Energy Fuels 2010, 24, 1222–1233 Published on Web 02/04/2010
: DOI:10.1021/ef9010109
Contaminant Estimates and Removal in Product Gas from Biomass Steam Gasification Hong Cui,* Scott Q. Turn, Vheissu Keffer, Donald Evans, Thai Tran, and Michael Foley Hawaii Natural Energy Institute, University of Hawaii at Manoa, 1680 East-West Road, POST 109, Honolulu, Hawaii 96822 Received September 10, 2009. Revised Manuscript Received January 12, 2010
Permanent gas species, tar compounds, sulfur compounds, and ammonia produced from a bench-scale (∼1 kg/h) fluidized-bed biomass gasifier were analyzed. Two commercial Ni-based catalysts and one commercial ZnO sorbent were evaluated under varied conditions by quantifying contaminants from the reactor inlet and outlet with specific sampling and analysis methods. The Ni catalysts targeted tar destruction and ammonia reduction, and the ZnO sorbent was selected for sulfur compound removal. Tar components were identified by gas chromatography-mass spectrometry (GC-MS) and quantified by GC-flame ionization detector (FID). A total of 13 compounds (gC6) were identified in raw product gas, principally “lighter tar” species with an average concentration of 15.5 g m-3 (dry gas basis). For tar species that were not detected by GC, a gravimetric method was used to quantify the portion of “heavier tar” (5.3 g m-3 dry gas basis). These data are raw gas tar concentrations for the gasifier-operating conditions used for the remainder of the tests. The performance of two commercial Ni catalysts were evaluated by comparing the concentrations of both “lighter tar” and “heavier tar” after the raw gas passed through the tarreforming reactor. Concentrations of hydrogen sulfide (H2S), carbonyl sulfide (COS), and thiophene (C4H4S) in the raw, dry, product gas averaged 93, 1.7, and 2.2 ppmv, respectively. C4H4S and two additional sulfur compounds, benzothiophene and one unidentified compound (UN1), were found in the tar-trapping solution. Removal of sulfur compounds using the ZnO sorbent at varied temperatures and gas hourly space velocities (GHSVs) was investigated. The primary sulfur component, H2S, was reduced to less than 1 ppmv; COS was not reduced significantly; and C4H4S concentrations were not affected at all. The average NO and ammonia concentrations were determined to be 8.2 and 2662 ppmv in the dry gas, respectively. Both were successfully converted to permanent gas species by Ni catalysts.
temperature, gasifying agent, equivalence ratio, residence time, etc.2 Devi et al.3 classified tar compounds into five groups based on their solubility and condensability: gas chromatography (GC)-undetectable, heterocyclic, light aromatic, light polyaromatic, and heavy polyaromatic. Class 1 and class 5 tars could be the main reason for the condensation problems; class 2 tars are potential water contaminants because of their high aqueous solubility; and class 4 tars were hard to remove even after severe catalytic treatment. Similarly, Hasler and Nussbaumer4 defined the tars as particles, heavy tars, polycyclic aromatic hydrocarbons (PAHs), phenols, and water-soluble organic residue. The heavy tar was determined gravimetrically without determination of the chemical composition at the evaporation temperature of 155 °C. The determination of heavy tars was regarded as an excellent indicator for the assessment of the gas products for the internal combustion engines, even though more than 70% of the heavy tars were unidentified. On the basis of the final application and objective, the tar definition, classification method, and analysis may vary. Tar samples have been collected using an impinger train with an organic solvent as a trapping solution and solid-phase
1. Introduction Biomass gasification is an attractive technology that converts biomass to synthesis gas (syngas) for the production of renewable liquid transportation fuels. In general, the raw product gas contains minor and trace contaminants, including particles, tar, H2S, NH3, and halides, as well as other species. Among these, tar compounds and sulfur and nitrogen contaminants are the three most harmful species present in the raw product gas and are priorities for removal by downstream processing. Tar is the most undesirable component, in both internal combustion engines and chemical synthesis reactions. A general definition of tar is that of a complex mixture of condensable hydrocarbons. Other definitions of tar have been given based on the sampling and detection methods. On the basis of the biomass reaction regimes, Milne et al.1 classified tar products as “primary”, “secondary”, and “tertiary”. The primary tars are fragments of the original material (cellulose, hemicellulose, and lignin) and can react further to produce secondary and tertiary tars at the same or higher temperature. Therefore, the tar amount and species are mostly dependent upon the gasification fuels and operating conditions, such as *To whom correspondence should be addressed. Telephone: 808-9565397. Fax: 808-956-2344. E-mail:
[email protected]. (1) Milne, T. A.; Evans, R. J.; Abatzoglou, N. Biomass gasifier “tars”: Their nature, formation, destruction, and tolerance limits in energy conversion devices. Making a Business from Biomass in Energy, Environment, Chemicals, Fibers and Materials; Proceedings of the 3rd Biomass Conference of the Americas, Montreal, Quebec, Canada; Overend, R. P., Chornet, E., Eds.; Elsevier Science, Inc.: New York, Aug 24-29, 1997; Vol. 1, pp 729-738. r 2010 American Chemical Society
(2) Kinoshita, C. M.; Wang, Y.; Zhou, J. Tar formation under different biomass gasification conditions. J. Anal. Appl. Pyrolysis 1994, 29 (2), 169–181. (3) Devi, L.; Ptasinski, K. J.; Janssen, F. J. J. G.; van Paasen, S. V. B.; Bergman, P. C. A.; Kiel, J. H. A. Catalytic decomposition of biomass tars: Use of dolomite and untreated olivine. Renewable Energy 2005, 30 (4), 565–587. (4) Hasler, P.; Nussbaumer, T. Sampling and analysis of particles and tars from biomass gasifiers. Biomass Bioenergy 1999, 18 (1), 61–66.
1222
pubs.acs.org/EF
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
4,5
20
ppm reported by Proll et al. NH3 content in the gasification product gas depends upon the nitrogen content of the feedstock21 and the gasifier conditions. Zhou et al.16 reported a ∼58% decrease in NH3 content when the gasifier temperature was raised from 700 to 900 °C for sawdust gasification and an ∼80% decrease when the gasifier temperature was increased from 750 to 900 °C for Leucaena gasification. NH3 can be removed by scrubbing processes20 or catalytic elimination. The process of catalytic elimination of NH3 in biomass gasification has been studied by several groups.9,22,23 The Ni-based catalyst, which is active for tar reduction, can decompose the ammonia at high temperatures.22,24,25 NH3 is commonly determined using impinger sampling trains loaded with dilute acid solution to absorb the NH3, followed by analysis of the trapping solution using an ion-selective electrode (ISE). Hydrogen sulfide (H2S) is the primary sulfur species converted from sulfur present in the feedstock under biomass gasification conditions. Minor amounts of COS are also formed. These gaseous sulfur compounds deactivate most catalysts used in downstream units for tar cracking and the water-gas shift (WGS) reaction. Zinc oxide (ZnO) is the most commonly used sorbent for H2S removal because of its favorable sulfidation thermodynamics. The removal of H2S is accomplished by reacting ZnO and H2S to produce ZnS and water.26 Many efforts have been made on the characterization of reaction and process for packed beds of ZnO employed for H2S and COS removal27-31 but are limited in a steam-rich environment,31-33 such as biomass steam gasification or hydrogasification.
