Desulfurization of coke oven gas from the coking coke blended with a sorbent and waste plastic
Abstract A new way to implement the simultaneous reuti-lization of solid waste, the desulfurization of coke oven gas (COG), and even the desulfurization of coke by the co-coking of coking coal (CC) and waste plastic (WP) blended with a sorbent is proposed; the evolution of H 2S and the removal ef? ciency of H 2S from COG during the co-coking process were investigated in a lab-scale cylindrical reactor. The experimental results indicated that for the coking of CC blended with ZnO, Fe 2O 3, or blast furnace dust (BFD) as a sorbent, the instantaneous concentration of H 2S in COG was lower than 500 mg/m 3 (which meets the technical speci? cation requirement of the Chinese Cleaner Produc-tion Standard–Coking Industry, HJ/T 126-2003) when the molar ratio between the key component of the sorbent and the volatile S in CC or the CC/WP blend, n Zn +Fe /n S , was about 1.2 for ZnO and Fe 2O 3, but not for BFD under the same conditions, suggesting that ZnO and Fe 2O 3 are prom-ising sorbents, but that BFD must be treated chemical or thermally before being used as a sorbent because of the size and complicated nature of the in? uence of its phase/chem-ical composition on its desulfurization ability. However, for the co-coking of CC and WP blended with ZnO as a sorbent, n Zn +Fe /n S must increase to 1.4 and 1.7 for 100/2 and 100/5 blends of CC/WP, respectively, to ensure a satisfactory ef? -ciency for H 2S removal from COG.
Key words Coking · Coking coal · Waste plastic · Sorbent · Desulfurization Introduction
During the coking of coal, sulfur in coal can transfer to coke
oven gas (COG) in the gaseous forms of H 2S, COS, or CS 2, for example, or remain as residue in tar and/or coke. Con-ventionally, sulfur compounds in COG are removed by spraying ammonia water into COG in special desulfuriza-tion equipment, which increases investment and operating costs; any sulfur remaining in the coke negatively affects the quality of the coke. Therefore, to maximize economic bene? t and minimize environmental liability, it is necessary to develop a new method to remove sulfur from COG and coke. Guo et al.1 and Ye et al.2 found that a satisfying H 2S removal ef? ciency from COG could be achieved at very low cost by mixing a ZnO-based additive into the coal. Tang
et al.3 reported that the in?
uence of sorbent addition on coke quality would be diminished, or even negated, by maintaining a reducing atmosphere in the coking chamber at the stage of coke formation.
Metallurgical dust (MD) and aging-resistant, undegrad-able waste plastic (WP) are usually regulated as solid wastes because of their potential hazard to the environment. For-tunately, the principles of sustainable development and advanced technologies pave the way for recycling MD and WP. For example, MD containing metal oxides (e.g., ZnO
and Fe 2O 3) can be recycled as an additive for sulfur removal from COG.1,2 WP, especially thermoplastic polyole?
ns [e.g., Polyethene (PE), Polypropene (PP), and Polystyrene (PS)] can be recycled by introducing them to coal blends for coke manufacture, because it would be more expensive to dispose of them otherwise, and even more important is that the cracking of such polyole? ns tends to produce different hydrocarbons, H 2, CO, and residual carbon, for example, to supplement those of the coal.4
Theoretically, if WP and coal were simultaneously
treated together with a sorbent (e.g., ZnO, Fe 2O 3, or MD) at the stage of semicoke formation, S in coal and WP could be converted mainly to H 2S in COG and captured by the sorbent to form sul? des to remain in the semi-coke; subse-quently the sul? des could be reduced by reducing gases
J Mater Cycles Waste Manag (2007) 9:7–14 ? Springer 2007
DOI 10.1007/s10163-006-0165-6
Zhao Rongfang · Ye Shufeng · Xie Yusheng · Chen Yunfa
Desulfurization of coke oven gas from the coking of coking coal blended with a sorbent and waste plastic
Z. Rongfang · Y. Shufeng (*) · X. Yusheng · C. Yunfa
Institute of Process Engineering, Chinese Academy of Sciences, Beijing, 100080, China Tel. +86-10-62588029; Fax +86-10-62542803
e-mail: sfye@https://www.sodocs.net/doc/aa10801465.html,
Z. Rongfang
Jiangxi Key Laboratory of Surface Engineering, Jiangxi Science and
Technology Normal University, Nanchang, China
Part of this paper was presented at the International Symposium on
EcoTopia Science 2005 (ISET05), Aug 8–9, 2005, Chikusa-ku, Nagoya, Japan Received: June 30, 2006 / Accepted: October 16, 2006
8
(e.g., H2 and CO) in the COG generated at the later stages of coke formation to form H2S and leave the coking chamber. Thus reutilization of solid wastes (i.e., WP and MD), desulfurization of COG, or even desulfurization of coke might be simultaneously accomplished during the coking process.
