- Email: [email protected]
Ammonia yield from gasification of biomass and coal in fluidized bed reactor
Fuel 117 (2014) 917–925
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Ammonia yield from gasification of biomass and coal in fluidized bed reactor M. Jeremiáš a,b,⇑, M. Pohorˇely´ a,b, P. Bode c, S. Skoblia d, Z. Benˇo d, K. Svoboda a a
Institute of Chemical Process Fundamentals, Academy of Science of the Czech Republic, Rozvojová 135, 165 02 Praha 6-Suchdol, Czech Republic Department of Power Engineering, Institute of Chemical Technology Prague, Technická 5, 166 28 Praha 6, Czech Republic c University of Twente, Postbus 217, 7500 AE Enschede, The Netherlands d Department of Gas, Coke and Air Protection, Institute of Chemical Technology Prague, Technická 5, 166 28 Praha 5, Czech Republic b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
We investigated the conversion rate
of fuel-N into ammonia. We adjusted ammonia sampling and
analysis procedure. We measured the effect of limestone
and CO2 addition.
a r t i c l e
i n f o
Article history: Received 13 August 2013 Received in revised form 2 October 2013 Accepted 8 October 2013 Available online 26 October 2013 Keywords: NH3 Limestone Dolomite NOx Gasifying agent
a b s t r a c t Gas from the gasification of coal and biomass can be used for combined production of heat and power or for the synthesis of alternative fuels. In both cases, ammonia, which is created by the conversion of fuelbound nitrogen, can cause technological difficulties. Therefore, ammonia concentration in the produced gas has to be monitored, or, at least, estimated to take appropriate measures for its abatement. In this paper, we report in detail the analysis of ammonia content in the produced gas as well as results of several experiments of fluidized-bed gasification of Spanish coal, South African coal, Rheinish coal and woody biomass. We focused on the evaluation and comparison of ammonia yield based on the following variables: (1) different fuels, (2) the addition of dolomitic limestone into the fluidized bed and (3) the use of carbon dioxide instead of steam in the gasifying agent. The conversion rate of fuel nitrogen into gas was surprisingly high for the less-reactive coals and within presumed levels for Rheinish coal and biomass. The presence of dolomitic limestone in the fluidized bed substantially increased the conversion rate of fuel-nitrogen into ammonia. The use of different gasifying agents had an ambiguous effect on ammonia yield. Ó 2013 Elsevier Ltd. All rights reserved.
Abbreviations: CEM, controlled evaporation and mixing system; ER, equivalence ratio; IC, ion chromatography; LHV, lower heating value; MFC, mass flow controller; PTFE, polytetrafluoroethylene (Teflon). ⇑ Corresponding author at: Institute of Chemical Process Fundamentals, Academy of Science of the Czech Republic, Rozvojová 135, 165 02 Praha 6-Suchdol, Czech Republic. Tel.: +420 731 570 803; fax: +420 220 920 661. E-mail addresses: [email protected], [email protected] (M. Jeremiáš), [email protected] (M. Pohorˇely´). 0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2013.10.009
1. Introduction Biomass and coal can be processed into gaseous fuel via gasification in a fluidized bed reactor. The produced gas can be used as fuel for gas engines with internal combustion – in this case it can be called ‘producer gas’ – or as a raw material for synthesis of
918
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
various chemical products, such as methanol or Fischer-Tropsch fuels – in this case it may be referred to as ‘synthesis gas’. Gaseous nitrogenous compounds are regarded as an unwanted component of the gas. In the case of producer gas, they reduce the operating lifetime of oil in the internal combustion engine [1] and they can act as precursors of formation of nitrogen oxides [2,3]. For example, trials with 300–1120 ppmv NH3 showed that, during gas turbine combustion, 20–50% of ammonia is converted into NOx [4]. However, the conversion rate can vary in a relatively broad range according to experimental conditions [3]. In synthesis gas applications (such as methanol production or Fischer–Tropsch synthesis) ammonia can contribute to the formation of amines on the synthesis catalyst or it can poison the catalyst itself. The presence of amines in the final products (methanol, liquid fuels) is not permitted in the internationally accepted specifications and their removal is possible only with an ion exchanger [5]. Moreover, where chlorine is present (typically when gasifying coal) ammonia will combine with chlorides to form ammonium chloride at about 250–280 °C, which can form deposits on gas coolers or gas filters [5]. Thus, the concentration of nitrogen compounds in both the producer gas and synthesis gas has to be monitored and minimized. However, the nitrogenous-compounds-content is not usually the main issue regarding gas purity. Tar and soot removal, desulfurization and dehalogenation take priority over the removal of N-compounds; thus, the reduction of nitrogenous compounds is usually a by-product of some other means applied; for example, oil or water quench for tar removal, in which the nitrogenous compounds are absorbed into water phase [1,6]. Consequently, the nitrogenous compounds are usually not fully removed from the producer gas. Their more efficient removal can be achieved through installing another gas cleaning unit, but this may render smaller plants uneconomic [1]. Lower emissions can also be achieved by adjusting the operating parameters of the gasifier or of the gas cleaning devices. For this purpose, both theoretical and empirical knowledge of the behavior of nitrogenous compounds during gasification and gas cleaning stages is required. During gasification, fuel nitrogen is liberated mainly in the form of ammonia, cyanides, thiocyanates, molecular nitrogen, nitrogen oxides and various aromatic compounds, while a smaller part of the nitrogen is retained in solid char [7]. The yield of these compounds and their ratios depend on various parameters of the gasification process; such as, the type of gasifier, temperature, equivalence ratio (ER) and catalysts. In fluidized bed reactors, longer gas-char contact time results in higher NH3/HCN ratio [7]; consequently, NH3 is a vastly predominant nitrogenous compound in the resulting gas [8]. The yield of ammonia is influenced by the gasification process parameters whose effects are summarized in the literature [9– 11]. In this study, we want to contribute to the knowledge base of nitrogen behavior during gasification by presenting the data obtained during fluidized-bed gasification of coal and biomass at 850 °C at atmospheric pressure. We focused on evaluating and comparing the ammonia yield based on the three following variables: (1) different fuels, (2) the addition of dolomitic limestone into the fluidized bed and (3) the use of different gasifying agent mixtures (H2O + O2, CO2 + O2, CO2 + H2O + O2). 1.1. Ammonia yield from gasification of different types of coal and biomass The yield of nitrogenous compounds during gasification is dependent on the nitrogen forms in the fuel. Leppälathi and Koljonen [7] published a comprehensive review about the forms of nitrogen in coal, peat and wood. It is generally assumed that the majority of nitrogen in coals is in pyrrolic and pyridinic form, and, on the other hand, nitrogen in living organic tissues is bound
mainly in proteins, aminoacids and alkaloids. Consequently, in peat and wood gasification, most of fuel nitrogen is liberated during the pyrolysis stage and, in coal gasification at low temperatures (<1200 K), most of the nitrogen is retained in char after pyrolysis and released during the char gasification stage. Leppälathi and Koljonen also report the rate of fuel-N to NH3 conversion for different fuels during fluidized bed gasification of lignite, sub-bituminous and bituminous coals and wood by oxygen and steam. During the gasification of lignite and sub-bituminous coal at high pressures (0.8 and 1.5 MPa), approximately 60% of fuel-N was converted into NH3. Even, a conversion of 93% was observed with a highly volatile coal. With bituminous coal, the ammonia conversion was significantly lower (0.6%–19%). In wood gasification, the NH3 conversion was 72%–97% (at pressure 0.4– 1.0 MPa). However, lower gasification pressure can lead to a significantly lower NH3 yield. [7] In allothermal steam gasification of wood in the Guessing (FICFB) plant, up to 70% of the fuel nitrogen was found as NH3 in the producer gas [1]. The concentration of nitrogen in fuel can also affect its conversion into gas; with the increase of concentration of chemically bound nitrogen in fuel, its conversion rate decreased [3]. The innovation of this study is the comparison of the yield of ammonia during atmospheric fluidized-bed gasification of less reactive Spanish coal and South African coal, reactive Rheinish coal and woody biomass. To our best knowledge, such a comparison is missing in the available literature. 1.2. Dolomitic limestone in the fluidized bed The use of catalytically active dolomitic limestone during gasification promotes char conversion, changes product gas composition, and reduces tar yield [12]. Inevitably, it will also affect the yield of ammonia. The presence of a catalyst can affect both the creation of ammonia and its decomposition. The decomposition of ammonia into H2 and N2 occurs over a reforming catalyst at high temperatures (1073 K) [13]. For example, Simmel et al. [14] report a 100% conversion rate on a nickel catalyst and a 53% conversion rate on dolomite in a laboratory-scale fixed-bed tube reactor at temperature of 900–910 °C, residence time 0.2–0.3 s and pressure 1 bar (in simulated synthesis gas). On the other hand, Leppälahti et al. [15] stated that the limestone and dolomite did not have any catalytic effect in the decomposition of ammonia, but they reduced the nitrogen cyanide content in the producer gas [15]. Pinto et al. [11] tested the use of dolomite directly in fluidized bed. They report that dolomite did not change the concentration of NH3 in the producer gas from coal gasification, but it increased the conversion of fuel-N into NH3 (this discrepancy is caused by the increase of overall producer gas yield on dolomite) [11]. This indicates that a directly used catalyst enhances the formation of NH3 from fuel, probably by increasing the conversion rate of char and by increasing the concentration of hydrogen in the system. This enhancement probably prevails over ammonia-decomposition effect. In this article, we compare the yield of ammonia from the gasification of reactive Rheinish coal or biomass either on sand in the fluidized bed or on the mixture(s) of sand and dolomitic limestone in the fluidized bed. To our best knowledge, such a comparison of ammonia yield on different mixtures of sand and dolomitic limestone has not been published. 1.3. The use of different gasifying agents Gasification in steam generates large amounts of H radicals as an intermediate between the reactions of steam with char, drastically enhancing the formation of NH3 [16]. The use of CO2 as
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
gasifying agent lowers the partial pressure of steam in the reactor; thus, it limits the generation of H radicals. Therefore, the highest NH3 yield can be expected when the mixture O2 + H2O is used as gasifying agent. In this article, we try to compare the yield of ammonia from biomass on the following mixtures of gasifying agents: H2O + O2, CO2 + O2 and CO2 + H2O + O2. 2. Experimental 2.1. Fluidized bed gasifier The fluidized-bed reactor (see Fig. 1) is an electrically heated 2200 mm high tube with inner diameter of 51.1 mm in the lower section and 99.0 mm in the upper section (freeboard) made of high-temperature resistant stainless steel (AISI 446). The 24 mm tall size reduction area between the lower and higher sections is placed 545 mm above the grate. Its maximal operating temperature is 980 °C at reducing conditions. The preheated (up to 500 °C) gasification agent fluidizes the loose inert material in the lower section of the reactor. Electrical heating with maximal output of 10 kW consists of three independent heating elements along the height of the reactor. The heating elements can be regulated separately. The temperature inside the reactor is measured by two K-type thermocouples; the first one is immersed in the center of the fluidized bed and the second one is placed at the central part of the freeboard.
