Environ. Sci. Technol. 2008, 42, 2509–2514
Influence of Calcium Content of Biomass-Based Materials on Simultaneous NOX and SO2 Reduction SARMA V. PISUPATI* AND SUMEET BHALLA Energy and Mineral Engineering Department, The Pennsylvania State University, 110 Hosler Building, University Park, Pennsylvania 16802
Received August 3, 2007. Revised manuscript received December 15, 2007. Accepted January 7, 2008.
Pyrolysis products of biomass (bio-oils) have been shown to cause a reduction in NOx emissions when used as reburn fuels in combustion systems. When these bio-oils are processed with lime, calcium is ion-exchanged and the product is called BioLime. BioLime, when introduced into a combustion chamber, pyrolyzes and produces volatile products that reduce NOx emissions through reburn mechanisms. Simultaneously, calcium reacts with SO2 to form calcium sulfate and thus reduces SO2 emissions. This paper reports the characterization of composition and pyrolysis behavior of two BioLime products and the influence of feedstock on pyrolysis products. Thermogravimetric analysis (TGA) and 13C-CP/MAS NMR techniques were used to study the composition of two biomassbased materials. The composition of the pyrolysis products of BioLime was determined in a laboratory scale flow reactor. The effect of BioLime composition on NOx and SO2 reduction performance was evaluated in a 146.5 kW pilot-scale, down fired combustor (DFC). The effect of pyrolysis gas composition on NOx reduction is discussed. The TGA weight loss curves of BioLime samples in an inert atmosphere showed two distinct peaks corresponding to the decomposition of light and heavy components of the BioLime and a third distinct peak corresponding to secondary thermal decomposition of char. The study also showed that BioLime sample with lower content of residual lignin derivatives and lower calcium content produced more volatile compounds upon pyrolysis in the combustor and achieved higher NOx reduction (15%). Higher yields of pyrolysis gases increased the NO reduction potential of BioLime through homogeneous gas phase reactions. Calcium in BioLime samples effectively reduced SO2 emissions (60–85%). However, addition of higher calcium content to the BioLime samples also appeared to inhibit the volatile yield and thereby lowered the NOx reduction.
Introduction Pressure on utilities is increasing as a result of Title 1 of the Clean Air Act Amendments of 1990 and the Clear Skies Act to reduce the NOx emissions at cost significantly lower than the current selective catalytic reduction (SCR) technology. Several technologies such as low excess air firing (LEA), staged combustion or the use overfire air (OFA), low NOx burners, * Corresponding author e-mail:
[email protected]. 10.1021/es0719430 CCC: $40.75
Published on Web 02/19/2008
2008 American Chemical Society
selective catalytic reduction (SCR), and selective noncatalytic reduction (SNCR) are available for NOx reduction. Sulfur dioxide can be controlled by technologies such as dry sorbent injection, spray dryers, scrubbers etc. Many methods involving biomass and biomass-based products are being explored to reduce not only SO2 and NOx emissions, but also carbon dioxide emissions. Bio-oil produced from pyrolysis of biomass is simultaneously reacted with air and lime/water slurry to ion-exchange calcium with acidic functional groups to produce a product called BioLime. BioLime is a registered trademark product of Dynamotive Technologies Corporation, Vancouver, Canada. The BioLime can be fired in combination with traditional combustion fuels to reduce sulfur dioxide (SO2) and nitrogen oxides (NOx) emissions because of low nitrogen and sulfur content and also reduced greenhouse gases (N2O and CO2). Since BioLime is an emulsion of biomass pyrolysis oil and calcium, upon introduction into a high temperature combustor, pyrolysis products react with NOx through a series of homogeneous and heterogeneous reactions, reducing it to N2. This NOx reduction is similar to that of a “reburn” mechanism. The reburning process (1–8) is reasonably well understood for coal and natural gas as reburn fuels. However pyrolysis products of biomass and other biomass-based materials have also been shown to cause NOx reduction by mechanisms similar to reburning (1). Biomass and biomass-based products offer the advantage of not only reducing SO2 and NOx emissions because of low inherent nitrogen and sulfur contents, but also reducing greenhouse gas emissions (CO2 and N2O) (9, 10). Composition of biomass (lignin, cellulose, and