- Email: [email protected]
The fate of fuel-nitrogen during gasification of biomass in a pressurised fluidised bed gasifier
Fuel 86 (2007) 611–618 www.fuelfirst.com
The fate of fuel-nitrogen during gasification of biomass in a pressurised fluidised bed gasifier Q-Z. Yu *, C. Brage, G-X. Chen 1, K. Sjo¨stro¨m Department of Chemical Engineering and Technology/Chemical Technology, KTH – Royal Institute of Technology, SE-100 44 Stockholm, Sweden Received 6 December 2005; received in revised form 7 August 2006; accepted 8 August 2006 Available online 8 September 2006
Abstract The distribution of fuel-nitrogen in gases, tar and char from gasification of biomass in a pressurised fluidised bed gasifier was investigated. Four species of biomass: birch, Salix, Miscanthus and Reed canary grass were gasified at 0.4 MPa and 900 C. Oxygen-enriched nitrogen was used as fluidising agent. As a reference, gasification of Daw Mill coal was also carried out under the same experimental conditions. The experimental results illustrate that both the nature of the original fuels and the chemical structure of the nitrogen in the fuel have influence on the distribution of fuel-nitrogen in gases (NH3, HCN, NO), tar and char under the employed experimental conditions. The present work also shows that the types of nitrogen heterocyclic compounds (NHCs) in the tar from different kinds of biomass are the same and the major compound is pyridine. However, the distribution of the various NHCs in the tar from the four species of biomass varies: the higher the content of fuel-nitrogen, the higher the concentration of two-ring NHCs in the tar. An effective method for extracting NHCs from the acidic absorption of the product gas was introduced in the present work. The method makes use of solid phase extraction (SPE) by a silica-based C18 tube to extract the NHCs which subsequently were analysed by gas chromatography (GC) with flame ionisation detection (FID). The recovery and reproducibility of the SPE technique for NHCs is discussed. 2006 Elsevier Ltd. All rights reserved. Keywords: Gasification; Biomass; Nitrogen compounds
1. Introduction Currently, biomass accounts for 14–15% of worldwide energy consumption. The biomass combustion process yields heat, which is used to fulfil basic energy requirements for homes and various industries in developing countries [1]. The biomass energy plantation and conversion systems are an effective source of renewable energy and have had a substantial impact on reducing the global CO2 emission budget [2]. Since energy from biomass is an effective substitute for fossil fuels, it has also been widely used in devel*
Corresponding author. Tel.: +46 8 7908252; fax: +46 8 108579. E-mail address: [email protected] (Q-Z. Yu). 1 Dr. G-X. CHEN worked on all the gasification experiments and reviewed the paper. He passed away two years ago. This paper is dedicated to the memory of Dr. G-X. Chen for his contribution to the biomass thermochemical research field. 0016-2361/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2006.08.007
oped countries over the last century [3]. In recent years, the decreasing availability of fossil fuels has become a growing concern and using fossil fuels is faced with stricter environmental regulations [4]. Therefore, a much greater interest in the use of biomass energy as an alternative to fossil fuels has developed. Among the biomass thermochemical conversion processes, pyrolysis, gasification and combustion, biomass gasification should be explored further since biomass gasification can form a solid basis of new promising energy conversion technologies. Among these are fuel cells and integrated gasification combined cycle (IGCC) installations. Conditioning and upgrading the biomass gasification product gas can make it a suitable feed for methanol or Fischer–Tropsch liquid synthesis [5]. During solid fuel thermochemical conversion the fuel-nitrogen is converted to ammonia (NH3), cyanide (HCN), isocyanic acid (HNCO), nitrogen oxides (NOx and N2O), molecular nitrogen (N2) and organic nitrogen.
612
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
Part of the fuel-nitrogen remains in the char. NH3, HCN, HNCO and organic nitrogen in the product gas pass through end-use systems, where they can poison catalysts or fuel cell internals. It is well known that NH3, HCN, HNCO and organic nitrogen are the precursors to the emissions of nitrogen oxides (NOx and N2O). Nitrogen oxides are pollutants causing photochemical smog, acid rain, the greenhouse effect and ozone depletion [6]. Therefore an intensive study on the formation of the precursors of nitrogen oxides during fuel thermochemical conversion is going on in many countries. However, the mechanisms for the formation of the precursors of nitrogen oxides during thermochemical conversion processes of solid fuels are still not clear, especially in the case of biofuels. The purpose of the present work is to investigate the correlation between the characteristics of the raw materials and the distribution of fuel-nitrogen in products (disregarding molecular nitrogen) under the studied experimental conditions. Two cultivated energy crops, Reed canary grass and Miscanthus, and a fast growing energy forest specie, Salix, as well as birch wood were employed. Daw Mill coal of UK origin was used as a reference. The factors influencing fuelnitrogen conversion to the gases (HCN, NH3, NO) or retention in the char as well as the distribution of individual NHCs in the tar are discussed in brief. A solid phase extraction (SPE) method for extracting the NHCs from the acidic absorption of the product gas was established and is presented.
cals for preparation of calibration standards and internal standards were used as received, i.e. at a minimum purity of 98–99% (Sigma–Aldrich, Sweden). SPE ENVI–C18 tubes (Supelco, Inc., Bellefonte, US) required for sample preparation were loaded with 500 mg of octadecyl-bonded endcapped silica (surface area 400–600 m2 g1, average ˚ ). particle size 50–60 lm and average pore diameter 60–70 A 2.2. Sampling
2. Experimental
Fig. 1 shows the sampling system for gaseous and liquid nitrogen compounds. The first two traps containing 0.1 M H2SO4 (2 · 40 ml) were used for sampling of both NH3 and NHCs. HCN was absorbed in the following four bottles of which two contained 0.1 M NaOH (2 · 50 ml in the second and the third bottle). The last four traps containing dichloromethane (DCM) were used to wash out the residual tar from the gas flow before the gas entered the volume measurement meter. A stream of the product gas from the top of the gasifier passed through a bypass pipe, which was warmed up to about 250 C, and bubbled through the sampling system. The gas volume measurement meter was placed at the end of the sampling system to record the total volume of sampled gas. The gas flow was controlled at about 5 l min1. The sampling time was about 1 h. A char sample was collected at the end of each experiment (about 1 h) after the temperature and the pressure in the gasifier had been reduced to room temperature and atmospheric pressure.
2.1. Materials
2.3. Sample preparation
The feedstocks used were four species of biomass (1– 3 mm, 5.3% moisture): birch, Salix, Miscanthus, Reed canary grass, and Daw Mill coal of UK origin (1– 1.35 mm, 7.3% moisture). The chemical composition of the fuels is listed in Table 1. The sample source was a pressurised fluidised bed reactor (silver sand as bed material) operated at 900 C and 0.4 MPa. Oxygen-enriched nitrogen was used as fluidising agent. The concentration of oxygen in the fluidising agent controlled the temperature of the bed in the experiments [7,8]. The solvent required for sample preparation, chemi-
The acidic solution in the first two impingers was extracted with a small portion of DCM in order to remove phenolic and neutral tar fractions. The water phase was poured into a funnel equipped with a folded filter paper. The filtrate was then transferred into a 100 ml volumetric bottle and brought to volume with 0.1 M H2SO4. A SPE method was used to remove any trace of organic compounds from the filtrate in order to protect the electrodes that were used for measuring ammonia. A SPE ENVI-C18 tube was conditioned with a 1 ml methanol followed by three tube fillings of distilled water. A part of the
Table 1 The chemical composition (wt.%, dba) of the fuels Fuel
C
H
N
O
S
Ash
Fixed carbon
Volatiles
Heating value (MJ/kg,dafb)
Birch Salix Miscanthus Reed canary grass Daw Mill coal
49.0 48.1 47.9 44.2 73.8
6.1 6.2 6.3 6.0 4.9
0.1 0.4 0.7 0.8 1.3
44.5 43.9 43.8 41.7 12.3
<0.01 0.036 0.11 0.21 1.63
0.4 1.6 2.8 8.8 6.5
n.d.c 14.9 n.d. n.d. 58.5
n.d. 82.6 78.5d 73.5d 35.6
19.3 19.2 17.0 17.6 30.4
a b c d
db: dry basis. daf: dry and ash free. n.d.: not determined. Source: Ref. [40].
