In Situ Fourier Transform Infrared (FTIR) Investigation of CO2

Jul 22, 2008 - Current CO2 capture technologies are highly energy-intensive and introduce a large penalty into the cost of energy production from foss...
79 downloads 6 Views 352KB Size
3070

Energy & Fuels 2008, 22, 3070–3079

In Situ Fourier Transform Infrared (FTIR) Investigation of CO2 Adsorption onto Zeolite Materials Robert W. Stevens, Jr.,†,‡ Ranjani V. Siriwardane,*,† and Jennifer Logan† National Energy Technology Laboratory, United States Department of Energy, 3610 Collins Ferry Road, Morgantown, West Virginia 26507, and Parsons, Post Office Box 618, South Park, PennsylVania 15129 ReceiVed March 24, 2008. ReVised Manuscript ReceiVed May 30, 2008

The adsorption of CO2 onto five zeolite materials (13X, WEG, AGP, 4A, and 5A) was studied by in situ infrared spectroscopy at 1 atm as a function of the pretreatment temperature (120 and 350 °C) and adsorption temperature (30 and 120 °C). Adsorbed CO2 surface species identified in the current work include physisorbed CO2, bidentate carbonate, bridged bidentate carbonate, monodentate carbonate, and carboxylate. Both pretreatment temperature and CO2 adsorption temperature affected the type and amount of adsorbed CO2 species formed. Materials pretreated at 350 °C, as opposed to 120 °C, had more surface adsorption sites available as evidenced from the resulting more intense IR bands. Physisorbed CO2 was the most abundant species observed. Bridged bidentate carbonate was found to be more stable than bidentate carbonate. Tests involving both CO2 and H2O showed that the two species competed for the same adsorption sites on the zeolite surface.

Introduction The majority of the energy needs in the world are supplied through use of fossil fuels. Fossil fuel combustion is one of the leading sources of CO2 emissions, the major contributor to the greenhouse gas concentration in the atmosphere. A growing concern exists for the increasing CO2 concentration in the atmosphere and its effect on global warming. Current CO2 capture technologies are highly energy-intensive and introduce a large penalty into the cost of energy production from fossil fuels. Improved technologies for CO2 capture are therefore necessary to generate power economically while eliminating or minimizing CO2 emissions to the atmosphere. The capture and separation of CO2 can be accomplished by using solvents, cryogenic techniques, membranes, and solid sorbents. Among the potential solutions, pressure swing adsorption (PSA) and temperature swing adsorption (TSA) techniques over solid sorbents are promising. Liquid amine-based and cryogenic technologies are energy-intensive. In addition, corrosion problems have to be dealt with when amines are used in the process.1 Therefore, improved technologies have to be developed. For the capture technology to be effective, however, durable, regenerable sorbents must be developed that possess a high CO2 adsorption selectivity, large CO2 capacity, and high CO2 adsorption and desorption rates. A number of examples of amine-modified solid sorbents exist in the literature.2–4 These materials, however, require near* To whom correspondence should be addressed. Telephone: (304) 2854513. Fax: (304) 285-0903. E-mail: [email protected]. † United States Department of Energy. ‡ Parsons. (1) Veawab, A.; Tontiwachwuthikul, P.; Chakma, A. Ind. Eng. Chem. Res. 1999, 38, 3917. (2) Gray, M. L.; Soong, Y.; Champagne, K. J.; Pennline, H.; Baltrus, J. P.; Stevens, R. W., Jr.; Khatri, R.; Chuang, S. S. C.; Filburn, T. Fuel Process. Technol. 2005, 86, 1449. (3) Huang, H. Y.; Yang, R. T.; Chinn, D.; Munson, C. L. Ind. Eng. Chem. Res. 2003, 42, 2427. (4) Leal, O.; Bolivar, C.; Ovalles, C.; Garcia, J. J.; Espidel, Y. Inorg. Chim. Acta 1995, 240, 183.

