ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Regular Article
Removal of Hydrogen Sulfide and Ammonia by Goethite-Rich Limonite in the Coexistence of Coke Oven Gas Components
Yuuki MochizukiAyumu OgawaNaoto Tsubouchi
Author information
JOURNALS OPEN ACCESS FULL-TEXT HTML

2017 Volume 57 Issue 3 Pages 435-442

Details
Abstract

Hydrogen sulfide (H2S) removal and catalytic ammonia (NH3) decomposition performance of limonite in the presence of coke oven gas (COG) components has been studied in a cylindrical quartz reactor at 300–850°C under a high space velocity of 51000 h−1 to develop a novel hot gas cleanup method. The H2S removal behavior in 50% H2/He depends on the temperature, with high performance observed at lower temperature. An investigation of the removal behavior of H2S in the presence of COG components (CH4, CO, CO2 and H2O) at 400°C reveals that CH4 does not affect the removal performance. On the other hand, the coexistence of CO drastically decreases the H2S removal performance. However, the addition of 5% H2O to 50% H2/30% CH4/5% CO/He dramatically improves the H2S removal performance, whereas the performance is low at 5% CO2 with 50% H2/30% CH4/5% CO/He. In addition, the H2S breakthrough curve strongly depends on the space velocity.

The limonite catalyst achieves almost complete decomposition of NH3 in He at 850°C until 240 min. When the decomposition run is performed in the presence of COG components, the coexistence of 30% CH4 deactivates limonite with significant formation of deposited carbon. On the other hand, the addition of 5% CO2, 5% H2O or 5% CO2/5% H2O to 50% H2/30% CH4/5% CO improves the catalytic activity without carbon deposition, and >99% conversion of NH3 to N2 is maintained until 240 min.

1. Introduction

Coke oven gas (COG) produced during coal carbonization is generally used as a fuel source for coke ovens and other combustion units in iron steel making process. COG has recently attracted much attention as a low-cost source of hydrogen and value-added products for the chemical industry.1) However, COG contains a variety of impurities, such as hydrogen sulfide (H2S) and ammonia (NH3), which are detrimental in some industrial applications. In conventional processes, the hot raw COG (>800°C) must be quenched to near room temperature using aqueous ammonia solution to remove tarry materials contained in the COG.1) On the other hand, the catalytic or noncatalytic reforming of COG is a promising technology that effectively utilizes the energy of the hot COG.2,3,4,5) Upon coal carbonization, some of the nitrogen and sulfur present in coal are retained in the solid phase, and the rest is released as volatile-N (N2, NH3, HCN and tar-N) and -S (H2S, COS, CS2 and tar-S) species.6) It is well known that NH3, HCN, tar-N, H2S, organic-S and tar-S not only serve as sources of NOx and SOx but also are catalyst poison materials used for COG reforming and/or gas tube corrosion.

Our research group has been working on vapor deposition of gaseous-tar, which is contained in COG, into the pores of limonite or cold-bonded pellets to develop carbon-containing iron materials (composite) with enhanced strength and reducibility for blast furnaces.7) A vapor deposition temperature of 350°C is optimum for obtaining completely filled mesopores in limonite or cold-bonded pellets using carbonaceous materials derived from tar via vapor infiltration. However, during vapor deposition treatment, N and S in the feed gases are likely to transfer from the gas-phase to the composite, which may cause high N and S contents in the resulting composite. When such a composite is used for blast furnaces, the N and S species in the composite may have adverse effects on the reduction rate, strength or exhaust gas purification equipment. N and S in COG mainly exist as NH3 and H2S, with contents of approximately 1 vol% and 3000 ppmv, respectively. Therefore, it is very important to develop a removal method for gaseous-N and -S species in COG.

Our research group has been investigating the catalytic decomposition of NH3 or model tar-N compounds (pyridine or pyrrole) using inexpensive iron catalysts.8,9,10,11,12,13,14,15,16,17,18) We have found that fine particles of metallic iron (α-Fe) formed from low-value iron ore (limonite) give high activity in the decomposition of NH3 and model tar-N compounds. Limonite can achieve almost complete decomposition of 2000 ppmv NH3 in inert or simulated fuel gas from coal gasification at 500–850°C.10) In addition, we have recently shown that the limonite-derived α-Fe can provide stable activity for the decomposition of 100 ppmv C5H5N to N2 in He, with fuel gas or COG components at 500–850°C.17,18) On the other hand, iron oxide is known to be effective for the removal of H2S in fuel gas at 300–700°C.19,20,21,22,23) Thus, it is possible that limonite-derived α-Fe will be a good adsorbent for H2S in COG. However, only a few studies have reported H2S removal in COG by iron-based adsorbents, such as iron oxide and/or iron-bearing sorbents combined with other metals.24,25,26,27) It is well-known that a typical COG is composed of 54–59% H2, 24–28% CH4, 4–7% CO, 3–5% CO2, 1–3% CO2 and 1–3% H2O with impurities, but the influence of the individual COG components on H2S removal has not been investigated in previous reports.24,25,26,27)