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
3
4
7
613
8
11
2 5
1
10
6 9
Fig. 1. Sampling system for nitrogen compounds in product gas: (1) Heated line (200–250 C) to sampling needle valve (stainless steel); (2) glass tee (200– 250 C); (3) NH3 and NHCs absorption bottles immersed in ice-water; (4) HCN absorption bottles; (5) water-cooled super condenser; (6) round bottom flask; (7) tar impinger bottles (40 ml DCM in each) immersed in CO2/C2H5OH–H2O; (8) tar impinger bottles (40 ml DCM in each) immersed in solid CO2; (9) soap gas-flow meter; (10) integral gas volume meter; (11) ventilator.
filtrate was taken (10–60 ml, the volume depending on the NH3 concentration in the filtrate) and passed through the conditioned SPE ENVI-C18 tube with the help of a water pump. An aliquot of the elution (5–50 ml) was transferred to a 50 ml volumetric bottle by a volumetric pipette and the solution brought to volume with 0.1 M H2SO4 (if needed). The acidic solution was prepared for the NH3 analysis. In order to prepare the sample for determination of NHCs, another portion of the filtrate (5–10 ml) taken by a volumetric pipette was poured into an Erlenmeyer flask with a stopper; 2,4,6-trimethylpyridine (141 lg in 50 ll DCM) as an internal standard for GC analysis of NHCs was added to the flask using a syringe. Then the solution was adjusted to about pH 14 with 3 ml 1 M NaOH. Subsequently, the alkaline solution was slowly passed through a conditioned SPE ENVI-C18 tube (see above) with the help of a water pump. The flow rate should not exceed 5 ml/min (dropwise flow is best). After all of the water in the sample had eluted, the tube was additionally dried for 5 min with full vacuum by a water pump. The NHCs were then eluted by 3 · 500 ll of DCM/iso-propanol (1:1, V/V). Finally, the elution was dried by passing through a small column containing 3 g anhydrous Na2SO4. If there was any water visible in the eluate, it was cautiously removed before using the anhydrous Na2SO4 column. The alkaline solution in the four bottles in the gas washing system was filtered and brought to a given volume with 0.1 M NaOH in a volumetric bottle. This alkaline solution was then used for HCN analysis.
2.4. Analysis NH3 was determined using an ISE (ion-selective electrode). The prepared 50 ml sample was poured into a 100 ml beaker containing a magnetic spinning bar and 3 ml of pH-adjusting ISA (Orion Research INC., USA) was added. The treated sample was then immediately measured by a pH/ISE instrument, model 710A (UNICAM/ Orion), using an ISE gas (NH3) sensing electrode (cali-
brated with 1–500 ppm standard NH3) and an ATC probe (automatic temperature compensation). It should be noted that the NH3 yields quantified in this study may also include some contribution from the hydrolysis (in the absorption solution) of HNCO which has been observed as a nitrogen-containing species from the pyrolysis of coal and biomass [9,10]. However, the contribution to NH3 from the hydrolysis of HNCO in the bubbler may be very small because the yield of HNCO in the product gas was always much smaller than the yields of HCN and NH3 [11]. Furthermore, De Jong has stated that HNCO was never detected even by means of a high resolution FTIR spectrophotometer from the pressurised gasification of biomass under Delft experimental conditions [12]. A GC with FID was used to analyse the NHCs. The instrument (CHROMPACK CP9010) was equipped with an automatic liquid sampler (CHROMPACK CP 9002). GC control and data handling was achieved with a PCbased chromatography program. A constant flow of helium carrier gas through the column (CP-Sil 8 CB, 25 m · 0.25 mm ID, 0.5 lm film thickness) was automatically maintained using electronic pressure control. The temperatures of the injector (split 53:1) and detector were maintained at 300 C. The oven was programmed to run from 80 C (3 min) to 280 C (3 min) at 8 C min1; a 1 ll sample was injected. HCN was measured calorimetrically using Merck Microquant reagent M14798. According to the literature [11], the high solubility of HCN in the acidic solution could cause an underestimation of HCN yield when the product gas passed through the sampling system. In comparison the HCN yield is much lower than the NH3 yield under the studied experimental conditions (Table 2), therefore error is negligible. Nitrogen remaining in the char was analysed by a professional laboratory, MIKRO KEMI AB, using a so-called MK2003 method. The char sample was incinerated in an oven at a temperature of around 1000 C. Briefly, the nitrogen and nitrogen compounds were analysed by a GC with packed column and TCD detector.
614
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
Table 2 The distribution of fuel-nitrogen in the gases, tar and char Raw material
Reed canary grass
Miscanthus
Salix
Daw mill coal
% of fuel N in char % of fuel N in NH3 % of fuel N in HCN % of fuel N in NO % of fuel N in NHC The ratio (%) of N(HCN) to N(NH3)a The ratio (%) of N(NO) to N(NH3)b
0.7 34.3 0.10 0.20 1.30 0.29
9.4 12.7 0.12 0.04 0.37 0.94
0.0 24.4 0.22 0.66 0.67 0.90
34.1 7.5 0.10 0.10 n.d. 1.33
0.58
0.31
2.70
1.33
a b
The ratio of distribution of fuel-nitrogen between HCN and NH3. The ratio of distribution of fuel-nitrogen between NO and NH3.
3. Results and discussion The distribution of fuel-nitrogen between the gases (excluding molecular nitrogen), tar and char from gasification of three species of biomass and Daw Mill coal at 900 C and 0.4 MPa using oxygen-enriched nitrogen as fluidising agent is given in Table 2. 3.1. Fuel-nitrogen converted to NH3 Different pathways leading to the formation of NH3 during pyrolysis of solid fuels have been suggested: • decomposition of amines (proteins and amino acids) [13–20]; • thermal cracking reactions of tar and char [11,14,16,21– 25]; • hydrogenation and hydrolysis of HCN [21,26,27]. Inspection of the data in Table 2 shows that a larger percentage of nitrogen in the three species of biomass forms NH3 than does the nitrogen in Daw Mill coal. Nitrogen is bound to solid fuel in various ways. Studies by X-ray photoelectron spectroscopy (XPS) suggest that nitrogen in coal is present in heterocyclic structures, predominantly of the pyrrolic and pyridinic types [19,25]. In contrast to coal, nitrogen in wood is mainly in the form of protein and free amino acid [28]. According to the literature, the amine groups are believed to form NH3 directly at relatively low temperatures during pyrolysis [13]. Since biomass has a larger share of nitrogen as amines than coal has, there should be no doubt that the higher content of amine in biomass favours the production of NH3 in biomass gasification. Aho et al. have suggested that the O/N ratio of the fuel might act as an indicator of the ratio of nitrogen bound in side-chains of the aromatic structures in the fuel. Much of the nitrogen in the side-chains may exist in amino form [29]. From Table 1 it may be seen that biomass has a higher O/N ratio than coal, indicating that the share of the amino form of the nitrogen in biomass is higher than that of the
nitrogen in coal. As the amino compounds exhibit the highest conversion to NH3 among the three types of fuel-nitrogen compounds, pyridinic type, pyrrole type and amino compounds [19], this also supports the observed larger percentage conversion to NH3 for nitrogen from the three species of biomass as compared to nitrogen from Daw Mill coal. The tar from thermochemical processing of solid fuels has some opportunity to undergo secondary gas-phase cracking thereby releasing the nitrogen into the gas phase, primarily as HCN, NH3, HNCO [25]. In general, high temperature secondary pyrolysis reaction of tars yield HCN as primary product, and HCN may act as a precursor for NH3 [14,30]. Researchers have observed that NH3 could be formed from the HCN hydrolysis and the reaction of HCN with hydrogen although it may not be the main way to convert fuel-nitrogen to NH3 [21,23,26,27]. There are several major ways for the generation of active hydrogen and free radicals in the char: the substituting groups might be broken off from the (hydro) aromatic ring systems; dehydrogenation of the aliphatics or dehydrogenation of hydroaromatic ring systems and the condensation reactions between the aromatic ring systems [22]. Since biomass contains more aliphatic groups than coal does, the concentration of H radicals from the gasification of biomass could be higher than that from the gasification of coal. Therefore the heterogeneous hydrogenation of HCN forming NH3 on the surface of the char from three species of biomass could be favoured under our experimental conditions. Thermal cracking of char may also be an important course for NH3 formation. After scrutiny of the experimental data of several investigators, Hill et al. concluded that the appearance of fuel-nitrogen as NH3 in the gas phase might be important during char oxidation [14]. Xie et al. described that a large amount of NH3 was formed by thermal cracking of nascent char based on their experiment results [24]. Li et al. believed that direct hydrogenation of the nitrogen compounds in the pyrolysis of char particles is the main route for the formation of NH3. They explained that the formation of NH3 from the thermal cracking of char could be considered as a ‘‘self-gasification’’ process of the char by the active hydrogen generated in situ within the char [22]. The presence of H atoms and reactive radicals in the reacting solid fuels/char mass enhances the decomposition of N-containing heteroaromatic ring systems, resulting in the release of simple gas-phase nitrogen-containing species [31]. These observations seem consistent with the results of the present study. Table 2 illustrates that the percentage of fuel-nitrogen in NH3 has a strong inverse relation to the percentage of fuel-nitrogen remaining in the char. The formation of NH3 from secondary radical reactions in the char might be controlled by the thermochemical reactivity and the surface area of the char as well as the concentration of the H radicals [22,24]. Chars from gasification of biomass have higher thermochemical reactivity, larger surface area and a larger concentration of possible