The primary goal associated with this project was to investigate the feasibility of the co-coking of coal and WP blended with a sorbent, and to propose a new way to imple-ment simultaneously the reutilization of solid waste, the desulfurization of COG, and even the desulfurization of coke during the co-coking process. The speci? c goals were to ? nd universal and quantitative descriptions of the follow-ing: (1) H2S evolution from the co-coking of coal and WP;
(2) the desulfurization effects of different sorbents on COG generated during the coking of coal; (3) the removal ef? -ciency of H2S from COG during the co-coking of coal and WP blended with a sorbent; (4) and the in? uence of WP and sorbent on the quality of the coke.
The present article focuses on the evolution of H2S and its removal ef? ciency from COG during the co-coking process performed in a lab-scale cylindrical coking chamber [inside diameter (i.d.) 40 mm, length 200 mm]. Moreover, the WP used in this study were mainly common thermoplas-tics (e.g., PE, PP, and PS), whereas Polyvinylchloride (PVC), Polyethylene terephthalate (PET), and other waste plastics (e.g., automobile shredder residue, ASR) were not applied because PVC is related to the emission of dioxin to some extent, ASR consists mainly of thermosetting plastic and is dif? cult to melt and pyrolyze, and PET bottles are easily sorted from other waste plastic and can be subjected to monomer recycling by decomposition in a supercritical methanol or glycolic environment.
Experimental
Materials
Coking coal (CC) from Beijing Coking and Chemistry Plant with a moisture content of 3.12% on an accepted basis was ? rst dried in an oven at 80°C with N2 as an antioxidant to remove free water and was then ground to a narrow particle size range in preparation for proximate and ultimate analy-sis (≤0.125 mm) and for the coking study (≤2.36 mm).
WP powder/beads, obtained from Nanchang Organic Chemical Plant, was a mixture consisting mainly of PE, PP, PS, and dechlorinated PVC and was prepared to have average particle sizes of ≤0.25 mm for proximate and ulti-mate analysis and ≤2.36 mm for the coking test.
Proximate analysis of CC and WP was carried out accord-ing to the GB3715-91 Standard of China. Sulfur content was measured on a PerkinElmer PE2400 series II elemental analyzer (Norwalk, Connecticut, USA), while the chlorine content was measured using a Jiangfen WK-2D microcou-lometer (Jiangsu, China). Table 1 presents the proximate and ultimate analyses of CC and WP.
Blast furnace dust (BFD) was taken from a cyclone sep-arator at a plant operated by Ji’nan Iron and Steel Company. ZnO and Fe2O3were commercial reagents (analytical grade) from Tianjin No.3 Chemical Reagent Factory (Tianjin, China). The chemical composition of BFD was measured on a Jobin Yvon JY38 atomic emission spec-troscope (Longjumeau, France) as well as by conventional chemical analysis; the phase composition analysis was performed on a Philips X’pert PRO MPD X-ray diffractom-eter (XRD, Eindhoven, Netherlands) and identi? ed by comparing the measured XRD patterns with JCPDS cards. Table 2 shows the chemical and phase composition of BFD.
Particle size distribution was analyzed on a Coulter LS230 particle size analyzer (Fullerton, California, USA). The average particle sizes of BFD, ZnO, and Fe2O3were 100 μm, 2.5 μm, and 2.5 μm, respectively.
Experimental apparatus
The co-coking properties of CC, WP, and sorbent were investigated in the experimental apparatus shown schemat-ically in Fig. 1, which included (1) a ceramic electric furnace with a stainless cylindrical reactor (40 mm i.d., 200 mm in length), (2) an S-type thermocouple and a temperature program controller, (3) a tar and water separation and col-lection system, (4) and a gas collection and analysis system.