919
axis of the fuel feeding pneumatic transport tube at the entrance to the gasifier is about 88 mm above the grate. 2.3. Gasifying agent preparation Individual gasses (N2, O2, H2O and CO2) are fed from gas cylinders by MFCs (El-flowÒ F-201 AV) to the mixing device together with distilled water (fed from a N2-pressurized tank by a MFC Liqui-Flow™ L23). All the gas and liquid streams are heated up to 200 °C and mixed in CEM W303A. The mixture is heated up to 500 °C in an electrical pre-heater and enters the reactor below the grate. 2.4. Sampling of the producer gas The top part (‘the head’) of the gasification reactor is equipped with vertical outlets to monitor pressure and to sample the gas for off-line analysis of major gas components and tars. Here, the samples of gas for determination of H2, O2, N2, CO, CO2, CH4, and minor organic gasses are collected (a detailed description of these analyses as well as the experimental procedures can be found elsewhere [19–21]). After leaving the reactor, hot raw producer gas is dedusted in a hot cyclone (with operational temperature of 400–500 °C) and then is fed into the exhaust. The sampling point for ammonia determination is in the wall of the exhaust tube downstream the cyclone. 2.5. Ammonia sampling and analysis
2.2. Fuel feeding A two-chamber slide feeder (see again Fig. 1) doses the fuel into a stream of nitrogen, which transports it to the fluidized bed of the reactor. The flow of nitrogen is maintained by MFC (El-flowÒ F-201 AV by Bronkhorst). The fuel-feeding rate can be controlled by changing the frequency of the sliding of the feeder or by installing a PTFE plate with different dimensions of cylindrical chambers. A detailed description of the fuel-feeding device can be found elsewhere [17,18]. The feeding line is cooled by a water cooler at the entrance into the gasifier in order to prevent fuel sintering. The
Ammonia sampling and analysis procedure was developed on the basis of a Czech technical standard used for monitoring ammonia emissions from stationary pollutant sources, namely CˇSN 834728 (1986). This norm is in accordance with the norm of the Environmental Protection Agency (EPA) CTM-027 for collection and analysis of ammonia in stationary sources (1982). These methods are, basically, composed of an absorption step, in which the ammonia is absorbed into an acidic solution, and a subsequent off-line analysis of the absorbed ammonia content. To determine the concentration of ammonia in the solution, ion chromatography,
Fig. 1. Experimental facility.
920
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
ion selective electrodes, colorimetry or distillation with subsequent acido-basic titration can be used. However, the method (based on these standards) had to be adjusted to be applicable for the analysis of ammonia content in producer (or synthesis) gas. Several authors discussed the ammonia determination during gasification [1,2,8,11,15,16,22]. In our opinion, the most inspiring study was published by Norton et al. [23].
consequent analyses. An empty impinger was placed at the end of the sampling train to protect the vacuum pump and the wet type gas meter. The main problem during sampling was the desublimation of light tar and the condensation in the impingers that occurred at low temperatures (the impingers are immersed in an ice bath). These tars blocked gas distributors and bubble tubes and worsened the pressure drop of the sampling line during sample collection. To minimize this problem, the gas flow had to be controlled and adjusted continuously during sampling and the sampling line and impingers had to be washed by an organic solvent before each other sampling. Usually, we collected the ammonia from (at least) 30 l of the producer gas during approx. 30 min into approx. 300 ml of the absorption solution. Another 50–100 ml of the absorption solution was used to wash up the impingers and to wash out traces of ammonia that could remain in connecting lines. Afterwards, all of the solution was weighed and stored in a refrigerator at 5 °C.
2.5.1. Ammonia absorption step Inspired by the above-mentioned standards and literature, we adapted the configuration of the sampling line to preserve its functionality while simplifying the sampling process as much as possible. In contrast to the sampling line described by Norton et al. [23], we did not use a hot filter (450 °C) to remove solid particles, because the fly-ash particles can catalyze the NH3 decomposition. Prior to sampling, the gas was driven through a hot cyclone (operating at 450–550 °C) with a theoretical collection efficiency of 100% for particle size of d > 40 lm, and approx. 50% for size of d 5 lm [24]. The temperature of the producer gas downstream the hot cyclone at the level of the sampling point was 380–430 °C. Gas was sampled through a stainless steel sampling probe into a PTFE tube and then into a filter container with a microfibers glass filter. The length of the PTFE tube between the sampling probe and the filter container was approx. 8 cm. At 3 the sampling gas flow of approx. 1.7 dmn min1, the temperature of the gas at the exit of the filter container was always above 120 °C. This ensured that also heavy tars were filtered out by the microfiber glass filter (Munktell, MG 227/1/60 with declared efficiency 99.9% for particles d > 0.3 lm). After passing through the filter, the gas flowed downwards through a polyurethane tube (30 cm), where it was cooled down to ambient temperature. The water that condenses in this tube trickled down by the force of gravitation into the first impinger (Fig. 2). For fuels with higher nitrogen content, i.e., coals, the gas passed through three impingers with 0.04 M H2SO4 solution. For biomass, which, compared to coals, has low nitrogen content, 0.01 M H2SO4 solution was used. Concentrations of H2SO4 in the solutions were chosen with the intent to safely absorb all ammonia on one hand, but, on the other hand, not to acidify the solution too much in order to prevent some possible interferences with excessive SO24 -during the IC analysis. The volume of the solution was approx. 100 ml in each impinger. During the initial experiments, we also measured the ammonium content in each individual impinger in order to prove that all ammonia is securely absorbed. 98% of all of captured ammonium was found in the first impinger and 2% in the second one. In the third impinger, the ammonium concentration was below the detection limit of our analytical method (distillation and back titration). This is in accordance with the results measured by Norton et al. [23]. After an ordinary measurement, liquid from all three impingers was mixed into one sample together with the solution used to wash the sample line after each collection resulting in about 350–400 ml of the solution being collected and stored for
2.5.2. Absorbed ammonium analysis Informed by the comparison of colorimetry, titration, and ISE described by Norton et al. [23], we chose titration with a preceding distillation step (similar to Kjeldahl method) as the most reliable method for ammonia analysis. The distillation step in the Parnas–Wagner distillation apparatus is, however, time consuming (1 sample per 30–45 min) and requires good operational skills and focus. Moreover, about 100 g of the sample is consumed by each distillation, which limits the repeatability of such analysis. Therefore, prior to the analysis of the real samples, we had prepared several artificial samples, which we used to train our skills of distilling and titrating while polishing the method until we measured exactly the amount of ammonia that we had been introducing into the artificial samples. The final polished procedure manual is in the supplement of this article. We also analyzed some (non-distilled) samples by colorimetry, but the results obtained were not reproducible, and we did not use this method any further. The fact that we had not filtered the samples prior to the analysis can be a possible explanation for this method’s failure. As the distillation and titration analysis is time demanding, we also examined the possibility of analyzing the collected samples by ion chromatography (by an external laboratory on Dionex ICS5000) even if Norton et al. [23] reported erroneous results obtained by this method. They reported that acid strength had a strong effect on the results. Thus, we prepared an absorption solution with the minimal possible acidity (0.01 M H2SO4 for biomass and 0.04 M H2SO4 for coal) and we diluted this solution further by deionized water (1:100 and 1:1000) before analysis. In order to test the feasibility of ion chromatography (IC), we first prepared artificial samples with known concentrations of ammonia with appropriate H2SO4 background, and into some samples we added a typical mixture of organic compounds that are normally present in producer gas (benzene, toluene, phenol, naphthalene and anthracene). Even
From gasifier
Vent Ash Soot filter filter
Vacuum pump Gas meter
Empty impinger
Ice bath
Empty impinger 0.1M0.04 H2SO4 (0.01)impingers M H SO impingers 2
4
Fig. 2. Ammonia sampling train.
921
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
though the results obtained by IC were not precise, they were comparable with the expected values and introducing the organic compounds did not seem to affect the analysis. Throughout our experiments, we compared the results of analyzing real samples by distillation and titration and by IC. Unfortunately, different results were obtained by each of these methods. The comparison (Fig. 3) shows a linear dependency of the results, which was, however, different for each experiment; above all, the two data sets in the low values show this discrepancy (the low values were measured during biomass gasification). As we trust distillation and titration more, we concluded that the IC analysis could be useful when comparing individual results of a data set, but not in the assessment of the absolute results. Nevertheless, we used this knowledge to our advantage and, because of lack of time during the experimental campaign, we performed only two analyses by distillation and titration within a data set (collected during one experiment) and used the IC analysis for quick scanning of relative values of all the samples in the data set. After finishing the experimental campaign, we analyzed the remaining samples by distillation and titration in order to verify the results obtained by IC and linear recalculation, and it worked. In this article, only verified data by titration analysis are given; however, the combination of the two methods enables us to reduce the time needed for a quick analysis while maintaining high credibility of the results. 2.5.3. Overall experimental error The ammonia sampling started after about 1 h after the beginning of the dosing of fuel to the reactor at steady-state conditions.