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
H radicals within the char precursors than the chars from coal [8,32,33]. These factors may explain the higher distribution ratio of fuel-nitrogen in NH3 obtained from the gasification of three species of biomass, compared with the gasification of Daw Mill coal under our experimental conditions. However, the mechanism of NH3 formation during thermochemical conversion of solid fuels, especially biomass, is very complicated and no clear picture is available. Therefore further study is necessary. 3.2. The ratio of fuel-nitrogen in HCN to fuel-nitrogen in NH3 The fuel-nitrogen released in the volatiles from the primary pyrolysis of the fuels subsequently undergoes secondary reactions resulting in the formation of the NOx precursors such as HCN and NH3. It has been observed that the distribution of different nitrogen functional groups in the tar may lead to different rates of the secondary reactions and different products [31,34]. The thermally more stable NHCs may be hydrogenated slowly by radicals to NH3 while the thermally less stable NHCs are mainly responsible for the formation of HCN [22,34]. In Table 3, the data illustrate that the distribution of individual NHCs in the biomass tars. The order of the ratio of two-ring NHCs to one-ring NHCs in the biomass tar is Reed canary grass > Miscanthus > Salix > birch according to the results of the present study. As two-ring NHCs are more stable than one-ring NHCs, a tar containing a higher concentration of two-ring NHCs is expected to yield a larger amount of NH3. Meanwhile the nature of the char also has a strong link to the conversion of fuel nitrogen to NH3, as mentioned in the previous section. Comparing the experimental results from the gasification of three species of biomass (Reed canary grass, Miscanthus, Salix) in Tables 2 and 3, it may be seen that the gasification of Reed canary grass produced the tar with the highest ratio of two-ring NHCs to one-ring NHCs, and a char with a low percentage of fuel-nitrogen. Therefore the Table 3 The distribution of NHCs in the tar (relative mol.%) Compound
Pyridine 2-Picoline 3-Picoline 2-Ethylpyridine Lutidine 3-Ethylpyridine Quinoline Iso-quinoline The ratio of two-ring NHCs to one-ring NHCs
615
ratio of N(HCN) to N(NH3) is the lowest in the product gas from gasification of Reed canary grass among the three species of biomass (see Table 2). 3.3. The ratio of fuel-nitrogen in NO to fuel-nitrogen in NH3 The present work also showed that among the three species of biomass Salix gave the highest ratio of fuel-nitrogen in NO to fuel-nitrogen in NH3 when they were gasified under the same conditions (Table 2). It is well known that nitrogen is absorbed by the roots of plants in the form of þ þ NO 3 and NH4 . NO3 and NH4 then move from the roots þ to leaves and young stems. Portions of NO 3 and NH4 also become stored in the xylem as they are transported through þ the xylem tissue. However the bulk of NO 3 and NH4 are deposited in leaves. NO3 is converted to amino acids or proteins and stored in the leaves and stems. A small portion of fuel-nitrogen is found in the xylem, predominantly þ as NO 3 and, to a lesser extent, as NH4 . Leaves contain þ more NH4 , and xylem contains more NO 3 . According to the observation of Masutani et al., the amount of nitrate nitrogen ðNO 3 Þ in the feedstock will impact NO yield [35]. Since the fuel-nitrogen in Salix leads to a larger share of NO than the two energy crops during gasification it could be suspected that Salix may contain more NO 3 than the two energy crops. 3.4. The distribution of NHCs in the biomass tar The distribution of NHCs in the tar of four species of biomass is shown in Table 3. The results indicate that the types of the NHCs were similar and that the major compound was pyridine from the gasification of the four species of biomass under the employed experimental conditions. The order of the concentration of two-ring NHCs in the tar is Reed canary grass > Miscanthus > Salix > birch. It seems that the order of the content of nitrogen in the four species of biomass has the same tendency as the concentration of two-ring NHCs in the tar (see Table 1). The conclusion is that the higher the content of nitrogen in the biomass, the higher the concentration of two-ring NHCs in the tar. 3.5. Fuel-nitrogen in char
Raw material Birch
Reed canary grass
Miscanthus
Salix
82.6 8.3 5.9 0.0 0.0 0.0 3.2 0.0 3.3
73.9 4.8 3.7 0.0 0.0 0.0 16.0 1.6 21.4
69.8 8.3 7.5 0.0 0.8 0.0 11.3 2.2 15.6
84.6 3.9 3.3 0.0 0.0 0.0 6.1 2.1 8.9
Comparing the results in Table 2, it seems that the percentage of the nitrogen in three species biomass retained in the char is much lower than for the nitrogen in Daw Mill coal retained in the char. The phenomenon may be explained by the differences in the feedstocks; the different chemical constituents of the fuel-nitrogen in the raw materials. During devolatilisation, fuel-nitrogen is divided between char nitrogen and the volatiles, mainly in the tar and as NH3 and HCN [30]. Pyrolysis of coal by various techniques has shown that the fuel-nitrogen retention in the char is high [36]. Model studies on coal combustion have predicted that the distribution of the nitrogen in coal
616
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
between volatiles and char is linked to the nature of the coal and the combustion environment [14]. However, the extent of fuel-nitrogen released from the original fuel was roughly proportional to the content of volatile matter in the original fuel at the devolatilisation stage [37]. As shown in Table 1 the content of volatiles in three species of biomass is higher than in Daw Mill coal. It could therefore result in the nitrogen in three species of biomass being released with a larger proportion in the volatiles than for the nitrogen in Daw Mill coal at the devolatilisation stage. According to literature [15,34], the chemical structures of the nitrogen in the fuels are also an important factor influencing the evolution of organic nitrogen into the volatiles and char during secondary thermal conversion. It is well known that the different chemical structures of organic nitrogen compounds have different thermochemical decomposition temperatures. For example, pyrrole and pyridine in their pure forms are not significantly decomposed until about 1200–1300 K [16]. In contrast to this, most amino acids are thermally decomposed at 500 C (773 K) [17,18]. Even high molecular weight amines such as polyamide (nylon) is completely decomposed at 873 K and the main products detected below this temperature are carbon oxides, NH3, and water [38]. The constituents of the nitrogen in coal have been widely investigated. An XPS spectrum study on coal found that most of the nitrogen in coal occurs in pyrrolic and pyridinic structures [21,30]. The amount and the composition of nitrogen in biomass are affected by several factors, such as the types and species of plants, storage conditions and processing methods after harvest, etc. In general, most of the nitrogen in a plant is stored in the leaves and shoots in the form of amino acids [35]. However, almost 100% of the nitrogen in wood bark is in the pyrrolic form according to Ha¨ma¨la¨inen [19]. Since a considerable part of the nitrogen in biomass may be found as amines, it is understandable that the decomposition and release of biomass nitrogen is easier and faster than that of the nitrogen in coal during thermochemical conversion processes. Therefore the percentage of the nitrogen in three species of biomass retained in the char is much lower than that of the nitrogen in Daw Mill coal retained in the char. Fuel-nitrogen conversion during gasification in a pressurised fluidised bed is very complicated. The discussion in the present paper is limited to the characteristics of the original fuels. There are also many other factors influencing the fate of fuel-nitrogen during a thermochemical process. The mineral matter in the fuel can also play a role in the distribution of fuel-nitrogen between the individual products. Alkali, alkaline earth and transition metal cations could change the distribution of volatile products during coal pyrolysis. Ohtsuka has shown that small amounts of minerals (for example iron and calcium), probably in their ion-exchanged forms, dramatically change the fate of the nitrogen in coal and catalyse conversion of the nitrogen in coal to N2 [39]. Therefore an intensive and comprehensive investigation of the fuel-nitrogen chemistry in the
Table 4 The recovery and reproducibility of NHCs extracted using SPE C18 tube Compound
Mean value (lg)
Recovery (%)
Relative STDEVa
Pyridine 2-Picoline 3-Picoline 2-Ethylpyridine Lutidin 3-Ethylpyridine Quinoline Iso-quinoline
39.90 44.00 39.73 38.28 27.81 38.27 49.87 56.15
106.4 101.9 99.8 109.4 90.3 101.0 95.2 94.7
2.3 3.1 2.4 5.0 5.1 4.5 2.9 1.8
a
Relative standard deviation (%).
Table 5 The recovery of synthetic NHC samples with different concentrations Compound
A
B
C
D
Pyridine 2-Picoline 3-Picoline 2-Ethylpyridine Lutidine 3-Ethylpyridine Quinoline Iso-quinoline
107.5 104.6 102.5 109.4 93.9 105.9 93.9 96.8
108.6 111.1 109.9 110.6 101.5 105.8 89.7 109.1
110.3 105.6 111.2 113.2 104.8 107.2 94.9 110.1
99.1 99.5 101.6 101.7 100.6 107.0 93.4 95.5
Eluted compounds were analysed by capillary GC in triplicate. Sample A: concentration = 6.718 mg/ml, sample size 336 lg. Sample B: concentration = 13.59 mg/ml, sample size 680 lg. Sample C: concentration = 20.38 mg/ml, sample size 1.019 mg. Sample D: concentration = 20.38 mg/ml, sample size 4.076 mg.
