Environ. Sci. Technol. 2003, 37, 2596-2599
A Biomass-Supported Na2CO3 Sorbent for Flue Gas Desulfurization HONGSHAN SHANG, TI OUYANG, FAN YANG, AND YUAN KOU* College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
gas in the NOXSO process, but a relatively higher temperature and suitable catalyst are generally needed 400-600 °C
Na2SOx/γ-Al2O3 + reductant 9 8 H 2S + catalyst
Na2CO3 (x ) 3, 4)
In fact the sulfite/sulfate is first reduced to H2S, which then subsequently reacts with excess SO2
SO2 + 2H2S f 3S + 2H2O
A novel sorbent for SO2 removal has been investigated. The sorbent is obtained by conventional incipient wetness impregnation of abandoned biomaterials (straw or dried leaves) with an aqueous solution of Na2CO3. A material with the composition 80 wt % Na2CO3/straw shows a desulfurization activity which is both higher and faster than that of the reference sample Na2CO3/γ-Al2O3. The breakthrough and stoichiometric SO2 adsorption efficiencies for 80 wt % Na2CO3/straw reach 48.9% and 80.6%, respectively, at a temperature of 80 °C. The adsorption efficiencies are almost constant in the temperature range 70 to 300 °C. According to IR and XPS analysis the main products observed on the spent sorbent are sulfite below 150 °C and sulfate at 300 °C. The Na2CO3 in 80 wt % Na2CO3/straw can potentially be recycled by the oxidation of the straw with concomitant reduction of the sulfite species to elemental sulfur, making the proposed process CO2 neutral.
Introduction Traditional SO2 removal processes using calcium-based sorbents are so-called wet-processes since large amounts of water have to be used. In dry areas such as those in North China, the lack of water resources limits the application of wet processes in many power stations although the processes are easy to construct and can be operated with viable capital and running costs. Calcium-based sorbents can also be used in semidry or dry processes, but their efficiency is generally lower than that of the wet processes and they are harder to fit as an add-on to an existing combustion system. Therefore, there is considerable interest in studying other materials as potential recyclable solid sorbents in dry desulfurization processes. Although many different types of solid sorbents have been investigated, including activated carbon/coke (1, 2), supported copper oxide (3-6), and supported sodium carbonate (7-15), the key question is always whether the process is sufficiently economic and environmentally friendly in comparison with current wet processes. For example, in the NOXSO process (10, 11), Na2CO3/γ-Al2O3 has been employed as a regenerable sorbent for SO2. The chemistry involved in such a process must include two key steps, which can be shown schematically as follows: (1) enrichment of SO2 80-150 °C
SO2 + Na2CO3/γ-Al2O3 98 Na2SO3/γ-Al2O3 + CO2 Na2SO3/γ-Al2O3 + 1/2 O2 f Na2SO4/γ-Al2O3 (2) reduction of sulfite/sulfate to elemental sulfur. In that case not only is a reducing agent essential, for example natural * Corresponding author phone: (8610) 6275-7792; fax: (8610) 62751708; e-mail:
[email protected]. 2596
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
The outlook for actual application of these processes in current power stations is not promising due to the complex nature of the processes involved, the high cost of the sorbents, and the large amount of energy required. An alternative cheap, easily obtainable and disposable/recyclable sorbent is clearly a prerequisite for a highly efficient, industrially acceptable desulfurization system. We have proposed (16, 17) that utilization of biomaterials or abandoned biomaterials (abbreviated in this paper as BIOM), of which the major component is cellulose (C6H10O5)n, is a creative new approach to reach this goal. Examples of BIOM include farmland straw, rice husks and stalks, dried branches and leaves, waste paper, and similar materials. Solid castoffs in city waste also contain many biomaterials such as dried branches and leaves. Some of the BIOM is recycled in the soil and some is collected for use as a domestic fuel. However, in China other biomass materials (e.g. stubble in wheat fields) are still burnt in situ, generating CO2 and other forms of pollution without leading to any reduction in comsuption of fossil fuels. Instead of being destroyed by atmospheric oxidation in this way, it should be possible for the BIOM to act as the reductant for sodium sulfite/sulfate. Oxidation of the BIOM will regenerate Na2CO3 and also avoid CO2 release. The strategy proposed is described schematically below: 80-110 °C
Na2CO3/BIOM + SO2 98 Na2SO3/BIOM + CO2 calcination
Na2SO3/BIOM 98 S + Na2CO3 + H2O In this paper, we show that Na2CO3, a traditional solid sorbent precursor, can be supported on BIOM, generating a very efficient SO2 sorbents (16, 17). We will show in a subsequent paper that, after enrichment of SO2 by Na2CO3/ BIOM, the reduction of S4+ to S can indeed be realized by BIOM itself. This new approach combines energy efficiency, waste treatment, and flue gas desulfurization in one process.
