Catal Surv Asia (2007) 11:134–144 DOI 10.1007/s10563-007-9022-4
Catalytic Systems for the H2S Wet Oxidation at room Temperature Eun-Ku Lee Æ Kwang-Deog Jung Æ Yong-Gun Shul
Received: 15 July 2007 / Accepted: 15 August 2007 / Published online: 12 September 2007 Springer Science+Business Media, LLC 2007
Abstract The catalytic wet oxidation process is the most attractive process for small-scale hydrogen sulfide (H2S) removal from natural gas. The catalytic wet oxidation process is anticipated to be cost effective and simple so that it can be used for treating sour gases containing small amounts of H2S and can be easily operated even in isolated sites. The development of effective catalyst is the key technology in the wet catalytic oxidation of H2S. The scale of operation for the process has to be flexible so its use will not be limited by the flow rates of the gas to be treated. The heterogeneous catalytic wet oxidation of H2S has been attempted on activated carbons, but the H2S removal capacity still shows the low removal efficiency. The catalytic wet oxidation of H2S was studied over Fe/MgO for an effective removal of H2S. In order to develop a sulfur removal technology, one has to know what surface species of catalyst are the most active. This article discusses the following systematic studies: (i) the catalytic preparation to disperse Fe metal well on MgO support for enhancing H2S removal capacity, (ii) the effect of the catalytic morphology
E.-K. Lee Central Research Institute, Hyosung Corporation, Hogye-Dong 183, Dongan-Ku, Anyang-Si, Kyonggi-Do 431-080, Republic of Korea K.-D. Jung (&) Department of Environmental and Process Technology Division, Korea Institute of Science and Technology, Hawolgok-Dong 39-1, Seongbukgu, Seoul 139-791, Republic of Korea e-mail:
[email protected] Y.-G. Shul Department of Chemical Engineering, Yonsei University, 134 Shinchon-Dong, Seodaemoon-Gu, Seoul 120-749, Republic of Korea
123
on the activity of Fe/MgO for the H2S wet oxidation, (iii) the influence of precursor and support on the activity of Fe/ MgO for catalytic wet oxidation of H2S to sulfur. Keywords Catalytic wet oxidation Fe/MgO Catalytic morphology Surface species Catalytic deactivation
1 Introduction Hydrogen sulfide (H2S) is present in natural gas, refinery streams, and coal gas, naturally or as a byproduct of processing operations. It has to be removed from these sour gases because H2S is toxic and corrosive. A number of processes have been invented to remove H2S from sour gases [1–6]. An elaborate introduction and information on the sour gases and H2S treating processes can be found in many handbooks and literatures [7, 8]. Here the disadvantages of the current processes are analyzed according to a classification of the primary mechanism, e.g., absorption, adsorption, and conversion. Obviously, burning H2S leads to SO2 emission and air pollution. As environmental protection policies become tighter, the absorption and burning, suitable for sweetening the sour gas that contains small amounts of H2S or is trapped from isolated sites, is being banned. Adsorption occurs when the H2S-containing gas stream is passed through a fixed bed of metal oxides such as iron oxide or zinc oxide, where H2S reacts to form either the metal sulfides or sulfur. The slow reaction rate in this process restricts its use to treat gas streams of limited flow rate and low concentration of H2S. The disposal of the used adsorbent causes secondary environmental pollution. Adsorption using molecular sieves may enable the absorbent to be regenerated; however, the H2S-enriched gas resulting from the regeneration needs to be treated.
Catalytic Systems for the H2S Wet Oxidation
135
H2 S þ 1=2O2 ¼ S þ H2 O
ð1Þ
4
H2S removal capacity, gH2S/gcat
The conversion processes, among which the modified Claus process is the mostly widely used, convert H2S into elemental sulfur, saleable product. The burning unit of the Claus process, where 1/3 H2S is converted to SO2 for the Claus reaction, consumes a large amount of energy. The tail gas treatment technology, following the Claus process, is required to produce sweet gas, because the conversion in the Claus process is thermodynamically limited. Such conversion processes are economically justifiable only when sulfur production is large. The technology to be developed in this study is based on the following reaction.