thermochemical conversion processes of biomass is necessary. 3.6. SPE method A sampling and analysis method for NHCs in the product gas from gasification of biomass is presented here. No report has been found of the SPE method ever being used to extract NHCs from the acidic absorption of the product gas. Table 4 shows the recovery and reproducibility of a synthetic NHCs sample extracted from 0.1 M H2SO4 absorption by a SPE C18 tube. The number of SPE runs was 9. Eluted compounds were analysed by GC in triplicate. The RSD of the SPE method was in the range of 1.8–5.1%. The recovery was further evaluated using samples spiked at concentrations of 6.718 mg/ml, 13.59 mg/ ml and 20.38 mg/ml. Table 5 illustrates the recovery for individual NHCs ranging from 90% to 113%. To summarize, it may be concluded from the results that SPE has provided a quick and efficient handling technique for extracting NHCs from the absorption solution. 4. Conclusions The present study shows that the characteristics of the original fuel (such as volatility, content of fuel-nitrogen, and the chemical structure of the fuel-nitrogen) have strong
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
linkage to the distribution of fuel-nitrogen in the products of gasification. The percentage of the nitrogen in Daw Mill coal remaining in the char after gasification is higher than for the nitrogen in three species of biomass. NH3 was in the present investigation the main product (excluding molecular nitrogen) from fuel-nitrogen conversion during gasification of solid fuels. It seems that the content of fuel-nitrogen in the raw material is not the key factor determining the formation of NH3. However, the intrinsic reactivity and internal surface area of the nascent char, as well as the concentration of active hydrogen in the products of the gasification process play important roles in the formation of NH3. The distribution ratio of fuel-nitrogen in HCN/ NH3 seems complicatedly linked to several factors. The less fuel-nitrogen remaining in the char and the higher the ratio of two-ring NHCs to one-ring NHCs in the tar, the lower the ratio of fuel-nitrogen in HCN/NH3 observed under the conditions employed. In the biomass fuels, the composition of the fuel-nitrogen may impact on the distribution ratio of fuel-nitrogen in NO and NH3. The higher content of NO 3 in the biomass might be obtained the higher ratio of NO/NH3 in the product gas when the biomass was gasified under the same condition. The conversion of the nitrogen in biomass to NHCs was verified and quantified. The results indicate that the types of the NHCs were similar and the major compound was pyridine. However, the higher the content of the nitrogen in the biomass, the higher the ratio of two-ring NHCs to one-ring NHCs in the tar. The SPE method was applied to extract the low concentration of NHCs from the acidic absorption of the product gas. The result shows that using SPE (C18) to extract NHCs from the acidic absorption of the product gas provided a quick and efficient handling technique. Acknowledgements The work was financially supported by the Swedish National Board for Industrial and Technological Development (NUTEK) and the European Commission within the framework of the JOULE III programme. We wish also to acknowledge Dr. Christina Ho¨rnell for her reviewing, revising and improving the linguistic quality of the paper. References [1] Liang XH, Kozinski JA. Numerical modeling of combustion and pyrolysis of cellulosic biomass in thermogravimetric systems. Fuel 2000;79:1477. [2] Komiyama H, Mitsumori T, Yamaji K, Yamada K. Assessment of energy systems by using biomass plantation. Fuel 2001;80:707. [3] Werther J, Saenger M, Hartge EU, Ogada T, Siagi Z. Combustion of agricultural residues. Progress Energy Combust Sci 2000;26:1. [4] De Jong W, Pirone A, Wojtowicz MA. Pyrolysis of Miscanthus giganteus and wood pellets: TG-FTIR analysis and reaction kinetics. Fuel 2003;82:1139. [5] Dayton D. A review of the literature on catalytic biomass tar destruction. NREL/TP-510-32815 National Renewable Energy Laboratory, CO, USA.
617
[6] Mukadi L, Guy C, Legros R. Prediction of gas emissions in an internally circulating fluidized-bed combustor for treatment of industrial solid wastes. Fuel 2000;79:1125. [7] Chen G-X. The reactivity of char from rapid heating processes under pressure. Ph.D. Thesis, Royal Institute of Technology—KTH, Stockholm, Sweden, 1998. [8] Chen G-X, Yu Q-Z, Brage C, Rose´n C, Sjo¨stro¨m K. Experimental study on cogasification of biomass and coal. Survey of Combustion Research in Sweden, Chalmers University of Technology, Go¨teborg, 21–22 October 1998. p. 94. [9] Hansson KM. Formation of HNCO, HCN, and NH3 from the pyrolysis of bark and nitrogen-containing model compounds. Combust Flame 2004;137:265. [10] Nelson PF, Li C-Z, Ledesma E. Formation of HNCO from the rapid pyrolysis of coal. Energy Fuels 1996;10:264. [11] Tan L-L, Li C-Z. Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part I. Effects of reactor configuration on the determined yields of HCN and NH3 during pyrolysis. Fuel 2000;79:1883. [12] De Jong W. Nitrogen compounds in pressurised fluidised bed gasification of biomass and fossil fuels. PhD Thesis, Technische Universiteit Delft, The Netherlands, 2005. [13] Johnsson JE. Formation and reduction of nitrogen oxides in fluidized-bed combustion. Fuel 1994;73:1398. [14] Hill SC, Smoot LD. Modeling of nitrogen oxides formation and destruction in combustion systems. Progress Energy Combust Sci 2000;26:417. [15] Kambara S, Takarada T, Toyoshima M, Kato K. Relation between functional forms of coal nitrogen and NOx emissions from pulverized coal combustion. Fuel 1995;74:1247. [16] Perry ST, Fletcher TH, Solum MS, Pugmire RJ. Modeling nitrogen evolution during coal pyrolysis based on a global free-radical mechanism. Energy Fuels 2000;14:1094. [17] Basiuk VAJ. Pyrolysis of valine and leucine at 500 C: identification of less-volatile products using gas chromatography–Fourier transform infrared spectroscopy–mass spectrometry. Anal Appl Pyrolysis 1998;47:127. [18] Douda J, Basiuk VAJ. Pyrolysis of amino acids: recovery of starting materials and yields of condensation products. Anal Appl Pyrolysis 2000;56:113. [19] Ha¨ma¨la¨inen JP, Aho MJ, Tummavuori JL. Formation of nitrogen oxides from fuel-N through HCN and NH3: a model-compound study. Fuel 1994;73:1894. [20] Friebel J, Ko¨psel RFW. The fate of nitrogen during pyrolysis of German low rank coals—a parameter study. Fuel 1999;78:923. [21] Leppa¨lahti J. Formation of NH3 and HCN in slow-heating-rate inert pyrolysis of peat, coal and bark. Fuel 1995;74:1363. [22] Li C-Z, Tan L-L. Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part III. Further discussion on the formation of HCN and NH3 during pyrolysis. Fuel 2000;79:1899. [23] Li C-Z, Tan LL. Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part II. Effects of experimental conditions on the yields of NOx and SOx precursors from the pyrolysis of a Victorian brown coal. Fuel 2000;79:1891. [24] Xie KC, Lin J-Y, Li W-Y, Cheng L-P, Feng J, Zhao W. Formation of HCN and NH3 during coal macerals pyrolysis and gasification with CO2. Fuel 2005;84:271. [25] Ledesma EB, Li C-Z, Nelson PF, Mackie JC. Release of HCN, NH3 and HNCO from thermal gas-phase cracking of coal pyrolysis tar. Energy Fuels 1998;12:536. [26] Scha¨fer S, Bonn B. Hydrolysis of HCN as an important step in nitrogen oxide formation in fluidised combustion. Part 1. Homogeneous reactions. Fuel 2000;79:1239. [27] Amure O, Hanson S, Cloke M, Patrick JW. The formation of ammonia in air-blown gasification: does char-derived NO act as a precursor? Fuel 2003;82:2139. [28] Hansson K-M. Principles of biomass pyrolysis with emphasis on the formation of the nitrogen-containing gases HNCO, HCN and NH3.
618
[29]
[30] [31]
[32]
[33]
[34]
Q-Z. Yu et al. / Fuel 86 (2007) 611–618 Ph.D. Thesis, Chalmers University of Technology, Go¨teborg, Sweden, 2003. Aho MJ, Ha¨ma¨la¨inen JP, Tummavuori JL. Conversion of peat and coal nitrogen through HCN and NH3 to nitrogen oxides at 800 C. Fuel 1993;72:837. Ninomiya Y, Dong Z, Suzuki Y, Koketsu J. Theoretical study on the thermal decomposition of pyridine. Fuel 2000;79:449. Li C-Z, Buckley AN, Nelson PF. Effects of temperature and molecular mass on the nitrogen functionality of tars produced under high heating-rate conditions. Fuel 1998;77:157. Shim H-S, Hurt RH. Thermal annealing of chars from diverse organic precursors under combustion-like conditions. Energy Fuels 2000;14:340. Sjo¨stro¨m K, Chen G-X, Yu Q-Z, Brage C, Rose´n C. Promoted reactivity of char in co-gasification of biomass and coal: synergies in the thermochemical process. Fuel 1999;78:1189. Nelson PF, Kelly MD, Wornat MJ. Conversion of fuel-nitrogen in coal volatilise to NOx precursors under rapid heating conditions. Fuel 1991;70:403.
[35] Zhou J, Masutani SM, Ishimura DM, Turn SQ, Kinoshita CM. Release of fuel-bound nitrogen during biomass gasification. Ind Eng Chem Res 2000;39:626. [36] Wang W, Brown SD, Hindmarsh CJ, Thomas KM. NOx release and reactivity of chars from a wide range of coals during combustion. Fuel 1994;73:1381. [37] Middleton SP, Patrick JW, Walkel A. The release of coal nitrogen and sulphur on pyrolysis and partial gasification in a fluidized bed. Fuel 1997;76:1195. [38] Leichtnam JN, Schwartz DRJ. The behaviour of fuel-nitrogen during fast pyrolysis of polyamide at high temperature. Anal Appl Pyrolysis 2000;55:255. [39] Tsubouchi N, Ohshima Y, Xu C, Ohtsuka Y. Enhancement of N2 formation from the nitrogen in carbon and coal by calcium. Energy Fuels 2001;15:158. [40] Sipila¨ K. Research into thermochemical conversion of biomass into fuels, chemicals and fibres, In: Chartier Ph, Beenackers AACM, Grassi G, editors. The 8th E.C. conference biomass for energy, environment, agriculture and industry, Vienna, Austria, 3–5 October 1994. p. 156.