Materials and Methods Materials and Sorbent Preparation. The wheat straw employed as the BIOM in this work was from the farmland in Shandong Province, China. The dried leaves were from the campus of Peking University. The composition of the wheat straw was 40.50% C, 5.58% H, and 45.26% O based on elemental analysis (Elementar Vario EL, Germany). The major component was cellulose (C6H10O5)n, the formula of which is shown in Figure 1. The composition of the dried leaves was 45.02% C, 5.78% H, and 36.27% O. A typical preparation procedure for an 80 wt % Na2CO3/ straw sorbent was as follows. A sample of straw (2 g) was cut into pieces below 0.28 mm in length and impregnated with aqueous Na2CO3 solution (20 g Na2CO3/100 g H2O) (40 mL). The mixture was kept at room temperature for several days until most of the water had evaporated and was then dried at 110 °C for 24 h. Several samples of Na2CO3/BIOM with 10.1021/es021026o CCC: $25.00
2003 American Chemical Society Published on Web 04/25/2003
TABLE 1. Physical Properties of the Sorbents Used
FIGURE 1. Structural formula of cellulose. different loadings were prepared by varying the concentration of the Na2CO3 solution prior to impregnating the BIOM. The reference sorbent Na2CO3/γ-Al2O3 was prepared according to the literature procedure (14). Methods. The SO2 removal experiments were carried out under atmospheric pressure, using a fixed-bed quartz reactor, in which the sorbent (0.33 g) was supported on a quartz net of medium porosity (80 mesh). Some silica sand (30-60 mesh) was put on the surface of the quartz net to support the sorbent, the particle size of which was below 0.63 mm (>30 mesh). The flow rate was 40 mL/min. The simulated flue gas contained 1960 ppm SO2 (99.9%) with N2 (99.999%) as the balance gas. An FT-IR spectrometer (Vector 22, Bruker) with an on-line cell was used to monitor the SO2 concentrations before and after the reactor as shown in Figure 2. The concentrations were calibrated from a standard curve in the range 50-2000 ppm, obtained by means of blank experiments in 100 ppm steps based on the peak area of the SO2 absorption at 1360 cm-1. The adsorption efficiency η (breakthrough/ stoichiometric/saturation) (18, 19) of the sorbent is given by η ) fs/fc × 100%, where fc is the amount in moles of Na2CO3 in the sorbent and fs is the amount in moles of SO2 adsorbed by the sorbent. For breakthrough efficiency, fs ) v‚t, where v is the flow rate in moles per minute and t is the breakthrough time in minutes. For stoichiometric efficiency, fs ) fc - ∫ttbs v(t)dt, where ts is stoichiometric time in minutes. The Na/S ratio is given by r ) 2 fc/fs and is theoretically equal to 2. Characterization of the sorbents was performed by XPS (VG ESCA LAB 5, Al KR-1486.6 eV), XRD (Dmax-2000, Rigaku, λ (Cu KR1) ) 1.54 Å), SEM (JSSM-6301F, JEOL), ATR-IR (Vector 22, Bruker), and BET (ASAP 2010 V3.02) methods.
Results and Discussion Characterization of Fresh Sorbents. Although supported Na2CO3 is frequently considered as an SO2 sorbent, bulk Na2CO3 itself is not effective because it only has a very limited pore volume and thus a very small surface area. Dispersion of Na2CO3 at the level of 80 wt % on the surface of straw leads to a major increase in both its surface area (20 times) and pore volume (13 times). There is only a relatively small decrease in the modal pore diameter (from 43.3 to 30.5). Thus it is clear that the surface properties of Na2CO3 change considerably after being supported on the surface of the straw and its ability to absorb SO2 should be greatly enhanced. Compared with the straw, inorganic oxide supports such as γ-Al2O3 have a much large surface area (>150 m2/g). The
sample
surface areaa (m2/g)
pore vol. (cm3/g)
modal pore diameter (nm)
density (g/cm3)
Na2CO3 straw 80 wt % Na2CO3/straw 20 wt % Na2CO3/γ-Al2O3
0.6 2.3 12.1 34.4
0.0066 0.0046 0.093 0.14
43.3 8.0 30.5 16.0
2.53 0.68 1.15 1.60
a Determined using N as absorbate and helium as carrier gas at a 2 temperature of 78 K, using a sample dried at a temperature of 150 °C for 30 min.