Table 1.Proximate and ultimate analysis of coking coal and waste plastic
Proximate analysis (w%)Ultimate
analysis
(w ad%)Particle size (mm)
Moisture, M ad Ash, A d Volatile, V daf S Cl
CC0.7610.4525.760.62–≤2.36a
≤0.125b WP0 3.4286.19 4.33 1.08≤2.36a
≤0.25b CC, coking coal; WP, waste plastic; ad, air dry basis; d, dry basis; daf, dry and ash-free basis
a For coking study
b For proximate and ultimate analysis
9
Experimental method
After a sample was introduced into the reactor and air in the reactor system had been displaced with an inert gas, the reactor was externally heated at a predetermined heating rate to a predetermined ? nal temperature, which was main-tained for 2 h. During the coking process, the evolved vola-tile material passed through a condenser to trap coal tar and water and then the noncondensable gases were released to the gas collector; meanwhile, the H 2S concentration was measured using a gas detection tube that enabled easy
instant analysis of H 2S at set intervals and the volume ?
ow rate of COG was recorded. After the experiment, the coke was collected for S and Zn content analysis using a Jobin Yvon JY38 atomic emission spectroscope, and ash content analysis was done using a Setaram Labsys (Caluire, France) thermogravimetry analyzer.Experimental conditions
In an industrial-scale coke oven, after coal is introduced into the coking chamber, heat ? ows successively but slowly to the center of the chamber, creating diverse temperature pro? les in different parts of the chamber, as shown in Fig. 2a,5 thus forming different coke formation zones, i.e., the coal zone, the preplastic zone, the plastic zone, the solidi? cation zone, and the coke zone.5,6
Our previous research on pyrolysis of CC indicated that active thermal decomposition of the CC takes place over the temperature range 400°–650°C, accompanied by the release of the majority of volatile compounds, including water, COG, and coal tar.7 Furthermore, the release of volatile compounds greatly depends on the heating rate; in detail, the slower the heating rate is, the more completely the CC pyrolyzes, and the more volatile compounds evolve. The isotherms of coal in different parts of an industrial-scale coking chamber (Fig. 2b) show that the residence times of coal for temperature curves No. 2–5 in Fig. 2a over temperature range 400°–650°C are 1.23 h, 2.78 h, 3.46 h, and 1.52 h respectively,5 meaning that the heating rate increases in the sequence No.4 < No.3 < No.5 < No.2; the linear heating rate indicated by temperature curve No. 4 within the corresponding temperature range means that it is easy to minimize the lag between the indicated and actual tem-perature. So, temperature curve No. 4 and a ? nal tempera-ture of 1050°C, which was maintained for 2 h, were adopted in this lab-scale investigation.
Table 2. Chemical and phase composition of ZnO, Fe 2O 3, BFD, and C-BFD Element
Composition (wt%)ZnO
Fe 2O 3
BFD
C-BFD
Fe
Hematite (Fe 2O 3): 99.0
Hematite (Fe 2O 3) Fe (II): 4.53Hematite (Fe 2O 3): 20.01
Magnetite (Fe 3O 4) Fe (III):10.43Wustite (FeO)
Metallic iron (Fe): 0.12a ΣFe: 15.08ZnO Zincite: 99.0
Zincite: 8.05Zincite: 10.68SiO 2Quartz: 4.43Quartz: 5.86Al 2O 3 3.00 3.98CaO 5.607.43MgO 1.35 1.65PbO 1.87
2.14
C Graphite: 22.08Graphite: 0.31H 2O
2.86
BFD, blast furnace dust; C-BFD, BFD calcined at 700°C for 3 h with air as an oxidant a
Calculated by difference
scrubber
meter
2solution
H S analysis
trapper Fig. 1. Schematic diagram of the experimental apparatus. The pipe work from the reactor to the tar trapper was wrapped with thermal insulation tape. a , cooling water inlet; b , cooling water outlet; Cd(ac. c.)2, Cd(CH 3COO)2
10
tration of H 2S in COG are listed in Table 3. In this study, H 2S removal ef? ciency is considered to be satisfactory if the
instantaneous concentration of H 2S in COG during the coking process is lower than 500 mg/m 3.Results and discussion
Ef? uent rate of COG from coking of CC
The actual temperature curve of the reactor heated at the
indicated rate of temperature curve No. 4 in Fig. 2a and the ef? uent rate of COG under the aforementioned conditions (see Fig. 3) indicate that the maximum lag between the actual and the indicated temperature was about 15°C and that COG was released over the temperature range 350°–690°C, with a maximum ef? uent rate of 2.82 l/(min·kg CC ) at 490°C and a total volume yield of 325 l/kg CC . The total volume yield of COG on a per unit weight of CC basis is obtained using the following equation:
V Q t
t
COG d =∫0
(1)
where, V COG is the total volume yield of COG, l/kg CC ; Q is the volume ef? uent rate of COG, l/(min·kg CC ); t is time, min.