The fate of fuel-nitrogen during gasification of biomass in a pressurised fluidised bed gasifier Q-Z. Yu *, C. Brage, G-X. Chen 1, K. Sjo¨stro¨m Department of Chemical Engineering and Technology/Chemical Technology, KTH – Royal Institute of Technology, SE-100 44 Stockholm, Sweden Received 6 December 2005; received in revised form 7 August 2006; accepted 8 August 2006 Available online 8 September 2006
Abstract The distribution of fuel-nitrogen in gases, tar and char from gasification of biomass in a pressurised fluidised bed gasifier was investigated. Four species of biomass: birch, Salix, Miscanthus and Reed canary grass were gasified at 0.4 MPa and 900 C. Oxygen-enriched nitrogen was used as fluidising agent. As a reference, gasification of Daw Mill coal was also carried out under the same experimental conditions. The experimental results illustrate that both the nature of the original fuels and the chemical structure of the nitrogen in the fuel have influence on the distribution of fuel-nitrogen in gases (NH3, HCN, NO), tar and char under the employed experimental conditions. The present work also shows that the types of nitrogen heterocyclic compounds (NHCs) in the tar from different kinds of biomass are the same and the major compound is pyridine. However, the distribution of the various NHCs in the tar from the four species of biomass varies: the higher the content of fuel-nitrogen, the higher the concentration of two-ring NHCs in the tar. An effective method for extracting NHCs from the acidic absorption of the product gas was introduced in the present work. The method makes use of solid phase extraction (SPE) by a silica-based C18 tube to extract the NHCs which subsequently were analysed by gas chromatography (GC) with flame ionisation detection (FID). The recovery and reproducibility of the SPE technique for NHCs is discussed. 2006 Elsevier Ltd. All rights reserved. Keywords: Gasification; Biomass; Nitrogen compounds
1. Introduction Currently, biomass accounts for 14–15% of worldwide energy consumption. The biomass combustion process yields heat, which is used to fulfil basic energy requirements for homes and various industries in developing countries [1]. The biomass energy plantation and conversion systems are an effective source of renewable energy and have had a substantial impact on reducing the global CO2 emission budget [2]. Since energy from biomass is an effective substitute for fossil fuels, it has also been widely used in devel*
Corresponding author. Tel.: +46 8 7908252; fax: +46 8 108579. E-mail address: [email protected] (Q-Z. Yu). 1 Dr. G-X. CHEN worked on all the gasification experiments and reviewed the paper. He passed away two years ago. This paper is dedicated to the memory of Dr. G-X. Chen for his contribution to the biomass thermochemical research field. 0016-2361/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2006.08.007
oped countries over the last century [3]. In recent years, the decreasing availability of fossil fuels has become a growing concern and using fossil fuels is faced with stricter environmental regulations [4]. Therefore, a much greater interest in the use of biomass energy as an alternative to fossil fuels has developed. Among the biomass thermochemical conversion processes, pyrolysis, gasification and combustion, biomass gasification should be explored further since biomass gasification can form a solid basis of new promising energy conversion technologies. Among these are fuel cells and integrated gasification combined cycle (IGCC) installations. Conditioning and upgrading the biomass gasification product gas can make it a suitable feed for methanol or Fischer–Tropsch liquid synthesis [5]. During solid fuel thermochemical conversion the fuel-nitrogen is converted to ammonia (NH3), cyanide (HCN), isocyanic acid (HNCO), nitrogen oxides (NOx and N2O), molecular nitrogen (N2) and organic nitrogen.
612
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
Part of the fuel-nitrogen remains in the char. NH3, HCN, HNCO and organic nitrogen in the product gas pass through end-use systems, where they can poison catalysts or fuel cell internals. It is well known that NH3, HCN, HNCO and organic nitrogen are the precursors to the emissions of nitrogen oxides (NOx and N2O). Nitrogen oxides are pollutants causing photochemical smog, acid rain, the greenhouse effect and ozone depletion [6]. Therefore an intensive study on the formation of the precursors of nitrogen oxides during fuel thermochemical conversion is going on in many countries. However, the mechanisms for the formation of the precursors of nitrogen oxides during thermochemical conversion processes of solid fuels are still not clear, especially in the case of biofuels. The purpose of the present work is to investigate the correlation between the characteristics of the raw materials and the distribution of fuel-nitrogen in products (disregarding molecular nitrogen) under the studied experimental conditions. Two cultivated energy crops, Reed canary grass and Miscanthus, and a fast growing energy forest specie, Salix, as well as birch wood were employed. Daw Mill coal of UK origin was used as a reference. The factors influencing fuelnitrogen conversion to the gases (HCN, NH3, NO) or retention in the char as well as the distribution of individual NHCs in the tar are discussed in brief. A solid phase extraction (SPE) method for extracting the NHCs from the acidic absorption of the product gas was established and is presented.
cals for preparation of calibration standards and internal standards were used as received, i.e. at a minimum purity of 98–99% (Sigma–Aldrich, Sweden). SPE ENVI–C18 tubes (Supelco, Inc., Bellefonte, US) required for sample preparation were loaded with 500 mg of octadecyl-bonded endcapped silica (surface area 400–600 m2 g1, average ˚ ). particle size 50–60 lm and average pore diameter 60–70 A 2.2. Sampling
2. Experimental
Fig. 1 shows the sampling system for gaseous and liquid nitrogen compounds. The first two traps containing 0.1 M H2SO4 (2 · 40 ml) were used for sampling of both NH3 and NHCs. HCN was absorbed in the following four bottles of which two contained 0.1 M NaOH (2 · 50 ml in the second and the third bottle). The last four traps containing dichloromethane (DCM) were used to wash out the residual tar from the gas flow before the gas entered the volume measurement meter. A stream of the product gas from the top of the gasifier passed through a bypass pipe, which was warmed up to about 250 C, and bubbled through the sampling system. The gas volume measurement meter was placed at the end of the sampling system to record the total volume of sampled gas. The gas flow was controlled at about 5 l min1. The sampling time was about 1 h. A char sample was collected at the end of each experiment (about 1 h) after the temperature and the pressure in the gasifier had been reduced to room temperature and atmospheric pressure.
2.1. Materials
2.3. Sample preparation
The feedstocks used were four species of biomass (1– 3 mm, 5.3% moisture): birch, Salix, Miscanthus, Reed canary grass, and Daw Mill coal of UK origin (1– 1.35 mm, 7.3% moisture). The chemical composition of the fuels is listed in Table 1. The sample source was a pressurised fluidised bed reactor (silver sand as bed material) operated at 900 C and 0.4 MPa. Oxygen-enriched nitrogen was used as fluidising agent. The concentration of oxygen in the fluidising agent controlled the temperature of the bed in the experiments [7,8]. The solvent required for sample preparation, chemi-
The acidic solution in the first two impingers was extracted with a small portion of DCM in order to remove phenolic and neutral tar fractions. The water phase was poured into a funnel equipped with a folded filter paper. The filtrate was then transferred into a 100 ml volumetric bottle and brought to volume with 0.1 M H2SO4. A SPE method was used to remove any trace of organic compounds from the filtrate in order to protect the electrodes that were used for measuring ammonia. A SPE ENVI-C18 tube was conditioned with a 1 ml methanol followed by three tube fillings of distilled water. A part of the
Table 1 The chemical composition (wt.%, dba) of the fuels Fuel
C
H
N
O
S
Ash
Fixed carbon
Volatiles
Heating value (MJ/kg,dafb)
Birch Salix Miscanthus Reed canary grass Daw Mill coal
49.0 48.1 47.9 44.2 73.8
6.1 6.2 6.3 6.0 4.9
0.1 0.4 0.7 0.8 1.3
44.5 43.9 43.8 41.7 12.3
<0.01 0.036 0.11 0.21 1.63
0.4 1.6 2.8 8.8 6.5
n.d.c 14.9 n.d. n.d. 58.5
n.d. 82.6 78.5d 73.5d 35.6
19.3 19.2 17.0 17.6 30.4
a b c d
db: dry basis. daf: dry and ash free. n.d.: not determined. Source: Ref. [40].
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
3
4
7
613
8
11
2 5
1
10
6 9
Fig. 1. Sampling system for nitrogen compounds in product gas: (1) Heated line (200–250 C) to sampling needle valve (stainless steel); (2) glass tee (200– 250 C); (3) NH3 and NHCs absorption bottles immersed in ice-water; (4) HCN absorption bottles; (5) water-cooled super condenser; (6) round bottom flask; (7) tar impinger bottles (40 ml DCM in each) immersed in CO2/C2H5OH–H2O; (8) tar impinger bottles (40 ml DCM in each) immersed in solid CO2; (9) soap gas-flow meter; (10) integral gas volume meter; (11) ventilator.
filtrate was taken (10–60 ml, the volume depending on the NH3 concentration in the filtrate) and passed through the conditioned SPE ENVI-C18 tube with the help of a water pump. An aliquot of the elution (5–50 ml) was transferred to a 50 ml volumetric bottle by a volumetric pipette and the solution brought to volume with 0.1 M H2SO4 (if needed). The acidic solution was prepared for the NH3 analysis. In order to prepare the sample for determination of NHCs, another portion of the filtrate (5–10 ml) taken by a volumetric pipette was poured into an Erlenmeyer flask with a stopper; 2,4,6-trimethylpyridine (141 lg in 50 ll DCM) as an internal standard for GC analysis of NHCs was added to the flask using a syringe. Then the solution was adjusted to about pH 14 with 3 ml 1 M NaOH. Subsequently, the alkaline solution was slowly passed through a conditioned SPE ENVI-C18 tube (see above) with the help of a water pump. The flow rate should not exceed 5 ml/min (dropwise flow is best). After all of the water in the sample had eluted, the tube was additionally dried for 5 min with full vacuum by a water pump. The NHCs were then eluted by 3 · 500 ll of DCM/iso-propanol (1:1, V/V). Finally, the elution was dried by passing through a small column containing 3 g anhydrous Na2SO4. If there was any water visible in the eluate, it was cautiously removed before using the anhydrous Na2SO4 column. The alkaline solution in the four bottles in the gas washing system was filtered and brought to a given volume with 0.1 M NaOH in a volumetric bottle. This alkaline solution was then used for HCN analysis.