dispersion of Na2CO3 on the surface may improve the surface properties of Na2CO3 but generally results in a significant decrease in surface area. Indeed, the surface area of 20 wt % Na2CO3/γ-Al2O3 used as the reference in this work is reduced to only 34.4 m2/g. In contrast, however, it can be seen from Table 1 that the surface area of 80 wt % Na2CO3/ straw is about five times larger than the straw itself and the volume is some 20 times larger than that of the straw, suggesting a strong interaction between Na2CO3 and the surface of straw. XRD analysis provides further evidence for this strong interaction. Figure 3 shows the XRD patterns of straw, 80 wt % Na2CO3/straw, pure Na2CO3 treated under the same conditions as that of 80 wt % Na2CO3/straw (dissolved in water, kept at room temperature until most of the water had evaporated and then dried at 110 °C for 24 h) and untreated pure Na2CO3. The diffraction pattern of the unsupported treated Na2CO3 shows no difference from that of untreated Na2CO3. However, it is interesting that 80 wt % Na2CO3/straw shows a completely different pattern, indicating the formation of two new Na2CO3 species, Na3(CO3)(HCO3)‚2H2O (JCPDS 291447) and Na2CO3‚H2O (JCPDS 8-448) on the surface. Dispersion of sodium carbonate on the surface of straw clearly favors the growth of hydrated sodium carbonate species. Figure 4 shows the IR spectra of pure Na2CO3, wheat straw, and fresh 80 wt % Na2CO3/straw. It can be seen that the pure Na2CO3 exhibits two strong absorption peaks at 1425 and 879 cm-1 which can be assigned to CO32- vibrations. In the spectrum of 80 wt % Na2CO3/ straw, however, the peak at 1425 cm-1 shows a shoulder at 1409 cm-1. The peak at 879 cm-1 shifts to 865 cm-1, and two shoulders at 902 and 849 cm-1 also appear. Although the detailed structure is unclear, it is likely that the environment of CO32- has changed significantly and that the new adsorption peaks can be assigned to the formation of activated Na2CO3 on the surface. In addition, 80 wt % Na2CO3/straw exhibits a peak at 1683 cm-1, indicative of the formation of HCO-3 on the surface (20). When 80% Na2CO3 is treated with SO2 at 80 °C, the sulfur is almost completely converted into sulfite as shown by the presence of an adsorption peak at 964 cm-1 (21). As the temperature at which the sorbent is treated with SO2 is increased, the amount of sulfate formed
FIGURE 2. Schematic diagram of the experimental apparatus and FTIR online analysis system. VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2597
TABLE 2. SO2 Removal Efficiency of Different Sorbents at 80 °Ca breakthrough sorbent Na2CO3 80 wt % Na2CO3/γ-Al2O3 20 wt % Na2CO3/γ-Al2O3 80 wt % Na2CO3/straw 80 wt % Na2CO3/dried leaves
efficiency time (mol %) (h) 0 7.6 37.2 48.9 42.2
Na/S
0 0.9 26.6 1.1 5.38 5.8 4.09 5.1 4.72
stoichiometric Na/S ) 2 efficiency (mol %)
Na/S
8.6 24.0 74.9 80.6 75.8
23.2 8.34 2.67 2.48 2.63
a Reaction conditions: 0.33 g sorbent, 80 °C, total flow rate 40 mL min-1, SO2 1960 ppm, balance N2.
FIGURE 3. XRD patterns of sorbents. ∆Na2CO3 (Natrite, JCPDS37451); × Na3(CO3)(HCO3)‚2H2O (JCPDS29-1447);‚Na2CO3‚H2O(JCPDS8448).