3
2
1
0 0
5
10
15
20
25
30
35
Fe content, wt %
This reaction will be carried out in one or more vessels such that H2S in the gas stream will be converted into sulfur; and the latter remains trapped in the liquid. Compared to the modified Claus process, this technology could be more economical since the operation of the units will be simpler. The flexibility of the process allows it to be applicable to isolated gas reservoirs and to the recovery of separated flare gases. Moreover, this process will allow the Claus reaction to be utilized directly, eliminating the H2S absorption-regeneration units and the tail-gas problem. It is desirable to find a highly effective catalyst for the wet catalytic oxidation of H2S to elemental sulfur at room temperature. Furthermore, it will be worthwhile to suggest a redox mechanism and reason of deactivation during the reaction. The present article discusses the catalyst development based on our recent effects.
2 Selective Oxidation of H2S to Elemental Sulfur with Fe/MgO Catalyst A new type catalyst based on Fe/MgO was developed to improve H2S in the wet catalytic oxidation of H2S to element sulfur at room temperature [9]. The Fe/MgO catalytic system, prepared by the impregnation method, showed the high removal capacity of H2S with 1.0 gsulfur/gcat in air and 2.6 gsulfur/gcat in O2. The isolated Fe ion was proposed to be active site for the reaction [10]. In this study, it was confirmed that part of MgO was dissolved in acidic iron nitrate solution and then, the dissolved basic Mg ions were co-precipitated with Fe ions on the undissolved MgO spontaneously. The characterization of the resulting catalysts showed that Fe ions were highly dispersed on MgO support for enhancing H2S removal capacity. Figure 1 shows the H2S removal capacities of Fe/MgO catalysts with different Fe loadings. In experiments, feed gases (H2S: 5 ml/min, O2: 100 ml/min) are introduced into the stirred slurry reactor with 1.5 L of distilled water and
Fig. 1 H2S removal capacities of Fe/MgO catalysts with different Fe loadings in H2S wet catalytic oxidation. H2S flow rate: 5 ml/min, O2 flow rate: 100 ml/min
3 g of catalyst. The H2S removal capacity was obtained by calculating the total amount of H2S removed up to 50% of the H2S removal efficiency. No sulfur oxides are detected in the exit gas streams, indicating that total oxidation of H2S–SO2 can be prevented in the wet catalytic oxidation. The removal capacity in H2S oxidation increased up to 15 wt% Fe, and then decreased with further increase in Fe loading. The removal capacity of H2S for 15 wt% Fe/MgO was 3.74 gsulfur/gcat, while that for 1 wt% Fe/MgO was 0.71 gsulfur/gcat. It is interesting to note that the H2S removal capacity was maximized at 15 wt% Fe/MgO on the catalysts prepared by the impregnation method of salts solution, while it was at 6 wt% Fe/MgO on the catalysts prepared by the incipient wet impregnation method [9]. During the preparation, the acidic Fe solution dissolves basic MgO support partly to increase the pH of the solution. The co-precipitation of the dissolved Fe salt and the dissolved MgO occurs at the condition above the pH of 6.0. Therefore, Fe can be dispersed on MgO by the impregnation of the support suspension with salts solution better than by the incipient impregnation. In order to investigate changes in surface areas caused by different Fe loadings, the BET surface areas of Fe/MgO catalysts were measured. Table 1 shows the physical properties of Fe/MgO samples. As the loading of Fe increases from 1 wt% to 15 wt%, the BET surface area increases from 23.2 m2/g to 86.8 m2/g, and then decreases to 44.9 m2/g with a further increase in Fe loading. The observed results in surface area with an increase in Fe loading is similar to that reported previously for the Fe/ MgO catalysts prepared by an impregnation method [11]. With increasing Fe loading, the surface area of the catalyst is increased, indicating FeO–MgO mixed oxide phases are developed on the surface. Accordingly, plugging of the
123
136 Table 1 Physical characterization of Fe/MgO Samples
E.-K. Lee et al.