0123456789101112131415
100200300400500600700800900
100011005
4
3
21T e m p e r a t u r e ( o
C )
Coking time ( hr )
2
4
6
8
10
12
14
050
100
150
200
D i s t a n c e t o s u r f a c e o f t h e c h a m b e r (m m )
Coking time (h)
(b) Isotherm of coal
(a) Typical temperature curves
Fig. 2. Typical temperature curves (a ) and isotherms (b ) of coal at
different parts of an industrial-scale coking chamber. 1, surface; 2,
adjacent to surface; 3, 130–140 mm from the center; 4, 50–60 mm from center; 5, the center Table 3. Standard limits on concentration of H 2S in coke oven gas from the Chinese Cleaner
Production Standard – Coking Industry Usage
Standard
International advanced
National advanced National basic City gas ≤20 mg/m 3≤20 mg/m 3≤20 mg/m 3Others
≤200 mg/m 3
≤500 mg/m 3
≤500 mg/m 3
Criterion for removal ef? ciency of H 2S from COG
According to the Chinese Environmental Protection Indus-try Standard HJ/T 126-2003,8 i.e., the Cleaner Production Standard – Coking Industry, the technical speci? cation requirement for COG quality and the limit on the concen-
200
400
600
800
1000
0.0
0.61.21.8
2.4
3.0
E f f l u e n t r a t e o f C O G (L /m i n /k g (C C ))
Time (min)0300
600
900
1200
T e m p e r a t u r e (o
C )
Fig. 3. Actual temperature curve and ef? uent rate of coke oven gas (COG ) during the coking of coking coal (CC )
11
H 2S evolution during the co-coking of CC and WP
Figure 4 shows the instantaneous concentration of H 2S in COG from the coking of CC and the co-coking of CC and WP at CC/WP weight ratios of 100/2 and 100/5. For the coking of CC, only one peak can be seen over the basic pyrolysis temperature range of 440°–650°C, with a maximum H 2S concentration of about 11 500 mg/m 3 at 523°C. However, for the co-coking of 100/2 and 100/5 blends of CC/WP, peaks can be seen over a wider temperature range of 305°–690°C, i.e., one peak with a maximum of 54 000 mg/m 3 at 400°C for the 100/5 blend and two peaks with maxima of
24 500 mg/m 3 at 393°C and 10
000 mg/m 3 at 517°C for the 100/2 blend. In addition, due to the high absolute percent-age of CC in the blends, hardly any difference was noted among the ef? uent rates of COG from the co-coking of blends with different ratios.
It is generally accepted that the thermal decomposition kinetics of both coal and plastic are consistent with the free-radial mechanism.4 Our previous research on the co-pyroly-sis of CC and WP indicated that the active thermal decomposition temperature range of CC overlaps partially with that of WP,7 thereby suggesting the possibility of inter-radical interactions among radicals derived from CC and WP during the thermal cotreatment process, e.g., co-pyrol-ysis or co-coking.9 In other words, radicals derived from CC become stabilized by hydrogen abstraction from those derived from WP, which is usually considered to be a poten-tial hydrogen donator owing to its relatively high H content, and thereby the thermal co-treatment of CC and WP is promoted.
Figure 5 shows the ? tting of a polynomial expression to the instantaneous concentration of H 2S for the 100/2 blend, suggesting that there should theoretically be two H 2S con-centration peaks for the blends, one attributed to the release
of H 2S derived from S in the WP and the other from the CC. However, the different proportions of WP in the blends result in different degrees of interaction, especially in the hydrogen transfer between radicals derived from the CC and WP, so the two H 2S concentration peaks overlap for the 100/5 blend, but not for the 100/2 blend.
Based on the H 2S concentration and the volume ef?
uent rate of COG, the cumulative weight of H 2S in COG (W H 2S ,COG ) and the cumulative percent of S in the blend converted to H 2S in COG (X CC or X CC/WP ) can be obtained using the following expressions:W C Q t
t
H S,COG H S 22d =∫0
(2)
X W M M W X W M M CC H S,COG S
H S S,CC CC/WP H S,COG S
H S 2222for CC
=??×=
??100%W S,CC/WP
for CC/WP blend
×100%
(3)
where W H 2S,COG is the cumulative weigh of H 2S in COG, mg/kg CC ; C H 2S is the instantaneous concentration of H 2S in
COG, mg/m 3; Q is the volume ef?
uent rate of COG, l/(min·kg CC ); t is time, min; X CC is the cumulative weight percent of S in CC converted to H 2S in COG, wt%; X CC/WP is the total cumulative weight percent of S in CC/WP blend converted to H 2S in COG, wt%; M H 2S and M S are the molec-ular weights of H 2S and S respectively, g/mol; W S,CC is the total weight of S in CC, mg/kg CC ; and W S,CC/WP is the total weight of S in CC/WP blend, mg/kg CC . The experimental value for X CC was 29.1% and X CC/WP was 38.0% and 46.8% for 100/2 and 100/5 blends, respectively.