During each experiment, 2–3 samples were taken. In the results, we present the mean values. The possible experimental error is difficult to estimate, because various influences are on the experimental side, the sampling and the analysis side. On the analysis side (i.e. the distillation of the sample and consecutive back titration), the analysis error is less then 0.05% relative. The error during sampling could be estimated only by a parallel sampling of several samples, which would be quite difficult; but, because of the volume of the sampled gas (at least 30 l at normal conditions), the relative error is diminished. Moreover, the yield of ammonia changes slightly during the experiment. Estimated overall relative error of the presented sampling and analysis is 1.5%. 2.6. Fuels Four fuels selected within the Fecundus project (cited in acknowledgements) were gasified in the 10 kWth fluidized bed reactor. Their proximate and ultimate analysis is depicted in Table 1 with highlighted nitrogen content. Also the Cl content of the fuels is important, because the ammonia concentration in the gas can decrease through the formation of NH4Cl. Nevertheless, chlorine concentration is negligible in comparison with the nitrogen content. Because of higher ash level in the coals, the content of catalytically active components (Fe and Ca) can catalyze the nitrogen reactions. Elemental composition of the ashes of coals measured by X-ray fluorescence on spectrometer ARL 9400 XP is presented in Table 2. 2.7. Experimental conditions All the fuels were gasified at similar conditions (Table 3). Despite some variations, the experimental conditions of each
Table 2 Ash components in oxidized form measured by X-ray fluorescence. Possibly catalytically active components are highlighted.
Fig. 3. Comparison of NHþ 4 concentration measured in the absorption solution by distillation and titration and by ion chromatography.
(wt.%)
Spanish coal
South African Coal
SiO2 A12O3 Fe2O3 MgO CaO Na2O K2O MnO TiO2
58 31 5.2 1.0 1.1 0.41 2.4 0.10 0.71
52 38 3.1 0.84 3.4 0.42 0.73 n.d. 1.8
Table 1 Proximate and ultimate fuel analysis. Fuel (raw)
Spanish coal
South African coal
Rhenish coal
Biomass
Water (% wt) Total combustibles (% wt) Ash (% wt) Volatile combustibles (% wt)
5.9% 47% 47% 18%
5.9% 83% 11% 26%
12% 85% 3.9% 44%
9.8% 89% 0.88% 75%
Fixed carbon (% wt.)
29%
57%
41%
15%
HHV (MJ/kg) LHV (MJ/kg)
14 14
27 26
23 22
18 16
C H N O
36% 2.4% 0.77% 7.2%
66% 3.9% 1.7% 11%
56% 4.3% 0.71% 23%
44% 5.4% 0.17% 40%
S Cl F Br
0.82% 0.040% 0.012% n.a.
0.40% 0.023% 0.013% n.a.
0.22% 0.036% 0.0042% 0.0008%
0.017% 0.011% 0.0030% n.a.
922
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
Table 3 Experimental conditions (the gasses volume is stated at 25 °C and 101 325 Pa. Fuel
Spanish Coal
South African Coal
Rhenish coal
Biomass
Gasifying agent Material of the fluidized beda Reactor temperature (°C) Dry fuel feeding rate (g h1) Raw fuel feeding rate (g h1)
H2O + O2 Sand 850 ± 5 1100 1170
H2O + O2 Sand 850 ± 5 1080 1150
H2O + O 2 Sand s + l(l:l) 850 ± 5 1070 1140 1200 1300
H2O + O2 Sand s + l(l:l) s 850 ± 5 1280 1320 1420 1450
Steam (m3h1) CO2(m3h1) O2(m3h1) additional N2(m3h1) N2 pneumatic transport (m3/h Total gas inlet (m3/h)
1.36 – 0.19 0.45 1.50 3.50
1.36 – 0.34 0.30 1.50 3.50
1.36 – 0.30 0.34 1.50 3.50
1.36 – 0.33 0.31 1.50 3.50
1.36 – 0.28 0.37 1.35 3.36
ER H2O/C (mol/mol) CO2/C (mol/mol)
0.20 1.70 –
0.20 0.94 –
0.20 1.13 –
0.21 1.07 –
0.22 1.22 –
1290 1430
CO2 + O2 s + l(l:l) 850 ± 5 1320 1450
1.36 – 0.27 0.37 1.50 3.50
1.36 – 0.27 0.37 1.50 3.50
0.20 1.17 –
0.20 1.21 –
s + 1 (3:1)
1290 1430
CO2 + H2O + O2 s + l(l:l) s+1(3:1) 850 ± 5 1320 1290 1450 1430
– 1.36 0.27 0.37 1.50 3.50
– 1.35 0.27 0.37 1.50 3.50
0.41 0.91 0.27 0.37 1.50 3.46
0.41 0.95 0.27 0.37 1.50 3.55
0.20 0.13b 1.04
0.20 0.15b 1.06
0.20 0.45 0.69
0.20 0.47 0.74
s + 1(3:1)
Gasifying agent
a b
s = s and, l = limestone. The value is higher than zero, because some moisture is coming to the reactor within the fuel.
experiment can be assumed to be almost identical; hence, the ammonia content and yield can be safely compared. The experiments with Spanish coal and South-African coal were performed only with (inert) silica sand in the fluidized bed. To test the effect of adding catalytic material, we performed experiments with Rheinish coal and biomass with the mixture of silica sand and dolomitic limestone. Two different ratios (1:1 and 1:3) of dolomitic limestone and silica sand in the fluidized bed were tested with biomass as the fuel. The detailed composition and properties of the fluidized bed materials can be found elsewhere [25]. 3. Results Ammonia yield is closely connected with the producer gas composition and with the total fuel-to-gas conversion. Thus, we first introduce a summary of the yield of main gasses from different fuels with different ratios of dolomitic limestone in the fluidized bed and with different gasifying agent mixtures (Fig. 4). In the same chart, the carbon conversion is depicted. To give a detailed overview of the relative partial pressures of individual gasses in the producer gas mix, its composition is presented in Fig. 5. This dry and N2-free composition together with the LHV of the gas (also dry and N2-free) is a basic technical parameter, which is included mainly for a potential future comparison with data from other sources. And, finally, the concentration of the ammonia in the dry and N2-free producer gas is depicted in Fig. 6 alongside the conversion rate of fuel-nitrogen to ammonia. This value is independent of the actual gas yield and it can be used for basic estimation of the ammonia yield from a specific fuel under specific conditions. 4. Discussion 4.1. Ammonia yield from gasification of different types of coal and biomass As was mentioned in the introduction, the nitrogen in coals is stored mainly in forms belonging to the non-volatile part of the fuel. Therefore, the rate of NH3 yield from coal gasification should be related to the rate of char gasification. Non-reactive coals are supposed to have a very low rate of conversion of fuel-nitrogen into ammonia. However, in the case of Spanish Coal and South African coal (presented in the charts of Figs. 4 and 6) this did not
happen. The carbon conversion of Spanish coal was only 41% and, in the case of South African coal, only 39%, which is a very low value compared to the other studied fuels (see Fig. 4). The high rate of conversion of fuel nitrogen into ammonia during the gasification of these two fuels (Spanish Coal and South African coal; Fig. 6) can be attributed to any of the following: (1) the distribution of fuel-N into volatile and non-volatile parts of the fuel could be different from the assumptions stated in the introduction. (2) There could be a high content of catalytically active components in the ash (Ca, Fe), or (3) the retention time of the char in the fluidized bed is quite long. The content of the possibly catalytically active components in the ash was quite low (see Table 2) and the ammonia yield slightly increased during the experiment. Therefore, we are inclined towards the third explanation: the relatively high conversion of fuel-N into ammonia was probably caused by a long retention time of the char and by its accumulation in the fluidized bed with the release of the nitrogen from its structures. Concerning the reactive coal (Rheinish coal), the measured conversion of fuel-N into ammonia is in accordance with our assumptions. High conversion of the char (the carbon conversion was 78% on the sand fluidized bed and 89% on the mixture of limestone and sand in the fluidized bed) is with accordance with the high conversion of the fuel-nitrogen into ammonia (see Fig. 6). The conversion of biomass nitrogen into ammonia ranged from 43% to 57%, which is a lower value than reported in the literature. On the other hand, the biomass we used is not a typical fuel for gasification and contains a very small amount of nitrogen. 4.2. Dolomitic limestone in the fluidized bed We examined the effect of adding dolomitic limestone into the fluidized bed when gasifying the reactive coal and biomass. The addition had a positive impact on the conversion of fuel-N into ammonia, both in the case of Rheinish coal and wood; however, the use of different gasifying media had a definite influence on the ammonia yield. If we focus only on the use of H2O + O2, the presence of the dolomitic limestone in the fluidized bed (with the ratio 1:1 of sand to limestone) increased the conversion rate of fuel-N into NH3 by approx. 10% (see Fig. 6). This fact is in accordance with Pinto et al. [11] and suggests that the dolomitic limestone improves the conversion rate of fuel-N into NH3 by affecting the whole gasification process including the char conversion, steam conversion, gas composition etc.
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
Fig. 4. Raw producer gas yield and fuel-carbon conversion into producer gas.
Fig. 5. Producer gas composition and LHV (dry and N2-free).
Fig. 6. Concentration of NH3 and the rate of conversion of fuel nitrogen into ammonia in producer gas.