adsorption (SPA) technology. Tar compounds have also been measured online by a molecular-beam mass spectrometer (MBMS).6 The liquid recovered using the impinger method is analyzed for tar compounds by GC and gravimetric methods. GC analysis can provide rich information about tar composition and concentration but is not suitable for compounds with higher boiling points. Thus, the GC method is suitable for “lighter” tar compounds with boiling points matched to the temperature range of the instrument, and the gravimetric method is suited for “heavier” compounds with higher boiling points. Both methods have their limitation; therefore, it is better to combine them to characterize and evaluate the whole tar sample. For the fluidized bed gasifier,7 the total content of tar identifiable by GC is usually less than 20 g m-3 and the amount of heavy tar compounds is usually less than 3 g m-3, with a total tar content of ∼20 g/m-3. Catalytic tar removal from syngas is one of most effective processes and extensively reported in the literature. Nickel catalysts have been found to be very effective.8-14 Ammonia (NH3) is the most significant species from fuel nitrogen conversion in biomass gasification and is the precursor to NOx emissions in downstream burners, gas engines, or gas turbines.15 The formation of NH3 has been widely studied in biomass and coal gasification.16-19 In general, the concentration of NH3 in the gasification product gas ranges from 500 to 30 000 ppm reported by Zhou et al.16 and from 1100 to 1700 (5) Brage, C.; Yu, Q.; Chen, G.; Sjoestroem, K. Use of amino phase adsorbent for biomass tar sampling and separation. Fuel 1997, 76 (2), 137–142. (6) Carpenter, D. L.; Deutch, S. P.; French, R. J. Quantitative measurement of biomass gasifier tars using a molecular-beam mass spectrometer: Comparison with traditional impinger sampling. Energy Fuels 2007, 21 (5), 3036–3043. (7) Simell, P.; Stahlberg, P.; Kurkela, E.; Albrecht, J.; Deutsch, S.; Sjostrom, K. Provisional protocol for the sampling and anlaysis of tar and particulates in the gas from large-scale biomass gasifiers. Version 1998. Biomass Bioenergy 2000, 18 (1), 19–38. (8) Kimura, T.; Miyazawa, T.; Nishikawa, J.; Kado, S.; Okumura, K.; Miyao, T.; Naito, S.; Kunimori, K.; Tomishige, K. Development of Ni catalysts for tar removal by steam gasification of biomass. Appl. Catal., B 2006, 68 (3-4), 160–170. (9) Wang, W.; Padban, N.; Ye, Z.; Olofsson, G.; Andersson, A.; Bjerle, I. Catalytic hot gas cleaning of fuel gas from an air-blown pressurized fluidized-bed gasifier. Ind. Eng. Chem. Res. 2000, 39 (11), 4075–4081. (10) Simell, P. Catalytic hot gas cleaning of gasification gas. VTT Publ. 1997, 330, 1–68. (11) Aznar, M. P.; Caballero, M. A.; Gil, J.; Martin, J. A.; Corella, J. Commercial steam reforming catalysts to improve biomass gasification with steam-oxygen mixtures. 2. Catalytic tar removal. Ind. Eng. Chem. Res. 1998, 37 (7), 2668–2680. (12) Sato, K. Development of catalytic reforming of tar in biomass gasification processes. Baiomasu Kagaku Kaigi Happyo Ronbunshu 2008, 3, 12–13. (13) Dou, B.; Zhang, M.; Gao, J.; Shen, W.; Sha, X. High-temperature removal of NH3, organic sulfur, HCl, and tar component from coalderived gas. Ind. Eng. Chem. Res. 2002, 41 (17), 4195–4200. (14) Corella, J.; Toledo, J. M.; Aznar, M.-P. Improving the modeling of the kinetics of the catalytic tar elimination in biomass gasification. Ind. Eng. Chem. Res. 2002, 41 (14), 3351–3356. (15) Yu, Q. Z.; Brage, C.; Chen, G. X.; Sjoestroem, K. The fate of fuelnitrogen during gasification of biomass in a pressurized fluidized bed gasifier. Fuel 2006, 86 (4), 611–618. (16) Zhou, J.; Masutani, S. M.; Ishimura, D. M.; Turn, S. Q.; Kinoshita, C. M. Release of fuel-bound nitrogen during biomass gasification. Ind. Eng. Chem. Res. 2000, 39 (3), 626–634. (17) McKenzie, L. J.; Tian, F.-J.; Guo, X.; Li, C.-Z. NH3 and HCN formation during the gasification of three rank-ordered coals in steam and oxygen. Fuel 2008, 87 (7), 1102–1107. (18) McKenzie, L. J.; Tian, F.-J.; Li, C.-Z. NH3 formation and destruction during the gasification of coal in oxygen and steam. Environ. Sci. Technol. 2007, 41 (15), 5505–5509. (19) Torres, W.; Pansare, S. S.; Goodwin, J. G., Jr. Hot gas removal of tars, ammonia, and hydrogen sulfide from biomass gasification gas. Catal. Rev.-Sci. Eng. 2007, 49 (4), 407–456.
(20) Proell, T.; Siefert, I. G.; Friedl, A.; Hofbauer, H. Removal of NH3 from biomass gasification producer gas by water condensing in an organic solvent scrubber. Ind. Eng. Chem. Res. 2005, 44 (5), 1576–1584. (21) Turn, S. Q.; Kinoshita, C. M.; Ishimura, D. M.; Zhou, J. The fate of inorganic constituents of biomass in fluidized bed gasification. Fuel 1998, 77 (3), 135–146. (22) Mojtahedi, W.; Ylitalo, M.; Maunula, T.; Abbasian, J. Catalytic decomposition of ammonia in fuel gas produced in pilot-scale pressurized fluidized-bed gasifier. Fuel Process. Technol. 1995, 45 (3), 221–236. (23) Wang, W.; Padban, N.; Ye, Z.; Andersson, A.; Bjerle, I. Kinetics of ammonia decomposition in hot gas cleaning. Ind. Eng. Chem. Res. 1999, 38 (11), 4175–4182. (24) Nassos, S.; Svensson, E. E.; Boutonnet, M.; Jaeras, S. G. The influence of Ni load and support material on catalysts for the selective catalytic oxidation of ammonia in gasified biomass. Appl. Catal., B 2007, 74 (1-2), 92–102. (25) Corella, J.; Toledo, J. M.; Padilla, R. Catalytic hot gas cleaning with monoliths in biomass gasification in fluidized beds. 3. Their effectiveness for ammonia elimination. Ind. Eng. Chem. Res. 2005, 44 (7), 2036–2045. (26) Topsoe, H. Sulfur removal methods. In Handbook of Fuel Cells; Fundamentals, Technology and Applications; Vielstich, W., Ed.; John Wiley and Sons, Ltd.: New York, 2003; Vol. 3, Chapter 15, pp 177-189. (27) Sasaoka, E.; Taniguchi, K.; Uddin, M. A.; Hirano, S.; Kasaoka, S.; Sakata, Y. Characterization of reaction between ZnO and COS. Ind. Eng. Chem. Res. 1996, 35 (7), 2389–2394. (28) Sasaoka, E.; Taniguchi, K.; Hirano, S.; Uddin, M. A.; Kasaoka, S.; Sakata, Y. Catalytic activity of ZnS formed from desulfurization sorbent ZnO for conversion of COS to H2S. Ind. Eng. Chem. Res. 1995, 34 (4), 1102–1106. (29) Sasaoka, E.; Hirano, S.; Kasaoka, S.; Sakata, Y. Characterization of reaction between zinc oxide and hydrogen sulfide. Energy Fuels 1994, 8 (5), 1100–1105. (30) Li, L.; King, D. L. H2S removal with ZnO during fuel processing for PEM fuel cell applications. Catal. Today 2006, 116 (4), 537–541. (31) Kim, K.; Jeon, S. K.; Vo, C.; Park, C. S.; Norbeck, J. M. Removal of hydrogen sulfide from a steam-hydrogasifier product gas by zinc oxide sorbent. Ind. Eng. Chem. Res. 2007, 46 (18), 5848–5854. (32) Kwon, K. C.; Park, Y.; Gangwal, S. K.; Das, K. Reactivity of sorbents with hot hydrogen sulfide in the presence of moisture and hydrogen. Sep. Sci. Technol. 2003, 38 (12 and 13), 3289–3311. (33) Novochinskii, I. I.; Song, C.; Ma, X.; Liu, X.; Shore, L.; Lampert, J.; Farrauto, R. J. Low-temperature H2S removal from steam-containing gas mixtures with ZnO for fuel cell application. 1. ZnO particles and extrudates. Energy Fuels 2004, 18 (2), 576–583.