In this paper, therefore, we focus on investigating H2S removal and catalytic NH3 decomposition performance using an Australian limonite in the presence of COG components to develop a novel gas cleaning method for removing H2S and NH3.

2. Experimental

2.1. Sample

An Australian natural limonite ore composed of about 70% goethite (α-FeOOH) was employed in the present study. The metal composition of the limonite was Fe, 44; Si, 9.4; Al, 7.2; Mg, 0.15; and Ca, 0.07 mass%-dry.17) The as-received limonite was sieved to select the 250–500 μm size fractions, and the Brunauer-Emmett-Teller (BET) specific surface area was measured as 40 m2/g.

2.2. Hydrogen Sulfide Removal and Ammonia Decomposition

The experiments to evaluate H2S removal and catalytic NH3 decomposition performance were carried out in a cylindrical fixed-bed quartz reactor (8 mm i.d.) under ambient pressure. The details of the experimental apparatus have been described elsewhere.17) The temperature was controlled with a K-type thermocouple attached to the exterior surface of the reactor. Approximately 0.25 g of limonite was first charged into the reactor with quartz wool, and a flow of high-purity He (>99.99995%) was then passed through the reactor until the concentration of N2 in the experimental system decreased to less than 20 ppmv. After taking precautions against leakage, the reactor was heated electrically to 500°C. At this temperature, the He was replaced with high-purity H2 (>99.9999%), and limonite was reduced by H2 for 2 h. After the reduction pretreatment, the atmosphere was restored to He, and the reactor was held at a temperature of 300–850°C. The H2S removal or NH3 decomposition experiments began by passing 3000 ppmv H2S or 1 vol% NH3 diluted with He and simulated COG (50% H2/He, 50% H2/30% CH4/He, 50% H2/30% CH4/5% CO/He, 50% H2/30% CH4/5% CO/5% CO2/He, 50% H2/30% CH4/5% CO/5% H2O/He or 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He) over the catalyst bed. The space velocity (SV) was maintained at 51000 h−1 throughout each run, unless otherwise noted.

2.3. Gas Analysis

In the H2S removal experiments, the H2S concentration in the exit gas was measured using a standard detector tube (Gastec) at arbitrary times to obtain the breakthrough curves. The breakthrough and saturation points were defined as Ci/C0 = 0.1 and 0.9, respectively. Here, C0 and Ci are the initial concentration and the H2S concentration in the exit gas at arbitrary time, respectively. The extent of sulfidation was calculated based on the iron content in limonite and the results of the breakthrough curves.

In the NH3 decomposition runs, the amount of N2 produced during the decomposition of NH3 was measured at 3 min intervals with a high-speed micro gas chromatograph (GC; Agilent) equipped with a thermal conductivity detector. The conversion of NH3 to N2 was calculated based on the amount of NH3 inlet and the N2 concentration at the reactor exit.

2.4. Characterization

Powder X-ray diffraction (XRD; Shimadzu) measurements of the samples (as-received, after H2 reduction and after H2S removal or NH3 decomposition) were performed with Mn-filtered Fe-Kα radiation. To avoid rapid oxidation of the α-Fe particles upon exposure to laboratory air, the limonite after H2 reduction, H2S removal or NH3 decomposition was passivated using 1% O2/He at room temperature and then recovered from the reactor.10)

Temperature-programmed oxidation (TPO) was carried to quantitate the formation amounts of Fe3C and C deposition during H2S removal or NH3 decomposition runs. In a TPO run, the limonite in the reactor after H2S removal or NH3 decomposition was first quenched to room temperature in a stream of high purity He and then heated at 10°C/min to 950°C in 10% O2/He. The concentrations of CO and CO2 evolved in this process were monitored using the micro GC. Some samples after TPO were also subjected to XRD measurements.

