions at 30 °C. It is known that primary alkanol amines, like MEA, have higher absorption rates in aqueous systems than secondary and tertiary amines (Hagewiesche et al., 1995), and this may explain the higher adsorption capacity of AC-MEA compared to AC-MDEA and AC-DEA at 30 and 70 °C. When the temperature is increased to 70 °C, all of the impregnated samples present higher adsorption capacities than the non-impregnated counterpart (Table 4). As previously described for AC-MDEA (Fig. 3), for all of the impregnated samples, except AC-MEA, the CO2 uptake increases with increasing temperature and reaches a maximum at 100 °C or 75 °C, probably indicating a chemical adsorption process. However, when increasing the temperature further to 120 °C, the adsorption capacities decrease compared to those reported at lower temperatures, and the CO2 capacities for the AC-DEA and AC-MEA samples are lower than those reported for the non-impregnated sample (4.2 and 5.5 vs. 7.7 mg CO2/g sorbent). For the AC-MM sample (sequential loading of MDEA followed by MEA), the CO2 capacity observed at 120 °C seems to be much higher than a simple additive effect (23.7 mg CO2/g sorbent compared to only 16.1 and 5.5 mg CO2/g sorbent for AC-MDEA and AC-MEA). This has been previously observed in aqueous absorption processes, where addition of MEA to an aqueous solution of MDEA increased significantly the rate of absorption of CO2 (Hagewiesche et al., 1995). The higher adsorption capacity at 30 and 70 °C for the amine impregnated activated carbons was probably due to a combination of physical adsorption inherent from the parent sample and chemical adsorption of the loaded amine groups. As previously discussed for the parent sample (Fig. 2), its adsorption capacity decreased with temperature, as expected from a physical adsorption process (Arenillas et al., 2005). The main reaction responsible for CO2 chemical adsorption with amine is believed to be carbamate formation (Satyapal et al., 2001) that can be pro-

moted with an increase in temperature (Arenillas et al., 2005). Therefore, for the impregnated samples, as the temperature increases, the contribution from physical adsorption drops off quickly and probably offsets any gain in the chemical adsorption enhanced by the increase in temperature. Moreover, the decrease in CO2 adsorption capacities at higher temperatures may also be related to the volatilization of the amines used. This could be the case for the MEA impregnated sample, as the flash point of the MEA is 85 °C. 3.6. Implications of this research Fly ash derived sorbents represent a potential alternative for CO2 capture to the existing methods using specialized activated carbons and molecular sieves, which tend to be very expensive and hinder the viability of the CO2 sorption process due to economic constraints. For these new precursors to compete effectively with commercial sorbents, they must be inexpensive, and be easily converted into high surface area materials (Zhang et al., 2003). Fly ash carbons meet satisfactorily all of these conditions. Firstly, they can be easily obtained from the utility industries as a byproduct. Secondly, the conventional production of activated carbons consists of a two-step process, which includes a devolatilization of the raw materials, followed by an activation step. In contrast, for fly ash carbons only a one-step activation process is required, since it has already gone through a devolatilization process while in the combustor (Zhang et al., 2003). The carbon in the fly ash can be concentrated to contents >97% by using acid digestion methods. This de-ashing process not only removes the inorganic ash from the carbon, but also changes the porous structure and surface properties of the resultant carbon. Hence, de-ashed fly ash samples with high surface area and CO2 adsorption capacities can be produced by acid washing. Furthermore, the one-step steam activation can significantly increase the surface area and, therefore, provide a large chemical uptake for the fly ash samples without de-ashing. The sample FA1–AC-c produced at 850cC for 90 min activation has a CO2 adsorption capacity of 40.3 mg/g at 30 °C. The impregnation of MEA, DEA, and MDEA chemicals on activated sample can reduce the surface area and the total pore volume of the activated carbons due to the blocking of both micro- and mesopores. Previous work published using alcohol amines at loading levels between 20% and 65 wt% on carbon molecular sieves has only reported CO2 capacities of around 25 mg CO2/g sorbent at 25 °C (Zinnen et al., 1989). In contrast, the AC-MEA sample in this work presents a much higher CO2 capacity of 68.6 mg CO2/g sorbent. Therefore, chemically attached amino groups in fly ash derived sorbents may have a great potential when used in flue gases for CO2 capture and the selection of the chemical is a critical step.