923
924
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
Moreover, it can catalyze the transformation of other nitrogenous compounds into ammonia. Limestone can improve the decomposition reactions (as is reported by Simell et al. [26]), but this decomposition is less influential than the above-mentioned effect. Interesting is also the effect of increasing the limestone concentration in the fluidized bed from 25% to 50%. In the case of gasification with H2O + O2, such an increase had only a negligible effect; however, in the case of gasification with CO2 + O2, the rate of nitrogen conversion increased by 10%, and, in the case of CO2 + H2O + O2, even by 14%. It seems that for every gasifyingagent mixture there exists a specific minimal concentration of limestone in the fluidized bed that increases discretely the ammonia yield. 4.3. The use of different gasifying media The use of CO2 instead of H2O in the gasifying agent had a different effect when using 25% of dolomitic limestone in the sand fluidized bed and when using 50% of limestone. Thus, the effect of different gasifying agent is closely connected with the catalytic activity of the material of the fluidized bed as is mentioned in the Section 4.2. For the fluidized bed of sand and limestone (3:1), the conversion of fuel–N into ammonia decreased by 8%, when fully substituting the H2O by CO2, and even by 11% in the case of partial substitution of H2O by CO2 (see dashed line in Fig. 6), which is in accordance with the expectations stated in chapter 1.3. On the other hand, for a catalytically stronger material of the fluidized bed (the ratio of 1:1), the conversion only slightly increased with the substitution of H2O by CO2 (full line in Fig. 6). We suggest that there exists some critical mass of limestone in the fluidized bed. This critical mass of limestone that caused a discrete jump of ammonia yield by almost 10% was achieved by using a mixture of sand and limestone with the ratio of 3:1 when H2O + O2 was used as gasifying agent. However, when CO2 was used as gasifying agent, this boundary was crossed somewhere in the range of concentration 25–50% of limestone in the fluidized bed. 5. Conclusion We measured the conversion rate of fuel-bound nitrogen into ammonia during gasification of Spanish coal, South African coal, Rheinish coal and woody biomass. The rate was surprisingly high for less-reactive coals (61% for Spanish coals and 52% for South African Coal). For the reactive Rheinish coal, the rate was 74–83% and for wood biomass it was 43–57%. The conversion of fuel-N into ammonia increased substantially (by approx. 10%) when the gasification was catalyzed by dolomitic limestone in fluidized bed. When steam in the gasifying agent was substituted with carbon dioxide, there was a different effect on the rate of fuel-N to ammonia conversion when 25% of limestone and 50% of limestone was in the fluidized bed. However, in both cases the concentration of ammonia in the dry gas was diminished because of the dilution by CO2. In this study, the variation of composition of the gasifying agent was examined for a fuel with low content of nitrogen (0.17 wt%) and different results for different concentrations of limestone in the fluidized bed were obtained. Further work on this theme should be directed towards the use of fuels with higher content of nitrogen and on the clarification of the use of CO2 as a gasifying agent. Nomenclature All volume units of the gas and other relevant values are reported at 101.325 kPa and 25 °C.
Dry gas yield is the volume of the gas (at 101.325 kPa and 25 °C) related to mass flow of dry fuel (1). Total gas yield was calculated from the N2 balance (known volume in the inlet and known concentration in the dry gas at the outlet). The yields of specific gases were calculated from the total gas yield multiplied by their concentrations in the dry producer gas.
gas yield ¼ V_ dry
_ dry fuel gas =m
ð1Þ
Steam yield was calculated from H and O elemental balances. Lower heating value (LHV) of the produced gas was calculated as the sum of products of heating value of measured gas components (LHVi) and their concentrations (ui) in the producer gas, according to the equation (2). The lower heating values of the gas components were adopted from the norm EN ISO 6976. The tarry compounds were not included in the computation.
LHV ¼
n X
ui LHV i
ð2Þ
i¼1
Cold gas efficiency is the ratio of the chemical energy of the dry producer gas (based on LHVd) to the chemical energy stored in the dry fuel (LHVd). Equivalence ratio (ER) is the molar ratio of the total input of O2 by the gasifying agent related to O2 theoretically needed for total combustion of the fuel into CO2 and H2O (and SO2, etc.; the real oxygen demand will be higher because of thermodynamic and kinetic limitations). For example, the ER of 0.2 means that 20% of the oxygen theoretically needed for total combustion is fed into the reactor. Steam to fuel carbon ratio (H2O/C) is the molar ratio of steam fed into the reactor by gasifying agent and/or as the fuel moisture related to the elemental carbon in the fuel. Carbon dioxide to carbon ratio (CO2/C) is the molar ratio of carbon dioxide fed into the reactor by gasifying agent related to elemental carbon in the fuel.
Acknowledgements The authors appreciate the financial support from EU Research Fund for Coal and Steel, Grant RFCR-CT-2010-0009 ‘Fecundus’, The support of Ministry for Education, Youth and Sport of the Czech Republic, Projects Nos. 7C11009 and 6046137304 and specific university research projects Nos. 20/2013, 21/2012, 21/2011 and 21/ 2010. Appendix A. Supplementary material Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.fuel.2013.10.009. References [1] Pröll T, Siefert IG, Friedl A, Hofbauer H. Removal of NH3 from biomass gasification producer gas by water condensing in an organic solvent scrubber. Ind Eng Chem Res 2005:1576–85. [2] Hämäläinen JP, Aho MJ, Tummavuori JL. Formation of nitrogen oxides from fuel-N through HCN and NH3: a model-compound study. Fuel 1994;73:1894–8. [3] Svoboda K, Hartman M. Formation of NOx in fluidized bed combustion of model mixtures of liquid organic compounds containing nitrogen. Fuel 1991;70:865–71. [4] Kelsall GJ, Smith MA, Cannon MF. Low emissions combustor development for an industrial gas turbine to utilize LCV fuel gas. J Eng Gas Turbines Power 1994;116:559. [5] Higman C, Burgt M van der. Gasification. Gulf Professional Publishing; 2008. [6] Rabou LPLM, Zwart RWR, Vreugdenhil BJ, Bos L. Tar in biomass producer gas, the energy research center of The Netherlands (ECN) experience: an enduring challenge. Energy Fuels 2009;23:6189–98.
M. Jeremiáš et al. / Fuel 117 (2014) 917–925 [7] Leppälahti J, Koljonen T. Nitrogen evolution from coal, peat and wood during gasification: literature review. Fuel Process Technol 1995;43:1–45. [8] Wilk V, Hofbauer H. Conversion of fuel nitrogen in a dual fluidized bed steam gasifier. Fuel 2013;106:793–801. [9] Devi LG, Ptasinski KJ, Janssen FJJG. A review of the primary measures for tar elimination in biomass gasification processes. Biomass Bioenergy 2003;24:125–40. [10] Leppälahti J. Formation and behavior of nitrogen compounds in an IGCC process. Bioresour Technol 1993;46:65–70. [11] Pinto F, Lopes H, André RN, Gulyurtlu I, Cabrita I. Effect of catalysts in the quality of syngas and by-products obtained by co-gasification of coal and wastes. 1. Tars and nitrogen compounds abatement. Fuel 2007;86:2052–63. [12] Devi LG, Ptasinski KJ, Janssen FJJG, van Paasen SVB, Bergman PC, Kiel JH. Catalytic decomposition of biomass tars: use of dolomite and untreated olivine. Renew Energy 2005;30:565–87. [13] Tunå P, Brandin J. Selective catalytic oxidation of ammonia by nitrogen oxides in a model synthesis gas. Fuel 2013;105:331–7. [14] Simell PA, Kurkela EA, Ståhlberg P, Hepola JO. Catalytic hot gas cleaning of gasification gas. Catal Today 1996;27:55–62. [15] Leppälahti J, Simell PA, Kurkela EA. Catalytic conversion of nitrogen compounds in gasification gas. Fuel Process Technol 1991;29:43–56. [16] Chang L, Xie Z, Xie K-C, Pratt KC, Hayashi J, Chiba T, et al. Formation of NOx precursors during the pyrolysis of coal and biomass. Part VI. Effects of gas atmosphere on the formation of NH3 and HCN. Fuel 2003;82:1159–66.
925
[17] Pohorˇely´ M, Svoboda K, Hartman M. Feeding small quantities of particulate solids. Powder Technol 2004;142:1–6. [18] Hartman M, Pohorˇely´ M, Trnka O. Transport velocities of different particulate materials in pneumatic conveying. Chem Pap 2006;60:74–7. [19] Pohorˇely´ M, Vosecky´ M, Hejdová P, Puncˇochárˇ M, Skoblia S, Staf M, et al. Gasification of coal and PET in fluidized bed reactor. Fuel 2006;85:2458–68. [20] Svoboda K, Pohorˇely´ M, Jeremiáš M, Kameníková P, Hartman M, Skoblia S, et al. Fluidized bed gasification of coal–oil and coal–water–oil slurries by oxygen– steam and oxygen–CO2 mixtures. Fuel Process Technol 2012;95:16–26. [21] Šyc M, Pohorˇely´ M, Jeremiáš M, Vosecky´ M, Kameníková P, Skoblia S, et al. Behavior of heavy metals in steam fluidized bed gasification of contaminated biomass. Energy Fuels 2011;25:2284–91. [22] Yu Q-Z, Brage C, Chen G-X, Sjöström K. The fate of fuel-nitrogen during gasification of biomass in a pressurised fluidised bed gasifier. Fuel 2007;86:611–8. [23] Norton GA, Brown RC. Wet chemical method for determining levels of ammonia in syngas from a biomass gasifier. Energy Fuels 2005;19:618–24. [24] Seville JPK. Gas cleaning in demanding applications. Blackie Acad Prof 1997. [25] Hartman M, Svoboda K, Pohorˇely´ M, Šyc M, Jeremiáš M. Attrition of dolomitic lime in a fluidized-bed at high temperature. Chem Pap 2013;67:164–72. [26] Simell PA, Hepola JO, Krause AOI. Effects of gasification gas components on tar and ammonia decomposition over hot gas cleanup catalysts. Fuel 1997;76:1117–27.