1223
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
The fluidized-bed gasifier has been described elsewhere.34 For convenience, the brief structure and operation parameters are described below. The gasifier reactor is constructed of 310 stainless-steel pipes, with a bed diameter of 89 mm and a freeboard diameter of 152 mm. The reactor is externally heated by four, 4 kW heaters. Pressure taps, thermocouples, and probe access ports are located along the height of the reactor. Fuel was fed to the reactor from a sealed fuel hopper via a variable speed metering screw. Nitrogen and steam were used as fluidizing agents for the tests. Nitrogen was added to the windbox below the distributor by a mass flow controller. Steam was directed to the windbox from a steam generator that received water from a calibrated precision metering pump. Nitrogen was used as an inert trace gas to permit the calculation of gas yields and control fluidization. The bed material consisted of alumina beads with diameters in the range of 210-420 μm. Flow exits the reactor and passes through a heated, silicon carbide filter, 60 mm in diameter, 490 mm in length, and with a pore size of 3 μm (Pall Process Filtration Corporation, Cortland, NY). Filter elements have been operated at temperatures ranging from 600 to 850 °C. At the exit of the filter, the product gas can be directed through a series of downstream vessels for contaminant removal or destruction. A provision is also made to divert the entire product gas stream or a fraction of it through a heat exchanger to remove condensate followed by disposal in a swirl burner. The diversion point is located upstream from the vessels and allows the gas flow rate through the downstream vessels to be varied. A sampling point is also located at the exit of the filter vessel to provide a slip stream of the product gas for characterization. The gas-conditioning system downstream of the heated filter consists of three vessels in series. A provision has been made to have all three in line during a test or to bypass individual units, with the method of operation depending upon the test objectives. The first vessel downstream of the filter is a 76 mm inner diameter, fixed bed of solid sorbent that can be operated at temperatures up to 800 °C. A practical lower temperature limit of 300 °C is mandated by the condensation point of tar compounds. At the exit (port 2) of the first vessel, a provision is made to direct a slip stream of the gas to a sampling system for characterization. The first vessel has been used to test removal of sulfur compounds, which are poisons for catalysts commonly used for tar and methane destruction, thus the location of the bed prior to the secondary tar-cracking vessel. After exiting the sorbent bed, the product gas stream passes through two preheaters to raise the gas temperature to ∼800 °C prior to entering the catalytic tar-cracker unit. The tar-cracker unit is a 76 mm inner diameter, fixed bed equipped with heating elements that allow for control of the bed temperature up to 850 °C. At the exit (port 3) of the tar-cracking unit, a provision is made to direct a slip stream of the gas to a sampling system for characterization. The product gas stream exiting from port 4 of the catalytic tar-cracking unit is directed to the WGS reactor. This is a 76 mm inner diameter, fixed bed of shift catalyst equipped with heating elements that allow for control of the bed temperature up to 600 °C. This provides the capability to upgrade the hydrogen content of the product gas by converting carbon monoxide present in the gas stream according to the WGS reaction. Gas exiting the third vessel passes through a heat exchanger to remove condensate and can then be directed to the flare for disposal or to a compressor for gas storage. Stored gas can be used for hydrogen purification studies using pressure swing adsorption (PSA) or other appropriate technologies using real syngas from biomass gasification, such as Fischer-Tropsch synthesis or other liquid fuel or chemical production. These units are available but not presented in Figure 1.
Table 1. Results of Analyses on the L. leucocephala Sample Obtained from the Fuel Lot Harvested in Waimanalo, HI Proximate Analysis (wt % Dry Basis) ash volatiles fixed carbon
0.87 83.10 16.03
Ultimate Analysis (wt % Dry Basis) C H O (by difference) N S
53.08 6.13 39.56 0.31 0.05
Element Analysis of Ash (wt % Dry Basis)a SiO2 Al2O3 TiO2 Fe2O3 CaO MgO Na2O K2O P2O5 SO3 Cl CO2 higher heating value (MJ/kg)
2.91 0.48 0.09 0.73 31.59 10.96 2.71 20.53 7.18 1.68 4.04 17.10 19.5
a
The ash was calcined at 600 °C prior to analysis.
An advanced gasification system requires an integrated unit for gas cleanup to provide qualified gas products for the downstream applications, such as the synthesis of chemicals and liquid fuels, internal combustion engine, or fuel-cell power generation. It is important, therefore, to estimate the level of these contaminants in the gas phase and then evaluate and improve the current existing recovery technologies and processes. In this paper, a commercial ZnO sorbent and two commercial tar-cracking catalysts (Ni-based) were employed for downstream processing of the gasifier product gas. The composition of gasification product gas was characterized by H2, CO, CH4, CO2, C2H4, C2H6, H2S, COS, C4H4S, NH3, NO, and tar compounds. The fates of tar, ammonia, and sulfur in the gas stream, before and after conditioning, were also reported. 2. Experimental Section 2.1. Fuel Preparation. Leucaena leucocephala, a leguminous tree grown for fuel and fodder in many parts of the tropics, was harvested from the Waimanalo Experiment Station of the University of Hawaii and used as fuel in the current gasification project. The trees were delimbed, and the bolt wood was chipped. Chips were dried in a forced ambient air drying bed until equilibrium moisture content was reached, about 10% dry basis. The chips were sampled and subjected to ultimate (C, H, O, N, S, and Cl), proximate (volatile, fixed carbon, and ash), heating value, and elemental ash analyses (Si, Al, Ti, Fe, Ca, Mg, Na, K, P, S, Cl, and CO2) by Hazen Research, Inc., Golden, CO, as shown in Table 1. The fuel was hammer-milled to pass a screen with a 3 mm hole diameter prior to gasification. 2.2. Biomass Gasification System. A schematic of the HNEI bench-scale biomass gasification system is shown in Figure 1. It consists of an augur screw feeder, fluidized-bed reactor, silicon carbide filter, H2S sorbent bed, tar cracker, and WGS reactor. (34) Turn, S.; Kinoshita, C.; Zhang, Z.; Ishimura, D.; Zhou, J. An experimental investigation of hydrogen production from biomass gasification. Int. J. Hydrogen Energy 1998, 23 (8), 641–648.