On the other hand, neglecting any interactions between CC and WP during the co-coking process, X CC/WP can be expressed as the simple linear addition of X CC and X WP :X S W S W S W X S W X CC WP CC CC WP WP CC CC CC WP WP WP
+()=+
(4)where, S CC , W CC , and X CC are the S content of CC, the weight percent of CC in the blend, and the percent of S in 200
400600800
1000
015000
30000
45000
60000
C o n c e n t r a t i o n o f H 2S i n C O G (m g /m 3
)
Temperature (o
C)
Fig. 4. Concentration of H 2S in COG during coking of CC and co-coking of CC and waste plastic (WP ) with 100/2 and 100/5 weight ratios of CC/WP
200
400600800
1000
05000
100001500020000
25000C o n c e n t r a t i o n o f H 2S i n C O G (m g /m 3
)
Temperature (o
C)
Fig. 5. Polynomial ? tting for the H 2S concentration for a 100/2 weight ratio of CC/WP
12
CC converted to H 2S in COG, respectively, while S WP , W WP , and X WP are the corresponding variables for WP. Thus for the 100/2 and 100/5 blends, Eq. 4 can be presented as the following simultaneous equations with two unknown variables,
38.0%0.62%100 4.33%20.62%100 4.33%2CC WP ××+×()
=××+××x x
(4a)46.8%0.62%100 4.33%0.62%100 4.33%CC WP ××+×()
=××+××55x x
(4b)The solutions of these equations are X WP = 95.0% and X CC = 30.1%; the former is within the theoretical bound of 0 ≤ X WP ≤ 100% and so is acceptable, but the latter is greater than the experimental result of X CC = 29.1%, consequently verifying the slight interaction between CC and WP, espe-cially the donation of H to CC by WP during the co-coking process.
Ef? ciency of H 2S removal from COG during coking of CC
Taking account of the fact that the major sulfur pollutant in COG is H 2S (up to 90%) with small amounts of COS, CS 2, and others, the theoretical amount of sorbent required to absorb sulfur pollution is calculated on the basis that 40% of S in the CC and 100% in the WP are volatile and are released to COG during the co-coking process. It must be pointed out that BFD has abundant metal oxides having the ability to capture sulfur, but only zinc and iron are taken as key components because of their excellent desulfuriza-tion ef? ciencies 10 and the limited contributions of other minor elements.
Figure 6 shows the instantaneous concentrations of H 2S in COG during the coking of CC with ZnO, Fe 2O 3, or BFD as a sorbent with a molar ratio between the key component
of the sorbent and volatile S in CC or CC/WP blend (n Zn +Fe /n S ) of about 1.2 (Table 4). It is evident that the maximum H 2S concentration in the whole process of coking is lower than 500 mg/m 3 for ZnO and Fe 2O 3, meaning a satisfactory ef? ciency for the removal of H 2S from COG. However, BFD exhibits a lower desulfurization ability under the same conditions, with maximal H 2S concentrations of about 2000 mg/m 3 at 523°C and 1750 mg/m 3 at 930°C, and a cumu-lative weight of H 2S in COG of about 40 mg.
