- Email: [email protected]
Preparation of zinc ferrite in the presence of carbon material and its application to hot-gas cleaning
Fuel 83 (2004) 661–669 www.fuelfirst.com
Preparation of zinc ferrite in the presence of carbon material and its application to hot-gas cleaning Na-oki Ikenaga*, Yousuku Ohgaito, Hiroaki Matsushima, Toshimitsu Suzuki Department of Chemical Engineering, Faculty of Engineering, Kansai University 3-3-35 Yamate, Suita, Osaka 564-8680, Japan Received 17 March 2003; revised 29 July 2003; accepted 1 August 2003; available online 1 December 2003
Abstract In order to develop an efficient absorbent of H2S in coal gasification, zinc ferrite (ZnFe2O4) was prepared in the presence of carbon materials such as activated carbon (AC), activated carbon fiber (ACF), and Yallourn coal (YL). The absorption behavior of absorbents for H2S was examined using a fixed-bed flow type reactor equipped with a quadrupole mass spectrometer. Carbon material-supported ZnFe2O4 exhibited larger desulfurization capacity for H2S than unsupported ferrites. They could efficiently remove H2S from 4000 ppm levels in a simulated coal gasification gas to less than 1 ppm at 500 8C. The absorption capacity of H2S with ZnFe2O4/AC, ZnFe2O4/ACF, and ZnFe2O4/YL exhibited nearly 100% of stoichiometric amount of loaded metal species. They could be regenerated by an air oxidation in O2 – Ar (50 vol%) at 450 8C for 30 min. The regenerated ferrite can be used for repeated absorption of H2S with a very slight decrease in the absorption capacity. q 2003 Elsevier Ltd. All rights reserved. Keywords: Hydrogen sulfide; Zinc ferrite; Activated carbon; Coal; Preparation of ferrite; Hydrogen sulfide absorbent
1. Introduction Since energy consumption would progressively increase in developing countries, development of alternate clean fuel for current petroleum or new energy sources are important research area. For this purpose, natural gas and coal are the most promising sources, and their uses are increasing currently. Use of a large amount of coal causes various environmental problems such as a large amount of CO2 and several pollutants emission. Development of clean and efficient advanced coal utilization technology is of current importance. Increasing attention is being paid to integrated coal gasification combined cycle (IGCC) process and gasification-molten carbonate fuel cell (MCFC) technology as one of the most promising technologies. For these technologies, several thousands ppm of hydrogen sulfide containing in an effluent gas from gasifier must be efficiently removed to several ppm for IGCC and , 1 ppm for MCFC. Wet processes using basic organic solvents are commercially established technologies, and further investigation is in progress [1,2]. However, hot gas from the gasifier must be * Corresponding author. Tel.: þ 81-6-6368-0792; fax: þ81-6388-8869. E-mail address: [email protected] (N. Ikenaga). 0016-2361/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2003.08.019
cooled down below 100 8C in order to apply an organic solvent as an absorbent. This looses energy contained as latent heat in the hot gas [3,4]. Hot gas cleaning has been conducted by using various inorganic absorbents such as activated carbon [5 – 7], ZnO [8,10], Fe2O3 [8,9,11], CuO [8,11], Mn2O3 [8,12 –14], and ZnFe2O4 [15 – 17]. For example, Slimane et al. used four kinds of manganese oxides such as MnO, Mn3O4, Mn2O3, and MnO2 for desulfurization of hot coal-derived fuel gases and reported the reduction and sulfidation thermodynamics and kinetic characteristics of these absorbents in detail [12 –14]. They systematically studied copper, iron, manganese, and zinc oxide as bases for developing regenerable absorbent for hydrogen sulfide absorption at moderate temperatures. These absorbents can remove hydrogen sulfide from an effluent gas below 10 ppm. For fuel cell use, however, further reduction of hydrogen sulfide is required. Many researchers have been trying to develop high performance absorbent for hydrogen sulfide removal such as zinc titanate [18–25], zinc ferrite doped with several kinds of metal (V, Ti, and Cu) [26,27], and activated carbonsupported zinc ferrite [28]. Lew et al. indicated that Zn–Ti–O absorbent exhibited the high activity for hydrogen sulfide removal and the sulfidation of Zn –Ti –O proceeded by
662
N. Ikenaga et al. / Fuel 83 (2004) 661–669
the same mechanism as ZnO sulfidation [18]. However, the reactivity of zinc titanate absorbent towards hydrogen sulfide gradually decreased during sulfidation – regenaration tests [20]. In the present work, activated carbon-, activated carbon fiber-, and coal-supported zinc ferrite were prepared at a low temperature in order to obtain a highly active absorbent for H2S removal, which can be used repeatedly. Characterization of zinc ferrite prepared here was carried out by using powder X-ray diffraction and BET method. The absorption behavior of the absorbent was examined by using a fixedbed flow type reactor. Regeneration of the used absorbent for H2S removal was investigated.
was calcined at 500 – 800 8C for 2 h in air to obtain zinc ferrite [28].
2. Experimental
2.4. Absorption of H2S
2.1. Materials
The breakthrough curves were obtained by using a fixedbed flow type reactor made of quartz. After placing a 10 mg of absorbent and 40 mg of silica sand as a diluent in the reactor (i.d. 4 mm, 235 mm length), an inert gas (Ar) was introduced until the absorbent reached to the absorption temperature (300 – 700 8C), and H2S (5 ml/min) diluted with nitrogen to 0.96 vol% and Ar (5 ml/min) were supplied to the reactor (SV, 1400 h21). Exit gas was led to a quadrupole mass spectrometer (Q-mass), and the hydrogen sulfide concentrations of the tail gas were observed. H2 (17 vol%) and/or CO (33 vol%) were added to the inlet gas, in order to simulate coal gasification gas composition.
Activated carbon (AC, Wako Pure Chemical Industries, Ltd., 100-mesh pass, surface area 993 m2/g), activated carbon fiber (ACF, AD’ALL, Co., Ltd., surface area 2000 m2/g), Australian Yallourn brown coal (YL, C 66.3; H 4.70; N 0.48; S 0.26 daf%; Ash 2.0 d%, 100-mesh pass, surface area 5 m2/g) were used as a support. Zn(NO3)2· 6H2O, Fe(NO3)3·9H2O, urea, and aqueous ammonia (ca 28%) (Wako Pure Chemical Industries, Ltd.) were used to prepare absorbents in this study.
2.3. Characterization of absorbents The absorbents freshly prepared here and those after absorption of H2S were analyzed by powder X-ray diffraction using an X-ray diffractometer (Shimadzu XRD6000) with monochromatized Cu Ka radiation. Crystallite sizes of zinc ferrites were calculated from diffraction peak at 358 according to Scherrer’s equation. Surface areas of absorbents were measured by BET method using a Micromeritics Gemini 2375, applying adsorption isotherms of nitrogen at 2 196 8C.