1224
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Figure 1. Schematic diagram of the bench-scale fluidized-bed biomass gasifier system. Table 2. Typical Gasifier Operating Conditions and Primary Process Parameters
Electronic signals from thermocouples, pressure transducers, online gas analyzers, and gas meters are processed using two, 32 channel multiplexer amplifiers (model SCXI-1100, National Instruments, Austin, TX) and a 12 bit, analog to digital converter board (National Instruments, Austin, TX) controlled by a personal computer. 2.3. Test Procedure and Conditions. In preparation for a test, bed material and fuel were weighed and placed in the reactor and fuel hopper, respectively. The fuel feed rate was also calibrated over a range of screw speeds prior to each test. The gasifier bed was preheated to the desired temperature using the external heaters, with air flowing through the reactor until a stable temperature profile was obtained. The gas flow was then switched from air to the desired mixture of steam and nitrogen. After the reactor was once again stabilized, the feeder was turned on and the test begun. Flow was directed through the bypass line to the flare during the initial period of feeding until the gasifier had attained a steady operating condition, normally about 2 h. When steady conditions were attained, flow was directed to downstream vessels as desired. Quantitative sampling, as described below, was performed when the system had attained a steady temperature and product gas distribution. At the conclusion of the test, fuel, fluidizing gases, and system heaters were shutdown simultaneously. After the system had cooled, the remaining fuel was removed from the feed hopper, char was recovered from the ceramic filter, and bed material was removed from the reactor. Each was weighed and sampled. A sample stream could be extracted from the exit of the hightemperature filter or any of the three downstream vessels for the analysis of tar (gC6) species and ammonia. Each required a different sampling technique, and only one sample was acquired at any given time. All of the analytes were separated from the gas stream using impingers filled with liquid-trapping solution designed to extract the target analyte from the gas stream and hold it in solution. The dry gas exiting the impingers passed through a vacuum pump, a coalescing filter, and a gas meter that indicated the gas flow rate and provided a record of the total gas volume drawn through the impinger set during the sampling period.
average reactor temperature (°C) fuel feeding rate (kg h-1) steam feed rate (kg h-1) steam/dry biomass ratio ceramic filter temperature (°C) sulfur sorbent bed (°C) cracker temperature (°C)
800 1.0 2.0 2.0 800 350-700 800-850
A summary of typical gasifier operating conditions is provided in Table 2. As described in Figure 1, the reactor is equipped with electric heaters to maintain temperature. An average of four thermocouples inserted in the bed above the fuel feeding point was calculated to be 800 °C. Average feed rates of fuel and steam were 1.0 and 2.0 kg/h, respectively, providing an average steam/ fuel ratio of 2.0. Nitrogen flowed to the reactor at a rate of 12 L min-1 to aid in fluidization. The high-temperature ceramic filter was operated at an average temperature of 800 °C during the test. The sulfur sorption bed was operated at the temperature range of 350-700 °C. The tar cracker was operated at 800 and 850 °C for two types of Ni catalysts, respectively. Table 3 summarizes the gasifier tests, sample inventory, and data availability. 2.4. Sampling and Analytical Methods. The process stream from the outlet of the gasifier or any of the downstream vessels could be quantitatively sampled to determine the permanent gas species, sulfur and nitrogen species, tar species, and ammonia in the product gas. The sampling program for a particular test depended upon the test objective. The bulk gas stream that was directed through the bypass line to the flare first passed through a water-cooled heat exchanger followed by an ice-cooled condensate separator and a coalescing filter. The cool, dry, gas stream was directed through a gas meter that indicated the gas flow rate and provided a record of the total gas volume over the duration of the test. A description of the sampling and analytical methods for each component follows. 2.4.1. Permanent Gas and NOx. A sample of the product gas exiting the hot gas filter or any of the three downstream vessels 1225
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Table 3. Summary of Samples Reported in This Papera sample number
080129
080206
080213
080228
080313
080318
080403
test objective
blank
ZnO
ZnO
ZnO
ZnO
blank
G91
duration of fuel feeding (h) permanent gas NO tar
5.71
5.38
6.95
6.83
5.95
6.03
þ þ 3 blank samples
þ þ 2 samples with 1 blank -
þ þ -
-
þ þ 3 samples with no blank -
þ þ 3 samples
NH3
þ þ 2 samples with no blank -
-
6 samples
sulfur gas sulfur in trapping solution
3 blank samples
þ 2 samples with no blank
þ 3 samples with no blank
þ 2 samples with 1 blank
þ 3 samples
þ -
a
080408
080411
080417
C11-NK
G91
6.22
ceramic ball 6.08
6.45
7.37
þ þ 2 samples with 1 blank 4 samples with 2 blank þ 2 samples with 1 blank
þ þ 2 samples with 1 blank 4 samples with 2 blank þ 2 samples with 1 blank
þ þ 2 samples with 1 blank
þ þ -
4 samples with 2 blank þ 2 samples with 1 blank
4 samples with 2 blank þ -
þ, data are available; -, data are unavailable.
primary standard (AccuStandard, New Haven, CT) was used to produce a calibration curve for 13 individual tar species: benzene, pyridine, toluene, guaiacol, phenol, indene, naphthalene, acenaphthylene, fluorine, phenanthrene, anthracene, fluoranthene, and pyrene. A solution of decane in acetone was prepared as an internal standard (ISD) and was used for evaluating results and computing a response factor for estimating concentrations of unidentified peaks in sample chromatograms. Each sample was analyzed twice: once as an undiluted sample and again as a sample diluted to 10% strength using acetone. This technique was used to ensure that all analytes were evaluated at concentrations that fell within the calibration range of 1-200 μg/mL. A typical GC-FID chromatogram of the tar present in the trapping solution is shown in Figure 2. 2.4.2.2. Gravimetric Analysis. A total of 10 mL of tar-trapping solution was pipetted into an evaporating dish and put into a fume hood at room temperature overnight. The following day, the dishes were placed on a sand-bath heating plate at 105-110 °C for 3 h and then cooled to room temperature in a desiccator for 3 h. The residue remaining in the evaporating dish was the gravimetric tar. Two parallel determinations were made for each sample. 2.4.3. Ammonia. Ammonia was sampled using a single impinger that employed dilute sulfuric acid (0.1 N) as a trapping solution. The impinger was placed in an ice bath. The sample flow through the impinger was maintained at ∼2 L/min using a vacuum pump, and a total of 40 L of dry gas was extracted for each sample. Two impingers in series were used in preliminary tests, and results showed that the first impinger captured >99% NH3 of the total. Thereafter, only a single impinger was used for NH3 collection. Similar capture efficiency has been reported elsewhere.35 After the test was completed, the trapping solution was recovered from each of the three impinger sets and analyzed using an ion-selective electrode (ISE) (part EW-27502-00, Cole Parmer Instrument Co., Vernon Hills, IL). A primary calibration standard was prepared by dissolving reagent-grade NH4Cl in 0.1 N H2SO4 solution. Lower concentration calibration standards were produce by serially diluting the primary standard. 2.4.4. Sulfur. Samples of the product gas could be extracted from the exit of the hot gas filter or any of the three downstream vessels. After the sample passed through a condenser and coalescing filter, it was directed to a GC (model 2014, Shimadzu, Kyoto, Japan) equipped with a capillary column (Rtx-1, 60 m 0.53 mm 7 μm, Restek Corporation) and sulfur chemiluminescence detector (SCD) (Sievers model 355, GE Analytical Instruments, Boulder, CO). Samples were injected into GC by