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Ammonia yield from gasification of biomass and coal in fluidized bed reactor M. Jeremiáš a,b,⇑, M. Pohorˇely´ a,b, P. Bode c, S. Skoblia d, Z. Benˇo d, K. Svoboda a a
Institute of Chemical Process Fundamentals, Academy of Science of the Czech Republic, Rozvojová 135, 165 02 Praha 6-Suchdol, Czech Republic Department of Power Engineering, Institute of Chemical Technology Prague, Technická 5, 166 28 Praha 6, Czech Republic c University of Twente, Postbus 217, 7500 AE Enschede, The Netherlands d Department of Gas, Coke and Air Protection, Institute of Chemical Technology Prague, Technická 5, 166 28 Praha 5, Czech Republic b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
We investigated the conversion rate
of fuel-N into ammonia. We adjusted ammonia sampling and
analysis procedure. We measured the effect of limestone
and CO2 addition.
a r t i c l e
i n f o
Article history: Received 13 August 2013 Received in revised form 2 October 2013 Accepted 8 October 2013 Available online 26 October 2013 Keywords: NH3 Limestone Dolomite NOx Gasifying agent
a b s t r a c t Gas from the gasification of coal and biomass can be used for combined production of heat and power or for the synthesis of alternative fuels. In both cases, ammonia, which is created by the conversion of fuelbound nitrogen, can cause technological difficulties. Therefore, ammonia concentration in the produced gas has to be monitored, or, at least, estimated to take appropriate measures for its abatement. In this paper, we report in detail the analysis of ammonia content in the produced gas as well as results of several experiments of fluidized-bed gasification of Spanish coal, South African coal, Rheinish coal and woody biomass. We focused on the evaluation and comparison of ammonia yield based on the following variables: (1) different fuels, (2) the addition of dolomitic limestone into the fluidized bed and (3) the use of carbon dioxide instead of steam in the gasifying agent. The conversion rate of fuel nitrogen into gas was surprisingly high for the less-reactive coals and within presumed levels for Rheinish coal and biomass. The presence of dolomitic limestone in the fluidized bed substantially increased the conversion rate of fuel-nitrogen into ammonia. The use of different gasifying agents had an ambiguous effect on ammonia yield. Ó 2013 Elsevier Ltd. All rights reserved.
Abbreviations: CEM, controlled evaporation and mixing system; ER, equivalence ratio; IC, ion chromatography; LHV, lower heating value; MFC, mass flow controller; PTFE, polytetrafluoroethylene (Teflon). ⇑ Corresponding author at: Institute of Chemical Process Fundamentals, Academy of Science of the Czech Republic, Rozvojová 135, 165 02 Praha 6-Suchdol, Czech Republic. Tel.: +420 731 570 803; fax: +420 220 920 661. E-mail addresses: [email protected], [email protected] (M. Jeremiáš), [email protected] (M. Pohorˇely´). 0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2013.10.009
1. Introduction Biomass and coal can be processed into gaseous fuel via gasification in a fluidized bed reactor. The produced gas can be used as fuel for gas engines with internal combustion – in this case it can be called ‘producer gas’ – or as a raw material for synthesis of
918
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
various chemical products, such as methanol or Fischer-Tropsch fuels – in this case it may be referred to as ‘synthesis gas’. Gaseous nitrogenous compounds are regarded as an unwanted component of the gas. In the case of producer gas, they reduce the operating lifetime of oil in the internal combustion engine [1] and they can act as precursors of formation of nitrogen oxides [2,3]. For example, trials with 300–1120 ppmv NH3 showed that, during gas turbine combustion, 20–50% of ammonia is converted into NOx [4]. However, the conversion rate can vary in a relatively broad range according to experimental conditions [3]. In synthesis gas applications (such as methanol production or Fischer–Tropsch synthesis) ammonia can contribute to the formation of amines on the synthesis catalyst or it can poison the catalyst itself. The presence of amines in the final products (methanol, liquid fuels) is not permitted in the internationally accepted specifications and their removal is possible only with an ion exchanger [5]. Moreover, where chlorine is present (typically when gasifying coal) ammonia will combine with chlorides to form ammonium chloride at about 250–280 °C, which can form deposits on gas coolers or gas filters [5]. Thus, the concentration of nitrogen compounds in both the producer gas and synthesis gas has to be monitored and minimized. However, the nitrogenous-compounds-content is not usually the main issue regarding gas purity. Tar and soot removal, desulfurization and dehalogenation take priority over the removal of N-compounds; thus, the reduction of nitrogenous compounds is usually a by-product of some other means applied; for example, oil or water quench for tar removal, in which the nitrogenous compounds are absorbed into water phase [1,6]. Consequently, the nitrogenous compounds are usually not fully removed from the producer gas. Their more efficient removal can be achieved through installing another gas cleaning unit, but this may render smaller plants uneconomic [1]. Lower emissions can also be achieved by adjusting the operating parameters of the gasifier or of the gas cleaning devices. For this purpose, both theoretical and empirical knowledge of the behavior of nitrogenous compounds during gasification and gas cleaning stages is required. During gasification, fuel nitrogen is liberated mainly in the form of ammonia, cyanides, thiocyanates, molecular nitrogen, nitrogen oxides and various aromatic compounds, while a smaller part of the nitrogen is retained in solid char [7]. The yield of these compounds and their ratios depend on various parameters of the gasification process; such as, the type of gasifier, temperature, equivalence ratio (ER) and catalysts. In fluidized bed reactors, longer gas-char contact time results in higher NH3/HCN ratio [7]; consequently, NH3 is a vastly predominant nitrogenous compound in the resulting gas [8]. The yield of ammonia is influenced by the gasification process parameters whose effects are summarized in the literature [9– 11]. In this study, we want to contribute to the knowledge base of nitrogen behavior during gasification by presenting the data obtained during fluidized-bed gasification of coal and biomass at 850 °C at atmospheric pressure. We focused on evaluating and comparing the ammonia yield based on the three following variables: (1) different fuels, (2) the addition of dolomitic limestone into the fluidized bed and (3) the use of different gasifying agent mixtures (H2O + O2, CO2 + O2, CO2 + H2O + O2). 1.1. Ammonia yield from gasification of different types of coal and biomass The yield of nitrogenous compounds during gasification is dependent on the nitrogen forms in the fuel. Leppälathi and Koljonen [7] published a comprehensive review about the forms of nitrogen in coal, peat and wood. It is generally assumed that the majority of nitrogen in coals is in pyrrolic and pyridinic form, and, on the other hand, nitrogen in living organic tissues is bound
mainly in proteins, aminoacids and alkaloids. Consequently, in peat and wood gasification, most of fuel nitrogen is liberated during the pyrolysis stage and, in coal gasification at low temperatures (<1200 K), most of the nitrogen is retained in char after pyrolysis and released during the char gasification stage. Leppälathi and Koljonen also report the rate of fuel-N to NH3 conversion for different fuels during fluidized bed gasification of lignite, sub-bituminous and bituminous coals and wood by oxygen and steam. During the gasification of lignite and sub-bituminous coal at high pressures (0.8 and 1.5 MPa), approximately 60% of fuel-N was converted into NH3. Even, a conversion of 93% was observed with a highly volatile coal. With bituminous coal, the ammonia conversion was significantly lower (0.6%–19%). In wood gasification, the NH3 conversion was 72%–97% (at pressure 0.4– 1.0 MPa). However, lower gasification pressure can lead to a significantly lower NH3 yield. [7] In allothermal steam gasification of wood in the Guessing (FICFB) plant, up to 70% of the fuel nitrogen was found as NH3 in the producer gas [1]. The concentration of nitrogen in fuel can also affect its conversion into gas; with the increase of concentration of chemically bound nitrogen in fuel, its conversion rate decreased [3]. The innovation of this study is the comparison of the yield of ammonia during atmospheric fluidized-bed gasification of less reactive Spanish coal and South African coal, reactive Rheinish coal and woody biomass. To our best knowledge, such a comparison is missing in the available literature. 1.2. Dolomitic limestone in the fluidized bed The use of catalytically active dolomitic limestone during gasification promotes char conversion, changes product gas composition, and reduces tar yield [12]. Inevitably, it will also affect the yield of ammonia. The presence of a catalyst can affect both the creation of ammonia and its decomposition. The decomposition of ammonia into H2 and N2 occurs over a reforming catalyst at high temperatures (1073 K) [13]. For example, Simmel et al. [14] report a 100% conversion rate on a nickel catalyst and a 53% conversion rate on dolomite in a laboratory-scale fixed-bed tube reactor at temperature of 900–910 °C, residence time 0.2–0.3 s and pressure 1 bar (in simulated synthesis gas). On the other hand, Leppälahti et al. [15] stated that the limestone and dolomite did not have any catalytic effect in the decomposition of ammonia, but they reduced the nitrogen cyanide content in the producer gas [15]. Pinto et al. [11] tested the use of dolomite directly in fluidized bed. They report that dolomite did not change the concentration of NH3 in the producer gas from coal gasification, but it increased the conversion of fuel-N into NH3 (this discrepancy is caused by the increase of overall producer gas yield on dolomite) [11]. This indicates that a directly used catalyst enhances the formation of NH3 from fuel, probably by increasing the conversion rate of char and by increasing the concentration of hydrogen in the system. This enhancement probably prevails over ammonia-decomposition effect. In this article, we compare the yield of ammonia from the gasification of reactive Rheinish coal or biomass either on sand in the fluidized bed or on the mixture(s) of sand and dolomitic limestone in the fluidized bed. To our best knowledge, such a comparison of ammonia yield on different mixtures of sand and dolomitic limestone has not been published. 1.3. The use of different gasifying agents Gasification in steam generates large amounts of H radicals as an intermediate between the reactions of steam with char, drastically enhancing the formation of NH3 [16]. The use of CO2 as
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
gasifying agent lowers the partial pressure of steam in the reactor; thus, it limits the generation of H radicals. Therefore, the highest NH3 yield can be expected when the mixture O2 + H2O is used as gasifying agent. In this article, we try to compare the yield of ammonia from biomass on the following mixtures of gasifying agents: H2O + O2, CO2 + O2 and CO2 + H2O + O2. 2. Experimental 2.1. Fluidized bed gasifier The fluidized-bed reactor (see Fig. 1) is an electrically heated 2200 mm high tube with inner diameter of 51.1 mm in the lower section and 99.0 mm in the upper section (freeboard) made of high-temperature resistant stainless steel (AISI 446). The 24 mm tall size reduction area between the lower and higher sections is placed 545 mm above the grate. Its maximal operating temperature is 980 °C at reducing conditions. The preheated (up to 500 °C) gasification agent fluidizes the loose inert material in the lower section of the reactor. Electrical heating with maximal output of 10 kW consists of three independent heating elements along the height of the reactor. The heating elements can be regulated separately. The temperature inside the reactor is measured by two K-type thermocouples; the first one is immersed in the center of the fluidized bed and the second one is placed at the central part of the freeboard.