2.2. Preparation of absorbent for H2S 2.2.1. Homogeneous precipitation method Into a mixed solution of Fe(NO3)3 and Zn(NO3)2 (150 mmol/l, 80 ml), 0.04 mmol of urea and 0.3 g of AC, ACF, or YL were added. The mixture was allowed to keep 90 8C for 2 h. During this stage, urea was decomposed to NH3 and CO to precipitate Zn(OH)2 and Fe(OH)3. AC, ACF, or YL coal was filtered off together with precipitates. Then the filtrate was dried at 70 8C under vacuum. Dried Zn(OH)2 and Fe(OH)3 on the support (AC, ACF, YL coal) were calcined at 400 or 500 8C under air flow for 2 h. 2.2.2. Coprecipitation method A 0.3 g of AC or YL coal was added into a mixed solution of Fe(NO3)3 (20 mmol/l, 100 ml) and Zn(NO3)2 (10 mmol/l, 100 ml), and then into this suspension aqueous ammonia was added until pH of the supernatant reaches 10. The combined precipitates was separated and dried in vacuo. The mixture was calcined at 500 8C under air flow for 2 h to obtain zinc ferrite. 2.2.3. Impregnation method A 0.3 g of AC or YL was added to a mixed solution of iron nitrate (III) (83 mmol/l, 30 ml) and zinc nitrate (40 mmol/l, 30 ml). After standing for overnight, the mixture was dried by evaporation at 70 8C, and the mixture
3. Results and discussion 3.1. Preparation and characterization of absorbents for H2S removal Table 1 summarizes surface area, the content of oxides, and the average crystallite size ðd311 Þ of various absorbents prepared on AC, ACF, and YL char prepared by various methods. Mixed oxides prepared by various methods were subjected to XRD analyses to determine oxide structures. Results of XRD patterns of typical absorbents are shown in Fig. 1. ZnFe2O4 prepared by dry process at 1000 8C showed sharp diffraction peaks ascribed to ZnFe2O4 (JCPDS card no 22-1012) with 343 nm of average crystallite size evaluated from the half height width of XRD peaks at 2u ¼ 358 according to Scherrer’s equation. ZnFe2O4 prepared without carbon materials by coprecipitation method at 500 8C (Sample C) showed relatively broad diffraction peaks of ZnFe2O4. When the zinc ferrites were prepared in the presence of AC and ACF by homogeneous precipitation and impregnation methods at 500 8C, diffraction peaks ascribed to ZnO and Fe2O3 were observed (Samples E, F, and H), in addition to ZnFe2O4. In the previous paper [28], we have reported that zinc ferrite could be produced on activated carbon at a lower calcinations temperature of 300 8C.
N. Ikenaga et al. / Fuel 83 (2004) 661–669
663
Table 1 Surafce area and oxide content of various sorbents Sample
Sorbent
Method
Calcination temp. (8C)
Phase determined by XRD
Crystallite size (nm)
S.A. (m2/g)
Oxide content (wt%)
Absorption capacitya (%)
A B C D E F G H I J K L
ZnFe2O4 Fe2O3 ZnFe2O4 ZnFe2O4 ZnFe2O4/ACF ZnFe2O4/AC ZnFe2O4/AC ZnFe2O4/AC Fe2O3/AC ZnFe2O4/YL ZnFe2O4/YL ZnFe2O4/YL
Dry process PPT CoPPT CoPPT HomoPPT HomoPPT CoPPT IMP IMP HomoPPT CoPPT IMP
1000 500 500 300 400 500 500 500 500 500 500 500
ZnFe2O4 Fe2O3 ZnFe2O4 Amorphous ZnFe2O4, ZnO, Fe2O3 ZnFe2O4, ZnO, Fe2O3 ZnFe2O4 ZnFe2O4, ZnO, Fe2O3 Fe2O3 ZnFe2O4 ZnFe2O4 ZnFe2O4
343 95 8 – 10 20 9 22 90 13 5 10
17 40 41 134 1190 317 463 50 67 51 58 50
– – – – 10.7 66.2 56.9 96.8 99.0 93.1 95.5 86.0
49 94 48 116 145 113 84 100 113 98 97 90
Carried out using a fixed-bed reactor (SV, 1400 h21; H2S, 4000 ppm; H2, 17 vol%; absorption temperature, 500 8C) PPT, precipitation method; CoPPT, coprecipitation method; HomoPPT, homogeneous precipitation method; IMP, impregnation method. a
The ferrites prepared in the presence of YL coal showed only diffraction peaks ascribed to ZnFe2O4 with 5– 13 nm of average crystallite size (Samples J, K, and L). In the cases of ferrites loaded on carbon material, the oxide content in the total absorbent was evaluated based on the weight decreases of the sample, by burning the organic part using a thermogravimetric balance in a flowing air. Surface areas of zinc ferrites prepared in the presence of ACF and AC (Samples D, F, and G) were higher than those of zinc ferrites (Samples A –C), which were prepared without carbon materials. Surface area of ZnFe2O4/ACF reached 1190 m2/g due to a large amount of remained high surface area ACF. Surface areas decreased to 50 m2/g with a decrease in the amount of carbon remaining (1 –14%). In the preparation of ZnFe2O4 in the presence of ACF and AC by homogeneous precipitation and coprecipitation methods, since Zn(OH)2 and Fe(OH)3 covered the surface of ACF and AC, the large amount of carbon seemed to be remained in the absorbent after the calcination.
3.2. Absorption behavior of H2S The performance of an absorbent can be evaluated in terms of both the absorption capacity of the absorbent and the absorption rate of H2S by using a fixed-bed flow type reactor, where the concentrations of H2S in the tail gas are an important factor in addition to the breakthrough time. Low concentration of H2S in the tail gas indicates that the absorbent has a high absorption rate and a low equilibrium concentration of H2S. Fig. 2 shows examples of breakthrough behavior of H2S using various absorbents measured at an absorption temperature of 500 8C. The horizontal axis of Fig. 2 designates the absorption capacity, which was calculated by assuming that all zinc and iron in the ferrite
would be transformed into ZnS and FeS as follows ZnFe2 O4 þ 3H2 S þ H2 ! ZnS þ 2FeSðFe12x S or FeS2 Þ þ 4H2 O Absorption capacity ð%Þ ¼
Amount of absorbed H2 S ðmolÞ=3 £ 100 Amount of ferrite ðmolÞ
The vertical axis of Fig. 2 indicates H2S concentration in the effluent. The absorption capacity of respective absorbent was estimated by extrapolating a slope of breakthrough curve to H2S concentration 0 ppm in the effluent, and the absorption capacity of each absorbent is shown in the last column of Table 1. The absorption capacity of several samples exceeded 100% because a part of the iron was transformed into FeS2. The formation of FeS2 was not confirmed with XRD analysis after the reaction, but the peak ascribed to FeS2 was observed with XPS analysis. The absorption capacities of H2S (48 – 49%) with unloaded ZnFe2O4 prepared by dry process and coprecipitation method (Samples A and C) were lower than those (. 85%) of loaded ZnFe2O4. In addition, when unloaded ZnFe2O4 prepared by dry process (Sample A) was used for H2S absorption, high H2S concentration of 5 – 10 ppm was observed in the effluent gas. Fe2O3 and Fe2O3/AC (Samples B and I) exhibited high absorption capacities of 94 and 113%, respectively. However, in the breakthrough curve H2S concentration of 10– 20 ppm was obtained in the effluent gas through out the run due to thermodynamic reasons [10]. These results indicate that Fe2O3 absorbent is not suitable for complete removal of H2S, and that zinc in the ferrite contributes to the high absorption rate of H2S. On the other hand, the absorption capacities of H2S increased to nearly 100% or higher with ZnFe2O4 prepared
664
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 1. XRD patterns of absorbents prepared by various methods.