3. Results and Discussion

3.1. Effect of COG Components on H2S Removal

This section focuses on the effect of coexisting of H2, CH4, CO, CO2 and H2O on H2S removal by reduced-limonite. The changes in the breakthrough curves of reduced-limonite with temperature in 50% H2/He were first examined, and the results are summarized in Table 1. In addition, the relationship between the extent of sulfidation and H2S concentration in the exit gas based on the breakthrough curves are presented in Fig. 1. In the H2S breakthrough curve at 300°C, a removal extent of 100% was measured until about 30 min, and the H2S concentration in the exit gas increased after the breakthrough point at 40 min with increasing time on stream. The saturation point was observed at 65 min. Although similar tendencies were measured at 400, 500 and 600°C, the breakthrough points decreased with increasing temperature and were observed at 35, 30 and 24 min, respectively. On the other hand, at 700°C, the removal extent of H2S decreased to 85% at 1 min after the experimental run start, and this value was maintained until 15 min. Subsequently, the H2S removal extent decreased slightly to 75% until to 40 min, and then decreased with increasing time on stream, reaching saturation at 80 min. Although a similar tendency was observed at 800°C, the removal extent of H2S at 2 min after the experimental run start was 67%, which was smaller than that of 700°C. The above-mentioned breakthrough points corresponded to sulfidation extents of 0.75, 0.65, 0.55 and 0.45 at 300, 400, 500 and 600°C, respectively, as seen in Fig. 1, and these values decreased with increasing temperature. At 700 and 800°C (Fig. 1(b)), the reduced-limonite attained breakthrough points immediately after the runs started. These results suggest that H2S removal by reduced-limonite in a reduction atmosphere (50% H2/He) depends on the temperature and has higher efficiency at lower temperatures.

Table 1. Summaries of breakthrough and saturation point at breakthrough curves , the extent of sulfidation and crystalline forms of Fe before and after H2S removal in 50%H2/He.
Temperature (°C)Breakthrough point (min)Saturation point (min)Extent of Sulfidation (H2S/Fe)aIron species after H2S removalb
Reduced limoniteα-Fe
30040651.04Fe1−xS(m)
40035600.95Fe1−xS(m)
50030600.94Fe1−xS(m), FeS(w)
60024580.92Fe1−xS(m), FeS(m)
700<1800.93Fe1−xS(w), FeS(s)
800<1900.90FeS(s)

aAfter reaction. bXRD intensities designated by w (weak), m (medium), and s (strong).

Fig. 1.

Change in sulfidation behavior of limonite-reduced with temperature in 50% H2/He.

XRD measurements of limonite after the H2S removal runs were carried out to clarify the temperature dependencies of H2S removal in 50% H2/He. Table 1 also summarizes the results of the XRD measurements. The main peak attributed to Fe1−xS at approximately 55.8° was observed for the samples obtained after runs in 50% H2/He at 300–400°C. On the other hand, the intensity of this peak gradually decreased with temperature (300–600°C), whereas the main peak at 600–800°C was observed at approximately 55.15°. This peak is attributed to FeS, and this observation indicates that the form of adsorbed H2S on α-Fe changes with temperature. The H2S/Fe ratios at after reaction in 50% H2/He ranged from 0.9 to 1.05 (Table 1), and these values trended to decrease with increasing temperature. This result shows that almost all Fe in limonite was sulfidized by H2S, and the difference in removal performance with increasing temperature is due to different adsorption forms (Fe1−xS and FeS). The sulfidation extent of more than 1.0 at 300°C in Fig. 1 and Table 1 is due to analytical error. As mentioned in the Introduction, the optimum vapor deposition conditions for preparing the composites, in which the carbonaceous material fills the mesopores in de-hydrated limonite via vapor deposition of gaseous-tar, is 350°C. Therefore, 400°C was used for H2S removal by reduced-limonite in subsequently investigations.