hemicellulose content) and the pyrolysis conditions influence the distribution of pyrolysis products (gas, liquid, char) (11, 12) and hence the reduction of SO2 and NOx. Therefore, a better understanding of the composition of the feedstock will help in selecting appropriate feedstock that will increase the yield of pyrolysis gases that are responsible for reducing NOx and SO2 emissions. Individual components (lignin, cellulose, and hemicellulose) of biomass and bio-oils play a significant role in determining the pyrolysis behavior of the feedstock (13, 14). It is also reported that the way in which these components are bound is not as important as the actual amounts of individual components present in the biomass or bio-oils. The hydrocarbon radicals produced during pyrolysis of BioLime react with NOx and reduce NOx to molecular N2. The ion-exchanged Ca in an atomic state efficiently reacts with SO2 molecules to form CaSO4 (solid) which is easily removed from the combustor in a solid form. While it is beneficial to ion exchange more calcium for SO2 reduction, the effect of adding more calcium on pyrolysis products (hydrocarbons) produced during reburning and their effect of NOx reduction is not clear.
Objective of This Study The objective of this study is to gain a better understanding of the influence of calcium content of biomass-based products (BioLime) on the evolution of volatiles during pyrolysis, and their subsequent effect on NOx and SO2 reduction potential. In the present work, composition of two biomass-based materials was studied using TGA and CP/ MAS 13C NMR techniques. Samples were then flash pyrolyzed in a flow reactor to relate the yield of pyrolysis gases and char with the composition of bio-oil. BioLime samples were used in a 146.5 kW pilot-scale down fired combustor (DFC), and the NOx and SO2 reductions were determined. VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2509
TABLE 1. Composition of BioLime Samples BioLime I (dry basis) volatile matter (wt. %)a 68.73 ash (wt. %)a 22.80 fixed carbon (wt. %)a 8.47 HHV (MJ/kg) carbonb 48.5 hydrogenb 7.7 nitrogenb 0.23 b sulfur 0.01 oxygenb 43.5 H/C ratiob 1.91 O/C ratiob 0.67 a
Dry basis.
b
BioLime III (dry basis) 49.11 50.57 0.33 38.7 8.4 0.21 0.02 52.6 2.61 1.01
Dry ash free basis.
Experimental Section Two biomass-based materials (BioLime I and III) were obtained from the DynaMotive Technologies Corporation, Vancouver, Canada. BioLime samples I and III have approximately 7 and 14% wt. % of calcium. Both samples were highly viscous liquids. The dried samples were ground to pass though screen 60 mesh screen and stored in sample bottles. Compositional analysis of the BioLime samples is shown in Table 1. Thermogravimetric Analysis. A Perkin-Elmer 7 series thermal analysis system was used for the TGA studies on the two BioLime samples. Approximately 10 mg of BioLime sample was placed in a platinum crucible of the TGA furnace. The sample mass was chosen to ensure that vapor-phase mass transfer did not hinder BioLime pyrolysis. The furnace was heated from 30 to 1000 °C at a rate of 50 °C min-1 and then held at 1000 °C for 10 min. The methodology used was similar to that reported by Ghetti et al. (13), Vitolo and Ghetti (15), and Raveendran et al. (14) in characterizing biomass materials. All pyrolysis studies were carried out in an inert atmosphere of flowing nitrogen (150 sccm). Each test yielded a plot of weight loss versus time and temperature. The percentage weight loss was calculated based on the ash-free initial sample weight as a function of temperature. 13C CP/MAS NMR Analysis. 13C NMR has been used to study the structure of biomass and biomass-based materials. The solid state 13C NMR spectra, employing the techniques of cross polarization/magic angle spinning (CP/MAS) were collected for samples on a Chemagnetics M100 NMR spectrometer operating at field strength of 2.3 T. Approximately 250 mg of sample was loaded into a 9 mL ceramic rotor for the experiments. The rotor was then placed in a spectrometer and accelerated to a spinning rate of 3.5 kHz. Spectra were indirectly referred to tetramethylsilane (TMS). The peaks were quantified using GRAMS/32 software involving a mixture of Gauss and Lorz integration methods. Flash Pyrolysis in a Flow Reactor. Rapid pyrolysis of BioLime occurs when injected into a combustor at high temperature. Therefore, samples were flash pyrolyzed in a laboratory flow reactor to study the composition of the pyrolysis products released and to relate the yield of these pyrolysis gases and char to the sample composition. A Lindberg single-zone furnace with a 61.0 cm heating zone and capable of achieving and maintaining temperatures between 200 and 1200 °C was used in the study. A detailed description of the furnace can be seen elsewhere (16). Approximately 2 g of BioLime sample was used for each test. The samples were pyrolyzed at furnace temperatures of 1100 and 1200 °C in the reactor by using helium as a carrier gas at a total flow rate of 0.7 scmh. The peak temperatures were selected to match the optimum temperatures reported by Pisupati et al. (17) using the DFC. The pyrolysis products 2510