Previous reports have shown that the desulfurization activity of a sorbent depends not only on the sulfuration activities of the key components but also on the diffusion ef? ciency of H 2S into the interior of the sorbent particle. Both Fe 3O 4 and FeO have the ability to capture H 2S from the thermodynamic viewpoint, but it is dif? cult, or even impossible for H 2S to successfully diffuse into Fe 3O 4 and FeO particles from the kinetic viewpoint because of the spinel structure and high density of Fe 3O 4 and the incom-patible crystal lattice parameters between FeO and SO 2.11
BFD differs greatly from ZnO and Fe 2O 3 in two ways (Table 2). One is the phase composition of iron oxides (FeO x ), i.e., Fe 2O 3 is composed only of Fe 2O 3, whereas BFD contains Fe 2O 3, Fe 3O 4, and FeO. The other difference is the chemical composition, i.e., besides ZnO and FeO x , BFD contains many other oxides such as CaO, SiO 2, Al 2O 3, MgO, and up to 22.08% of C. Thus we can infer that the lower desulfurization ability of BFD might be ascribed to its special chemical and/or phase composition. To support the infer-ence, BFD was calcined at 700°C for 3 h with air as an oxidant to oxidize FeO x to Fe 2O 3 and to remove C. Meanwhile, to follow the change in the microstructure [e.g., Brunauer-Emmet-Teller (BET) surface area and pore size distribu-tion] induced by calcination treatment, BFD and C-BFD were analyzed by N 2 sorption at 77K using an Autosorb1 QuantaChrome apparatus (Florida, USA), then the speci? c surface area was determined by the BET method, and the pore size distribution for pores smaller than 250 nm was cal-culated by the Barret-Joyner-Halenda (BJH) method. The chemical and phase composition of C-BFD is shown in Table 2. The N 2 sorption analysis showed that the sorption/desorp-tion isotherms of BFD and C-BFD approach type II behav-ior in terms of the International Union of Pure and Applied Chemistry (IUPAC) classi? cation,12,13 and also that calcina-tion leads to a slightly increased BET surface area (from 1.51 m 2/g t o 2.50 m 2/g) a nd t otal p ore v olume (from 0.0061 cm 3/g to 0.0137 cm 3/g), ascribed mostly to the volume increase of macropores, which are not crucial for desulfurization.
The instantaneous concentration of H 2S in COG during coking of CC with C-BFD as a sorbent and with a n Zn +Fe /n S
200
400600800
1000
03000
6000
9000
12000
C o n c e n t r a t i o n o f H 2S i n C O G (m g /m 3
)
Temperature (o
C)
Fig. 6. Concentration of H 2S in COG during the coking of CC with ZnO, Fe 2O 3, blast furnace dust (BFD ), or BFD calcined at 700°C for 3 h (C-BFD) as sorbents. n Zn +Fe /n S , molar ratio of the key component of sorbent to the volatile S in CC
Table 4. Experimental conditions used in the coking process
Coking process Experimental conditions
CC Sorbent: ZnO, Fe 2O 3, BFD, C-BFD n Zn +Fe /n S : 0–1.2
CC/WP Sorbent: ZnO n Zn +Fe /n S : 0–1.4 for 100/2 0–1.7 for 100/5n Zn +Fe /n S , molar ratio of the key components of sorbent and volatile S in CC or CC/WP blend
13
value of 1.2 (Table 4) is also shown in Fig. 6. Evidently, the H 2S removal ef? ciency of C-BFD is satisfactory within the temperature range below 730°C and the cumulative weight of H 2S decreases to about 12.5 mg, but the maximum H 2S concentration is still about 1000 mg/m 3 at 930°C, further implying that the in? uence of CaO, SiO 2, and others on the desulfurization ability of BFD/C-BFD cannot be neglected.
Sulfur in coal takes primarily two distinct forms, pyritic sulfur and organic sulfur, and these two forms thermally decompose over different temperature ranges. In detail, the weakest aliphatic organic sulfur compounds decompose completely below 500°C, aromatic organic sulfur com-pounds decompose between 400° and 750°C, and the most stable organic thiophenic sulfur compounds begin decom-posing above 500°C, but are still evident even at 850°C. Pyritic sulfur begins decomposing at about 470°C to form pyrrhotite (Fe 1-x S, x = 0–0.17), then to FeS in a nonoxidative atmosphere above 650°C, accompanied by the formation of active sulfur, which might react with organic matter in coal to from new C–S bonds and remain in the char, or might capture H derived from coal to form H 2S in COG.14–19 These reactions can be summarized as follows:
nFeS nFe S S 00.1721n →+=?()?x x
(5)Fe S FeS S 1n ?→+x
(6)S Organic-H nH S Organic
n 2+→+
(7)S Organic S-Organic
n +→
(8)Moreover, thermal decomposition of sulfur in coal is in? uenced by inorganic mineral matter, for example, the still-solid alkaline-earth metal oxides (e.g., CaO) serve as a diluent to the softening CC during the coking process, reducing the swelling and thermoplastic properties of CC, affecting the mass and/or heat transport in the coking process, and thereby inhibiting the continuous removal of products formed from decomposition of FeS 2.20–24 Similarly, silicate tends to induce the easily removable organic sulfur compounds to convert to thermally stable organic sulfur compounds, i.e., thiophenic or condensed thiophenic compounds.24–26
Accordingly, given the above points, it seems that ZnO and Fe 2O 3 are promising sorbents for the coking of coal, but BFD must be treated chemical or thermally before being used as a sorbent because of the signi? cant and complicated in? uence of its phase/chemical composition on its desulfur-ization ability.