ambient temperatures to capture an appreciable amount of CO2 and therefore may not be efficient for capture of CO2 from higher temperature applications. Zeolites have shown potential for separating CO2 from gas mixtures and may be candidate sorbents for the PSA and/or TSA techniques. Furthermore, we previously showed that several zeolite samples can capture CO2 at moderate temperatures.5 The present work is a continuation of that study. The objective of the current study is to investigate the CO2capture mechanism over zeolites 13X, WEG 592, APG-II, 5A, and 4A through use of in situ Fourier transform infrared (FTIR) analysis in an effort to elucidate the relationship between the chemistry of CO2 capture and sorbent performance. FTIR analyses of CO2 adsorption onto zeolites have been reported by a number of authors.4,6–14 These studies, however, are largely based on low-temperature (e30 °C) and low-pressure (e1 torr) adsorption of CO2. FTIR studies of CO2 adsorption and/or desorption onto zeolites at temperatures above 100 °C and atmospheric pressure are not reported in the literature. In the current work, FTIR analyses of CO2 adsorption and desorption at 30 and 120 °C are reported. The work presented here is unique because it addresses moderatetemperature and atmospheric-pressure CO2 capture and the (5) Siriwardane, R. V.; Shen, M.-S.; Fisher, E. P.; Losch, J. Energy Fuels 2005, 19, 1153. (6) Chang, A. C. C.; Chuang, S. S. C.; Gray, M.; Soong, Y. Energy Fuels 2003, 17, 468. (7) Delaval, Y.; Lara, E. C. D. J. Chem. Soc., Faraday Trans. 1 1981, 77, 869. (8) Delaval, Y.; Seloudoux, R.; Lara, E. C. D. J. Chem. Soc., Faraday Trans. 1 1986, 82, 365. (9) Gallei, E.; Stumpf, G. J. Colloid Interface Sci. 1976, 55, 415. (10) Jacobs, P. A.; Van Cauwelaert, F. H.; Vansant, E. F. J. Chem. Soc., Faraday Trans. 1 1973, 69, 2130. (11) Jacobs, P. A.; Van Cauwelaert, F. H.; Vansant, E. F.; Uytterhoeven, J. B. J. Chem. Soc., Faraday Trans. 1 1973, 69, 1056. (12) Kamble, V. S.; Gupta, N. M.; Iyer, R. M. Indian J. Chem. 1990, 29A, 1089. (13) Lavalley, J. C. Catal. Today 1996, 27, 377. (14) Rege, S. U.; Yang, R. T. Chem. Eng. Sci. 2001, 56, 3781.

10.1021/ef800209a CCC: $40.75  2008 American Chemical Society Published on Web 07/22/2008

CO2 Adsorption onto Zeolite Materials

Energy & Fuels, Vol. 22, No. 5, 2008 3071

Table 1. Physical and Chemical Properties of Zeolite Samples zeolite property surface area (m2/g) pore diameter (Å) composition (wt %) Na Al Si Ca K Mg a

13X

WEG 592

710 10

625 10

11.7 14.2 18.2 0.5 0.2 1.2

APG-II 710 10

13.7 15.6 16.5 0.1 0.1 NDa

8.8 10.7 14.3 0.2 0.2 1.0

5A

4A

NDa 5

NDa 4

3.8 14.8 16.7 7.8 0.8 1.0

10.8 13.6 16.1 0.8 0.9 1.2

ND ) not determined.

relationship between the surface intermediate formation and capture performance.

Experimental Section Zeolites 13X (Z10-02) and 4A (Z4-01) were obtained from Su¨dChemie, Inc., while zeolites WEG 592 and APG-II were obtained from UOP LLC. Zeolite 5A was obtained from Aldrich Chemical Co. The samples were characterized for surface area and pore diameter using a Micromeritics model ASAP 2010 microporevolume analyzer; pore diameters were analyzed through use of the Horvath-Kawazoe (HK) method assuming cylindrical pore geometries, while surface areas were calculated using density functional theory (DFT). Inductively coupled plasma (ICP) analysis was used to determine the chemical makeup of the zeolite samples. Results of these analyses are summarized in Table 1. Prior to use, the as-received zeolite samples were ground into a powder to maximize spectral quality via diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analysis. For each test, a zeolite sample of approximately 20 mg was charged to a DRIFTS reactor cell (Thermo Electron product 0031-902). The DRIFTS cell is a cold-walled reactor that features temperature control and ZnSe windows for IR transmission within the 650-4000 cm-1 range. Infrared spectra were collected with a Thermo-Nicolet Nexus 670 FTIR instrument, which housed the DRIFTS reactor cell. Spectra were collected at a resolution of 4 cm-1 with 50 co-added scans using a liquid N2-cooled MCT/A detector. The N2, obtained from a liquid N2 cylinder and dried with a filter, and the 1 vol % CO2-N2 (1.01 vol % CO2 and 98.99 vol % N2, from Butler Gas) flows of gas to the cell were controlled via flow meters (Dwyer Instruments, model RMA-100-SSV) at a total flow rate of 40 sccm and a pressure of 1 atm. Two variables were used in testing: (i) pretreatment temperature and (ii) adsorption temperature. Prior to each experiment, the zeolite samples were pretreated in situ by heating to the pretreatment temperature (120 or 350 °C) and held at 1 atm in flowing N2 for 1 h. The pretreated sample was then cooled to the adsorption temperature (30 or 120 °C). The pretreated zeolite sample at the desired adsorption temperature served as the IR background for each experiment. Adsorption of CO2 onto the zeolite samples was started via terminating N2 flow followed by initiating the 1 vol % CO2-N2 flow. Sample IR spectra were collected during selected intervals during adsorption to observe changes in the sample surface (i.e., formation and/or destruction of adsorbed species). The adsorption phase of the test was conducted for approximately 30 min. After the adsorption phase of testing, the CO2-N2 flow was returned to pure N2 (40 sccm) to evaluate desorption of CO2 from the zeolite via FTIR analysis. Two experiments were also conducted with the zeolite 13X sample using moist CO2-N2. Moisture was introduced to the test cell through use of a H2O saturator (bubbler). Data analysis was conducted with Omnic software provided by Thermo Electron.