was directed to an online gas chromatograph (AutoSystem, Perkin-Elmer, Norwalk, CT), equipped with a 1.5 m 3 mm packed column (Carboxen 1000, Supelco, Bellefonte, PA) and a thermal conductivity detector (TCD). A sample was injected every 20 min by an automatic sampling valve, and the gas species H2, N2, CO, CH4, CO2, C2H4, and C2H6 in the gas sample were quantified. A sample gas stream was also directed to a three-channel, online, nondispersive infrared (NDIR) gas analyzer (model URAS 10E, Applied Automation/Hartmann and Braun, Bartlesville, OK), equipped to analyze CO, CO2, and CH4, and a continuous-flow, thermal conductivity detector (model CALDOS 5G, Applied Automation/Hartmann and Braun, Bartlesville, OK) which measured the product gas H2 concentration. An additional gas stream was directed to a chemiluminescent analyzer for the detection of oxides of nitrogen as NO (model 10-AR, Thermo Environmental Instruments, Franklin, MA). The online analyzers were calibrated prior to each test using certified zero and span gases. 2.4.2. Tar. Tar sampling and analysis methods are mostly modified from the International CEN/BT/TF143 Standard “Biomass Gasification;Tar and Particles in Product Gases; Sampling and Analysis” published in 2005. Tar species were sampled using a series of six impingers that employed 2-propanol (IPA) as a trapping solution in the first five impingers. Impingers 1, 2, and 4 were placed in a water bath maintained at 35-40 °C, and impingers 3, 5, and 6 were placed in an ice/rock salt bath maintained at -12 °C. The sample flow through the impinger set was maintained at ∼5 L/min using a vacuum pump, and a total of 120 L of dry gas was extracted for each sample. The trapping solution was recovered from each of the impinger sets and kept in refrigerated storage until analyzed. Both GC and gravimetric analyses (evaporation) were conducted to obtain complete information on product tar. GC provides information on “lighter tar”, named GC-detectable tar. GC may not detect some tar species with higher boiling points because of the instrument analysis parameters, such as the oven temperature and column. The evaporation or gravimetric method can provide supporting information. The gravimetric tar may include both GC-detectable and GCundetectable tar components but is expected to contain “heavier tar” species with higher boiling points than “lighter tar”. The two methods can thus be used to provide a more complete assessment of tar species characteristics and quantities. 2.4.2.1. Gas Chromatographic Analysis. The tar-trapping solution was analyzed using GC (Varian) equipped with a capillary column (Rtx-1, 60 m 0.53 mm 7 μm, Restek Corporation) and a mass spectrometer (2000 Mass Spec) and GC (AutoSystem, Perkin-Elmer, Norwalk, CT) equipped with the same capillary column and a flame ionization detector (FID). A 200 μg/mL
(35) Norton, G. A.; Brown, R. C. Wet chemical method for determining levels of ammonia in syngas from a biomass gasifier. Energy Fuels 2005, 19 (2), 618–624.
1226
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Figure 2. Typical GC-FID chromatogram for a tar sample collected from the product gas stream at the filter outlet.
an automatic sampling valve at 10 min intervals. Retention times of sulfur compound peaks were identified prior to testing using a permeation tube gas generator (model CO395, Kin-Tek Laboratories, La Marque, TX) with tubes for COS, thiophene, benzonthiophene, and other appropriate sulfur species. The GC sulfur response was calibrated prior to each test using a standard containing 100 ppm H2S in N2. After separation on the capillary column, all sulfur species are converted to SO2 in a burner system; thus, the SCD produces an equimolar response to sulfur. Provided that all sulfur compounds contain only a single S atom, the peak area generated in proportion to the H2S standard can be applied to other sulfur compounds as well. The trapping solution recovered from tar-sampling impingers was also analyzed offline by GC-SCD. 2.4.5. Water. The tar-trapping solutions were also used to determine the water content of the product gas. Samples of the trapping solution recovered from the impingers were analyzed for water using a Karl Fischer titrator (model DL31, MettlerToledo, Inc., Columbus, OH). The water content of the impinger trapping solution was related to the gas volume drawn through the impingers to calculate the partial pressure of water in the product gas.
Table 4. Permanent Gas Composition in Product Gas Exiting the Hot Gas Filter by GC Analysis (vol %, Dry, N2-Free Gas) test ID T080129 average standard deviation T080206 average standard deviation T080213 average standard deviation T080228 average standard deviation T080313 average standard deviation T080318 average standard deviation T080403 average standard deviation T080408 average standard deviation T080411 average standard deviation T080417 average standard deviation a
3. Results and Discussion
CO 12.9 0.3 12.7 0.3 14.1 1.9 10.4 1.5 13.5 2.1 11.8 4.3 13.9 0.6 12.4 0.4 12.3 0.6 13.0 1.0
CH4 CO2 C2H4 C2H6 6.0 0.4 6.0 0.4 8.6 0.9 5.8 0.6 7.4 0.9 7.1 0.9 7.5 0.6 6.8 0.5 6.7 0.4 7.1 0.8
30.2 1.2 30.3 1.2 28.7 1.3 22.6 0.1 26.5 0.7 28.8 2.0 27.7 0.2 26.7 0.2 27.2 1.0 25.3 4.5
1.0 0.1 0.7 0.1 0.8 0.3 0.6 0.5 1.4 0.6 1.4 0.5 1.1 0.3 1.0 0.3 1.0 0.3 1.1 0.5
nda nd nd nd nd nd 0.0 0.0 0.0 0.0 0.1 0.0 0.1 0.0 0.1 0.0 0.1 0.0 0.1 0.0
nd = not detected.
Because of the WGS reaction, the H2 and CO2 concentrations are increased with a concomitant decrease in the CO concentration. CH4 and the other hydrocarbons can also react with steam to form additional H2. Gas composition measurements made with the continuous, online gas analyzers are presented and compared against the GC measurements in Figures 3 and 4. The continuous gas measurements are presented from the commencement of fuel feeding until the end of the test and include periods when the flow from the gasifier was diverted through a catalytic tar-cracking unit from ∼11:40 to 14:30. Agreement between the two measurement methods is generally good. Note that both data were not corrected to a N2-free basis. Peaks in the CO2 trace are the result of pressure changes because of loading of a filter in the sampling system line (first peak) and as a result of pressure fluctuations when flow is initially directed to the tar-reforming reactor (second peak). Differences between the online analyzer and GC measurements are largely due to calibration, with the former using a single-point method, whereas the GC is calibrated with a multi-point approach. The online
3.1. Characterization of the Raw Product Gas Stream. The results of analysis of the raw product gas from the fluidized bed gasifier are presented below. 3.1.1. Permanent Gas Species. H2, CO, N2, CH4, CO2, and minor amounts of light hydrocarbons are the primary products in biomass gasification. The permanent gas species analyzed by GC are presented in Table 4. The gas composition exiting the hot filter unit was in the range of 47.8-60.6% H2, 10.4-14.1% CO, 5.8-8.6% CH4, 22.6-30.3% CO2, 0.6-1.4% C2H4, and 0.0-0.1% C2H6. Gas yields of 10 tests under the similar reactor conditions averaged 1.34 ( 0.17 m3 of dry gas [standard temperature and pressure (STP)] kg-1 of biomass. The effects of operating conditions, such as reactor temperature, equivalence ratio, and ratio of fuel/steam have been investigated in the previous work.34 Under the current conditions, the product gas has typically a higher H2 concentration and lower CO and CH4 concentrations. The increase of the H2 concentration can result from a higher ratio of steam/fuel (S/F = 2) and the WGS reaction (eq 1). CO þ H2 O f CO2 þ H2
H2 49.8 1.9 49.8 1.9 47.8 2.5 60.6 1.7 51.2 2.9 50.8 3.7 49.8 1.4 53.0 1.2 52.7 1.0 53.5 4.4
ð1Þ 1227
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Figure 3. Comparison of the product gas CO2 and CH4 measurements by continuous gas analyzers and GC. Sample points are (1) at the exit of the hot gas filter from 09:00 to 11:40 and from 14:30 to 15:06 and (2) at the exit of the tar-cracking unit from 11:40 to 14:30.