2.4. Analysis NH3 was determined using an ISE (ion-selective electrode). The prepared 50 ml sample was poured into a 100 ml beaker containing a magnetic spinning bar and 3 ml of pH-adjusting ISA (Orion Research INC., USA) was added. The treated sample was then immediately measured by a pH/ISE instrument, model 710A (UNICAM/ Orion), using an ISE gas (NH3) sensing electrode (cali-
brated with 1–500 ppm standard NH3) and an ATC probe (automatic temperature compensation). It should be noted that the NH3 yields quantified in this study may also include some contribution from the hydrolysis (in the absorption solution) of HNCO which has been observed as a nitrogen-containing species from the pyrolysis of coal and biomass [9,10]. However, the contribution to NH3 from the hydrolysis of HNCO in the bubbler may be very small because the yield of HNCO in the product gas was always much smaller than the yields of HCN and NH3 [11]. Furthermore, De Jong has stated that HNCO was never detected even by means of a high resolution FTIR spectrophotometer from the pressurised gasification of biomass under Delft experimental conditions [12]. A GC with FID was used to analyse the NHCs. The instrument (CHROMPACK CP9010) was equipped with an automatic liquid sampler (CHROMPACK CP 9002). GC control and data handling was achieved with a PCbased chromatography program. A constant flow of helium carrier gas through the column (CP-Sil 8 CB, 25 m · 0.25 mm ID, 0.5 lm film thickness) was automatically maintained using electronic pressure control. The temperatures of the injector (split 53:1) and detector were maintained at 300 C. The oven was programmed to run from 80 C (3 min) to 280 C (3 min) at 8 C min1; a 1 ll sample was injected. HCN was measured calorimetrically using Merck Microquant reagent M14798. According to the literature [11], the high solubility of HCN in the acidic solution could cause an underestimation of HCN yield when the product gas passed through the sampling system. In comparison the HCN yield is much lower than the NH3 yield under the studied experimental conditions (Table 2), therefore error is negligible. Nitrogen remaining in the char was analysed by a professional laboratory, MIKRO KEMI AB, using a so-called MK2003 method. The char sample was incinerated in an oven at a temperature of around 1000 C. Briefly, the nitrogen and nitrogen compounds were analysed by a GC with packed column and TCD detector.
614
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
Table 2 The distribution of fuel-nitrogen in the gases, tar and char Raw material
Reed canary grass
Miscanthus
Salix
Daw mill coal
% of fuel N in char % of fuel N in NH3 % of fuel N in HCN % of fuel N in NO % of fuel N in NHC The ratio (%) of N(HCN) to N(NH3)a The ratio (%) of N(NO) to N(NH3)b
0.7 34.3 0.10 0.20 1.30 0.29
9.4 12.7 0.12 0.04 0.37 0.94
0.0 24.4 0.22 0.66 0.67 0.90
34.1 7.5 0.10 0.10 n.d. 1.33
0.58
0.31
2.70
1.33
a b
The ratio of distribution of fuel-nitrogen between HCN and NH3. The ratio of distribution of fuel-nitrogen between NO and NH3.
3. Results and discussion The distribution of fuel-nitrogen between the gases (excluding molecular nitrogen), tar and char from gasification of three species of biomass and Daw Mill coal at 900 C and 0.4 MPa using oxygen-enriched nitrogen as fluidising agent is given in Table 2. 3.1. Fuel-nitrogen converted to NH3 Different pathways leading to the formation of NH3 during pyrolysis of solid fuels have been suggested: • decomposition of amines (proteins and amino acids) [13–20]; • thermal cracking reactions of tar and char [11,14,16,21– 25]; • hydrogenation and hydrolysis of HCN [21,26,27]. Inspection of the data in Table 2 shows that a larger percentage of nitrogen in the three species of biomass forms NH3 than does the nitrogen in Daw Mill coal. Nitrogen is bound to solid fuel in various ways. Studies by X-ray photoelectron spectroscopy (XPS) suggest that nitrogen in coal is present in heterocyclic structures, predominantly of the pyrrolic and pyridinic types [19,25]. In contrast to coal, nitrogen in wood is mainly in the form of protein and free amino acid [28]. According to the literature, the amine groups are believed to form NH3 directly at relatively low temperatures during pyrolysis [13]. Since biomass has a larger share of nitrogen as amines than coal has, there should be no doubt that the higher content of amine in biomass favours the production of NH3 in biomass gasification. Aho et al. have suggested that the O/N ratio of the fuel might act as an indicator of the ratio of nitrogen bound in side-chains of the aromatic structures in the fuel. Much of the nitrogen in the side-chains may exist in amino form [29]. From Table 1 it may be seen that biomass has a higher O/N ratio than coal, indicating that the share of the amino form of the nitrogen in biomass is higher than that of the
nitrogen in coal. As the amino compounds exhibit the highest conversion to NH3 among the three types of fuel-nitrogen compounds, pyridinic type, pyrrole type and amino compounds [19], this also supports the observed larger percentage conversion to NH3 for nitrogen from the three species of biomass as compared to nitrogen from Daw Mill coal. The tar from thermochemical processing of solid fuels has some opportunity to undergo secondary gas-phase cracking thereby releasing the nitrogen into the gas phase, primarily as HCN, NH3, HNCO [25]. In general, high temperature secondary pyrolysis reaction of tars yield HCN as primary product, and HCN may act as a precursor for NH3 [14,30]. Researchers have observed that NH3 could be formed from the HCN hydrolysis and the reaction of HCN with hydrogen although it may not be the main way to convert fuel-nitrogen to NH3 [21,23,26,27]. There are several major ways for the generation of active hydrogen and free radicals in the char: the substituting groups might be broken off from the (hydro) aromatic ring systems; dehydrogenation of the aliphatics or dehydrogenation of hydroaromatic ring systems and the condensation reactions between the aromatic ring systems [22]. Since biomass contains more aliphatic groups than coal does, the concentration of H radicals from the gasification of biomass could be higher than that from the gasification of coal. Therefore the heterogeneous hydrogenation of HCN forming NH3 on the surface of the char from three species of biomass could be favoured under our experimental conditions. Thermal cracking of char may also be an important course for NH3 formation. After scrutiny of the experimental data of several investigators, Hill et al. concluded that the appearance of fuel-nitrogen as NH3 in the gas phase might be important during char oxidation [14]. Xie et al. described that a large amount of NH3 was formed by thermal cracking of nascent char based on their experiment results [24]. Li et al. believed that direct hydrogenation of the nitrogen compounds in the pyrolysis of char particles is the main route for the formation of NH3. They explained that the formation of NH3 from the thermal cracking of char could be considered as a ‘‘self-gasification’’ process of the char by the active hydrogen generated in situ within the char [22]. The presence of H atoms and reactive radicals in the reacting solid fuels/char mass enhances the decomposition of N-containing heteroaromatic ring systems, resulting in the release of simple gas-phase nitrogen-containing species [31]. These observations seem consistent with the results of the present study. Table 2 illustrates that the percentage of fuel-nitrogen in NH3 has a strong inverse relation to the percentage of fuel-nitrogen remaining in the char. The formation of NH3 from secondary radical reactions in the char might be controlled by the thermochemical reactivity and the surface area of the char as well as the concentration of the H radicals [22,24]. Chars from gasification of biomass have higher thermochemical reactivity, larger surface area and a larger concentration of possible
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