Results and Discussion Potential modes of CO2 adsorption identified in previous work are shown in Scheme 1 below.15 CO2 Adsorption onto 13X. Figure 1 depicts FTIR spectra of the 120 °C pretreated 13X sample after exposure to CO2 at 30 °C. The spectra indicate that there is formation of bridged bidentate carbonate bands at 1688 and 1361 cm-1, a band at 1623 cm-1, a physisorbed CO2 band at 2353 cm-1, as well as a negative (loss) peak at 3693 cm-1. The observed peaks are in accordance with previous observations in the literature.10,14 Increased exposure time led to growth of all bands described as well as further loss at 3693 cm-1. Desorption in a pure N2 flow at ambient temperature led to the disappearance of the bridged bidentate carbonate bands and a decrease in the intensity of physisorbed CO2, which still remained even after 3 h. No recovery of the band at 3693 cm-1 that was removed during CO2 exposure was observed. The removal of a band at 3693 cm-1 suggests that CO2 may be interacting with surface OH, which is likely to be present on the surface of 13X after activation at 120 °C. Interaction of the CO2 with OH may lead to the formation of bicarbonate or formate. Bands at 3605, 1640, 1480, and 1235 cm-1 that were observed by Rege et al. during CO2 adsorption onto γ-alumina were assigned to bicarbonate.14 Other than the band at 1623 cm-1, no other bands corresponding to bicarbonate were observed during our analysis. Formate would be expected to yield bands at 1597 and 1377 cm-1, but these bands were not observed. The band at 1623 cm-1 may be assigned to a shifted monodentate carbonate.16 Figure 2 illustrates exposure of the 350 °C pretreated sample to CO2 at ambient temperature. The initial exposure led to the formation of bands corresponding to bridged bidentate carbonate at 1710 and 1364 cm-1, an OH bend at 1642 cm-1, physisorbed CO2 at 2356 cm-1, as well as bidentate carbonate at 1486 and 1427 cm-1. No loss of OH at 3692 cm-1 was observed. Increased exposure time led to the growth in intensity of all bands, formation of an OH band at 3708 cm-1, and formation of a band at 1686 cm-1 that is believed to be an OH bending mode of vibration. Further exposure led to a decrease in bridged bidentate carbonate and bidentate carbonate bands, leaving the surface dominated by physisorbed CO2 and bridged bidentate carbonate with bands at 1642 and 1686 cm-1, as well as a broad absorption band in the 3600-3800 cm-1 range, indicative of H2O. Desorption in flowing N2 at ambient temperature left only surface H2O and physisorbed CO2, which survived even after 6 h. The lack of a negative change in the OH region of the IR spectrum during adsorption of CO2 over the 350 °C pretreated zeolite suggests that the activation temperature likely drove off surface hydroxyl groups from the sample in the form of water, thereby yielding less OH present for the interaction with CO2. It is likely that the removal of surface OH during the higher activation temperature also led to an increased number of available adsorption sites for CO2, as evidenced by the observation of bridged bidentate carbonate over the 350 °C activated sample, whereas it was not observed over the 120 °C activated sample. The slow decay in the concentration of physisorbed CO2 during desorption at 30 °C suggests that the strength of the interaction is nearly equivalent to the thermal energy in the system at ambient temperature. (15) Davydov, A. A. Infrared Specroscopy of Adsorbed Species on the Surface of Transition Metal Oxides; John Wiley and Sons: Chichester, U.K., 1990. (16) Little, L. H. Infrared Spectra of Adsorbed Species; Academic Press: New York, 1966.