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 2. Breakthrough curves of H2S absorption using various absorbents obtained at 500 8C. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 5 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min (SV, 1400 h21, H2S, 4000 ppm).
in the presence of AC, ACF, and YL coal (Samples E – L). These absorbents showed no H2S leakage in the effluent gas (less than 1 ppm). Although the surface area of Sample G was higher than that of Sample F, the absorption capacity of the former was lower than that of the latter. This fact indicates that the surface area seems not to be an important factor for the high absorption capacity of H2S at a high temperature. The important factor for high absorption capacity of H2S and high absorption rate of H2S seems to be not micro porous structure that contributes increasing surface area but macro porous structure that provides rapid diffusion of H2S into the absorbent [29]. The changes in the XRD patterns of ZnFe2O4 prepared with and without YL coal after H2S absorption are shown in Fig. 3. In the case of ZnFe2O4 prepared with YL coal, all of
665
ZnFe2O4 was transformed into ZnS, Fe12xS or FeS (Sample J, Fig. 3(a)) with 98% of the absorption capacity. On the other hands, unloaded ZnFe2O4 (Sample C), which exhibited low absorption capacity of 48%, showed weak diffraction peaks of ZnFe2O4 in addition to peaks of iron and zinc sulfides as shown in Fig. 3(b). The result is consistent with the lower absorption capacity of the unloaded ferrite. Although Sample C has slightly smaller crystallite size of 8 nm than that (13 nm) of Sample J, the transformation of ZnFe2O4 into iron and zinc sulfides did not proceed completely. Therefore, it seems reasonable to consider that the rapid diffusion of H2S into the absorbent is an important factor for the H2S absorption. Several mixed oxides of lower Zn/Fe ratio (ZnxFe32xO4, x ¼ 0 – 0:8) were prepared with YL coal and the absorption behavior of them for H2S was examined. When Zn0.8Fe2.2O4 ðZn=Fe ¼ 0:36Þ was utilized to absorb H2S, the absorption capacity decreased to 73%. The absorption capacity of H2S considerably decreased to 20% with decreasing Zn/Fe ratio to 0.034 and H2S concentration of 10 –30 ppm was obtained in the effluent gas until the breakthrough point. Ferrite structure (ZnFe2O4, Zn=Fe ¼ 0:5) seems to be required for the high H2S absorption capacity. 3.3. Effects of absorption temperature on the absorption behavior of H2S To examine the effect of the absorption temperature on the absorption capacity of the absorbent, the absorption of H2S was carried out with ZnFe2O4/YL in the temperature ranges of 300 – 700 8C. The result is illustrated in Figs. 4 and 5. Fig. 4 shows the breakthrough curves obtained in the H2S absorption at 300, 500, and 700 8C. In Fig. 5,
Fig. 3. XRD patterns of ZnFe2O4 and ZnFe2O4/YL after H2S absorption.
666
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 4. Breakthrough curves of H2S absorption using ZnFe2O4/YL obtained at various temperatures. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min, Absorption temperature; 300, 500, and 700 8C (H2S concentration; 4000 ppm).
the absorption capacity and the concentrations of H2S in the tail gas were plotted against the absorption temperature. ZnFe2O4 prepared in the presence of YL coal exhibited relatively high absorption capacity of H2S of 96% at a low absorption temperature of 300 8C. However, H2S concentration of 5 ppm was obtained in the effluent gas until the breakthrough point. At the absorption temperatures of 400 and 450 8C, the absorption capacities of H2S increased to 117 and 119%, respectively, and the H2S concentration in the tail gas decreased to less than 1 ppm. When the absorption temperature was increased from 500 to 600 8C, the absorption capacity of H2S decreased, however, the absorption capacities were nearly 100% (103 and 101%, respectively). The absorption capacity of H2S decreased to
65% with increasing the absorption temperature to 700 8C. According to XRD analysis of the absorbent after the H2S absorption at 700 8C, the intensity ratio of ZnS to Fe12xS decreased as compared to those in the absorbents reacted at 400– 600 8C. Furthermore, the H2S concentration in the tail gas obtained at 700 8C increased more gradually after the breakthrough point (at an absorption capacity of 65%) as compared to that obtained at 500 8C (Fig. 4). Therefore, it seems reasonable that the decrease in the absorption capacity would be ascribed to the reduction of ZnS or ZnO in the ferrite with H2 to give metallic Zn, followed by evaporation of Zn [16,30]. Although in the front edge of the absorbent in a fixed-bed reactor H2S concentration is high, H2S concentration is very low at the tail end of the fixed-bed as shown previously. In such circumstances either reverse reaction to give H2S or reduction of ZnO could be possible. These results indicate that optimal temperature of H2S removal with ZnFe2O4/YL is 400 –600 8C. 3.4. Effects of calcination temperature of absorbent on the absorption behavior of H2S In order to elucidate the effect of the calcination temperature of the absorbent on the absorption behavior of H2S, the breakthrough curves of the YL coal-supported zinc ferrites calcined at 500– 800 8C were measured at absorption temperature of 500 8C. Fig. 6 shows the crystallite size, the surface area, the absorption capacity of ZnFe2O4/YL (Sample L), and the concentration of H2S in the tail gas as a function of calcination temperature. The absorption capacity of ZnFe2O4/YL gradually decreased with increasing the calcination temperature to 800 8C. With an increase in the reaction temperature, crystallite size of ZnFe2O4 determined by XRD analyses increased gradually below 700 8C and markedly at 800 8C and surface area of ZnFe2O4 decreased linearly until 800 8C. This seems to indicate that crystallite size and surface area of ZnFe2O4 are partly contributed for the high absorption capacity of H2S, when the absorbent was prepared by the same method. 3.5. Effects of concentration of H2, CO and space velocity on the absorption of H2S
Fig. 5. Relationship between absorption capacity and absorption temperature. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min (H2S concentration, 4000 ppm, Absorbent: ZnFe2O4/YL (Sample L) calcined at 500 8C).
In order to obtain further information on the H2S absorption performance of ZnFe2O4 prepared in the presence of YL coal (Sample L), composition of feed gas and feed rate were varied in the absorption of H2S. The result is shown in Fig. 7. In the absorption of 4000 ppm of H2S at SV ¼ 1400 h21 without CO, the absorption capacity was 90%. When CO was added into the inlet gas (4000 ppm of H2S, 17 vol% of H2, and 33 vol% of CO at SV ¼ 1400 h21 ; which was standard conditions), the absorption capacity increased to 103%. ZnFe2O4/YL exhibited the high absorption capacity, even though the absorption of H2S was carried out under
N. Ikenaga et al. / Fuel 83 (2004) 661–669
667
gas H2 and CO and the space velocity were increased 3/2 times as compared to those of standard conditions. These results clearly indicate that ferrite prepared with YL coal is an excellent absorbent for H2S from the view point of not only the absorption capacity but also the absorption rate. 3.6. Regeneration and repeated use of absorbent Regeneration and repeated use of absorbent for H2S removal is crucial to practical use. In order to ascertain the possibility of regeneration of absorbent after the absorption reaction of H2S, the temperature-programmed oxidation (TPO) of the H2S absorbed ZnFe2O4/YL (Sample L) and ZnFe2O4/AC (Sample H) was carried out using a mixed gas of Ar and O2 (50 vol%). The regeneration of sulfurized absorbent seems to proceed as follows; ZnS þ 2FeS þ 3O2 ! ZnFe2 O4 þ SO2
Fig. 6. Relationship between absorption capacity and calcination temperature. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min, Absorption temperature; 500 8C (H2S concentration; 4000 ppm, Absorbent: ZnFe2O4/YL ( Sample L )).
a reductive atmosphere of H2 and CO. The absorption capacity slightly decreased to 94% with increasing the space velocity from 1400 to 2800 h21, and 1 –2 ppm of H2S was observed in the effluent gas until the breakthrough point. However, high H2S absorption performance of ZnFe2O4/YL did not change, even though the concentrations of reducing
TPO profile of the H2S absorbed ZnFe2O4/YL is illustrated in Fig. 8. The main release was monitored by stronger SOþ ðm=z ¼ 48Þ and SOþ 2 ðm=z ¼ 64Þ from the used absorbent was confirmed with a mass spectroscopy in the temperature range of 300 –500 8C. From this result regeneration of H2S absorbed ZnFe2O4/YL was carried out at 450 8C for 30 min using the fixed-bed flow type reactor in oxygen –argon gases (O2, 50 vol%). The regeneration behavior of H2S absorbed ZnFe2O4/YL is illustrated in Fig. 9. From this figure it was confirmed that the first and the third regeneration of absorbent was completed, because the releases of SOþand þ SOþ 2 terminated until 250 s. The slight release of CO2 was observed at 250 s through burning coal char remained in the absorbent in the first regeneration, but COþ 2 was not observed in the third regeneration cycle. Fig. 10 shows the XRD patterns of ferrites prepared with YL coal before and after the H2S absorption and after four sulfurization – regeneration cycles. ZnFe2O4/YL exhibited diffraction peaks ascribed to ZnFe2O4 and very weak diffraction peaks of ZnO and Fe2O3. The average crystallite size ðd311 Þ
Fig. 7. Effect of space velocity on the absorption of H2S. (Absorbent: ZnFe2O4/YL (Sample L) calcined at 500 8C).