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
leaving the reactor were then passed through a filter paper to collect the char. The gases were collected in a 10 L TEDLAR bag and were later analyzed using a gas chromatograph (Arnel 0158 series model). Evaluation of NOx and SO2 Reduction. Brouwer et al. (18) have demonstrated the effective use of biomass and biomass-based materials as reburn fuel for reducing NOx. A series of tests was done DFC to evaluate the SO2 and NOx reduction potential of both BioLime I and III. A detailed description of the DFC is provided elsewhere (19). The combustor has a 50.8 cm internal diameter, and is 3.05 m high. The combustor is designed for a nominal thermal input of 88 kW, but can be increased up to 147 kW.
Results and Discussion The proximate analysis (Table 1) showed that the ash yield of BioLime III is approximately 2.2 times that for BioLime I. This corresponds to 14% by weight of calcium for wet BioLime III and 7% by weight of calcium for wet BioLime I. Both of the wet samples have approximately 53 wt. % water. BioLime I had higher volatile matter (approximately 1.4 times) than BioLime III on a dry basis. Liu et al. (6) in their study of optimal fuel characteristics for efficient reduction by coal reburning, have shown that NO reduction efficiency generally increases with the proximate volatile matter of the coal for the same conditions. Therefore, a higher yield of volatiles from BioLime I, when equal quantities of BioLime samples are fired, is likely to result in higher NO reduction by homogeneous reactions. The weight loss and rate of weight loss curves for BioLime I and III in nitrogen atmosphere as a function of temperature are shown in Figures 1 and 2, respectively. Proximate analysis showed that BioLime I and III samples had ash yield of 22.8 and 50.7%, respectively, on a dry basis and, hence, the weight loss curves on an ash free basis were examined. The weight loss curves for both BioLime samples can be divided into several zones which relate to the decomposition of different components of the samples. Various weight loss zones arise because biomass mainly consists of cellulose, hemicelluloses, lignin, and minor quantities of organic extractives, and inorganic mineral species. Hemicellulose component which accounts for 25–35 wt. % in different woods decomposes between 200 and 260 °C. Cellulose, which constitutes 40–50 wt. %, starts to decompose between 240 and 350 °C to produce anhydrocellulose and levoglucosan. Lignin is a major component that accounts for about 16–33% depending of the wood is the last one to decompose, and the decomposition ranges between 280 and 500 °C. Pyrolysis of lignin yields phenols, methanol, acetic acid, and water. BioLime is a product of bio-oil and lime, the original components of biomass would be altered. Bio-oil therefore, is a complex mixture of water, guaiacols, catecols, syringols, furancarboxaldehydes, isoeugenol, pyrones, acetic acid, formic acid, and other carboxylic acids (20). The rate of weight loss curves show decomposition of various components at different temperatures. The distribution of moisture and volatile matter as a percent of total ash free weight loss during pyrolysis is shown in Table 2. Evaluation of the weight loss of the samples between 166 and 900 °C, indicates that approximately 83% of the volatile matter obtained by proximate analysis is released in that interval. Knowing that individual components play a significant role in determing the pyrolysis characteristics of BioLime than the basic structure or degree of polymerization, the zonal distribution of volatiles released can be used to compare the amounts of individual components present in the BioLime. Data in Table 2 show that for BioLime I, products decomposing between 166 and 400 °C had higher contributions to the volatiles released (1.8 times)
FIGURE 1. Pyrolysis profile (in nitrogen flow) for BioLime I.