Figure 4. Comparison of the product gas CO and H2 measurements by continuous gas analyzers and GC. Sample points are (1) at the exit of the hot gas filter from 09:00 to 11:40 and from 14:30 to 15:06 and (2) at the exit of the tar-cracking unit from 11:40 to 14:30.
measurements are required for monitoring and control of the system. 3.1.2. Tar. Because it has a lower boiling point (80 °C) than other aromatics and may not be a problematic compound for clogging, benzene could be regarded as a separate compound and excluded from the definition of tar.7 However, benzene was the major component found in tar from the downdraft gasifier when this was analyzed by direct mass spectrometry.1 The assumption that secondary tar (lighter tar) was represented by benzene has been used in a wood gasifier model, a wood particle decomposition model,36 and a coal
particle model.37 In this paper, benzene was considered as a main component of tar and included in the GC-detectable tar. Table 5 presents the tar analysis data for seven individual samples taken at the exit of the hot gas filter over the course of 5 test days. The GC analysis shows that the average total tar mass concentration in the product gas (identified and unidentified species) was 15.5 ( 3.2 g m-3 dry gas and identified compounds accounted for >89% of the total. Benzene, naphthalene, and toluene are in greatest abundance and account for ∼80% of the GC-detectable tar total, with
(36) Peters, B.; Bruch, C. A flexible and stable numerical method for simulating the thermal decomposition of wood particles. Chemosphere 2001, 42 (5-7), 481–490.
(37) Gurgel Veras, C. A.; Saastamoinen, J.; Carvalho, J. A., Jr.; Aho, M. Overlapping of the devolatilization and char combustion stages in the burning of coal particles. Combust. Flame 1998, 116 (4), 567–579.
1228
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Table 5. Summary of GC-Detectable and Gravimetric Tars in Product Gas Exiting the Hot Gas Filter [g m-3 of Dry Gas (STP)] sample
080129 set 1
080129 set 2
080129 set 3
benzene pyridine toluene guaiacol phenol indene naphthalene acenaphthylene fluorene phenanthrene anthracene fluoranthene pyrene unidentified total percent identified
8.75 0.18 1.27 0.4 0.07 0.08 1.93 0.11 0.01 0.16 0.01 0.02 0.03 1.67 14.69 89
8.04 0.13 0.89 1.35 0.12 0.12 1.41 0.01 nd 0.06 nd nd 0.01 1.98 14.12 86
6.87 0.09 0.67 0.48 nd 0.11 0.92 0.29 0.19 0.72 1.24 0.22 0 2.79 14.59 81
5.62
7.8
3.8
total
080228 set 1
080403 set 1
GC-Detectable Tar 11.01 10.27 0.24 0.18 2.04 0.84 nd 0.05 0.29 0.09 0.06 0.03 5.44 2.42 0.33 0.11 0.08 0.03 0.36 0.16 0.06 0.01 0.08 0.02 0.1 0.03 1.76 1.4 21.85 15.64 92 91 Gravimetric Tar 5.64
5.78
080408 set 1
080411 set 1
average
standard deviation
11.17 0.18 0.93 0.1 0.08 0.02 2.33 0.1 0.02 0.15 0.01 0.02 0.03 1.26 16.4 92
8.21 0.09 0.34 0.06 nd 0.03 1.85 nd nd nd nd nd nd 0.86 11.44 92
9.19 0.16 1.00 0.35 0.09 0.06 2.33 0.14 0.05 0.23 0.19 0.05 0.03 1.67 15.53 89
1.65 0.06 0.54 0.48 0.10 0.04 1.47 0.13 0.07 0.24 0.46 0.08 0.03 0.61 3.19 4
4.23
3.89
5.25
1.42
average concentrations of 9.19, 2.33, and 1.00 g m-3 dry gas, respectively. Benzene contributed almost 60% of the total GC-detectable tar. Benzene and naphthalene were also reported as the dominant species under the air-blown gasification,2 and their concentrations varied with gasification parameters, including temperature, equivalence ratio, and residence time. In comparison to steam gasification, it seems that air-blown gasification produced a greater tar yield with a wider variety of species ranging from 15 to 86 g m-3 with 23 identified tar species. In the steam gasification, more steam should favor the tar destruction according the reactions: tar þ H2 O f nCO þ mH2
ð2Þ
The gravimetric tar concentration in the product gas was 5.25 ( 1.42 g m-3 on average, less than half that of the GCdetectable tar, indicating that “heavier tar” was formed to a lesser degree in fluidized-bed gasification than “lighter tar”. Note that the gravimetric and GC-detectable tar species are not mutually exclusive and each provides an index for partially characterizing the tar. 3.1.3. Sulfur-Containing Gas Species. Three sulfur species, hydrogen sulfide (H2S), carbonyl sulfide (COS), and thiophene (C4H4S), were identified in the product gas, as shown in Figure 5. Figure 6 shows the online concentrations of the three sulfur compounds released during test periods on different days. H2S is the dominant sulfur compound present in the gas phase, and COS and C4H4S are present at lower concentrations. Averaging these measurements and converting to a dry N2-free basis, H2S, COS, and C4H4S were present in concentrations of 93.3 ( 13.5, 1.7 ( 0.2, and 2.2 ( 0.5 ppmv, respectively. It is also found that H2S and COS concentrations were more stable than thiophene over the test period. In the initial period after fuel feeding began, thiophene had a higher concentration and required a longer time to reach steady conditions. Sulfur analysis in the tar-trapping solution was also conducted by injecting liquid samples into the GC-SCD. Three sulfur compounds were detected, and two were identified as thiophene and benzothiophene, as shown in Table 6. Thiophene was the dominant sulfur compound, with an average concentration of 5.58 ( 1.18 mg of S m-3, which was equal to 3.91 ( 0.83 ppmv. This value was comparable to the
Figure 5. Sulfur species identified by GC-SCD in the gas stream.