668
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 8. Temperature-programmed oxidation of sulfided ZnFe2O4/YL (Sample L). Oxidation condition: heating rate; 5 8C/min, Ar þ O2 (50 vol%) flow rate; 5 ml/min.
Fig. 10. XRD patterns of ZnFe2O4/YL before and after H2S absorption and after regeneration.
Fig. 9. Regeneration behavior of H2S absorbed ZnFe2O4/YL (Sample L) at 450 8C for 30 min.
of ZnFe2O4 slightly increased from 10 to 14 nm during the regeneration. The sulfurization–regeneration of ZnFe2O4/AC proceeded the same way as ZnFe2O4/YL. Therefore, ZnFe2O4/AC and ZnFe2O4/YL could successfully be regenerated from ZnS, FeS, and Fe12xS at 450 8C, although AC and YL coal was completely burned out during the regeneration at 450 8C in a flowing air and the amount of carbon material decreased to nearly nil. The carbon material was not required to regenerate the absorbent. When the regeneration of ZnFe2O4 without carbon material (Sample C) was examined, the absorption capacity increased gradually from 42 to 58% by repeated sulfurization – regeneration cycle. Since cracks were observed on the surface of this particle by Scanning Electron Microscope observation during the sulfurization– regeneration runs, the particles of this absorbent would tend to be fine dust. On the other hand, Fig. 11 illustrates the breakthrough curves of H2S that were measured at 500 8C using fresh ZnFe2O4/YL (Fig. 11(a)) and ZnFe2O4/AC (Fig. 11(b)), regenerated zinc ferrite after the first (R-1), the second (R-2), and the third (R-3) regeneration cycles. The high absorption capacity of H2S with regenerated ZnFe2O4/YL and ZnFe2O4/AC did not change significantly. These regenerated ZnFe2O4/YL and ZnFe2O4/AC also showed
N. Ikenaga et al. / Fuel 83 (2004) 661–669
669
temperature of 500 8C. Coal and activated carbon-loaded zinc ferrites could be regenerated by oxidation with Ar –O2 (50 vol%) at 450 8C for 30 min. The regenerated zinc ferrites can be used for repeated absorption of H2S with a very slight decrease in the absorption capacity. ZnFe2O4/YL and ZnFe2O4/AC are excellent absorbents for the H2S absorption, because macro porous structure on the surface of these ferrites might contribute the higher absorption capacity.
Acknowledgements This work was financially supported by a grant-in-aid for priority area (11218212) from the Ministry of Education, Science, Japan.
References [1] [2] [3] [4] [5] [6] [7] Fig. 11. Breakthrough curves of H2S absorption using YL coal and ACsupported ZnFe2O4 and regenerated YL coal and AC-supported ZnFe2O4. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min, Absorption temperature; 500 8C (H2S concentration; 4000 ppm).
no H2S leakage in the effluent gas until breakthrough point, and deterioration of H2S absorption performance of this absorbent was not shown in four sulfurization – regeneration cycles. These results clearly indicate again that ZnFe2O4/YL and ZnFe2O4/AC are excellent absorbents for the H2S absorption, because macro porous structure developed by pyrolysis gas of coal or gas from burning carbonaceous materials, and did not affect the surface area, on the surface of these ferrites might contribute the higher absorption capacity.
[8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]
4. Conclusion [23]
Zinc ferrites could be prepared in the presence of activated carbon and Yallourn coal by impregnation method, and calcined at 500 8C in air. Performance of the zinc ferrite for H2S absorption was evaluated using a fixedbed flow type reactor. These absorbents exhibited larger absorption capacity of H2S than that of unloaded ferrites and could remove 4000 ppm of H2S to less than a few ppm at space velocity of 1400 – 2800 h21 at the absorption
[24] [25] [26] [27] [28] [29] [30]
Wang R. J Chem Engng Jpn 2002;35:15. Quinn R, Appleby JB, Pez GP. Sep Sci Technol 2002;37:627. Slimance RB, Abbasian J. Ind Engng Chem Res 2000;39:1338. Ahmed MA, Garcia E, Alonso L, Palacios JM. Appl Surf Sci 2000; 156:115. Bandosz TJ. Carbon 1999;37:486. Cal MP, Strickler BW, Lizzio AA. Carbon 2000;38:1757. Cal MP, Strickler BW, Lizzio AA, Gangwal SK. Carbon 2000;38: 1767. Slimane RB, Abbasiann J. Adv Environ Res 2000;4:147. Sasaoka E, Ichio T, Kasaoka S. Energy Fuels 1992;6:603. Sasaoka E, Hirano S, Kasaoka S, Sakata Y. Energy Fuels 1994;8:763. Sasaoka E, Hatori M, Yoshimura H, Su C, Uddin MA. Ind Engng Chem Res 2001;40:2512. Slimane RB, Hepworth MT. Energy Fuels 1994;8:1175. Slimane RB, Hepworth MT. Energy Fuels 1994;8:1184. Slimane RB, Hepworth MT. Energy Fuels 1994;9:372. Pineda M, Fierro JLG, Palacios JM, Cilleuelo C, Ibarra JV. J Mater Sci 1995;30:6171. Pineda M, Palacios JM, Garcia E, Cilleuelo C, Ibarra JV. Fuel 1997; 76:567. Ahmed MA, Alonso L, Palacios JM, Cilleruelo C, Abanades JC. Solid State Ionics 2000;138:51. Lew S, Sarofim AF, Stephanopoulos MF. Ind Engng Chem Res 1992; 31:1890. Mojtahedi W, Salo K, Abbasian J. Fuel Process Technol 1994;37:53. Mojtahedi W, Abbasian J. Energy Fuels 1995;9:429. Mojtahedi W, Abbasian J. Energy Fuels 1995;9:782. Konttinen JT, Zevenhoven CAP, Hupa MM. Ind Engng Chem Res 1997;36:2332. Konttinen JT, Zevenhoven CAP, Hupa MM. Ind Engng Chem Res 1997;36:2340. Jothimurugesan K, Gangwal SK. Ind Engng Chem Res 1998;37:1929. Ozdemir S, Bardakci T. Sep Purif Technol 1999;16:225. Akyurtlu JF, Akyurtlu A. Gas Sep Purif 1995;9:17. Pineda M, Fierro JLG, Palacios JM, Cilleruelo C, Garcia E, Ibarra JV. Appl Surf Sci 1997;119:1. Ikenaga N, Chiyoda N, Matsushima H, Suzuki T. Fuel 2002;81:1569. Ohgaito Y, Matsushima H, Ikenaga N, Suzuki T. 224th ACS National Meeting. Environ Chem Dev Prep 2002;135. Jagtap SB, Weelock TD. Energy Fuels 1996;10:821.