H radicals within the char precursors than the chars from coal [8,32,33]. These factors may explain the higher distribution ratio of fuel-nitrogen in NH3 obtained from the gasification of three species of biomass, compared with the gasification of Daw Mill coal under our experimental conditions. However, the mechanism of NH3 formation during thermochemical conversion of solid fuels, especially biomass, is very complicated and no clear picture is available. Therefore further study is necessary. 3.2. The ratio of fuel-nitrogen in HCN to fuel-nitrogen in NH3 The fuel-nitrogen released in the volatiles from the primary pyrolysis of the fuels subsequently undergoes secondary reactions resulting in the formation of the NOx precursors such as HCN and NH3. It has been observed that the distribution of different nitrogen functional groups in the tar may lead to different rates of the secondary reactions and different products [31,34]. The thermally more stable NHCs may be hydrogenated slowly by radicals to NH3 while the thermally less stable NHCs are mainly responsible for the formation of HCN [22,34]. In Table 3, the data illustrate that the distribution of individual NHCs in the biomass tars. The order of the ratio of two-ring NHCs to one-ring NHCs in the biomass tar is Reed canary grass > Miscanthus > Salix > birch according to the results of the present study. As two-ring NHCs are more stable than one-ring NHCs, a tar containing a higher concentration of two-ring NHCs is expected to yield a larger amount of NH3. Meanwhile the nature of the char also has a strong link to the conversion of fuel nitrogen to NH3, as mentioned in the previous section. Comparing the experimental results from the gasification of three species of biomass (Reed canary grass, Miscanthus, Salix) in Tables 2 and 3, it may be seen that the gasification of Reed canary grass produced the tar with the highest ratio of two-ring NHCs to one-ring NHCs, and a char with a low percentage of fuel-nitrogen. Therefore the Table 3 The distribution of NHCs in the tar (relative mol.%) Compound
Pyridine 2-Picoline 3-Picoline 2-Ethylpyridine Lutidine 3-Ethylpyridine Quinoline Iso-quinoline The ratio of two-ring NHCs to one-ring NHCs
615
ratio of N(HCN) to N(NH3) is the lowest in the product gas from gasification of Reed canary grass among the three species of biomass (see Table 2). 3.3. The ratio of fuel-nitrogen in NO to fuel-nitrogen in NH3 The present work also showed that among the three species of biomass Salix gave the highest ratio of fuel-nitrogen in NO to fuel-nitrogen in NH3 when they were gasified under the same conditions (Table 2). It is well known that nitrogen is absorbed by the roots of plants in the form of þ þ NO 3 and NH4 . NO3 and NH4 then move from the roots þ to leaves and young stems. Portions of NO 3 and NH4 also become stored in the xylem as they are transported through þ the xylem tissue. However the bulk of NO 3 and NH4 are deposited in leaves. NO3 is converted to amino acids or proteins and stored in the leaves and stems. A small portion of fuel-nitrogen is found in the xylem, predominantly þ as NO 3 and, to a lesser extent, as NH4 . Leaves contain þ more NH4 , and xylem contains more NO 3 . According to the observation of Masutani et al., the amount of nitrate nitrogen ðNO 3 Þ in the feedstock will impact NO yield [35]. Since the fuel-nitrogen in Salix leads to a larger share of NO than the two energy crops during gasification it could be suspected that Salix may contain more NO 3 than the two energy crops. 3.4. The distribution of NHCs in the biomass tar The distribution of NHCs in the tar of four species of biomass is shown in Table 3. The results indicate that the types of the NHCs were similar and that the major compound was pyridine from the gasification of the four species of biomass under the employed experimental conditions. The order of the concentration of two-ring NHCs in the tar is Reed canary grass > Miscanthus > Salix > birch. It seems that the order of the content of nitrogen in the four species of biomass has the same tendency as the concentration of two-ring NHCs in the tar (see Table 1). The conclusion is that the higher the content of nitrogen in the biomass, the higher the concentration of two-ring NHCs in the tar. 3.5. Fuel-nitrogen in char
Raw material Birch
Reed canary grass
Miscanthus
Salix
82.6 8.3 5.9 0.0 0.0 0.0 3.2 0.0 3.3
73.9 4.8 3.7 0.0 0.0 0.0 16.0 1.6 21.4
69.8 8.3 7.5 0.0 0.8 0.0 11.3 2.2 15.6
84.6 3.9 3.3 0.0 0.0 0.0 6.1 2.1 8.9
Comparing the results in Table 2, it seems that the percentage of the nitrogen in three species biomass retained in the char is much lower than for the nitrogen in Daw Mill coal retained in the char. The phenomenon may be explained by the differences in the feedstocks; the different chemical constituents of the fuel-nitrogen in the raw materials. During devolatilisation, fuel-nitrogen is divided between char nitrogen and the volatiles, mainly in the tar and as NH3 and HCN [30]. Pyrolysis of coal by various techniques has shown that the fuel-nitrogen retention in the char is high [36]. Model studies on coal combustion have predicted that the distribution of the nitrogen in coal
616
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
between volatiles and char is linked to the nature of the coal and the combustion environment [14]. However, the extent of fuel-nitrogen released from the original fuel was roughly proportional to the content of volatile matter in the original fuel at the devolatilisation stage [37]. As shown in Table 1 the content of volatiles in three species of biomass is higher than in Daw Mill coal. It could therefore result in the nitrogen in three species of biomass being released with a larger proportion in the volatiles than for the nitrogen in Daw Mill coal at the devolatilisation stage. According to literature [15,34], the chemical structures of the nitrogen in the fuels are also an important factor influencing the evolution of organic nitrogen into the volatiles and char during secondary thermal conversion. It is well known that the different chemical structures of organic nitrogen compounds have different thermochemical decomposition temperatures. For example, pyrrole and pyridine in their pure forms are not significantly decomposed until about 1200–1300 K [16]. In contrast to this, most amino acids are thermally decomposed at 500 C (773 K) [17,18]. Even high molecular weight amines such as polyamide (nylon) is completely decomposed at 873 K and the main products detected below this temperature are carbon oxides, NH3, and water [38]. The constituents of the nitrogen in coal have been widely investigated. An XPS spectrum study on coal found that most of the nitrogen in coal occurs in pyrrolic and pyridinic structures [21,30]. The amount and the composition of nitrogen in biomass are affected by several factors, such as the types and species of plants, storage conditions and processing methods after harvest, etc. In general, most of the nitrogen in a plant is stored in the leaves and shoots in the form of amino acids [35]. However, almost 100% of the nitrogen in wood bark is in the pyrrolic form according to Ha¨ma¨la¨inen [19]. Since a considerable part of the nitrogen in biomass may be found as amines, it is understandable that the decomposition and release of biomass nitrogen is easier and faster than that of the nitrogen in coal during thermochemical conversion processes. Therefore the percentage of the nitrogen in three species of biomass retained in the char is much lower than that of the nitrogen in Daw Mill coal retained in the char. Fuel-nitrogen conversion during gasification in a pressurised fluidised bed is very complicated. The discussion in the present paper is limited to the characteristics of the original fuels. There are also many other factors influencing the fate of fuel-nitrogen during a thermochemical process. The mineral matter in the fuel can also play a role in the distribution of fuel-nitrogen between the individual products. Alkali, alkaline earth and transition metal cations could change the distribution of volatile products during coal pyrolysis. Ohtsuka has shown that small amounts of minerals (for example iron and calcium), probably in their ion-exchanged forms, dramatically change the fate of the nitrogen in coal and catalyse conversion of the nitrogen in coal to N2 [39]. Therefore an intensive and comprehensive investigation of the fuel-nitrogen chemistry in the
Table 4 The recovery and reproducibility of NHCs extracted using SPE C18 tube Compound
Mean value (lg)
Recovery (%)
Relative STDEVa
Pyridine 2-Picoline 3-Picoline 2-Ethylpyridine Lutidin 3-Ethylpyridine Quinoline Iso-quinoline
39.90 44.00 39.73 38.28 27.81 38.27 49.87 56.15
106.4 101.9 99.8 109.4 90.3 101.0 95.2 94.7
2.3 3.1 2.4 5.0 5.1 4.5 2.9 1.8
a
Relative standard deviation (%).
Table 5 The recovery of synthetic NHC samples with different concentrations Compound
A
B
C
D
Pyridine 2-Picoline 3-Picoline 2-Ethylpyridine Lutidine 3-Ethylpyridine Quinoline Iso-quinoline
107.5 104.6 102.5 109.4 93.9 105.9 93.9 96.8
108.6 111.1 109.9 110.6 101.5 105.8 89.7 109.1
110.3 105.6 111.2 113.2 104.8 107.2 94.9 110.1
99.1 99.5 101.6 101.7 100.6 107.0 93.4 95.5
Eluted compounds were analysed by capillary GC in triplicate. Sample A: concentration = 6.718 mg/ml, sample size 336 lg. Sample B: concentration = 13.59 mg/ml, sample size 680 lg. Sample C: concentration = 20.38 mg/ml, sample size 1.019 mg. Sample D: concentration = 20.38 mg/ml, sample size 4.076 mg.