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

4. Conclusions The objective of this paper is the development of activated fly ash derived sorbents for CO2 capture. Two fly ash carbon samples were collected and subjected to acid digestion to further reduce their ash concentrations. The samples were then activated using the protocols previously developed by the authors, and the resultant activated samples were amine impregnated. The parent, activated and treated samples were tested for CO2 capture at different temperatures. The one-step steam activation process at 850 °C significantly increased the surface area and total pore volume with surface area and total pore volumes up to 1075 m2/g and 0.774 ml/g, respectively. Extending the activation time, and correspondingly increasing the burn-off, seems to promote the formation of mesopores, probably due to the removal of pore walls and enlargement of micropores. The CO2 adsorption of the parent and activated fly ash carbons is typical of a physical adsorption process, where the CO2 uptakes decrease with increasing temperature. The impregnation process with MDEA, DEA, MEA, and MDEA + MEA, results in a decrease of the micro and mesopore volumes, resulting from pore filling effects. The mechanism of pore filling using different amines is not well understood, but it is assumed that the different pore filling effect was due to the difference in the molecular size and shape of the amines used. Amine (MDEA, DEA, MEA, and MDEA + MEA) impregnation can improve significantly the CO2 adsorption of the samples and their activated and de-ashed counterparts. The highest adsorption capacity at 30 and 70 °C for the amine impregnated activated carbons was probably due to a combination of physical adsorption inherent from the parent sample and chemical adsorption of the loaded amine groups. The CO2 adsorption capacities for the activated amine impregnated samples are higher than those previously published for fly ash carbons without activation (68.6 vs. 45 mg CO2/g sorbent) and for molecular sieves impregnated with alkanol amines (25 mg CO2/g sorbent at 25 °C). Acknowledgement This work was supported by the US DOE through the Combustion Byproducts Recycling Consortium (Project number 01-CBRC-E9). References ACAA 2005, 2005. Coal Combustion Product Production and Survey, http://www.acaa-usa.org/PDF/2005_CCP_Production_and_Use_Figures_Released_by_ACAA.pdf (accessed 1.11.06). Arenillas, A., Smith, K.M., Drage, T.C., Snape, C.E., 2005. CO2 capture using some fly ash-derived carbon materials. Fuel 84 (17), 2204–2210. ECOBA, 2003. Coal Combustion Product Production and Survey, http:// www.ecoba.com/index.html (accessed 1.11.06).