Preparation of zinc ferrite in the presence of carbon material and its application to hot-gas cleaning Na-oki Ikenaga*, Yousuku Ohgaito, Hiroaki Matsushima, Toshimitsu Suzuki Department of Chemical Engineering, Faculty of Engineering, Kansai University 3-3-35 Yamate, Suita, Osaka 564-8680, Japan Received 17 March 2003; revised 29 July 2003; accepted 1 August 2003; available online 1 December 2003
Abstract In order to develop an efficient absorbent of H2S in coal gasification, zinc ferrite (ZnFe2O4) was prepared in the presence of carbon materials such as activated carbon (AC), activated carbon fiber (ACF), and Yallourn coal (YL). The absorption behavior of absorbents for H2S was examined using a fixed-bed flow type reactor equipped with a quadrupole mass spectrometer. Carbon material-supported ZnFe2O4 exhibited larger desulfurization capacity for H2S than unsupported ferrites. They could efficiently remove H2S from 4000 ppm levels in a simulated coal gasification gas to less than 1 ppm at 500 8C. The absorption capacity of H2S with ZnFe2O4/AC, ZnFe2O4/ACF, and ZnFe2O4/YL exhibited nearly 100% of stoichiometric amount of loaded metal species. They could be regenerated by an air oxidation in O2 – Ar (50 vol%) at 450 8C for 30 min. The regenerated ferrite can be used for repeated absorption of H2S with a very slight decrease in the absorption capacity. q 2003 Elsevier Ltd. All rights reserved. Keywords: Hydrogen sulfide; Zinc ferrite; Activated carbon; Coal; Preparation of ferrite; Hydrogen sulfide absorbent
1. Introduction Since energy consumption would progressively increase in developing countries, development of alternate clean fuel for current petroleum or new energy sources are important research area. For this purpose, natural gas and coal are the most promising sources, and their uses are increasing currently. Use of a large amount of coal causes various environmental problems such as a large amount of CO2 and several pollutants emission. Development of clean and efficient advanced coal utilization technology is of current importance. Increasing attention is being paid to integrated coal gasification combined cycle (IGCC) process and gasification-molten carbonate fuel cell (MCFC) technology as one of the most promising technologies. For these technologies, several thousands ppm of hydrogen sulfide containing in an effluent gas from gasifier must be efficiently removed to several ppm for IGCC and , 1 ppm for MCFC. Wet processes using basic organic solvents are commercially established technologies, and further investigation is in progress [1,2]. However, hot gas from the gasifier must be * Corresponding author. Tel.: þ 81-6-6368-0792; fax: þ81-6388-8869. E-mail address: [email protected] (N. Ikenaga). 0016-2361/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2003.08.019
cooled down below 100 8C in order to apply an organic solvent as an absorbent. This looses energy contained as latent heat in the hot gas [3,4]. Hot gas cleaning has been conducted by using various inorganic absorbents such as activated carbon [5 – 7], ZnO [8,10], Fe2O3 [8,9,11], CuO [8,11], Mn2O3 [8,12 –14], and ZnFe2O4 [15 – 17]. For example, Slimane et al. used four kinds of manganese oxides such as MnO, Mn3O4, Mn2O3, and MnO2 for desulfurization of hot coal-derived fuel gases and reported the reduction and sulfidation thermodynamics and kinetic characteristics of these absorbents in detail [12 –14]. They systematically studied copper, iron, manganese, and zinc oxide as bases for developing regenerable absorbent for hydrogen sulfide absorption at moderate temperatures. These absorbents can remove hydrogen sulfide from an effluent gas below 10 ppm. For fuel cell use, however, further reduction of hydrogen sulfide is required. Many researchers have been trying to develop high performance absorbent for hydrogen sulfide removal such as zinc titanate [18–25], zinc ferrite doped with several kinds of metal (V, Ti, and Cu) [26,27], and activated carbonsupported zinc ferrite [28]. Lew et al. indicated that Zn–Ti–O absorbent exhibited the high activity for hydrogen sulfide removal and the sulfidation of Zn –Ti –O proceeded by
662
N. Ikenaga et al. / Fuel 83 (2004) 661–669
the same mechanism as ZnO sulfidation [18]. However, the reactivity of zinc titanate absorbent towards hydrogen sulfide gradually decreased during sulfidation – regenaration tests [20]. In the present work, activated carbon-, activated carbon fiber-, and coal-supported zinc ferrite were prepared at a low temperature in order to obtain a highly active absorbent for H2S removal, which can be used repeatedly. Characterization of zinc ferrite prepared here was carried out by using powder X-ray diffraction and BET method. The absorption behavior of the absorbent was examined by using a fixedbed flow type reactor. Regeneration of the used absorbent for H2S removal was investigated.
was calcined at 500 – 800 8C for 2 h in air to obtain zinc ferrite [28].
2. Experimental
2.4. Absorption of H2S
2.1. Materials
The breakthrough curves were obtained by using a fixedbed flow type reactor made of quartz. After placing a 10 mg of absorbent and 40 mg of silica sand as a diluent in the reactor (i.d. 4 mm, 235 mm length), an inert gas (Ar) was introduced until the absorbent reached to the absorption temperature (300 – 700 8C), and H2S (5 ml/min) diluted with nitrogen to 0.96 vol% and Ar (5 ml/min) were supplied to the reactor (SV, 1400 h21). Exit gas was led to a quadrupole mass spectrometer (Q-mass), and the hydrogen sulfide concentrations of the tail gas were observed. H2 (17 vol%) and/or CO (33 vol%) were added to the inlet gas, in order to simulate coal gasification gas composition.
Activated carbon (AC, Wako Pure Chemical Industries, Ltd., 100-mesh pass, surface area 993 m2/g), activated carbon fiber (ACF, AD’ALL, Co., Ltd., surface area 2000 m2/g), Australian Yallourn brown coal (YL, C 66.3; H 4.70; N 0.48; S 0.26 daf%; Ash 2.0 d%, 100-mesh pass, surface area 5 m2/g) were used as a support. Zn(NO3)2· 6H2O, Fe(NO3)3·9H2O, urea, and aqueous ammonia (ca 28%) (Wako Pure Chemical Industries, Ltd.) were used to prepare absorbents in this study.
2.3. Characterization of absorbents The absorbents freshly prepared here and those after absorption of H2S were analyzed by powder X-ray diffraction using an X-ray diffractometer (Shimadzu XRD6000) with monochromatized Cu Ka radiation. Crystallite sizes of zinc ferrites were calculated from diffraction peak at 358 according to Scherrer’s equation. Surface areas of absorbents were measured by BET method using a Micromeritics Gemini 2375, applying adsorption isotherms of nitrogen at 2 196 8C.