Ef? ciency of H 2S removal from COG during co-coking of CC and WP
Sulfur in WP volatilizes rapidly to form H 2S over a rela-tively narrow temperature range due to its higher S content and special degradation behavior, which results in the abso-lutely higher concentration of H 2S in COG during the co-coking of CC and WP; consequently only ZnO was used as a sorbent in this part of the study because of its higher desulfurization ability. The instantaneous concentration of H 2S in COG during the co-coking of CC and WP with 100/2
and 100/5 weight ratios of CC/WP and with ZnO as a sorbent (Table 4), is shown in Fig. 7 and indicates that n Zn +Fe /n S must increase to 1.4 and 1.7 for the 100/2 and 100/5 blends, respectively, to ensure a satisfactory removal ef? -ciency of H 2S from COG.
Nevertheless, it is generally accepted that the coke yield in the coking process is about 75%–80%,5 and that the majority of the inorganic mineral matter introduced into the coking chamber, including ash in CC, the incompletely uti-lized sorbent, and the products from the sulfuration of the sorbent, might remain in the coke as ash and have negative effects on the quality of the coke. Thus, much work remains to be done to study whether it is essentially feasible to implement the simultaneous desulfurization of COG and coke by the co-coking of CC and WP with a directly blended sorbent, and if it is feasible, to determine the optimum amount of WP and sorbent and to what extent the addition of sorbent affects the quality of the coke.
200
400
600
800
1000
C o n c e n t r a t i o n o f H 2S i n C O G (m g /m 3
)
Temperature (o
C)
C o n c e n t r a t i o n o f H 2S i n C O G (m g /m 3)
Temperature (o
C)
Fig. 7. Concentration of H 2S in COG during co-coking of CC and WP with 100/2 and 100/5 weight ratios of CC/WP with ZnO as sorbent
14
Conclusions
A new way to implement simultaneously the reutilization of solid waste, the desulfurization of COG, and even the desulfurization of coke by the co-coking of CC and WP blended with ZnO, Fe2O3, BFD, or C-BFD as a sorbent was proposed, and H2S evolution and the ef? ciency of H2S removal from COG from the co-coking process were inves-tigated in a lab-scale cylindrical reactor. The experimental results can be summarized as follows:
1. During the coking of CC, COG was released over the
temperature range 350°–690°C, with a maximum ef? u-ent rate of 2.82 l/(min·kg CC) at 490°C and a total volume yield of 325 l/kg CC. The maximum concentration of H2S in COG was about 11 500 mg/m3 at 523°C and the cumu-lative percent of S in CC converted to H2S in COG (X CC) was 29.1%.
2. During the co-coking of CC and WP at 100/2 and 100/5
weight ratios of CC/WP, H2S was released over the wider temperature range of 305°–690°C with a maximum H2S concentration of 54 000 mg/m3at 400°C for the 100/5 blend, but with maxima of 24 500 mg/m3at 393°C and
10 000 mg/m3 at 517°C for the 100/2 blend. Moreover, the
? tting of a polynomial curve to the concentration curve of H2S for the blend and the calculated result of X CC =
30.1% verify the slight interaction between the CC and
WP during the co-coking process.
3. During the coking of CC blended with a sorbent, the
instantaneous concentration of H2S in COG was lower than 500 mg/m3(which meets the requirement of the Chinese Cleaner Production Standard – Coking Indus-try, i.e., HJ/T 123-2003) when the molar ratio between the key component of the sorbent and the volatile S in CC or CC/WP blend, n Zn+Fe/n S, was about 1.2 for ZnO and Fe2O3, but not for BFD. This fact suggests that BFD must be treated chemical or thermally before being used as a sorbent because of the signi? cant and complicated in? uence of its phase/chemical composition on its desul-furization ability. However, for the co-coking of CC and WP blended with ZnO as a sorbent, n Zn+Fe/n S must increase to 1.4 and 1.7 for the 100/2 and 100/5 CC/WP blends, respectively, to ensure a satisfactory ef? ciency of H2S removal from COG.
Acknowledgments The authors gratefully acknowledge ? nancial support from the National Natural Science Foundation of China (No. 50104010) and from the Key Laboratory of Multi-phase Reactions, Institute of Process Engineering, Chinese Academy of Sciences (2002-11).