thermochemical conversion processes of biomass is necessary. 3.6. SPE method A sampling and analysis method for NHCs in the product gas from gasification of biomass is presented here. No report has been found of the SPE method ever being used to extract NHCs from the acidic absorption of the product gas. Table 4 shows the recovery and reproducibility of a synthetic NHCs sample extracted from 0.1 M H2SO4 absorption by a SPE C18 tube. The number of SPE runs was 9. Eluted compounds were analysed by GC in triplicate. The RSD of the SPE method was in the range of 1.8–5.1%. The recovery was further evaluated using samples spiked at concentrations of 6.718 mg/ml, 13.59 mg/ ml and 20.38 mg/ml. Table 5 illustrates the recovery for individual NHCs ranging from 90% to 113%. To summarize, it may be concluded from the results that SPE has provided a quick and efficient handling technique for extracting NHCs from the absorption solution. 4. Conclusions The present study shows that the characteristics of the original fuel (such as volatility, content of fuel-nitrogen, and the chemical structure of the fuel-nitrogen) have strong
Q-Z. Yu et al. / Fuel 86 (2007) 611–618
linkage to the distribution of fuel-nitrogen in the products of gasification. The percentage of the nitrogen in Daw Mill coal remaining in the char after gasification is higher than for the nitrogen in three species of biomass. NH3 was in the present investigation the main product (excluding molecular nitrogen) from fuel-nitrogen conversion during gasification of solid fuels. It seems that the content of fuel-nitrogen in the raw material is not the key factor determining the formation of NH3. However, the intrinsic reactivity and internal surface area of the nascent char, as well as the concentration of active hydrogen in the products of the gasification process play important roles in the formation of NH3. The distribution ratio of fuel-nitrogen in HCN/ NH3 seems complicatedly linked to several factors. The less fuel-nitrogen remaining in the char and the higher the ratio of two-ring NHCs to one-ring NHCs in the tar, the lower the ratio of fuel-nitrogen in HCN/NH3 observed under the conditions employed. In the biomass fuels, the composition of the fuel-nitrogen may impact on the distribution ratio of fuel-nitrogen in NO and NH3. The higher content of NO 3 in the biomass might be obtained the higher ratio of NO/NH3 in the product gas when the biomass was gasified under the same condition. The conversion of the nitrogen in biomass to NHCs was verified and quantified. The results indicate that the types of the NHCs were similar and the major compound was pyridine. However, the higher the content of the nitrogen in the biomass, the higher the ratio of two-ring NHCs to one-ring NHCs in the tar. The SPE method was applied to extract the low concentration of NHCs from the acidic absorption of the product gas. The result shows that using SPE (C18) to extract NHCs from the acidic absorption of the product gas provided a quick and efficient handling technique. Acknowledgements The work was financially supported by the Swedish National Board for Industrial and Technological Development (NUTEK) and the European Commission within the framework of the JOULE III programme. We wish also to acknowledge Dr. Christina Ho¨rnell for her reviewing, revising and improving the linguistic quality of the paper. References [1] Liang XH, Kozinski JA. Numerical modeling of combustion and pyrolysis of cellulosic biomass in thermogravimetric systems. Fuel 2000;79:1477. [2] Komiyama H, Mitsumori T, Yamaji K, Yamada K. Assessment of energy systems by using biomass plantation. Fuel 2001;80:707. [3] Werther J, Saenger M, Hartge EU, Ogada T, Siagi Z. Combustion of agricultural residues. Progress Energy Combust Sci 2000;26:1. [4] De Jong W, Pirone A, Wojtowicz MA. Pyrolysis of Miscanthus giganteus and wood pellets: TG-FTIR analysis and reaction kinetics. Fuel 2003;82:1139. [5] Dayton D. A review of the literature on catalytic biomass tar destruction. NREL/TP-510-32815 National Renewable Energy Laboratory, CO, USA.
617
[6] Mukadi L, Guy C, Legros R. Prediction of gas emissions in an internally circulating fluidized-bed combustor for treatment of industrial solid wastes. Fuel 2000;79:1125. [7] Chen G-X. The reactivity of char from rapid heating processes under pressure. Ph.D. Thesis, Royal Institute of Technology—KTH, Stockholm, Sweden, 1998. [8] Chen G-X, Yu Q-Z, Brage C, Rose´n C, Sjo¨stro¨m K. Experimental study on cogasification of biomass and coal. Survey of Combustion Research in Sweden, Chalmers University of Technology, Go¨teborg, 21–22 October 1998. p. 94. [9] Hansson KM. Formation of HNCO, HCN, and NH3 from the pyrolysis of bark and nitrogen-containing model compounds. Combust Flame 2004;137:265. [10] Nelson PF, Li C-Z, Ledesma E. Formation of HNCO from the rapid pyrolysis of coal. Energy Fuels 1996;10:264. [11] Tan L-L, Li C-Z. Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part I. Effects of reactor configuration on the determined yields of HCN and NH3 during pyrolysis. Fuel 2000;79:1883. [12] De Jong W. Nitrogen compounds in pressurised fluidised bed gasification of biomass and fossil fuels. PhD Thesis, Technische Universiteit Delft, The Netherlands, 2005. [13] Johnsson JE. Formation and reduction of nitrogen oxides in fluidized-bed combustion. Fuel 1994;73:1398. [14] Hill SC, Smoot LD. Modeling of nitrogen oxides formation and destruction in combustion systems. Progress Energy Combust Sci 2000;26:417. [15] Kambara S, Takarada T, Toyoshima M, Kato K. Relation between functional forms of coal nitrogen and NOx emissions from pulverized coal combustion. Fuel 1995;74:1247. [16] Perry ST, Fletcher TH, Solum MS, Pugmire RJ. Modeling nitrogen evolution during coal pyrolysis based on a global free-radical mechanism. Energy Fuels 2000;14:1094. [17] Basiuk VAJ. Pyrolysis of valine and leucine at 500 C: identification of less-volatile products using gas chromatography–Fourier transform infrared spectroscopy–mass spectrometry. Anal Appl Pyrolysis 1998;47:127. [18] Douda J, Basiuk VAJ. Pyrolysis of amino acids: recovery of starting materials and yields of condensation products. Anal Appl Pyrolysis 2000;56:113. [19] Ha¨ma¨la¨inen JP, Aho MJ, Tummavuori JL. Formation of nitrogen oxides from fuel-N through HCN and NH3: a model-compound study. Fuel 1994;73:1894. [20] Friebel J, Ko¨psel RFW. The fate of nitrogen during pyrolysis of German low rank coals—a parameter study. Fuel 1999;78:923. [21] Leppa¨lahti J. Formation of NH3 and HCN in slow-heating-rate inert pyrolysis of peat, coal and bark. Fuel 1995;74:1363. [22] Li C-Z, Tan L-L. Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part III. Further discussion on the formation of HCN and NH3 during pyrolysis. Fuel 2000;79:1899. [23] Li C-Z, Tan LL. Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part II. Effects of experimental conditions on the yields of NOx and SOx precursors from the pyrolysis of a Victorian brown coal. Fuel 2000;79:1891. [24] Xie KC, Lin J-Y, Li W-Y, Cheng L-P, Feng J, Zhao W. Formation of HCN and NH3 during coal macerals pyrolysis and gasification with CO2. Fuel 2005;84:271. [25] Ledesma EB, Li C-Z, Nelson PF, Mackie JC. Release of HCN, NH3 and HNCO from thermal gas-phase cracking of coal pyrolysis tar. Energy Fuels 1998;12:536. [26] Scha¨fer S, Bonn B. Hydrolysis of HCN as an important step in nitrogen oxide formation in fluidised combustion. Part 1. Homogeneous reactions. Fuel 2000;79:1239. [27] Amure O, Hanson S, Cloke M, Patrick JW. The formation of ammonia in air-blown gasification: does char-derived NO act as a precursor? Fuel 2003;82:2139. [28] Hansson K-M. Principles of biomass pyrolysis with emphasis on the formation of the nitrogen-containing gases HNCO, HCN and NH3.
618
[29]
[30] [31]
[32]
[33]
[34]
Q-Z. Yu et al. / Fuel 86 (2007) 611–618 Ph.D. Thesis, Chalmers University of Technology, Go¨teborg, Sweden, 2003. Aho MJ, Ha¨ma¨la¨inen JP, Tummavuori JL. Conversion of peat and coal nitrogen through HCN and NH3 to nitrogen oxides at 800 C. Fuel 1993;72:837. Ninomiya Y, Dong Z, Suzuki Y, Koketsu J. Theoretical study on the thermal decomposition of pyridine. Fuel 2000;79:449. Li C-Z, Buckley AN, Nelson PF. Effects of temperature and molecular mass on the nitrogen functionality of tars produced under high heating-rate conditions. Fuel 1998;77:157. Shim H-S, Hurt RH. Thermal annealing of chars from diverse organic precursors under combustion-like conditions. Energy Fuels 2000;14:340. Sjo¨stro¨m K, Chen G-X, Yu Q-Z, Brage C, Rose´n C. Promoted reactivity of char in co-gasification of biomass and coal: synergies in the thermochemical process. Fuel 1999;78:1189. Nelson PF, Kelly MD, Wornat MJ. Conversion of fuel-nitrogen in coal volatilise to NOx precursors under rapid heating conditions. Fuel 1991;70:403.
[35] Zhou J, Masutani SM, Ishimura DM, Turn SQ, Kinoshita CM. Release of fuel-bound nitrogen during biomass gasification. Ind Eng Chem Res 2000;39:626. [36] Wang W, Brown SD, Hindmarsh CJ, Thomas KM. NOx release and reactivity of chars from a wide range of coals during combustion. Fuel 1994;73:1381. [37] Middleton SP, Patrick JW, Walkel A. The release of coal nitrogen and sulphur on pyrolysis and partial gasification in a fluidized bed. Fuel 1997;76:1195. [38] Leichtnam JN, Schwartz DRJ. The behaviour of fuel-nitrogen during fast pyrolysis of polyamide at high temperature. Anal Appl Pyrolysis 2000;55:255. [39] Tsubouchi N, Ohshima Y, Xu C, Ohtsuka Y. Enhancement of N2 formation from the nitrogen in carbon and coal by calcium. Energy Fuels 2001;15:158. [40] Sipila¨ K. Research into thermochemical conversion of biomass into fuels, chemicals and fibres, In: Chartier Ph, Beenackers AACM, Grassi G, editors. The 8th E.C. conference biomass for energy, environment, agriculture and industry, Vienna, Austria, 3–5 October 1994. p. 156.