2.2. Preparation of absorbent for H2S 2.2.1. Homogeneous precipitation method Into a mixed solution of Fe(NO3)3 and Zn(NO3)2 (150 mmol/l, 80 ml), 0.04 mmol of urea and 0.3 g of AC, ACF, or YL were added. The mixture was allowed to keep 90 8C for 2 h. During this stage, urea was decomposed to NH3 and CO to precipitate Zn(OH)2 and Fe(OH)3. AC, ACF, or YL coal was filtered off together with precipitates. Then the filtrate was dried at 70 8C under vacuum. Dried Zn(OH)2 and Fe(OH)3 on the support (AC, ACF, YL coal) were calcined at 400 or 500 8C under air flow for 2 h. 2.2.2. Coprecipitation method A 0.3 g of AC or YL coal was added into a mixed solution of Fe(NO3)3 (20 mmol/l, 100 ml) and Zn(NO3)2 (10 mmol/l, 100 ml), and then into this suspension aqueous ammonia was added until pH of the supernatant reaches 10. The combined precipitates was separated and dried in vacuo. The mixture was calcined at 500 8C under air flow for 2 h to obtain zinc ferrite. 2.2.3. Impregnation method A 0.3 g of AC or YL was added to a mixed solution of iron nitrate (III) (83 mmol/l, 30 ml) and zinc nitrate (40 mmol/l, 30 ml). After standing for overnight, the mixture was dried by evaporation at 70 8C, and the mixture
3. Results and discussion 3.1. Preparation and characterization of absorbents for H2S removal Table 1 summarizes surface area, the content of oxides, and the average crystallite size ðd311 Þ of various absorbents prepared on AC, ACF, and YL char prepared by various methods. Mixed oxides prepared by various methods were subjected to XRD analyses to determine oxide structures. Results of XRD patterns of typical absorbents are shown in Fig. 1. ZnFe2O4 prepared by dry process at 1000 8C showed sharp diffraction peaks ascribed to ZnFe2O4 (JCPDS card no 22-1012) with 343 nm of average crystallite size evaluated from the half height width of XRD peaks at 2u ¼ 358 according to Scherrer’s equation. ZnFe2O4 prepared without carbon materials by coprecipitation method at 500 8C (Sample C) showed relatively broad diffraction peaks of ZnFe2O4. When the zinc ferrites were prepared in the presence of AC and ACF by homogeneous precipitation and impregnation methods at 500 8C, diffraction peaks ascribed to ZnO and Fe2O3 were observed (Samples E, F, and H), in addition to ZnFe2O4. In the previous paper [28], we have reported that zinc ferrite could be produced on activated carbon at a lower calcinations temperature of 300 8C.
N. Ikenaga et al. / Fuel 83 (2004) 661–669
663
Table 1 Surafce area and oxide content of various sorbents Sample
Sorbent
Method
Calcination temp. (8C)
Phase determined by XRD
Crystallite size (nm)
S.A. (m2/g)
Oxide content (wt%)
Absorption capacitya (%)
A B C D E F G H I J K L
ZnFe2O4 Fe2O3 ZnFe2O4 ZnFe2O4 ZnFe2O4/ACF ZnFe2O4/AC ZnFe2O4/AC ZnFe2O4/AC Fe2O3/AC ZnFe2O4/YL ZnFe2O4/YL ZnFe2O4/YL
Dry process PPT CoPPT CoPPT HomoPPT HomoPPT CoPPT IMP IMP HomoPPT CoPPT IMP
1000 500 500 300 400 500 500 500 500 500 500 500
ZnFe2O4 Fe2O3 ZnFe2O4 Amorphous ZnFe2O4, ZnO, Fe2O3 ZnFe2O4, ZnO, Fe2O3 ZnFe2O4 ZnFe2O4, ZnO, Fe2O3 Fe2O3 ZnFe2O4 ZnFe2O4 ZnFe2O4
343 95 8 – 10 20 9 22 90 13 5 10
17 40 41 134 1190 317 463 50 67 51 58 50
– – – – 10.7 66.2 56.9 96.8 99.0 93.1 95.5 86.0
49 94 48 116 145 113 84 100 113 98 97 90
Carried out using a fixed-bed reactor (SV, 1400 h21; H2S, 4000 ppm; H2, 17 vol%; absorption temperature, 500 8C) PPT, precipitation method; CoPPT, coprecipitation method; HomoPPT, homogeneous precipitation method; IMP, impregnation method. a
The ferrites prepared in the presence of YL coal showed only diffraction peaks ascribed to ZnFe2O4 with 5– 13 nm of average crystallite size (Samples J, K, and L). In the cases of ferrites loaded on carbon material, the oxide content in the total absorbent was evaluated based on the weight decreases of the sample, by burning the organic part using a thermogravimetric balance in a flowing air. Surface areas of zinc ferrites prepared in the presence of ACF and AC (Samples D, F, and G) were higher than those of zinc ferrites (Samples A –C), which were prepared without carbon materials. Surface area of ZnFe2O4/ACF reached 1190 m2/g due to a large amount of remained high surface area ACF. Surface areas decreased to 50 m2/g with a decrease in the amount of carbon remaining (1 –14%). In the preparation of ZnFe2O4 in the presence of ACF and AC by homogeneous precipitation and coprecipitation methods, since Zn(OH)2 and Fe(OH)3 covered the surface of ACF and AC, the large amount of carbon seemed to be remained in the absorbent after the calcination.
3.2. Absorption behavior of H2S The performance of an absorbent can be evaluated in terms of both the absorption capacity of the absorbent and the absorption rate of H2S by using a fixed-bed flow type reactor, where the concentrations of H2S in the tail gas are an important factor in addition to the breakthrough time. Low concentration of H2S in the tail gas indicates that the absorbent has a high absorption rate and a low equilibrium concentration of H2S. Fig. 2 shows examples of breakthrough behavior of H2S using various absorbents measured at an absorption temperature of 500 8C. The horizontal axis of Fig. 2 designates the absorption capacity, which was calculated by assuming that all zinc and iron in the ferrite
would be transformed into ZnS and FeS as follows ZnFe2 O4 þ 3H2 S þ H2 ! ZnS þ 2FeSðFe12x S or FeS2 Þ þ 4H2 O Absorption capacity ð%Þ ¼
Amount of absorbed H2 S ðmolÞ=3 £ 100 Amount of ferrite ðmolÞ
The vertical axis of Fig. 2 indicates H2S concentration in the effluent. The absorption capacity of respective absorbent was estimated by extrapolating a slope of breakthrough curve to H2S concentration 0 ppm in the effluent, and the absorption capacity of each absorbent is shown in the last column of Table 1. The absorption capacity of several samples exceeded 100% because a part of the iron was transformed into FeS2. The formation of FeS2 was not confirmed with XRD analysis after the reaction, but the peak ascribed to FeS2 was observed with XPS analysis. The absorption capacities of H2S (48 – 49%) with unloaded ZnFe2O4 prepared by dry process and coprecipitation method (Samples A and C) were lower than those (. 85%) of loaded ZnFe2O4. In addition, when unloaded ZnFe2O4 prepared by dry process (Sample A) was used for H2S absorption, high H2S concentration of 5 – 10 ppm was observed in the effluent gas. Fe2O3 and Fe2O3/AC (Samples B and I) exhibited high absorption capacities of 94 and 113%, respectively. However, in the breakthrough curve H2S concentration of 10– 20 ppm was obtained in the effluent gas through out the run due to thermodynamic reasons [10]. These results indicate that Fe2O3 absorbent is not suitable for complete removal of H2S, and that zinc in the ferrite contributes to the high absorption rate of H2S. On the other hand, the absorption capacities of H2S increased to nearly 100% or higher with ZnFe2O4 prepared
664
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 1. XRD patterns of absorbents prepared by various methods.