FIGURE 2. Pyrolysis profile (in nitrogen flow) for BioLime III. than for BioLime III. Similarly, BioLime III had a higher contribution from the lignin derivatives (1.4 times). A higher weight loss in the temperature zone 600–900 °C from secondary thermal cracking of char for BioLime III indicates that the higher lignin content present in the BioLime forms more char during pyrolysis. A higher sum of percent weight loss (1.6 times) for Zones II and III for BioLime I (Table 3) shows that more volatiles are released by primary pyrolysis from BioLime with the lower lignin content. The 13C NMR spectra for BioLime I and III are shown in Figures SI.1 and Figure SI.2, respectively, in the Supporting Information. The signal assignments for the chemical shifts in the spectra based on the values found in the literature (21–28) with relative intensities of the peaks are summarized in Table S1 (Supporting Information). These results show significant concentrations of oxygenated compounds which is also evident from the compositional analysis. A relative measure of lignin aromatic carbon content can be obtained by integrating from approximately 160 ppm to the low shielding side of the C-1 peak (approximately 109
ppm). A similar method to measure the lignin aromatic carbon has been successfully used by Haw et al. (29) for wood samples. The assumptions and definitions used in the calculations are shown in the Supporting Information. Intensity in the region from 160 to 109 ppm was calculated for both BioLime I and III, and is defined as IAromatic. The Ilig and weight % of lignin calculated indicate that Ilig for BioLimes 1 and III are 30.1 and 50.1%, respectively. Calculations also showed that weight % lignin for BioLimes I and III is 23.9 and 42%, respectively. These data show that BioLime III has 1.8 times higher lignin content than BioLime I. A higher lignin content for BioLime III would result in more char formation and a lower yield of gases from primary pyrolysis as seen in the TGA analysis. Since BioLime is flash pyrolyzed on injection into the combustor, primary pyrolysis is thus predominantly responsible for the release of volatiles, which then react with NO, reducing it to N2. BioLime with lower lignin content would, therefore, release more volatiles on flash pyrolysis, and result in a higher NO reduction. VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2511
TABLE 2. Distribution of Volatiles Released during Pyrolysis of BioLime I and III on Ash Free Basis zones
BioLime I (wt. %)
BioLime III (wt. %)
Zone I (∼100 °C) moisture Zone II (166–400 °C) evaporation of lighter and heavier fractions Zone III (400 –600 °C) components from lignin Zone IV (600–900 °C) secondary thermal cracking of char total volatiles released in Zone II and III total volatiles released (Zone II + Zone III + Zone IV)
8.4
5.8
36.0
19.1
24.3
34.3
31.4
40.8
60.2
53.4
91.6
94.2
FIGURE 3. Reduction in SO2 with BioLime I as a function of inlet SO2 concentration.