FIGURE 4. IR spectra of sorbents and spent sorbents. increases as shown by the intensity of a band at 1109 cm-1. Since sulfite is more readily reduced than sulfate to sulfur, this suggests that use of Na2CO3 at low temperature is preferable, since reduction of the sulfur is our long-term goal. 2598
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
FIGURE 5. Effect of temperature on SO2 removal for 80 wt % Na2CO3/straw as sorbent. ∆Breakthrough efficiency; bStoichiometric efficiency. Desulfurization Activity. A comparison of Na2CO3/straw, Na2CO3/γ-Al2O3, and pure Na2CO3 as sorbents for the removal of SO2 at 80 °C is given in Table 2. The breakthrough adsorption efficiency of 80 wt % Na2CO3/straw is 48.9% which is 5.4 times the value for 80 wt % Na2CO3/γ-Al2O3. The stoichiometric adsorption efficiency of 80 wt % Na2CO3/straw is 80.6% which is 2.4 times the value for 80 wt % Na2CO3/ γ-Al2O3. These data clearly show that the straw is a highly efficient support medium. The activity of the 80 wt % Na2CO3/straw is also significantly higher than 20 wt % Na2CO3/ γ-Al2O3, which is a typical sorbent employed in PPFB processes (14). As shown in Table 2, the breakthrough adsorption efficiency of 80 wt % Na2CO3/straw is 11.7% higher than that of the 20 wt % Na2CO3/γ-Al2O3, while the stoichiometric efficiency is 5.7% higher. Therefore, it can be concluded that 80 wt % Na2CO3/straw has the highest adsorption efficiency and stoichiometric adsorption. Calculations based on the data demonstrate that the high adsorption efficiencies and low Na/S ratio of 80 wt % Na2CO3/straw are due to faster adsorption of SO2 (V ) fs/t ) 1.72 × 10-4 mol SO2/h) than that observed on 20 wt % Na2CO3/ γ-Al2O3 (V ) 1.59 × 10-4 mol SO2/h). Table 2 also shows that the desulfurization activity of 80 wt % Na2CO3/dried leaves is almost the same as that of 80 wt % Na2CO3/straw. The breakthrough and stoichiometric efficiencies are about 42.2% and 75.8%, respectively, indicating that Na2CO3 supported on this type of abandoned biomaterial is also very efficient in terms of removal of SO2. Figure 5 shows the effect of temperature on the efficiency of SO2 removal for 80 wt % Na2CO3/straw. It can be seen that as the temperature is increased from 70 to 300 °C, both the breakthrough adsorption efficiency and the stoichiometric adsorption efficiency are almost constant, indicating that
FIGURE 6. Effect of Na2CO3 loading on the removal of SO2 at a temperature of 80 °C. bBreakthrough efficiency × stoichiometric efficiency.
(C6H10O5)n (shown in Figure 1) containing three hydroxyl groups (-OH) for each unit. The hydroxyl groups of straw can link to the oxygen atoms in CO32- by either a hydrogen bond or Van der waals forces, resulting in chemisorption or physisorption respectively on the surface. The results presented in this paper for the fresh 80 wt % Na2CO3/straw sorbent clearly indicate that chemisorption plays a key role for this material in the light of the observed growth of fine crystallites, the changes induced in the environment of the CO32- ions, and the activation of sodium carbonate. The Na2CO3 can interact with the surface of the biomass via hydrogen bonding in two ways, i.e., terminal or bridging modes as shown in Figure 7. In either case, the symmetry of the CO32ion will be significantly altered. Fine crystallites as well as possibly amorphous species are randomly formed at the hydrogen-bond-based interfaces. The surface area and pore volume increase markedly without any significant reduction in modal pore diameter. Most of the sodium ion sites are exposed and are thus directly accessible for reaction with SO2. Fast adsorption occurs, assisted by the water molecules retained in the crystallites. This is the main reason the 80 wt % Na2CO3/straw shows a high activity for SO2 removal.
Acknowledgments The authors wish to express their heartfelt thanks to Professors Xu-chang Xu and Chang-he Chen at Tsinghua University for helpful discussions and to an anonymous reviewer for many valuable comments. The project was supported by the Ministry of Science and Technology, China (G1999022202).