2.2 ( 0.5 ppmv of thiophene measured in the gas stream. The current results show that sulfur species with higher boiling points, such as benzothiophene, are removed in the gas conditioning train and will not be detected as permanent gas species. Liquid streams generated as part of the gas condition should be analyzed for the presence of sulfur compounds for completeness. 3.1.4. NH3 and NO. A total of 14 ammonia measurements were made at the outlet to the hot filter in varied tests. The average and standard deviation were 2662 ( 484 ppmv NH3 in the dry product gas, as shown in Table 7. The NO concentration in the raw gas stream was made with the continuous, online gas analyzer, which was characterized by an initial, 2 h, transient period after startup, followed by relatively constant values that averaged 8.3 ppmv. 3.2. Hot Gas Cleanup Activities. 3.2.1. Tar Contaminants Removal. Table 8 summarizes test results of tar destruction tests using a fixed-bed reactor containing inert alumina silicate, 12 mm spheres, and two commercial nickel-based catalysts, C11-NK and G91 from S€ ud Chemie. All tests were conducted in a temperature window of 800-850 °C and at two flow rates. For GC detectable tar, inlet concentrations of 12-16 g of total tar m-3 of dry gas were converted by C11-NK to levels less than 0.1 g of total tar m-3 of dry gas. G91 did not convert the tar as effectively and the outlet gas from the reactor 1229
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Figure 7 presents the effect of catalysts and gas hourly space velocity (GHSV) on the tar yields. The differences in conversion may be due in part to differences between the catalysts, but GHSV (m3 kg-1 h-1) also contributes. GHSV values for each catalyst condition are indicated on the figure, and the values for C11-NK are nearly twice those of G91. The low levels of tar measured in the outlet from the C11-NK tests (T080411) contained acenaphthylene, pyrene, and unidentified compounds, as shown specifically in Table 8, and the species present in greatest concentrations at the inlet, benzene, naphthalene, and toluene, were not detected at the outlet. The concentration of methane present in the dry fuel gas under steady gasifier operating conditions was ∼4%, as shown in Figure 3. Methane can also be catalytically converted by the catalytic steam reaction. Figure 3 shows that methane was completely converted (from 11:40 to 14:30) in the product gas and that the conversion of methane and tar species resulted in an increase in permanent gas species, CO and H2. 3.2.2. Sulfur Contaminants Removal. Because of low quantities of sulfur compounds (about 100 ppmv as H2S), ZnO was a suitable adsorbent for sulfur removal. The process is a common practice in the petrochemical industry and occurs according to the reaction ZnO þ H2 S f ZnS þ H2 O
ð3Þ
A commercial sorbent (ZnO) was acquired and used in a series of tests in the gasifier facility. A 200 cm3 sample was placed in the sulfur sorbent bed, and tests were conducted to determine the effects of GHSV and sorbent temperature. Figure 8 presents the results of tests conducted at a GHSV of 8036 h-1 and over a range of temperatures. The data indicate that, at temperatures of 350 and 400 °C, the ∼90 ppmv H2S in the inlet stream to the reactor is nearly completely removed, with bed outlet levels of ∼0.1 ppmv. The data also show that the concentration of COS is also reduced under these conditions, decreasing roughly 1 order of magnitude from 1 to 0.1 ppmv. Sorbent temperatures of 450 °C and higher resulted in greater concentrations of H2S and COS in the outlet stream. Thiophene (C4H4S) present in the inlet gas stream is largely unaffected by exposure to the sorbent over the range of temperatures tested. The effects of GHSV on H2S are summarized in Figure 9 for a series of tests conducted at a sorbent temperature of 400 °C. The flow rate through the reactor was varied while maintaining the sorbent bed temperature. H2S removal is most clearly affected by changes in GHSV, with decreasing removal resulting from increased flow rates. Similar results were reported by Kim et al.31 They noted that higher space velocities lead to more irregular gas-solid contact with reduced contact time, thereby limiting H2S removal by sorbent particles. COS shows a lesser effect, with all flow rates resulting in the 1.5 ppmv COS inlet concentration being reduced to ∼0.4 ppmv. Thiophene (C4H4S) concentrations are largely unaffected. The fate of sulfur compounds was investigated in this paper when the tar-cracking catalysts were used for tar removal, as shown in Figure 10. The concentrations of H2S, COS, and C4H4S measured at the inlet of the tarcracking bed were around 100, 1.8, and 2 ppmv, respectively.
Figure 6. Sulfur species concentration detected in gasification product gas (a, H2S; b, COS; c, C4H4S; without N2 deduction).
contained tar in the range of 0.92-2.5 g of total tar m-3 of dry gas. As a check to determine the effect of thermal cracking without catalyst, the bed was filled with 12 mm diameter, alumina silicate balls. The data indicate little effect on the tar yield but were still below the range of the data from measurements at the outlet of the filter. Similar to GCdetectable tar, C11-NK converted gravimetric tar present at 3.89 g of tar m-3 of dry gas to levels of 0.75-0.89 g m-3. Although the test conditions were not identical, G91 reduced the gravimetric tar concentration but did not perform as well as C11-NK, reducing 5.78 g of tar m-3 of dry gas to levels ranging from 0.73 to 2.14 g of tar m-3. The results also show there was no significant reduction of tar in the test conducted with the reactor packed with ceramic balls. Tar reductions in the catalyst tests were thus clearly attributed to the catalytic activity rather than thermal cracking. 1230
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Table 6. Sulfur Compounds in the Tar-Trapping Solution Samples (mg of S m-3 of Dry, N2-Free Product Gas) 080129
080228
080403
080408
080411
sulfur compounds
set 1
set 2
set 3
set 1
set 1
set 1
set 1
averagea
standard deviation
thiophene UN1 benzothiophene
6.63 1.04 1.16
5.13 1.91 0.49
1.78 1.15 0.31
5.19 nd 1.06
5.15 0.58 0.43
7.32 1.04 1.38
4.08 0.53 0.89
5.58 0.85 0.90
1.18 0.65 0.38
a
The average value excluded the data of sample 080129 set 3.
Table 7. NH3 Concentration in the Gasification Product Gas (ppmv, Dry and N2-Free Basis) 080318
080403
080408
080411
080417
test ID
set 1
set 2
set 3
set 4
set 5
set 6
set 1
set 2
set 1
set 2
set 1
set 2
set 1
set 2
average
NH3
2519
2465
2406
2134
2166
3213
2477
2647
2169
2451
2713
2709
3585
3617
2662 ( 484
Table 8. GC-Detectable and Gravimetric Tars in Product Gas at the Inlet and Outlet of the Tar Cracker [g m-3 (STP)]a sample 3
-1
GHSV (m kg
T080408, ceramic ball
T080411, C11-NK
-1
h )
1.06
850
850
2.61
5.22
inlet
800
800
inlet
850
benzene pyridine toluene guaiacol phenol indene naphthalene acenaphthylene fluorene phenanthrene anthracene fluoranthene pyrene unidentified total percent identified
11.17 0.18 0.93 0.1 0.08 0.02 2.33 0.1 0.02 0.15 0.01 0.02 0.03 1.26 16.4 92
10.91 0.13 0.42 0.05 0.02 0.12 1.64 0.03 nd 0.08 nd nd 0.01 1.25 14.66 91
GC-Detectable Tar 8.66 8.21 0.09 0.09 0.32 0.34 0.17 0.06 nd nd 0.03 0.03 1.12 1.85 nd nd nd nd 0.02 nd nd nd nd nd nd nd 1.28 0.86 11.69 11.44 89 92
nd 0.01 nd nd nd nd nd 0.01 nd nd 0.01 nd 0.01 0.08 0.12 33
nd nd nd nd nd nd nd 0.01 nd nd nd nd 0.01 0.07 0.09 22
10.27 0.18 0.84 0.05 0.09 0.03 2.42 0.11 0.03 0.16 0.01 0.02 0.03 1.4 15.64 91
0.46 nd 0.02 nd nd nd 0.03 nd nd nd nd nd nd 0.41 0.92 55
1.88 nd 0.06 nd nd nd 0.05 nd nd nd nd nd nd 0.51 2.5 80
4.23
3.63
Gravimetric Tar 3.84 3.89
0.89
0.75
5.78
0.73
2.14
a
inlet
0.53
cracker temperature (°C)
total
850
T080403, G91
nd = not detected.