919
axis of the fuel feeding pneumatic transport tube at the entrance to the gasifier is about 88 mm above the grate. 2.3. Gasifying agent preparation Individual gasses (N2, O2, H2O and CO2) are fed from gas cylinders by MFCs (El-flowÒ F-201 AV) to the mixing device together with distilled water (fed from a N2-pressurized tank by a MFC Liqui-Flow™ L23). All the gas and liquid streams are heated up to 200 °C and mixed in CEM W303A. The mixture is heated up to 500 °C in an electrical pre-heater and enters the reactor below the grate. 2.4. Sampling of the producer gas The top part (‘the head’) of the gasification reactor is equipped with vertical outlets to monitor pressure and to sample the gas for off-line analysis of major gas components and tars. Here, the samples of gas for determination of H2, O2, N2, CO, CO2, CH4, and minor organic gasses are collected (a detailed description of these analyses as well as the experimental procedures can be found elsewhere [19–21]). After leaving the reactor, hot raw producer gas is dedusted in a hot cyclone (with operational temperature of 400–500 °C) and then is fed into the exhaust. The sampling point for ammonia determination is in the wall of the exhaust tube downstream the cyclone. 2.5. Ammonia sampling and analysis
2.2. Fuel feeding A two-chamber slide feeder (see again Fig. 1) doses the fuel into a stream of nitrogen, which transports it to the fluidized bed of the reactor. The flow of nitrogen is maintained by MFC (El-flowÒ F-201 AV by Bronkhorst). The fuel-feeding rate can be controlled by changing the frequency of the sliding of the feeder or by installing a PTFE plate with different dimensions of cylindrical chambers. A detailed description of the fuel-feeding device can be found elsewhere [17,18]. The feeding line is cooled by a water cooler at the entrance into the gasifier in order to prevent fuel sintering. The
Ammonia sampling and analysis procedure was developed on the basis of a Czech technical standard used for monitoring ammonia emissions from stationary pollutant sources, namely CˇSN 834728 (1986). This norm is in accordance with the norm of the Environmental Protection Agency (EPA) CTM-027 for collection and analysis of ammonia in stationary sources (1982). These methods are, basically, composed of an absorption step, in which the ammonia is absorbed into an acidic solution, and a subsequent off-line analysis of the absorbed ammonia content. To determine the concentration of ammonia in the solution, ion chromatography,
Fig. 1. Experimental facility.
920
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
ion selective electrodes, colorimetry or distillation with subsequent acido-basic titration can be used. However, the method (based on these standards) had to be adjusted to be applicable for the analysis of ammonia content in producer (or synthesis) gas. Several authors discussed the ammonia determination during gasification [1,2,8,11,15,16,22]. In our opinion, the most inspiring study was published by Norton et al. [23].
consequent analyses. An empty impinger was placed at the end of the sampling train to protect the vacuum pump and the wet type gas meter. The main problem during sampling was the desublimation of light tar and the condensation in the impingers that occurred at low temperatures (the impingers are immersed in an ice bath). These tars blocked gas distributors and bubble tubes and worsened the pressure drop of the sampling line during sample collection. To minimize this problem, the gas flow had to be controlled and adjusted continuously during sampling and the sampling line and impingers had to be washed by an organic solvent before each other sampling. Usually, we collected the ammonia from (at least) 30 l of the producer gas during approx. 30 min into approx. 300 ml of the absorption solution. Another 50–100 ml of the absorption solution was used to wash up the impingers and to wash out traces of ammonia that could remain in connecting lines. Afterwards, all of the solution was weighed and stored in a refrigerator at 5 °C.
2.5.1. Ammonia absorption step Inspired by the above-mentioned standards and literature, we adapted the configuration of the sampling line to preserve its functionality while simplifying the sampling process as much as possible. In contrast to the sampling line described by Norton et al. [23], we did not use a hot filter (450 °C) to remove solid particles, because the fly-ash particles can catalyze the NH3 decomposition. Prior to sampling, the gas was driven through a hot cyclone (operating at 450–550 °C) with a theoretical collection efficiency of 100% for particle size of d > 40 lm, and approx. 50% for size of d 5 lm [24]. The temperature of the producer gas downstream the hot cyclone at the level of the sampling point was 380–430 °C. Gas was sampled through a stainless steel sampling probe into a PTFE tube and then into a filter container with a microfibers glass filter. The length of the PTFE tube between the sampling probe and the filter container was approx. 8 cm. At 3 the sampling gas flow of approx. 1.7 dmn min1, the temperature of the gas at the exit of the filter container was always above 120 °C. This ensured that also heavy tars were filtered out by the microfiber glass filter (Munktell, MG 227/1/60 with declared efficiency 99.9% for particles d > 0.3 lm). After passing through the filter, the gas flowed downwards through a polyurethane tube (30 cm), where it was cooled down to ambient temperature. The water that condenses in this tube trickled down by the force of gravitation into the first impinger (Fig. 2). For fuels with higher nitrogen content, i.e., coals, the gas passed through three impingers with 0.04 M H2SO4 solution. For biomass, which, compared to coals, has low nitrogen content, 0.01 M H2SO4 solution was used. Concentrations of H2SO4 in the solutions were chosen with the intent to safely absorb all ammonia on one hand, but, on the other hand, not to acidify the solution too much in order to prevent some possible interferences with excessive SO24 -during the IC analysis. The volume of the solution was approx. 100 ml in each impinger. During the initial experiments, we also measured the ammonium content in each individual impinger in order to prove that all ammonia is securely absorbed. 98% of all of captured ammonium was found in the first impinger and 2% in the second one. In the third impinger, the ammonium concentration was below the detection limit of our analytical method (distillation and back titration). This is in accordance with the results measured by Norton et al. [23]. After an ordinary measurement, liquid from all three impingers was mixed into one sample together with the solution used to wash the sample line after each collection resulting in about 350–400 ml of the solution being collected and stored for
2.5.2. Absorbed ammonium analysis Informed by the comparison of colorimetry, titration, and ISE described by Norton et al. [23], we chose titration with a preceding distillation step (similar to Kjeldahl method) as the most reliable method for ammonia analysis. The distillation step in the Parnas–Wagner distillation apparatus is, however, time consuming (1 sample per 30–45 min) and requires good operational skills and focus. Moreover, about 100 g of the sample is consumed by each distillation, which limits the repeatability of such analysis. Therefore, prior to the analysis of the real samples, we had prepared several artificial samples, which we used to train our skills of distilling and titrating while polishing the method until we measured exactly the amount of ammonia that we had been introducing into the artificial samples. The final polished procedure manual is in the supplement of this article. We also analyzed some (non-distilled) samples by colorimetry, but the results obtained were not reproducible, and we did not use this method any further. The fact that we had not filtered the samples prior to the analysis can be a possible explanation for this method’s failure. As the distillation and titration analysis is time demanding, we also examined the possibility of analyzing the collected samples by ion chromatography (by an external laboratory on Dionex ICS5000) even if Norton et al. [23] reported erroneous results obtained by this method. They reported that acid strength had a strong effect on the results. Thus, we prepared an absorption solution with the minimal possible acidity (0.01 M H2SO4 for biomass and 0.04 M H2SO4 for coal) and we diluted this solution further by deionized water (1:100 and 1:1000) before analysis. In order to test the feasibility of ion chromatography (IC), we first prepared artificial samples with known concentrations of ammonia with appropriate H2SO4 background, and into some samples we added a typical mixture of organic compounds that are normally present in producer gas (benzene, toluene, phenol, naphthalene and anthracene). Even
From gasifier
Vent Ash Soot filter filter
Vacuum pump Gas meter
Empty impinger
Ice bath
Empty impinger 0.1M0.04 H2SO4 (0.01)impingers M H SO impingers 2
4
Fig. 2. Ammonia sampling train.
921
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
though the results obtained by IC were not precise, they were comparable with the expected values and introducing the organic compounds did not seem to affect the analysis. Throughout our experiments, we compared the results of analyzing real samples by distillation and titration and by IC. Unfortunately, different results were obtained by each of these methods. The comparison (Fig. 3) shows a linear dependency of the results, which was, however, different for each experiment; above all, the two data sets in the low values show this discrepancy (the low values were measured during biomass gasification). As we trust distillation and titration more, we concluded that the IC analysis could be useful when comparing individual results of a data set, but not in the assessment of the absolute results. Nevertheless, we used this knowledge to our advantage and, because of lack of time during the experimental campaign, we performed only two analyses by distillation and titration within a data set (collected during one experiment) and used the IC analysis for quick scanning of relative values of all the samples in the data set. After finishing the experimental campaign, we analyzed the remaining samples by distillation and titration in order to verify the results obtained by IC and linear recalculation, and it worked. In this article, only verified data by titration analysis are given; however, the combination of the two methods enables us to reduce the time needed for a quick analysis while maintaining high credibility of the results. 2.5.3. Overall experimental error The ammonia sampling started after about 1 h after the beginning of the dosing of fuel to the reactor at steady-state conditions.