Figure 2 shows the influence of coexisting gases on the relationships between the extent of sulfidation and H2S concentration in the exit gas calculated from the H2S breakthrough curves measured at 400°C in different feed gas compositions. The breakthrough and saturation points of the H2S breakthrough curves and the extent of sulfidation at these points are summarized in Table 2. The breakthrough and saturation points in 50% H2/He were observed at sulfidation extent of 0.65 and 0.93, respectively (Fig. 2(a)). When 30% CH4 was added to 50% H2/He, the breakthrough and saturation points were observed at 32 and 65 min, corresponding to sulfidation extent of 0.68 and 1.0, respectively, as well as those in 50% H2/He, and the presence of CH4 did not significantly affect H2S removal. On the other hand, the breakthrough and saturation times in 50% H2/30% CH4/5% CO/He drastically decreased from 35 and 65 min in 50% H2/30% CH4/He to 10 and 38 min, respectively. The extent of sulfidation at the breakthrough and saturation points also decreased to 0.18 and 0.4, respectively, and the two curves ((b) and (c) in Fig. 2(a)) were quite different. The H2S removal performance drastically decreased in the presence of CO. Figure 2(b) also shows the results with 5% CO2, 5% H2O or both coexisted to 50% H2/30% CH4/5% CO/He. With 5% CO2 added, the breakthrough point was observed at 10 min, and sulfidation curves was similar to that of 50% H2/30% CH4/5% CO/He. These results indicate that the presence of CO2 did not affect H2S removal in 50% H2/30% CH4/5% CO/He. On the other hand, the removal performance was improved with 5% H2O addition, as seen in Fig. 2(b), and breakthrough and saturation points increased to 32 and 67 min from 10 and 38 min in presence 5% CO, respectively. As seen in Fig. 2(b), similar behavior was observed in 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He. The order of sulfidation extent in Fig. 2 and Table 2 is 50% H2/30% CH4/5% CO/He < 50% H2/30% CH4/5% CO/5% CO2/He < 50% H2/30% CH4/5% CO/5% H2O/He≒50% H2/30% CH4/5% CO/5% CO2/5% H2O/He≒50% H2/30% CH4/He. From these results, it was found that H2O addition can improve the H2S removal performance of limonite, whereas CO addition decreases the removal ability of reduced-limonite under these conditions.

Fig. 2.

Influence of coexisting gases on the relationships between the extent of sulfidation and H2S concentraction in the exit gas at 400°C. (a) In 50% H2/He, (b) In 50% H2/30% CH4/He, (c) In 50% H2/30% CH4/5% CO/He, (d) 50% H2/30% CH4/5% CO/5% CO2/He, (e) 50% H2/30% CH4/5% CO/5% H2O/He, (f) 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He.

Table 2. Summaries of breakthrough and saturation point at breakthrough curves and sulfidation curves with H2S/Fe ratio at 400°C.
AtmosphereBreakthrough curveExtent of sulfidation (H2S/Fe)
Breakthrough point (min)Saturation point (min)At breakthrough pointAt saturation pointAt after reaction
50%H2/He35650.650.930.95
50%H2/30%CH4/He32650.681.001.00
50%H2/30%CH4/5%CO/He10380.180.410.42
50%H2/30%CH4/5%CO/5%CO2/He10420.180.510.53
50%H2/30%CH4/5%CO/5%H2O/He32670.630.981.00
50%H2/30%CH4/5%CO/5%CO2/5%H2O/He28600.630.940.95

3.2. Possible Mechanisms of Influence of COG Components on H2S Removal Performance

XRD measurements of the samples recovered after reaction were carried out to investigate the difference in H2S breakthrough curves with coexisting gases at 400°C. The results are summarized in Table 3. In these samples, iron always takes the form of Fe1−xS, regardless of the gas composition used, and thus, it was difficult to clarify the influence of coexisting gases on the H2S removal performance from the XRD results alone. Therefore, TPO of the samples were investigated in 10% O2/He up to 950°C, as shown in Fig. 3. No CO and CO2 evolution was measured during TPO of the sample recovered after the 50% H2/30% CH4/He run. However, during TPO of the sample after the 50% H2/30% CH4/5% CO/He run, CO and CO2 were evolved above approximately 300°C (Fig. 3(a)), with main and shoulder peaks for the CO and CO2 formation rates at approximately 350 and 450°C, respectively. As similar profile was observed for sample recovered after the 50% H2/30% CH4/5% CO/5% CO2/He run (Fig. 3(b)). On the other hand, the peaks of the CO and CO2 formation rates from the samples recovered after the 50% H2/30% CH4/5% CO/5% H2O/He and 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He runs were considerably smaller than those for coexisting CH4/CO and CH4/CO/CO2 (Fig. 3(c)). Table 3 summaries of the amounts of CO and CO2 evolved during the TPO runs. The Fe1−xS observed in all samples after H2S treatment was completely transformed into Fe2O3 after TPO. The order of the amounts of CO and CO2 evolved during TPO is 50% H2/30% CH4/He < 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He < 50% H2/30% CH4/5% CO/5% H2O/He << 50% H2/30% CH4/5% CO/5% CO2/He ≤ 50% H2/30% CH4/5% CO/He. This order was almost the inverse of that for the amount of H2S removal.