2327

Gomez-Serrano, V., Fernandez-Gonzalez, M.C., Cuerda-Correa, E.M., Macias-Garcia, A., Alexandre-Franco, M.F., Rojas-Cervantes, M.L., 2004. Physico–chemical properties of low-rank coals. Thermal and demineralisation effects. Powder Technology 148 (1), 38–42. Gray, M.L., Soong, Y., Champagne, K.J., Baltrus, J., Stevens, R.W., Tochinda, P., Chuang, S.S.C., 2004. CO2 capture by amine-enriched fly ash carbon sorbents. Separation Purification Technology 35, 31–36. Hagewiesche, D.P., Ashour, S.S., Al-Ghawas, H.A., Sandall, O.C., 1995. Absorption of carbon-dioxide into aqueous blends of monoethanolamine and N-methyldietheanolamine. Chemical Engineering Science 50 (7), 1071–1079. Hassett, D.J., Eylands, K.E., 1999. Mercury capture on coal combustion fly ash. Fuel 78 (2), 243–248. Hill, R.L., Sarkar, S.L., Rathbone, R.F., Hower, J.C., 1997. An examination of fly ash carbon and its interactions with air entraining agent. Cement Concrete Res. 27 (2), 193–204. Kruk, M., Jaroniec, M., Gadkaree, K.P., 1997. Nitrogen adsorption studies of novel synthetic active carbons. J. Colloid Interface Sci. 192 (1), 250–256. Maroto-Valer, M.M., Schobert, H.H., 1998. Proc. 10th Int. Conf. Coal Sci., 909–912. Maroto-Valer, M.M., Taulbee, D.N., Hower, J.C., 1998. Application of density gradient centrifugation to the concentration of unburned carbon types from fly ash. Am. Chem. Soc. Division Fuel Chem. Preprints 43, 1014–1018. Maroto-Valer, M.M., Taulbee, D.N., Hower, J.C., 2001. Characterization of differing forms of unburned carbon present in fly ash separated by density gradient centrifugation. Fuel 80 (6), 795–800. Maroto-Valer, M.M., Andre´sen, J.M., Zhang, Y., 2002. In: Lankey, P.L., Anastas, P.T. (Eds.), Advancing Sustainability through Green Chemistry and Engineering, vol. 823. American Chemical Society Symposium Series, pp. 225–241 (Chapter 15). Maroto-Valer, M.M., Zhang, Y., Granite, E.J., Tang, Z., Penline, H.W., 2005a. Effect of porous structure and surface functionality on the mercury capacity of a fly ash carbon and its activated sample. Fuel 84 (1), 105–108. Maroto-Valer, M.M., Tang, Z., Zhang, Y., 2005b. CO2 capture by activated and impregnated anthracites. Fuel Processing Technology 86 (14–15), 1487–1502. McEnaney, B., Mays, T.J., 1989. Introduction to Carbon Science. In: Marsh, H. (Ed.). Butterworth, London (Chapter 5). Satyapal, S., Filburn, T., Trela, J., Strange, J., 2001. Performance and properties of a solid amine sorbent for carbon dioxide removal in space life support applications. Energy Fuels 15, 250–255. Seggiani, M., Vitolo, S., Narducci, P., 2003. Investigation on the porosity development by CO2 activation in heavy oil fly ashes. Fuel 82, 1441– 1450. Seggiani, M., Vitolo, S., De Filippis, P., 2005. Effect of pre-oxidation on the porosity development in a heavy oil fly ash by CO2 activation. Fuel 84, 1593–1596. Serre, S.D., Silcox, G.D., 2000. Adsorption of elemental mercury on the residual carbon in coal fly ash. Ind. Eng. Chem. Res. 39 (6), 1723– 1730. Sing, K.S.W., Everett, D.H., Haul, R.A.W., 1985. Reporting physisorption data for gas solid systems with special reference to the determination of surface-area and porosity. Pure Appl. Chem. 57 (4), 603–619. Siriwardane, R.V., Shen, M.S., Fisher, E.P., Poston, J.A., 2001. Adsorption of CO2 on molecular sieves and activated carbon. Energy Fuels 15 (2), 279–284. Smith, I.M., 1999. CO2 reduction-prospects for coal, IEA coal research. Soong, Y., Gray, M.L., Siriwardane, R.V., Champagne, K.J., Stevens, R.W., Tochinda, P., Chuang, S.S.C., 2001. Novel solid sorbents for CO2 capture. Prepr. Pap. Am. Chem. Soc. Div. Fuel Chem. 46 (1), 285–286. Steel, K.M., Patrick, J.W., 2004. Re-generation of hydrofluoric acid and selective separation of Si(IV) in a process for producing ultra-clean coal. Fuel Process. Technol. 86 (2), 179–190.

2328

M. Mercedes Maroto-Valer et al. / Waste Management 28 (2008) 2320–2328

Xu, X., Song, C., Andresen, J.M., Miller, B.G., Scaroni, A.W., 2002. Novel polyethylenimine-modified mesoporous molecular sieve of MCM-41 type as high-capacity adsorbent for CO2 capture. Energy Fuels 16 (6), 1463–1469. Yang, R., 1987. Gas separation by adsorption process. Butterworths, Boston, London, Durban, Singapore, Sydney, Toronto, Wellington. Yeh, J.T., Pennline, H.W., 2001. Study of CO2 absorption and desorption in a packed column. Energy Fuels 15 (2), 274–278.

Zhang, Y., Maroto-Valer, M.M., Lu, Z., Andre´sen, J.M., Schobert, H.H., 2003. Comparison of high-unburned-carbon fly ashes from different combustor types and their steam activated products. Energy Fuels 17 (2), 369–377. Zinnen, H.A., Nalette, T.A., Papale, W., Filburn, T., 1989. Carbon dioxide removal using aminated carbon molecular sieves. US patent No. 4 810, 266.