During each experiment, 2–3 samples were taken. In the results, we present the mean values. The possible experimental error is difficult to estimate, because various influences are on the experimental side, the sampling and the analysis side. On the analysis side (i.e. the distillation of the sample and consecutive back titration), the analysis error is less then 0.05% relative. The error during sampling could be estimated only by a parallel sampling of several samples, which would be quite difficult; but, because of the volume of the sampled gas (at least 30 l at normal conditions), the relative error is diminished. Moreover, the yield of ammonia changes slightly during the experiment. Estimated overall relative error of the presented sampling and analysis is 1.5%. 2.6. Fuels Four fuels selected within the Fecundus project (cited in acknowledgements) were gasified in the 10 kWth fluidized bed reactor. Their proximate and ultimate analysis is depicted in Table 1 with highlighted nitrogen content. Also the Cl content of the fuels is important, because the ammonia concentration in the gas can decrease through the formation of NH4Cl. Nevertheless, chlorine concentration is negligible in comparison with the nitrogen content. Because of higher ash level in the coals, the content of catalytically active components (Fe and Ca) can catalyze the nitrogen reactions. Elemental composition of the ashes of coals measured by X-ray fluorescence on spectrometer ARL 9400 XP is presented in Table 2. 2.7. Experimental conditions All the fuels were gasified at similar conditions (Table 3). Despite some variations, the experimental conditions of each
Table 2 Ash components in oxidized form measured by X-ray fluorescence. Possibly catalytically active components are highlighted.
Fig. 3. Comparison of NHþ 4 concentration measured in the absorption solution by distillation and titration and by ion chromatography.
(wt.%)
Spanish coal
South African Coal
SiO2 A12O3 Fe2O3 MgO CaO Na2O K2O MnO TiO2
58 31 5.2 1.0 1.1 0.41 2.4 0.10 0.71
52 38 3.1 0.84 3.4 0.42 0.73 n.d. 1.8
Table 1 Proximate and ultimate fuel analysis. Fuel (raw)
Spanish coal
South African coal
Rhenish coal
Biomass
Water (% wt) Total combustibles (% wt) Ash (% wt) Volatile combustibles (% wt)
5.9% 47% 47% 18%
5.9% 83% 11% 26%
12% 85% 3.9% 44%
9.8% 89% 0.88% 75%
Fixed carbon (% wt.)
29%
57%
41%
15%
HHV (MJ/kg) LHV (MJ/kg)
14 14
27 26
23 22
18 16
C H N O
36% 2.4% 0.77% 7.2%
66% 3.9% 1.7% 11%
56% 4.3% 0.71% 23%
44% 5.4% 0.17% 40%
S Cl F Br
0.82% 0.040% 0.012% n.a.
0.40% 0.023% 0.013% n.a.
0.22% 0.036% 0.0042% 0.0008%
0.017% 0.011% 0.0030% n.a.
922
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
Table 3 Experimental conditions (the gasses volume is stated at 25 °C and 101 325 Pa. Fuel
Spanish Coal
South African Coal
Rhenish coal
Biomass
Gasifying agent Material of the fluidized beda Reactor temperature (°C) Dry fuel feeding rate (g h1) Raw fuel feeding rate (g h1)
H2O + O2 Sand 850 ± 5 1100 1170
H2O + O2 Sand 850 ± 5 1080 1150
H2O + O 2 Sand s + l(l:l) 850 ± 5 1070 1140 1200 1300
H2O + O2 Sand s + l(l:l) s 850 ± 5 1280 1320 1420 1450
Steam (m3h1) CO2(m3h1) O2(m3h1) additional N2(m3h1) N2 pneumatic transport (m3/h Total gas inlet (m3/h)
1.36 – 0.19 0.45 1.50 3.50
1.36 – 0.34 0.30 1.50 3.50
1.36 – 0.30 0.34 1.50 3.50
1.36 – 0.33 0.31 1.50 3.50
1.36 – 0.28 0.37 1.35 3.36
ER H2O/C (mol/mol) CO2/C (mol/mol)
0.20 1.70 –
0.20 0.94 –
0.20 1.13 –
0.21 1.07 –
0.22 1.22 –
1290 1430
CO2 + O2 s + l(l:l) 850 ± 5 1320 1450
1.36 – 0.27 0.37 1.50 3.50
1.36 – 0.27 0.37 1.50 3.50
0.20 1.17 –
0.20 1.21 –
s + 1 (3:1)
1290 1430
CO2 + H2O + O2 s + l(l:l) s+1(3:1) 850 ± 5 1320 1290 1450 1430
– 1.36 0.27 0.37 1.50 3.50
– 1.35 0.27 0.37 1.50 3.50
0.41 0.91 0.27 0.37 1.50 3.46
0.41 0.95 0.27 0.37 1.50 3.55
0.20 0.13b 1.04
0.20 0.15b 1.06
0.20 0.45 0.69
0.20 0.47 0.74
s + 1(3:1)
Gasifying agent
a b
s = s and, l = limestone. The value is higher than zero, because some moisture is coming to the reactor within the fuel.
experiment can be assumed to be almost identical; hence, the ammonia content and yield can be safely compared. The experiments with Spanish coal and South-African coal were performed only with (inert) silica sand in the fluidized bed. To test the effect of adding catalytic material, we performed experiments with Rheinish coal and biomass with the mixture of silica sand and dolomitic limestone. Two different ratios (1:1 and 1:3) of dolomitic limestone and silica sand in the fluidized bed were tested with biomass as the fuel. The detailed composition and properties of the fluidized bed materials can be found elsewhere [25]. 3. Results Ammonia yield is closely connected with the producer gas composition and with the total fuel-to-gas conversion. Thus, we first introduce a summary of the yield of main gasses from different fuels with different ratios of dolomitic limestone in the fluidized bed and with different gasifying agent mixtures (Fig. 4). In the same chart, the carbon conversion is depicted. To give a detailed overview of the relative partial pressures of individual gasses in the producer gas mix, its composition is presented in Fig. 5. This dry and N2-free composition together with the LHV of the gas (also dry and N2-free) is a basic technical parameter, which is included mainly for a potential future comparison with data from other sources. And, finally, the concentration of the ammonia in the dry and N2-free producer gas is depicted in Fig. 6 alongside the conversion rate of fuel-nitrogen to ammonia. This value is independent of the actual gas yield and it can be used for basic estimation of the ammonia yield from a specific fuel under specific conditions. 4. Discussion 4.1. Ammonia yield from gasification of different types of coal and biomass As was mentioned in the introduction, the nitrogen in coals is stored mainly in forms belonging to the non-volatile part of the fuel. Therefore, the rate of NH3 yield from coal gasification should be related to the rate of char gasification. Non-reactive coals are supposed to have a very low rate of conversion of fuel-nitrogen into ammonia. However, in the case of Spanish Coal and South African coal (presented in the charts of Figs. 4 and 6) this did not
happen. The carbon conversion of Spanish coal was only 41% and, in the case of South African coal, only 39%, which is a very low value compared to the other studied fuels (see Fig. 4). The high rate of conversion of fuel nitrogen into ammonia during the gasification of these two fuels (Spanish Coal and South African coal; Fig. 6) can be attributed to any of the following: (1) the distribution of fuel-N into volatile and non-volatile parts of the fuel could be different from the assumptions stated in the introduction. (2) There could be a high content of catalytically active components in the ash (Ca, Fe), or (3) the retention time of the char in the fluidized bed is quite long. The content of the possibly catalytically active components in the ash was quite low (see Table 2) and the ammonia yield slightly increased during the experiment. Therefore, we are inclined towards the third explanation: the relatively high conversion of fuel-N into ammonia was probably caused by a long retention time of the char and by its accumulation in the fluidized bed with the release of the nitrogen from its structures. Concerning the reactive coal (Rheinish coal), the measured conversion of fuel-N into ammonia is in accordance with our assumptions. High conversion of the char (the carbon conversion was 78% on the sand fluidized bed and 89% on the mixture of limestone and sand in the fluidized bed) is with accordance with the high conversion of the fuel-nitrogen into ammonia (see Fig. 6). The conversion of biomass nitrogen into ammonia ranged from 43% to 57%, which is a lower value than reported in the literature. On the other hand, the biomass we used is not a typical fuel for gasification and contains a very small amount of nitrogen. 4.2. Dolomitic limestone in the fluidized bed We examined the effect of adding dolomitic limestone into the fluidized bed when gasifying the reactive coal and biomass. The addition had a positive impact on the conversion of fuel-N into ammonia, both in the case of Rheinish coal and wood; however, the use of different gasifying media had a definite influence on the ammonia yield. If we focus only on the use of H2O + O2, the presence of the dolomitic limestone in the fluidized bed (with the ratio 1:1 of sand to limestone) increased the conversion rate of fuel-N into NH3 by approx. 10% (see Fig. 6). This fact is in accordance with Pinto et al. [11] and suggests that the dolomitic limestone improves the conversion rate of fuel-N into NH3 by affecting the whole gasification process including the char conversion, steam conversion, gas composition etc.
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
Fig. 4. Raw producer gas yield and fuel-carbon conversion into producer gas.
Fig. 5. Producer gas composition and LHV (dry and N2-free).
Fig. 6. Concentration of NH3 and the rate of conversion of fuel nitrogen into ammonia in producer gas.