Table 3. Crystalline form of Fe after H2S removal in different atmosphere and the amount of CO and CO2 evolved during temperature programmed oxidation of samples after sulfidation.
AtmosphereIron species after H2S removalaEvolution amount (μmol/g)
COCO2
50%H2/30%CH4/HeFe1−xS(m)010
50%H2/30%CH4/5%CO/HeFe1−xS(m)2901400
50%H2/30%CH4/5%CO/5%CO2/HeFe1−xS(m)2101100
50%H2/30%CH4/5%CO/5%H2O/HeFe1−xS(m)40180
50%H2/30%CH4/5%CO/5%CO2/5%H2O/HeFe1−xS(m)3060

aXRD intensities designated by m (medium)

Fig. 3.

Formation rate and cumulative amount of CO or CO2 during temperature-programmed oxidation of samples after sulfidation: (a) In 50% H2/30% CH4/5% CO/He, (b) In 50% H2/30% CH4/5% CO/5% CO2/He, (c) In 50% H2/30% CH4/5% CO/5% H2O/He.

The reaction between metallic Fe and H2S is well known to occur according to Eqs. (1) and (2) to form pyrrhotite (Fe1−xS) and iron sulfide (FeS), and these reactions are supported by thermodynamic calculations because the corresponding standard Gibbs free energies (⊿G) at 400°C are −13 and −14 kcal/mol.   

Fe+ H 2 SF e 1-x S+xFe+ H 2 ( x=0.123 was used in this calculation ) (1)
  
Fe+ H 2 SFeS+ H 2 (2)
In addition, C deposition and Fe3C formation from CH4 occur according to Eqs. (3) and (4):   
C H 4 C+2 H 2 (3)
  
3Fe+C H 4 F e 3 C+2 H 2 (4)
The ⊿G400°C values for Eqs. (3) and (4) are 3.6 and 6.0 kcal/mol, respectively, and thus, C deposition or Fe3C formation via these is not thermodynamically favorable under these conditions. Therefore, the H2S breakthrough curve is similar to that in 50% H2/He. In other words, CH4 does not affect H2S removal by reduced-limonite under the present conditions. This is supported by the TPO results, in which no significant evolution of CO and CO2 was observed. In the case of 5% CO addition to 50% H2/30% CH4/He, the H2S removal performance decreased, as seen in Fig. 2(a). C deposition or Fe3C formation from CO may occur according to Eqs. (5) and (6).   
2COC+C O 2 (5)
  
3Fe+2COF e 3 C+C O 2 (6)
  
CO+ H 2 C+ H 2 O (7)
  
3Fe+CF e 3 C (8)
  
F e 3 C+3 H 2 S3FeS+C+3 H 2 (9)
As the ⊿G400°C values for Eqs. (5) and (6) are −13 and −10 kcal/mol, respectively, these reactions are likely to be favorable under the conditions applied. The driving force for Eq. (7) (⊿G400°C = −9.4 kcal/mol) is smaller than those for Eqs. (5) and (6). The ⊿G400°C value for Eq. (8) (2.4 kcal/mol) indicates that this reaction is also not thermodynamically favored. Although Fe3C was not detected by XRD in the sample recovered after the run in 50% H2/30% CH4/5% CO/He, significant CO and CO2 formation were observed during TPO of this sample, as seen in Fig. 3(a). Therefore, Eq. (9), which has a ⊿G400°C values of −45 kcal/mol, should also be considered. The decreased H2S removal performance in the presence of CO may be caused by deposition of C derived from disproportionation reactions of CO on reduced-limonite active sites, and C deposited from Eqs. (5), (6), (7) and (9) will be a main factor for the decreased H2S removal performance. The H2S breakthrough curve in 50% H2/30% CH4/5% CO/5% CO2/He was similar to that in 50% H2/30% CH4/5% CO/He. The Boudouard reaction (⊿G400°C = 13 kcal/mol), expressed as Eq. (10), is not thermodynamically favorable. In addition, if Fe3C forms under the present conditions, Eq. (11) also does not occur (⊿G400°C = 10 kcal/mol). Therefore, coexisting CO2 may not affect the removal performance of H2S. This speculation supported by the similar amounts of CO and CO2 evolved during TPO of the samples recovered from the 50% H2/30% CH4/5% CO/He and 50% H2/30% CH4/5% CO/5% CO2/He runs (Table 3).   
C+C O 2 2CO (10)
  