TABLE 3. Composition of Pyrolysis Gases in Volume % for BioLime I and III at 1100 °C gases hydrogen (H2) carbon dioxide (CO2) carbon monoxide (CO) methane (CH4) ethane (C2H6) ethylene (C2H4) acetylene (C2H2) total
BioLime I (Vol. %) BioLime III (Vol. %) 16.2 60.9 14.0 5.3 0.5 2.8 0.3 100
22.0 54.0 13.3 6.0 0.4 4.0 0.3 100
The composition of pyrolysis products collected by pyrolyzing BioLime samples at a peak temperature of 1100 °C as analyzed by the GC is shown in volume % in Table 3. Data showed that the total yield and the yield of individual gases were 1.7 times higher for BioLime I than for BioLime III. BioLime III produced 1.4 times higher amount of char than BioLime I. These results from flash pyrolysis were consistent with the trends observed in the TGA data (Table 2). BioLime III had higher calcium content (14 wt. % vs 7 wt. % for BioLime I). The influence of cations on pyrolysis products in lignite pyrolysis is documented in the literature. The presence of ion exchangeable cations such as Ca and Mg markedly reduced the volatile matter yield during pyrolysis (30). Calcium is a good catalyst for promoting secondary char forming reactions and therefore reduces the volatile yield. Otake and Walker (31) reported that cations associated with carboxyl groups on lignites affect the composition of the gas during pyrolysis. Amounts of CO2 and H2 released were found to increase during pyrolysis. For BioLime III, the yield of H2 was higher than BioLime I, but the yield of CO2 was lower. The heating rates employed by Otake and Walker (31) was 5 °C min-1, whereas Morgan and Scaroni (30) used a high heating rates in an entrained flow reactor. The wt. % of the total gas yield from flash pyrolysis for BioLime I and III (34.2 and 20.0%, respectively) closely matches the weight loss in Zone II and III (37.6 and 23.4%, respectively) of the TGA analysis data (Table 2). This suggests that primary pyrolysis may be primarily responsible for the volatiles evolved during the flash pyrolysis. The wt. % of gas and char yield based on wt. % of lignin calculated for BioLime I and III showed that biomass of lower lignin content produces a lighter pyrolysis product, which may be considered a better bio-oil for use as a reburn fuel. Higher yields of pyrolysis 2512
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
FIGURE 4. Reduction in SO2 with BioLime III as a function of inlet SO2 concentration. gases increase the NO reduction potential of the BioLime through homogeneous gas phase reactions. Detailed modeling of NOx reduction using BioLime is discussed by us elsewhere (16). Tests to evaluate SO2 and NOx reduction were performed in the DFC using flue gases generated from a natural gas flame doped with bottled SO2 and NOx. A total of 18 tests were performed (nine each with BioLime I and III). The inlet SO2 concentration (doped) ranged from 900 to 2400 ppm. Inlet NOx concentration was varied from 120 to 415 ppm. Figures 3 and 4 show the percent reduction of SO2 as a function of inlet SO2 concentration for BioLime I and III, respectively. The results show that the SO2 reduction ranged from 60 to 85% for BioLime I for Ca/S molar ratios ranging between 1 and 1.5. For every mole of sulfur fed (by doping) to the combustor, the feed rate of BioLime was adjusted such that 1–1.5 moles of Ca was fed. However, SO2 reduction was limited to 50–70% for BioLime III for Ca/S ratios ranging between 1 and 1.7. These results suggest that the average calcium utilization efficiency is lower for BioLime III. Figures 5 and 6 show NOx reduction with BioLime I and III, respectively. Different runs of BioLime III with varying atomization pressure in the combustor and varying stoichiometry were conducted. Higher atomization pressure resulted in increased momentums of the droplets thereby enhancing the mixing of reburn fuel with primary jet fuel and resulting in higher NOx reduction. This behavior follows the earlier studies conducted by Zarnescu and Pisupati (32), Pisupati et al. (33), and Simons et al. (34). It was observed that for BioLime I if the inlet NOx concentration is