923
924
M. Jeremiáš et al. / Fuel 117 (2014) 917–925
Moreover, it can catalyze the transformation of other nitrogenous compounds into ammonia. Limestone can improve the decomposition reactions (as is reported by Simell et al. [26]), but this decomposition is less influential than the above-mentioned effect. Interesting is also the effect of increasing the limestone concentration in the fluidized bed from 25% to 50%. In the case of gasification with H2O + O2, such an increase had only a negligible effect; however, in the case of gasification with CO2 + O2, the rate of nitrogen conversion increased by 10%, and, in the case of CO2 + H2O + O2, even by 14%. It seems that for every gasifyingagent mixture there exists a specific minimal concentration of limestone in the fluidized bed that increases discretely the ammonia yield. 4.3. The use of different gasifying media The use of CO2 instead of H2O in the gasifying agent had a different effect when using 25% of dolomitic limestone in the sand fluidized bed and when using 50% of limestone. Thus, the effect of different gasifying agent is closely connected with the catalytic activity of the material of the fluidized bed as is mentioned in the Section 4.2. For the fluidized bed of sand and limestone (3:1), the conversion of fuel–N into ammonia decreased by 8%, when fully substituting the H2O by CO2, and even by 11% in the case of partial substitution of H2O by CO2 (see dashed line in Fig. 6), which is in accordance with the expectations stated in chapter 1.3. On the other hand, for a catalytically stronger material of the fluidized bed (the ratio of 1:1), the conversion only slightly increased with the substitution of H2O by CO2 (full line in Fig. 6). We suggest that there exists some critical mass of limestone in the fluidized bed. This critical mass of limestone that caused a discrete jump of ammonia yield by almost 10% was achieved by using a mixture of sand and limestone with the ratio of 3:1 when H2O + O2 was used as gasifying agent. However, when CO2 was used as gasifying agent, this boundary was crossed somewhere in the range of concentration 25–50% of limestone in the fluidized bed. 5. Conclusion We measured the conversion rate of fuel-bound nitrogen into ammonia during gasification of Spanish coal, South African coal, Rheinish coal and woody biomass. The rate was surprisingly high for less-reactive coals (61% for Spanish coals and 52% for South African Coal). For the reactive Rheinish coal, the rate was 74–83% and for wood biomass it was 43–57%. The conversion of fuel-N into ammonia increased substantially (by approx. 10%) when the gasification was catalyzed by dolomitic limestone in fluidized bed. When steam in the gasifying agent was substituted with carbon dioxide, there was a different effect on the rate of fuel-N to ammonia conversion when 25% of limestone and 50% of limestone was in the fluidized bed. However, in both cases the concentration of ammonia in the dry gas was diminished because of the dilution by CO2. In this study, the variation of composition of the gasifying agent was examined for a fuel with low content of nitrogen (0.17 wt%) and different results for different concentrations of limestone in the fluidized bed were obtained. Further work on this theme should be directed towards the use of fuels with higher content of nitrogen and on the clarification of the use of CO2 as a gasifying agent. Nomenclature All volume units of the gas and other relevant values are reported at 101.325 kPa and 25 °C.
Dry gas yield is the volume of the gas (at 101.325 kPa and 25 °C) related to mass flow of dry fuel (1). Total gas yield was calculated from the N2 balance (known volume in the inlet and known concentration in the dry gas at the outlet). The yields of specific gases were calculated from the total gas yield multiplied by their concentrations in the dry producer gas.
gas yield ¼ V_ dry
_ dry fuel gas =m
ð1Þ
Steam yield was calculated from H and O elemental balances. Lower heating value (LHV) of the produced gas was calculated as the sum of products of heating value of measured gas components (LHVi) and their concentrations (ui) in the producer gas, according to the equation (2). The lower heating values of the gas components were adopted from the norm EN ISO 6976. The tarry compounds were not included in the computation.
LHV ¼
n X
ui LHV i
ð2Þ
i¼1
Cold gas efficiency is the ratio of the chemical energy of the dry producer gas (based on LHVd) to the chemical energy stored in the dry fuel (LHVd). Equivalence ratio (ER) is the molar ratio of the total input of O2 by the gasifying agent related to O2 theoretically needed for total combustion of the fuel into CO2 and H2O (and SO2, etc.; the real oxygen demand will be higher because of thermodynamic and kinetic limitations). For example, the ER of 0.2 means that 20% of the oxygen theoretically needed for total combustion is fed into the reactor. Steam to fuel carbon ratio (H2O/C) is the molar ratio of steam fed into the reactor by gasifying agent and/or as the fuel moisture related to the elemental carbon in the fuel. Carbon dioxide to carbon ratio (CO2/C) is the molar ratio of carbon dioxide fed into the reactor by gasifying agent related to elemental carbon in the fuel.
Acknowledgements The authors appreciate the financial support from EU Research Fund for Coal and Steel, Grant RFCR-CT-2010-0009 ‘Fecundus’, The support of Ministry for Education, Youth and Sport of the Czech Republic, Projects Nos. 7C11009 and 6046137304 and specific university research projects Nos. 20/2013, 21/2012, 21/2011 and 21/ 2010. Appendix A. Supplementary material Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.fuel.2013.10.009. References [1] Pröll T, Siefert IG, Friedl A, Hofbauer H. Removal of NH3 from biomass gasification producer gas by water condensing in an organic solvent scrubber. Ind Eng Chem Res 2005:1576–85. [2] Hämäläinen JP, Aho MJ, Tummavuori JL. Formation of nitrogen oxides from fuel-N through HCN and NH3: a model-compound study. Fuel 1994;73:1894–8. [3] Svoboda K, Hartman M. Formation of NOx in fluidized bed combustion of model mixtures of liquid organic compounds containing nitrogen. Fuel 1991;70:865–71. [4] Kelsall GJ, Smith MA, Cannon MF. Low emissions combustor development for an industrial gas turbine to utilize LCV fuel gas. J Eng Gas Turbines Power 1994;116:559. [5] Higman C, Burgt M van der. Gasification. Gulf Professional Publishing; 2008. [6] Rabou LPLM, Zwart RWR, Vreugdenhil BJ, Bos L. Tar in biomass producer gas, the energy research center of The Netherlands (ECN) experience: an enduring challenge. Energy Fuels 2009;23:6189–98.
M. Jeremiáš et al. / Fuel 117 (2014) 917–925 [7] Leppälahti J, Koljonen T. Nitrogen evolution from coal, peat and wood during gasification: literature review. Fuel Process Technol 1995;43:1–45. [8] Wilk V, Hofbauer H. Conversion of fuel nitrogen in a dual fluidized bed steam gasifier. Fuel 2013;106:793–801. [9] Devi LG, Ptasinski KJ, Janssen FJJG. A review of the primary measures for tar elimination in biomass gasification processes. Biomass Bioenergy 2003;24:125–40. [10] Leppälahti J. Formation and behavior of nitrogen compounds in an IGCC process. Bioresour Technol 1993;46:65–70. [11] Pinto F, Lopes H, André RN, Gulyurtlu I, Cabrita I. Effect of catalysts in the quality of syngas and by-products obtained by co-gasification of coal and wastes. 1. Tars and nitrogen compounds abatement. Fuel 2007;86:2052–63. [12] Devi LG, Ptasinski KJ, Janssen FJJG, van Paasen SVB, Bergman PC, Kiel JH. Catalytic decomposition of biomass tars: use of dolomite and untreated olivine. Renew Energy 2005;30:565–87. [13] Tunå P, Brandin J. Selective catalytic oxidation of ammonia by nitrogen oxides in a model synthesis gas. Fuel 2013;105:331–7. [14] Simell PA, Kurkela EA, Ståhlberg P, Hepola JO. Catalytic hot gas cleaning of gasification gas. Catal Today 1996;27:55–62. [15] Leppälahti J, Simell PA, Kurkela EA. Catalytic conversion of nitrogen compounds in gasification gas. Fuel Process Technol 1991;29:43–56. [16] Chang L, Xie Z, Xie K-C, Pratt KC, Hayashi J, Chiba T, et al. Formation of NOx precursors during the pyrolysis of coal and biomass. Part VI. Effects of gas atmosphere on the formation of NH3 and HCN. Fuel 2003;82:1159–66.
925
[17] Pohorˇely´ M, Svoboda K, Hartman M. Feeding small quantities of particulate solids. Powder Technol 2004;142:1–6. [18] Hartman M, Pohorˇely´ M, Trnka O. Transport velocities of different particulate materials in pneumatic conveying. Chem Pap 2006;60:74–7. [19] Pohorˇely´ M, Vosecky´ M, Hejdová P, Puncˇochárˇ M, Skoblia S, Staf M, et al. Gasification of coal and PET in fluidized bed reactor. Fuel 2006;85:2458–68. [20] Svoboda K, Pohorˇely´ M, Jeremiáš M, Kameníková P, Hartman M, Skoblia S, et al. Fluidized bed gasification of coal–oil and coal–water–oil slurries by oxygen– steam and oxygen–CO2 mixtures. Fuel Process Technol 2012;95:16–26. [21] Šyc M, Pohorˇely´ M, Jeremiáš M, Vosecky´ M, Kameníková P, Skoblia S, et al. Behavior of heavy metals in steam fluidized bed gasification of contaminated biomass. Energy Fuels 2011;25:2284–91. [22] Yu Q-Z, Brage C, Chen G-X, Sjöström K. The fate of fuel-nitrogen during gasification of biomass in a pressurised fluidised bed gasifier. Fuel 2007;86:611–8. [23] Norton GA, Brown RC. Wet chemical method for determining levels of ammonia in syngas from a biomass gasifier. Energy Fuels 2005;19:618–24. [24] Seville JPK. Gas cleaning in demanding applications. Blackie Acad Prof 1997. [25] Hartman M, Svoboda K, Pohorˇely´ M, Šyc M, Jeremiáš M. Attrition of dolomitic lime in a fluidized-bed at high temperature. Chem Pap 2013;67:164–72. [26] Simell PA, Hepola JO, Krause AOI. Effects of gasification gas components on tar and ammonia decomposition over hot gas cleanup catalysts. Fuel 1997;76:1117–27.