F e 3 C+C O 2 3Fe+2CO (11)

In contrast, in the cases of H2O coexisting in the feed gases (50% H2/30% CH4/5% CO/5% H2O/He or 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He), the H2S breakthrough curves showed the similar tendencies to that of 50% H2/30% CH4/He, and the performance of the sample with 5% CO was improved by H2O addition. In addition, the amounts of CO and CO2 evolved during TPO were significantly smaller than those with coexisting CO, as seen in Table 3. These results shows that C deposition is inhibited by H2O. Although the ⊿G400°C values of two water-gas reactions (6.0–9.4 kcal/mol) expressed as Eqs. (12) and (13) are not thermodynamically favorable, the water-gas shift reaction of Eq. (14) (⊿G400°C = −3.3 kcal/mol) may occur. Although Fe3C was not observed by XRD, the H2S removal performance may be proceeded by the reaction shown in Eq. (9), which has a ⊿G400°C values (−45 kcal/mol) that is smaller than those of Eqs. (15) and (16) (3.6–7.0 kcal/mol), even if Fe3C is formed under these present conditions. Thus it may be possible that H2O addition prevents C deposition according to Eqs. (5), (6), (7) due to the water-gas shift reaction (Eq. (14)), and as a result, the H2S removal performance is improved.   

C+ H 2 OCO+ H 2 (12)
  
C+2 H 2 OC O 2 +2 H 2 (13)
  
CO+ H 2 OC O 2 + H 2 (14)
  
F e 3 C+2 H 2 O3Fe+C O 2 +2 H 2 (15)
  
F e 3 C+ H 2 O3Fe+CO+ H 2 (16)

Figure 4 illustrated the effect of space velocity on the H2S breakthrough curves in 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He. Here, the space velocity was varied from 51000 to 5600 h−1. When the SV decreased to 5600 h−1, H2S was not detected in the exit gas until 120 min. The breakthrough increased more than 6-fold from 27 min at an SV of 51000 h−1 to 165 min at an SV of 5600 h−1. Therefore, the breakthrough curves strongly depend on SV.

Fig. 4.

Effect of space velocity on H2S breakthrough curves at 400°C in 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He.

3.3. Decomposition of Ammonia in the Presence of COG Components

This section focuses on the effect of CH4, CO, CO2 and H2O on limonite-catalyzed NH3 decomposition. These runs were carried out at 850°C to simulate the utilization of COG sensible heat for NH3 decomposition. Limonite-catalyzed N-containing species decomposition at other temperatures has been already reported at our previous reports.10,11,12,13,14,15,16,17,18) Figure 5 presents the effect of the presence of CH4, CO, CO2 and H2O on NH3 decomposition. The decomposition of NH3 was maintained until 240 min in He, whereas N2 conversion in 50% H2/30% CH4/He and 50% H2/30% CH4/5% CO/He dramatically decreased after 90 min. A significant amount of C deposition was found on the catalyst surface in the presence of CH4 and/or CO. As the formation of carbonaceous materials occurred under the present conditions, the catalyst deactivation observed in 50% H2/30% CH4 and 50% H2/30% CH4/5% CO (Fig. 5) may be ascribed to C deposition according to the Eqs. (3) and (4), with ⊿G850°C values of −7.8 and −8.0 kcal/mol, respectively. These ⊿G values are smaller than those of Eqs. (5), (6) and (7) (⊿G850°C = 6.3, 6.1 and 6.0 kcal/mol, respectively). On the other hand, the reaction show in Eq. (8) may also occur (⊿G850°C = −0.2 kcal/mol).

Fig. 5.

Effect of the presence of CH4, CO, CO2 and H2O on limonite-catalyzed NH3 decomposition at 850°C. (a) In He, (b) In 50% H2/30% CH4/He, (c) In 50% H2/30% CH4/5% CO/He, (d) In 50% H2/30% CH4/5% CO/5% CO2/He, (e) In 50% H2/30% CH4/5% CO/5% H2O/He, (f) In 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He.