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 2. Breakthrough curves of H2S absorption using various absorbents obtained at 500 8C. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 5 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min (SV, 1400 h21, H2S, 4000 ppm).
in the presence of AC, ACF, and YL coal (Samples E – L). These absorbents showed no H2S leakage in the effluent gas (less than 1 ppm). Although the surface area of Sample G was higher than that of Sample F, the absorption capacity of the former was lower than that of the latter. This fact indicates that the surface area seems not to be an important factor for the high absorption capacity of H2S at a high temperature. The important factor for high absorption capacity of H2S and high absorption rate of H2S seems to be not micro porous structure that contributes increasing surface area but macro porous structure that provides rapid diffusion of H2S into the absorbent [29]. The changes in the XRD patterns of ZnFe2O4 prepared with and without YL coal after H2S absorption are shown in Fig. 3. In the case of ZnFe2O4 prepared with YL coal, all of
665
ZnFe2O4 was transformed into ZnS, Fe12xS or FeS (Sample J, Fig. 3(a)) with 98% of the absorption capacity. On the other hands, unloaded ZnFe2O4 (Sample C), which exhibited low absorption capacity of 48%, showed weak diffraction peaks of ZnFe2O4 in addition to peaks of iron and zinc sulfides as shown in Fig. 3(b). The result is consistent with the lower absorption capacity of the unloaded ferrite. Although Sample C has slightly smaller crystallite size of 8 nm than that (13 nm) of Sample J, the transformation of ZnFe2O4 into iron and zinc sulfides did not proceed completely. Therefore, it seems reasonable to consider that the rapid diffusion of H2S into the absorbent is an important factor for the H2S absorption. Several mixed oxides of lower Zn/Fe ratio (ZnxFe32xO4, x ¼ 0 – 0:8) were prepared with YL coal and the absorption behavior of them for H2S was examined. When Zn0.8Fe2.2O4 ðZn=Fe ¼ 0:36Þ was utilized to absorb H2S, the absorption capacity decreased to 73%. The absorption capacity of H2S considerably decreased to 20% with decreasing Zn/Fe ratio to 0.034 and H2S concentration of 10 –30 ppm was obtained in the effluent gas until the breakthrough point. Ferrite structure (ZnFe2O4, Zn=Fe ¼ 0:5) seems to be required for the high H2S absorption capacity. 3.3. Effects of absorption temperature on the absorption behavior of H2S To examine the effect of the absorption temperature on the absorption capacity of the absorbent, the absorption of H2S was carried out with ZnFe2O4/YL in the temperature ranges of 300 – 700 8C. The result is illustrated in Figs. 4 and 5. Fig. 4 shows the breakthrough curves obtained in the H2S absorption at 300, 500, and 700 8C. In Fig. 5,
Fig. 3. XRD patterns of ZnFe2O4 and ZnFe2O4/YL after H2S absorption.
666
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 4. Breakthrough curves of H2S absorption using ZnFe2O4/YL obtained at various temperatures. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min, Absorption temperature; 300, 500, and 700 8C (H2S concentration; 4000 ppm).
the absorption capacity and the concentrations of H2S in the tail gas were plotted against the absorption temperature. ZnFe2O4 prepared in the presence of YL coal exhibited relatively high absorption capacity of H2S of 96% at a low absorption temperature of 300 8C. However, H2S concentration of 5 ppm was obtained in the effluent gas until the breakthrough point. At the absorption temperatures of 400 and 450 8C, the absorption capacities of H2S increased to 117 and 119%, respectively, and the H2S concentration in the tail gas decreased to less than 1 ppm. When the absorption temperature was increased from 500 to 600 8C, the absorption capacity of H2S decreased, however, the absorption capacities were nearly 100% (103 and 101%, respectively). The absorption capacity of H2S decreased to
65% with increasing the absorption temperature to 700 8C. According to XRD analysis of the absorbent after the H2S absorption at 700 8C, the intensity ratio of ZnS to Fe12xS decreased as compared to those in the absorbents reacted at 400– 600 8C. Furthermore, the H2S concentration in the tail gas obtained at 700 8C increased more gradually after the breakthrough point (at an absorption capacity of 65%) as compared to that obtained at 500 8C (Fig. 4). Therefore, it seems reasonable that the decrease in the absorption capacity would be ascribed to the reduction of ZnS or ZnO in the ferrite with H2 to give metallic Zn, followed by evaporation of Zn [16,30]. Although in the front edge of the absorbent in a fixed-bed reactor H2S concentration is high, H2S concentration is very low at the tail end of the fixed-bed as shown previously. In such circumstances either reverse reaction to give H2S or reduction of ZnO could be possible. These results indicate that optimal temperature of H2S removal with ZnFe2O4/YL is 400 –600 8C. 3.4. Effects of calcination temperature of absorbent on the absorption behavior of H2S In order to elucidate the effect of the calcination temperature of the absorbent on the absorption behavior of H2S, the breakthrough curves of the YL coal-supported zinc ferrites calcined at 500– 800 8C were measured at absorption temperature of 500 8C. Fig. 6 shows the crystallite size, the surface area, the absorption capacity of ZnFe2O4/YL (Sample L), and the concentration of H2S in the tail gas as a function of calcination temperature. The absorption capacity of ZnFe2O4/YL gradually decreased with increasing the calcination temperature to 800 8C. With an increase in the reaction temperature, crystallite size of ZnFe2O4 determined by XRD analyses increased gradually below 700 8C and markedly at 800 8C and surface area of ZnFe2O4 decreased linearly until 800 8C. This seems to indicate that crystallite size and surface area of ZnFe2O4 are partly contributed for the high absorption capacity of H2S, when the absorbent was prepared by the same method. 3.5. Effects of concentration of H2, CO and space velocity on the absorption of H2S
Fig. 5. Relationship between absorption capacity and absorption temperature. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min (H2S concentration, 4000 ppm, Absorbent: ZnFe2O4/YL (Sample L) calcined at 500 8C).
In order to obtain further information on the H2S absorption performance of ZnFe2O4 prepared in the presence of YL coal (Sample L), composition of feed gas and feed rate were varied in the absorption of H2S. The result is shown in Fig. 7. In the absorption of 4000 ppm of H2S at SV ¼ 1400 h21 without CO, the absorption capacity was 90%. When CO was added into the inlet gas (4000 ppm of H2S, 17 vol% of H2, and 33 vol% of CO at SV ¼ 1400 h21 ; which was standard conditions), the absorption capacity increased to 103%. ZnFe2O4/YL exhibited the high absorption capacity, even though the absorption of H2S was carried out under
N. Ikenaga et al. / Fuel 83 (2004) 661–669
667
gas H2 and CO and the space velocity were increased 3/2 times as compared to those of standard conditions. These results clearly indicate that ferrite prepared with YL coal is an excellent absorbent for H2S from the view point of not only the absorption capacity but also the absorption rate. 3.6. Regeneration and repeated use of absorbent Regeneration and repeated use of absorbent for H2S removal is crucial to practical use. In order to ascertain the possibility of regeneration of absorbent after the absorption reaction of H2S, the temperature-programmed oxidation (TPO) of the H2S absorbed ZnFe2O4/YL (Sample L) and ZnFe2O4/AC (Sample H) was carried out using a mixed gas of Ar and O2 (50 vol%). The regeneration of sulfurized absorbent seems to proceed as follows; ZnS þ 2FeS þ 3O2 ! ZnFe2 O4 þ SO2
Fig. 6. Relationship between absorption capacity and calcination temperature. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min, Absorption temperature; 500 8C (H2S concentration; 4000 ppm, Absorbent: ZnFe2O4/YL ( Sample L )).
a reductive atmosphere of H2 and CO. The absorption capacity slightly decreased to 94% with increasing the space velocity from 1400 to 2800 h21, and 1 –2 ppm of H2S was observed in the effluent gas until the breakthrough point. However, high H2S absorption performance of ZnFe2O4/YL did not change, even though the concentrations of reducing
TPO profile of the H2S absorbed ZnFe2O4/YL is illustrated in Fig. 8. The main release was monitored by stronger SOþ ðm=z ¼ 48Þ and SOþ 2 ðm=z ¼ 64Þ from the used absorbent was confirmed with a mass spectroscopy in the temperature range of 300 –500 8C. From this result regeneration of H2S absorbed ZnFe2O4/YL was carried out at 450 8C for 30 min using the fixed-bed flow type reactor in oxygen –argon gases (O2, 50 vol%). The regeneration behavior of H2S absorbed ZnFe2O4/YL is illustrated in Fig. 9. From this figure it was confirmed that the first and the third regeneration of absorbent was completed, because the releases of SOþand þ SOþ 2 terminated until 250 s. The slight release of CO2 was observed at 250 s through burning coal char remained in the absorbent in the first regeneration, but COþ 2 was not observed in the third regeneration cycle. Fig. 10 shows the XRD patterns of ferrites prepared with YL coal before and after the H2S absorption and after four sulfurization – regeneration cycles. ZnFe2O4/YL exhibited diffraction peaks ascribed to ZnFe2O4 and very weak diffraction peaks of ZnO and Fe2O3. The average crystallite size ðd311 Þ
Fig. 7. Effect of space velocity on the absorption of H2S. (Absorbent: ZnFe2O4/YL (Sample L) calcined at 500 8C).