H2S concentrations were not affected by the use of inert ceramic balls in the 850 °C bed but were adsorbed by the catalyst C11-NK, resulting in an exit gas concentration of ∼1 ppmv. The catalyst type and GHSV (m3 kg-1 h-1) affected the removal, with lower exit gas concentrations corresponding to lower values of GHSV and the use of catalyst C11-NK. There is no apparent explanation for the increase of the COS concentration when the raw gas passed through the inert ceramic balls in the 850 °C bed. The C11-NK catalyst used at low GHSV produced a lowering of the COS concentration to 0.03 ppmv. The concentration of thiophene (C4H4S) was completely removed by both Ni catalysts at the low flow rate through the catalyst beds. C11-NK produced lower concentrations than G91 at higher GHSV. The ceramic balls had little effect on thiophene (C4H4S) concentrations. Lower concentrations of sulfur compounds were found in the tar-trapping solutions sampled at the outlet of the Ni catalyst bed. The GC-SCD profile is shown in Figure 11 for those samples collected from the outlet of the tar cracker loaded with the G91 catalyst. The 080403 set 1 data present the sulfur compounds detectable by GC-SCD at the inlet of the tar cracker, and 080403 set 2 and 080403 set 3 data
represent the detected sulfur compounds at the outlet of the tar cracker with GHSV of 2.61 and 5.22 m3 kg-1 h-1. Results show that no sulfur compounds (thiophene, UN1, or benzothiophenes) were detected in the trapping solutions when the G91 catalyst was used for tar cracking. It was expected that sulfur adsorption occurred on the nickel catalysts in different chemical states depending upon the process conditions.38,39 Sulfur compounds occupying active sites on the nickel would result in deactivation of catalysts. 3.2.3. Nitrogen Contaminants Removal. NH3 removal methods were not the main focus of the current work, because it has been shown that Ni-based tar-cracking catalysts are effective in the destruction of ammonia, as described by the reaction 2NH3 f N2 þ 3H2
ð4Þ
(38) Hepola, J.; Simell, P. Sulfur poisoning of nickel-based hot gas catalysts in synthetic gasification gas. Stud. Surf. Sci. Catal. 1997, 111 (Catalyst Deactivation 1997), 471–478. (39) Hepola, J.; Simell, P. Sulfur poisoning of nickel-based hot gas cleaning catalysts in synthetic gasification gas. I. Effect of different process parameters. Appl. Catal., B 1997, 14 (3-4), 287–303.
1231
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
Figure 9. Efficiency of sulfur compound removal using a ZnO sorbent bed at 400 °C as a function of the wet GHSV.
Figure 7. Effect of catalysts and GHSV values on tar removal (a, GC-detectable tar; b, gravimetric tar).
Figure 8. Efficiency of sulfur compound removal using a ZnO sorbent bed as a function of the temperature at GHSV = 8036 h-1.
Figure 12 summarizes the results of the destruction of ammonia by exposure to the Ni-based tar-cracking catalysts, C11-NK and G91. Ammonia concentrations measured at the inlet of the tar-cracking bed were in the range of 2300-2700 ppmv. Ammonia concentrations were not affected by the use of inert ceramic balls in the 850 °C bed but were reduced by the Ni catalysts to values ranging from 36 to 1100 ppmv. Effects of catalyst type and GHSV (m3 kg-1 h-1) were observed, with lower concentrations corresponding to lower values of GHSV and catalyst C11-NK providing better performance. The concentration of oxides of nitrogen present in the dry fuel gas (reported as NO) under steady gasifier operating conditions was ∼6 ppmv (8.3 ppmv on a dry N2-free basis) at
Figure 10. Effect of tar-cracking catalysts on the concentrations of sulfur compounds in the product gas (a, H2S; b, COS; c, C4H4S).
the inlet of the tar cracker, dropping to less than 1 ppmv when either catalyst was used. 1232
Energy Fuels 2010, 24, 1222–1233
: DOI:10.1021/ef9010109
Cui et al.
C4H4S) species, ammonia, and NOx. One commercial ZnO sorbent and two commercial Ni-based tar-cracking catalysts were evaluated under varied conditions. The purpose of the sorbent is to remove sulfur compounds and the purpose of the catalysts is for tar destruction and ammonia reduction. Gas chromatographic and gravimetric methods were performed to evaluate tar speciation and yields. The gas chromatographic method identified 13 compounds in raw product gas, principally “lighter tar” species at an average concentration of 15.5 g m-3 dry gas. Gravimetric methods found “heavier tar” components present at 5.3 g m-3 of dry gas. Tar fractions identified by the two methods were not mutually exclusive but can be used to supplement one another. Concentrations of H2S, COS, and C4H4S averaged 93, 1.7, and 2.2 ppmv, respectively. Thiophene, benzothiophenes, and one unidentified compound were found in the tar-trapping solution. NO and ammonia concentrations were determined to be 8.2 and 2662 ppmv, respectively. A cleanup strategy was developed to remove contaminant species or convert them to a benign form. Tar and ammonia conversions were accomplished using nickel catalysts, and the ZnO sorbent was used for sulfur removal. It was found that methane (CH4), tar, and ammonia (NH3) were successfully converted to permanent gas species using nickel catalysts. Concentrations of sulfur compounds in the gas products were lowered by the Ni catalyst bed, indicating that sulfur compounds were adsorbed by the Ni active sites, a source of catalyst deactivation. Removal using a commercial ZnO sorbent at varied temperatures and GHSVs were investigated under a steam-rich environment. Under these conditions, the primary sulfur component, H2S, was reduced to less than 1 ppmv and COS was not reduced significantly. Thiophene (C4H4S) concentrations were not affected. Application of commercial catalysts and sorbents to biomass gasification remains a challenging task. Process conditions (gas compositions and contaminants including tar, sulfur species, steam content, and alkali metals) and scales differ from those of coal gasification or petrochemical upgrading. These features should be considered and merit further investigation.
Figure 11. Sulfur species identified by GC-SCD in the tar-trapping solution samples.
Figure 12. Efficiency of NH3 destruction by nickel catalysts G91 and C11-NK compared to inert ceramic balls at 850 °C.
4. Conclusions The objective of this project was to develop hot gas cleanup capabilities for the gasifier to support research on the production of hydrogen or syngas from biomass. The product gas stream at the outlet of the hot gas filter was characterized for concentrations of permanent gas species (H2, CO, CO2, CH4, C2H4, and C2H6) and contaminants including tar (principally, benzene, naphthalene, and toluene), sulfur (H2S, COS, and
Acknowledgment. This work was supported by the U.S. Department of Energy under Contract DE-FC36-04GO14248.
1233