Figure 5 also illustrates the effects of 5% CO2, 5% H2O or 5% H2O/5% CO2 coexisting to 50% H2/30% CH4/5% CO on NH3 decomposition. In 50% H2/30% CH4/5% CO/5% CO2/He, 50% H2/30% CH4/5% CO/5% H2O/He and 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He, the NH3 decompositions performance was maintained until 240 min, and >99% conversion of NH3 to N2 was observed. Moreover, no significant C deposition was found on the catalysts recovered after the experimental runs. These results show that the coexistence of CO2, H2O or CO2/H2O with 50% H2/30% CH4/5% CO can improve NH3 decomposition. As mentioned above, the deactivation of limonite for NH3 decomposition mainly occurs due to deposition of C according to Eqs. (3) and (4). From the thermodynamic calculation results, preventing C deposition in the presence of CO2 may occur by Eqs. (10) and (11) (⊿G850°C = −6.3 and −6.1 kcal/mol, respectively). On the other hand, in the case of H2O addition to 50% H2/30% CH4/5% CO, ⊿G850°C values for Eqs. (12) and (13) are almost −6.0 and −5.8 kcal/mol, respectively, and these reactions are more favorable than Eq. (14), which has a ⊿G850°C value of 0.2 kcal/mol. In addition, Eqs. (15) and (16) (⊿G850°C = −5.6 and −5.8 kcal/mol, respectively) are also thermodynamically favorable. Therefore, preventing C deposition and retransformation of Fe3C into metallic Fe, as shown in Eqs. (10), (12) and (13) and Eqs. (11), (15) and (16), respectively, may be the driving force for the high NH3 decomposition performance with 5% CO2, 5% H2O or 5% CO2/5% H2O coexisted to 50% H2/30% CH4/5% CO. Similar results have been reported for NH3 decomposition by limonite in a simulated air-blown coal gasification composition (10% CO2 or 3% H2O coexisted in 20% CO/10% H2).12)

3.4. Possible Mechanisms of Influence of COG Components on NH3 Decomposition

Figure 6 shows the XRD results for the catalysts after NH3 decomposition in each gas composition, and these results are summarized in Table 4. When limonite was reduced in H2, only α-Fe was observed (Fig. 6(a)). This signal was also observed for the catalysts recovered after the He runs, and transformation of the iron form was not observed. According to previous our studies, the catalytic decomposition of NH3 to N2 with reduced-limonite occurs due to the cyclic reaction via iron nitride (FexN) intermediates, according to Eqs. (17) and (18).8,9,10,11,12,13,14,16)   

xFe+N H 3 F e x N+1.5 H 2 (17)
  
F e x NxFe+0.5 N 2 (18)
On the other hand, crystallized carbon (CC) was detected at around 32° in the samples recovered after reaction in 50% H2/30% CH4/He, along with peaks corresponding to α-Fe and Fe3C. The peaks attributable to crystallized C and Fe3C were also observed for the in catalyst after the decomposition run in 50% H2/30% CH4/5% CO/He. The formation of crystallized C and Fe3C may indicate that deactivation of the catalyst occurs by Eqs. (4) and (8) (⊿G850°C = −8.0 and −0.2 kcal/mol, respectively), in the presence of 30% CH4. On the other hand, no measurable Fe3C signals were observed in the sample obtained after reaction in 50% H2/30% CH4/5% CO/5% CO2/He, and the diffraction peak of crystallized C was significantly smaller than those in 50% H2/30% CH4/He and 50% H2/30% CH4/5% CO/He. In addition, the intensity of the α-Fe signal was almost that same as that of the catalyst before the reaction. Similar results were observed for the catalysts after NH3 decomposition runs in 50% H2/30% CH4/5% CO/5% H2O/He and 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He. From these results, the formation of Fe3C and crystallized C may inhibit the NH3 decomposition performance by limonite because NH3 conversion drastically decreased in the 50% H2/30% CH4/He and 50% H2/30% CH4/5% CO/He runs, in which Fe3C and crystallized C were detected in the spent catalysts. According previous report about investigation of NH3 decomposition by limonite and commercial Fe3C in syngas, CO2 or H2O/syngas,12) it has been suggested that the disproportionation of CO in syngas and the subsequent formation of Fe3C are responsible for the catalyst deactivation. In addition, NH3 decomposition performance of commercial Fe3C is very small.12) Therefore, the influence of Fe3C formation on limonite-catalyzed NH3 decomposition may be greater than that of crystallized C in this study.
Fig. 6.

XRD profiles of catalysts after NH3 decomposition at 850°C in various atmosphere. (a) Reduced-limonite, (b) In He, (c) In 50% H2/30% CH4/He, (d) In 50% H2/30% CH4/5% CO/He, (e) In 50% H2/30% CH4/5% CO2/He, (f) In 50% H2/30% CH4/5% CO/5% H2O/He, (g) In 50% H2/30% CH4/5% CO/5% CO2/5% H2O/He.