3072 Energy & Fuels, Vol. 22, No. 5, 2008

SteVens et al. Scheme 1. Modes of CO2 Adsorption

Adsorption of CO2 at 120 °C onto the 120 °C pretreated sample is shown in Figure 3. Initial exposure of the sample to CO2 led to the formation of bridged bidentate carbonate bands at 1693 and 1362 cm-1, similar to the observations at 30 °C. The bridged bidentate carbonate bands did grow slightly in intensity with an increased time of exposure, while physisorbed CO2 could be seen at 2358 cm-1, extending beyond the gaseous CO2 doublet. No other carbonate-type adsorbates were observed. Purging the surface with N2 at 120 °C led to nearly complete removal of the bridged bidentate carbonate and physisorbed CO2 after 7.5 min. Figure 4 illustrates the adsorption of CO2 at 120 °C onto the 350 °C pretreated 13X. Exposure of the sample to the CO2 flow initially led to rapid formation of bridged bidentate carbonate at 1709 and 1364 cm-1, bidentate carbonate at 1481 and 1426 cm-1, physisorbed CO2 at 2355 cm-1, an OH band at 3691 cm-1, and a band at 1646 cm-1. Increased time of exposure led to growth of the bridged bidentate carbonate, the monodentate carbonate, and the OH bands, whereas the bidentate carbonate bands at 1481 and 1426 cm-1 decreased in intensity. Physisorbed CO2 grew quickly to a stable intensity after only a 0.5 min exposure. Desorption of the CO2 species via a N2 purge at 120 °C led to a significant decrease in bridged bidentate carbonate after 24.5 min, whereas the bands corresponding to

bidentate carbonate at 1481 and 1426 cm-1 as well as the OH band at 3691 cm-1 and band at 1646 cm-1 increased in intensity. The formation of bidentate carbonate over only the 350 °C pretreated samples suggests that the adsorbate forms only at higher levels of CO2 coverage. Because bridged bidentate carbonate requires two sites for adsorption while bidentate carbonate only requires one site, it would be expected that bidentate carbonate would be preferred at the lower surface coverages observed with the 120 °C pretreated samples. The presence of bridged bidentate carbonate during all conditions suggests that bridged bidentate carbonate is more stable than that of bidentate carbonate. Jacobs et al.10 reported that the growth of bidentate carbonate occurs at the expense of bridged bidentate carbonate. The present results show the opposite trend: bidentate carbonate at 1481 and 1426 cm-1 decreased in intensity as bridged bidentate carbonate at 1709 and 1364 cm-1 increased in intensity (Figures 2 and 4). The survival of bidentate carbonate after the purge suggests that its bonding to the surface sites is stronger than that of bridged bidentate carbonate, which decreased in intensity. Jacobs et al. reported also that bidentate carbonate is thermally more stable than bridged bidentate carbonate.10 From the results shown in Figures 1–4, a series of mechanistic steps to describe the interaction of CO2 with the zeolite surface

Figure 1. FTIR analysis of 1 vol % CO2-N2 adsorption at ambient temperature onto the 120 °C pretreated 13X zeolite. The purge phase was also conducted at ambient temperature under pure N2 flow. Adsorption times indicated are relative to the onset of CO2 exposure; purge times indicated are relative to the onset of the N2 purge phase.

Figure 2. FTIR analysis of 1 vol % CO2-N2 adsorption at ambient temperature onto the 350 °C pretreated 13X zeolite. The purge phase was also conducted at ambient temperature under pure N2 flow. Adsorption times indicated are relative to the onset of CO2 exposure; purge times indicated are relative to the onset of the N2 purge phase.

CO2 Adsorption onto Zeolite Materials

Figure 3. FTIR analysis of 1 vol % CO2-N2 adsorption at 120 °C onto the 120 °C pretreated 13X zeolite. The purge phase was also conducted at 120 °C under pure N2 flow. Adsorption times indicated are relative to the onset of CO2 exposure; purge times indicated are relative to the onset of the N2 purge phase.