668
N. Ikenaga et al. / Fuel 83 (2004) 661–669
Fig. 8. Temperature-programmed oxidation of sulfided ZnFe2O4/YL (Sample L). Oxidation condition: heating rate; 5 8C/min, Ar þ O2 (50 vol%) flow rate; 5 ml/min.
Fig. 10. XRD patterns of ZnFe2O4/YL before and after H2S absorption and after regeneration.
Fig. 9. Regeneration behavior of H2S absorbed ZnFe2O4/YL (Sample L) at 450 8C for 30 min.
of ZnFe2O4 slightly increased from 10 to 14 nm during the regeneration. The sulfurization–regeneration of ZnFe2O4/AC proceeded the same way as ZnFe2O4/YL. Therefore, ZnFe2O4/AC and ZnFe2O4/YL could successfully be regenerated from ZnS, FeS, and Fe12xS at 450 8C, although AC and YL coal was completely burned out during the regeneration at 450 8C in a flowing air and the amount of carbon material decreased to nearly nil. The carbon material was not required to regenerate the absorbent. When the regeneration of ZnFe2O4 without carbon material (Sample C) was examined, the absorption capacity increased gradually from 42 to 58% by repeated sulfurization – regeneration cycle. Since cracks were observed on the surface of this particle by Scanning Electron Microscope observation during the sulfurization– regeneration runs, the particles of this absorbent would tend to be fine dust. On the other hand, Fig. 11 illustrates the breakthrough curves of H2S that were measured at 500 8C using fresh ZnFe2O4/YL (Fig. 11(a)) and ZnFe2O4/AC (Fig. 11(b)), regenerated zinc ferrite after the first (R-1), the second (R-2), and the third (R-3) regeneration cycles. The high absorption capacity of H2S with regenerated ZnFe2O4/YL and ZnFe2O4/AC did not change significantly. These regenerated ZnFe2O4/YL and ZnFe2O4/AC also showed
N. Ikenaga et al. / Fuel 83 (2004) 661–669
669
temperature of 500 8C. Coal and activated carbon-loaded zinc ferrites could be regenerated by oxidation with Ar –O2 (50 vol%) at 450 8C for 30 min. The regenerated zinc ferrites can be used for repeated absorption of H2S with a very slight decrease in the absorption capacity. ZnFe2O4/YL and ZnFe2O4/AC are excellent absorbents for the H2S absorption, because macro porous structure on the surface of these ferrites might contribute the higher absorption capacity.
Acknowledgements This work was financially supported by a grant-in-aid for priority area (11218212) from the Ministry of Education, Science, Japan.
References [1] [2] [3] [4] [5] [6] [7] Fig. 11. Breakthrough curves of H2S absorption using YL coal and ACsupported ZnFe2O4 and regenerated YL coal and AC-supported ZnFe2O4. Condition: absorbent; 10 mg, Silica sand; 40 mg, Ar; 1 ml/min, N2 þ H2S (0.96 vol%); 5 ml/min, H2; 2 ml/min, CO; 4 ml/min, Absorption temperature; 500 8C (H2S concentration; 4000 ppm).
no H2S leakage in the effluent gas until breakthrough point, and deterioration of H2S absorption performance of this absorbent was not shown in four sulfurization – regeneration cycles. These results clearly indicate again that ZnFe2O4/YL and ZnFe2O4/AC are excellent absorbents for the H2S absorption, because macro porous structure developed by pyrolysis gas of coal or gas from burning carbonaceous materials, and did not affect the surface area, on the surface of these ferrites might contribute the higher absorption capacity.
[8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]
4. Conclusion [23]
Zinc ferrites could be prepared in the presence of activated carbon and Yallourn coal by impregnation method, and calcined at 500 8C in air. Performance of the zinc ferrite for H2S absorption was evaluated using a fixedbed flow type reactor. These absorbents exhibited larger absorption capacity of H2S than that of unloaded ferrites and could remove 4000 ppm of H2S to less than a few ppm at space velocity of 1400 – 2800 h21 at the absorption
[24] [25] [26] [27] [28] [29] [30]
Wang R. J Chem Engng Jpn 2002;35:15. Quinn R, Appleby JB, Pez GP. Sep Sci Technol 2002;37:627. Slimance RB, Abbasian J. Ind Engng Chem Res 2000;39:1338. Ahmed MA, Garcia E, Alonso L, Palacios JM. Appl Surf Sci 2000; 156:115. Bandosz TJ. Carbon 1999;37:486. Cal MP, Strickler BW, Lizzio AA. Carbon 2000;38:1757. Cal MP, Strickler BW, Lizzio AA, Gangwal SK. Carbon 2000;38: 1767. Slimane RB, Abbasiann J. Adv Environ Res 2000;4:147. Sasaoka E, Ichio T, Kasaoka S. Energy Fuels 1992;6:603. Sasaoka E, Hirano S, Kasaoka S, Sakata Y. Energy Fuels 1994;8:763. Sasaoka E, Hatori M, Yoshimura H, Su C, Uddin MA. Ind Engng Chem Res 2001;40:2512. Slimane RB, Hepworth MT. Energy Fuels 1994;8:1175. Slimane RB, Hepworth MT. Energy Fuels 1994;8:1184. Slimane RB, Hepworth MT. Energy Fuels 1994;9:372. Pineda M, Fierro JLG, Palacios JM, Cilleuelo C, Ibarra JV. J Mater Sci 1995;30:6171. Pineda M, Palacios JM, Garcia E, Cilleuelo C, Ibarra JV. Fuel 1997; 76:567. Ahmed MA, Alonso L, Palacios JM, Cilleruelo C, Abanades JC. Solid State Ionics 2000;138:51. Lew S, Sarofim AF, Stephanopoulos MF. Ind Engng Chem Res 1992; 31:1890. Mojtahedi W, Salo K, Abbasian J. Fuel Process Technol 1994;37:53. Mojtahedi W, Abbasian J. Energy Fuels 1995;9:429. Mojtahedi W, Abbasian J. Energy Fuels 1995;9:782. Konttinen JT, Zevenhoven CAP, Hupa MM. Ind Engng Chem Res 1997;36:2332. Konttinen JT, Zevenhoven CAP, Hupa MM. Ind Engng Chem Res 1997;36:2340. Jothimurugesan K, Gangwal SK. Ind Engng Chem Res 1998;37:1929. Ozdemir S, Bardakci T. Sep Purif Technol 1999;16:225. Akyurtlu JF, Akyurtlu A. Gas Sep Purif 1995;9:17. Pineda M, Fierro JLG, Palacios JM, Cilleruelo C, Garcia E, Ibarra JV. Appl Surf Sci 1997;119:1. Ikenaga N, Chiyoda N, Matsushima H, Suzuki T. Fuel 2002;81:1569. Ohgaito Y, Matsushima H, Ikenaga N, Suzuki T. 224th ACS National Meeting. Environ Chem Dev Prep 2002;135. Jagtap SB, Weelock TD. Energy Fuels 1996;10:821.