Table 4. Summarize of TPO results and Fe forms after NH3 decomposition run at different atmosphere.
NH3 decomposition atmosphereIron species after NH3 decompositionaEvolution amount (μmol/g)
COCO2
Heα-Fe (s)n.a.bn.a.b
50%H2/30%CH4/Heα-Fe(m), Fe3C(vw), Crystallized carbon(m)1070040300
50%H2/30%CH4/5%CO/Heα-Fe(w), Fe3C(w), Crystallized carbon(m)1530028700
50%H2/30%CH4/5%CO/5%CO2/Heα-Fe (s)0350
50%H2/30%CH4/5%CO/5%H2O/Heα-Fe (s)040
50%H2/30%CH4/5%CO/5%CO2/5%H2O/Heα-Fe (s)1040

aXRD intensities designated by w (weak), m (medium), and s (strong), bNot analyzed.

Figure 7 presents the CO and CO2 evolution profiles during TPO, and Table 4 summarizes the amounts of CO and CO2 evolved during the TPO runs. CO and CO2 formation started above 300°C in the sample recovered after the 50% H2/30% CH4/He run, with main and shoulder peaks observed in the range from 400 to 650°C (Fig. 7(a)). Similar profiles were observed for the sample obtained after the 5% CO coexisted to 50% H2/30% CH4/He run (Fig. 7(b)). The evolution rates of CO and CO2 and these cumulative amounts were quite different with those of H2S. This distinction may cause by the difference of deposited-C forms in H2S removal and NH3 decomposition runs. These forms are distinguished to amorphous-C derived from disproportionation reaction of CO in H2S removal run and crystallized-C derived from CH4 decomposition in NH3 decomposition run, respectively. On the other hand, the amount of CO and CO2 evolved during TPO of the sample recovered after in the 5% H2O and 5% CO2 coexisted to 50% H2/30% CH4/5% CO/He run were considerably smaller than those for the run in the presence of 30% CH4 and 5% CO. These results show that C deposition is not caused by deactivation of the catalyst for NH3 decomposition in the presence of 5% CO2 and 5% H2O.

Fig. 7.

Formation rate and cumulative amount of CO or CO2 during temperature-programmed oxidation of samples after NH3 decomposition: (a) In 50% H2/30% CH4/He, (b) In 50% H2/30% CH4/5% CO/He.

4. Conclusions

The removal of 3000 ppmv H2S and catalytic decomposition of 1 vol% NH3 in the presence of COG components with limonite rich in α-FeOOH has been studied in a cylindrical quartz reactor under a pressure of 0.1 MPa, temperature of 300–850°C and an SV of 51000 h−1. The principal conclusions are summarized as follows:

(1) Although the reduced-limonite shows high H2S removal performance in 50% H2/He, the performance decreases with increasing temperature. The XRD analyses showed that α-FeOOH-derived active α-Fe is transformed into Fe1−XS and FeS.

(2) In the presence of 5% CO, the H2S removal performance at 400°C decreases considerably due to carbon deposition. In addition, 5% CO2 addition to 50% H2/30% CH4/5% CO/He did not affect the H2S removal ability of reduced-limonite.

(3) The addition of 5% H2O to 50% H2/30% CH4/5% CO/He suppresses carbon formation and consequently improves the H2S removal performance of limonite significantly.

(4) Although the limonite catalyst achieves almost complete decomposition of NH3 in inert He at 850°C, presence of 30% CH4 or 30% CH4/5% CO in 50%H2/He deactivates limonite and causes appreciable formation of deposited carbon.

(5) When 5% CO2, 5% H2O or both is added to 50% H2/30% CH4/5% CO/He, the catalytic activity dramatically improves without carbon deposition. The NH3 to N2 conversion of >99% is maintained until 240 min.

(6) The XRD analyses reveal the transformation of α-FeOOH-derived active α-Fe into Fe3C in the presence of 30% CH4 and 30% CH4/5% CO. On the other hand, when 5% H2O, 5% CO2 or 5% CO2/5% H2O is added to 50% H2/30% CH4/5% CO/He, α-Fe is the only Fe form.

Acknowledgments

The present study was supported in part by the Iron and Steel Institute of Japan (ISIJ) Research Promotion Grant and the Steel Foundation for Environmental Protection Technology (SEPT). The authors acknowledge the supply of limonite from Kobe Steel Ltd and Mitsubishi Chemical Corp. in Japan.

References
 
© 2017 by The Iron and Steel Institute of Japan
feedback
Top