References
1. Guo ZC, Tang HQ, Ye SF (2001) Removal of sulfur in coke oven
gas by mixing ZnO-based additive into coal. Trans Nonferrous Met Soc China 11:931–936
2. Ye SF, Xie YS, Guo ZC, Chen YF (2001) Desulfurization ability
of metallurgical dusts from blast furnace ironmaking process. J Chem Eng Jpn 34:1103–1107
3. Tang HQ, Guo ZC, Yao Y, Liu YX (2002) Desulfurization of coke
oven gas by mixing ZnO powder into coal (in Chinese). J Fuel Chem Technol 30:209–213
4. Sakurovs R (2003) Interaction between coking coals and plastic
during co-pyrolysis. Fuel 82:1911–1916
5. Guo SC (1992) Chemical technology of coal. Chemical Industry
Press, Beijing
6. Swietoslawski W (c1942) Coke formation process and physico-
chemical properties of coals. Polish Institute of Arts and Sciences in America, New York
7. Zhao RF, Ye SF, Xie YS, Chen YF (2003) Study on co-pyrolysis
of coking-coal, waste plastic and metallurgical dust (in Chinese).
Chin J Environ Sci 24:28–33
8. State Environmental Protection Administration of China (2003)
Cleaner Production Standard – Coking Industry. HJ/T 126-2003 (implemented June 1, 2003)
9. Nomura S, Kato K, Nakagawa T (2003) The effect of plastic
addition on coal caking properties during carbonization. Fuel 82: 1775–1782
10. Hao JM, Wang SX, Lu YQ (2001) Handbook of coal-burning
SO2pollution control technologies. Chemical Industry Press, Beijing
11. Li SF (1987) Research on improving the desulfurization ef? ciency
of iron oxide (in Chinese). Fuel Chem Process 18:35–39
12. IUPAC (1985) Reporting physisorption data for gas–solid system
with special reference to the determination of surface area and porosity. Pure Appl Chem 57:603–619
13. Huang ZT (1994) Industrial catalysis. Chemical Industry Press,
Beijing
14. Daniel T, Parekh BK (1999) An integrated approach to minimizing
gypsum and pyrite wastes by conversion to marketable products.
In: University coal research contractors review conference.
University of Kentucky, USA, June 1–2
15. Hu HQ, Zhou Q, Zhu SW, Mayer B, Steffen K, Chen GH (2004)
Product distribution and sulfur behavior in coal pyrolysis. Fuel Process Technol 85:849–861
16. Miura K, Mae K, Shimada M, Minami H (2001) Analysis of forma-
tion rates of sulfur-containing gases during the pyrolysis of various coals. Energy Fuels 15:629–636
17. Mikhlin Y (2000) Reactivity of pyrrhotite sulfates: an electrochem-
istry study. Phys Chem Chem Phys 2:5672–5677
18. Cleyle PJ, Caley WF, Stewart L, Whitemay SG (1984) Decomposi-
tion of pyrite and trapping of sulfur in a coal matrix during pyrol-ysis of coal. Fuel 63:1579–1582
19. Xu L, Yang JL, Li YM, Liu ZY (2004) Dynamic and simultaneous
analyses of gaseous sulfur and hydrocarbon compounds released during pyrolysis of coal. J Anal Appl Pyrolysis 71:591–600
20. Khan MR, Jenkins RJ (1986) In? uence of added calcium com-
pounds on swelling, plastic and pyrolysis behavior of coal devolatil-ized at elevated pressures. Fuel 65:1203–1208
21. Khan MR, Jenkins RJ (1986) Swelling and plastic properties of
coal devolatilized at elevated pressures of H2 and He: In? uence of potassium and silicon additive. Fuel 65:1291–1299
22. Oztas NA, Yurum Y (2000) Pyrolysis of Turkish Zonguldak bitu-
minous coal. Part I: Effect of mineral matter. Fuel 79:1221–1227 23. Skodras G, Sakellaropouls GP (2002) Mineral matter effects in
lignite gasi? cation. Fuel Process Technol 77/78:151–158
24. Jackson WR, Francis P, Larkins MM, Douglas R, White N (1979)
Hydrogenation of brown coal. I. The effects of additional quanti-ties of the inorganic constituents. Fuel 58:281–284
25. Chen HK, Li BQ, Zhang BJ (2000) Decomposition of pyrite and
the interaction of pyrite with coal organic matrix in pyrolysis and hydropyrolysis. Fuel 79:1627–1631
26. Semra K (2003) Desulfurzation of a Turkish lignite under various
gas atmospheres by pyrolysis: effect of mineral matter. Fuel 82: 1509–1516