Figure 4. FTIR analysis of 1 vol % CO2-N2 adsorption at 120 °C onto the 350 °C pretreated 13X zeolite. The purge phase was also conducted at 120 °C under pure N2 flow. Adsorption times indicated are relative to the onset of CO2 exposure; purge times indicated are relative to the onset of the N2 purge phase.

are proposed and illustrated in Scheme 2. Adsorbed H2O on the surface of the zeolite may block the adsorption sites of the zeolite, inhibiting the formation of carbonates yet still allowing for formation of physisorbed CO2 (path 1). Removal of adsorbed H2O through heating may allow for the direct interaction between the incident CO2 molecule and surface hydroxyl groups, resulting in displacement and/or removal of the OH groups to form adsorbed water and carbonates (path 2); bridged bidentate carbonate10 is shown here because it is the most common adsorbate observed during our tests. Further heating of the zeolite sample may lead to the removal of the OH groups in the form of H2O prior to exposure to CO2, resulting in the formation of carbonates without the removal of OH during CO2 exposure (path 3); bidentate carbonate formation may be facilitated through this pathway. During our tests, the 350 °C pretreatment appeared to lead to a combination of both paths 2 and 3 in addition to physisorbed species, whereas pretreatment at 120 °C appeared to lead to a combination of paths 1 and 2.

Energy & Fuels, Vol. 22, No. 5, 2008 3073

Effect of H2O on CO2 Adsorption onto 13X. Because 13X has the highest CO2 capture capacity of the zeolites in the present study,5 we selected it for competitive CO2/H2O adsorption testing. Figure 5 illustrates the adsorption, in a sequential manner, of H2O and CO2 at 120 °C onto a 350 °C pretreated 13X. Exposure of the sample to the H2O-N2 flow initially led to the rapid formation of OH at 3688 cm-1 and a band centered at 1652 cm-1. Increased exposure time led to growth of OH at 3688 cm-1 and the broadband at 1652 cm-1, as well as the formation of an additional OH band at 3661 cm-1 and loss of OH (negative) at 3743 cm-1. Further exposure to H2O-N2 led to the disappearance of OH at 3688 cm-1, further loss of OH at 3743 cm-1, the growth of a shoulder at 1673 cm-1, and a shift of 1652 to 1639 cm-1. Initiation of the CO2 flow (along with H2O) led only to the appearance of gaseous CO2 bands; no physisorbed CO2 or formation of carbonate bands was observed after 10 min of exposure. Switching gas feeds to CO2-N2 (no H2O) led only to a decrease of 1673 cm-1 and a reappearance of OH at 3688 cm-1, which became less negative after more than 1 h of exposure. It is also noteworthy that, after H2O exposure, exposure to CO2 led to no adsorbed species, suggesting that H2O can block the adsorption sites of the zeolite sorbent. Figure 6 depicts simultaneous adsorption of CO2 and H2O over the 350 °C pretreated 13X at 120 °C. Exposure initially led to the formation of OH at 3690 cm-1 and a band at 1652 cm-1. Increased exposure time led to growth of OH at 3690 cm-1, formation of physisorbed CO2 at 2354 cm-1, bridged bidentate carbonate at 1710 and 1363 cm-1, bidentate carbonate at 1484 and 1429 cm-1, and a shoulder at 1686 cm-1. Further exposure led to the disappearance of physisorbed CO2, bridged bidentate carbonate, and bidentate carbonate, as well as loss of OH at 3743 and 3690 cm-1. After only 7 min of CO2-H2O exposure, the only surviving adsorbate was a band at 1652 cm-1, which is clearly a bending mode vibration for H2O. During CO2 adsorption in the absence of H2O, identical adsorbates were observed (previously discussed): OH at 3691 cm-1, bridged bidentate carbonate at 1709 and 1364 cm-1, bidentate carbonate at 1481 and 1426 cm-1, as well as the formation of a shoulder at 1686 cm-1. After formation, the carbonates remained while within the CO2 environment. Upon purging with N2 at 120 °C, the bidentate carbonate at 1481 and 1426 cm-1 and OH at 3691 cm-1 survived after nearly 30 min, while bridged bidentate carbonate at 1709 and 1364 cm-1 nearly vanished. The competitive adsorption illustrated by Figure 6 shows that, not only can H2O block the adsorption sites, but it can also displace adsorbed CO2 species from the surface even while in a CO2-rich environment. This result agrees with a previous study of X-type zeolites at low pressure (