Literature Cited
FIGURE 7. Adsorption model for CO32- on the surface of straw. the desulfurization activity of 80 wt % Na2CO3/straw can be maintained over a wide temperature range. The effect of Na2CO3 loading on the breakthrough efficiency of Na2CO3/ straw is shown in Figure 6. The efficiency is low until the Na2CO3 loading exceeds 40%. At this point, the desulfurization activity, as shown by values of both breakthrough and stoichiometric efficiency, increases markedly. The breakthrough efficiency is significantly further improved with increasing Na2CO3 loading and reaches a maximum at 80 wt % Na2CO3. Characterization of Spent Sorbent. XPS analysis of the 80 wt % Na2CO3/straw used at 80 °C shows a single peak at a binding energy 167.8 eV corresponding to S4+ (22) in its S2p spectrum while the sorbent used at 300 °C shows two peaks, i.e., a new peak at 169.6 eV corresponding to S6+ (22) is also present. The S4+/S6+ ratio based on peak area is 0.16, indicating that most of the adsorbed SO2 species has been oxidized at the higher temperature. These results are consistent with the results from infrared spectroscopy (Figure 4) discussed above. Na2CO3/Biomass Reaction Model. To understand the unique microstructure and the efficient activation of Na2CO3 on the surface of the biomass, we propose a surface model at the atomic level as shown in Figure 7. It is known that the main chemical component of straw is cellulose
(1) Armor, J. N. Appl. Catal. B 1992, 1, 221-256. (2) Tsuji, K.; Shiraishi, I. Fuel 1997, 76, 549-560. (3) Centi, G.; Passarini, N.; Perathoner, S.; Riva, A. Ind. Eng. Chem. Res. 1992, 31, 1947-1955. (4) Yoo, K. S.; Kim, S. D.; Park, S. B. Ind. Eng. Chem. Res. 1994, 33, 1786-1791. (5) Yoo, K. S.; Jeong, S. M.; Kim, S. D.; Park, S. B. Ind. Eng. Chem. Res. 1996, 35, 1543-1549. (6) Macken, C.; Hodnett, B. K. Ind. Eng. Chem. Res. 2000, 39, 38683874. (7) Keener, T. C.; Davis, W. T. JAPCA 1984, 34, 651-654. (8) Mocek, K.; Beruto, D. Mater. Chem. Phys. 1986, 14, 219-227. (9) Muzio, L. J.; Offen, G. R. JAPCA 1989, 39, 1206-1209. (10) Ma, W. T.; Haslbeck, J. L. Environ. Prog. 1993, 12, 163-168. (11) Snip, O. C.; Wood, M.; Korbee, R.; Schouten, J. C.; Bleek, V. D. Chem. Eng. Sci. 1996, 51, 2021-2029. (12) Xu, G.; Gao, S.; Kato, K. Trans. Inst. Chem. Eng. 1999, 77, 77-87. (13) Xu, G.; Wang, B.; Kato, K. J. Chem. Eng. Jpn. 1999, 32, 82-90. (14) Xu, G.; Luo, G.; Akamatsu, H.; Kato, K. Ind. Eng. Chem. Res. 2000, 39, 2190-2198. (15) Guldur, C.; Dogu, G.; Dogu, T. Chem. Eng. Proc. 2001, 40, 1, 13-18. (16) Kou, Y.; Shang, H. S.; Yang, F.; Yan, Z. China Pat. Appl. 2001, Number: 01115505. (17) Kou, Y.; Shang, H. S.; Yang, F.; Yan, Z. China Pat. Appl. 2001, Number: 01115506. X. (18) Mocek, K.; Stejskalova, K.; Bach, P.; Lippert, E.; Bastl, Z.; Spirovova, I.; Erdos, E. Collect. Czech. Chem. Commun. 1996, 61, 825-836. (19) Stejskalova, K.; Bach, P.; Lippert, E.; Mocek, K. Collect. Czech. Chem. Commun. 1997, 62, 387-391. (20) Kyoko, K. B.; Kazuhiro, S.; Hironori, A. Appl. Catal. A: General 1997, 165, 391-409. (21) Wang, Y.; Mohammed, A. B.; Lavalley, J. C. Appl. Catal. B: Environ. 1998, 16, 279-290. (22) Stejskalova, K.; Spirovova, I.; Lippert, E. Appl. Surf. Sci. 1996, 103, 509-516.
Received for review November 5, 2002. Revised manuscript received February 19, 2003. Accepted March 3, 2003. ES021026O VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2599