Sample
Fe content (wt%)
SBET (m2/g)
Pore volume (cm3/g)
˚) Average pore diameter (A
1Fe/MgO
1
23.2
0.16
225
4Fe/MgO
4
27.3
0.09
142
6Fe/MgO
6
44.4
0.18
167
15Fe/MgO
15
86.8
0.25
118
30Fe/MgO
30
44.9
0.15
135
micro pores of MeO seems unlikely even at the highest Fe loading, Other morphology changes or the change in the surface chemical species may be possible. It is found that H2S removal capacities of Fe/MgO catalysts are proportional to their BET surface areas. Figure 2 shows XRD patterns for the Fe/MgO catalysts after calcinations at 460 C. The sharp diffraction peaks at 2h = 36.9, 42.9, 62.3, 74.6, and 78.6 are ascribed to MgO support (JCPDS 4-829). Fe/MgO catalysts with below 15 wt% Fe show the peaks attributable to MgO and no evidence for Fe2O3. In the case of the sample containing 30 wt% Fe, newly appeared peaks are attributed to crystallized a-Fe2O3 (JCPDS 24-0072). Therefore, it can be deduced that the decrease of H2S removal capacity results from the formation of the crystalline a-Fe2O3. Figure 3 shows temperature-programmed reduction (TPR) profiles. For the sample with 1 wt% Fe, the peak at c.a. 560 C is observed, corresponding to the reduction of Fe3+ species in Fe/MgO catalyst [12]. With increasing Fe loading up to 15 wt%, the peak slightly shifts from 560 C to 580 C. The further Fe loading, 30 wt% Fe/MgO, shows the peak at 600 C with a shoulder peak at 670 C, indicating that new phases can be formed. The peak at 560– 580 C can be assigned to the reduction of Fe3+ in well Fig. 2 X-ray diffraction patterns of (a) 1 wt% Fe/MgO, (b) 4 wt% Fe/MgO, (c) 6 wt% Fe/MgO, (d) 15 wt% Fe/MgO, and (e) 30 wt% Fe/MgO
dispersed small domains of Fe2O3 which are not observed in XRD patterns. The TPR peak at 670 C can be assigned to the reduction of Fe3+ in crystalline a-Fe2O3. The formation of crystalline a-Fe2O3 may result in the shift of the peak temperature for reduction The XRD patterns support the identification of these phases for the loading of 30 wt% Fe. As shown Fig. 2, well dispersed a-Fe2O3 phases increase up to a Fe loading of 15 wt%. The steady increase in H2S removal capacity seen in Fig. 1 up to a Fe loading of 15 wt% suggests that well dispersed a-Fe2O3 are active phases for H2S oxidation. A considerable decrease in the H2S removal capacity occurs with increase in the loading from 15 wt% to 30 wt% Fe. XRD shows that the characteristic peaks of crystalline Fe2O3 appear with increase in the Fe loading from 15 wt% to 30 wt%. Therefore, it can be suggested that the reducibility of Fe3+ in Fe/MgO can be related to the H2S removal capacity of Fe/MgO catalysts. The XPS spectra were obtained for the observation of changes in the catalyst surface after reactions. Figure 4 shows the XPS spectra of the Fe/MgO catalysts after 4, 12, and 30 h reaction. XPS peaks of Fe 2p3/2 for Fe3+, Fe2+, and Fe are located at 710.95, 709.85, and 706.70 eV, respectively [13, 14]. Iron component in a fresh Fe/MgO
I n t e n sity ( a .u .)
(e) (d) (c) (b)
(a)
20
30
40
50
2θ
123
60
70
80
Catalytic Systems for the H2S Wet Oxidation
137
Fig. 3 TPR profiles of (a) 1 wt% Fe/MgO, (b) 4 wt% Fe/ MgO, (c) 6 wt% Fe/MgO, (d) 15 wt% Fe/MgO, and (e) 30 wt% Fe/MgO
Intensity (a.u.)
(e) (d) (c) (b) (a)
0
200
400
600
800
1000
Temperature (°C)
catalyst is in the state of Fe3+. Fe valence state in Fe/MgO catalyst changes from Fe3+ to Fe2+ during the reaction. Such results indicate that H2S can reduce Fe3+ to Fe2+ at room temperature. It has been reported that liquid redox processes use the reduction/oxidation cycle, as the followings [15, 16]. H2 S þ 2Fe3þ ! 2Hþ þ 2Fe2þ þ S
ð1Þ
1=2O2 þ 2Fe2þ þ 2Hþ ! 2Fe3þ þ H2 O
ð2Þ
The Fe3+ is reduced to Fe2+ by H2S, where the Fe2+ is regenerated back to Fe3+ via a re-oxidation reactions Fig. 4 XPS spectra of (a) Fe/ MgO after 4 h reaction, (b) Fe/ MgO after 12 h reaction, and (c) Fe/MgO after 30 h reaction
involving oxygen on the heterogeneous catalyst as shown in Fe/MgO [17]. The reduction degree of the Fe/MgO catalyst during the reaction can be governed by the relative magnitudes of the reduction rate by H2S with the oxidation rate by oxygen.
3 Effect of Iron Precursor Iron supported on magnesia was suggested to be a good catalyst for catalytic wet oxidation of H2S [9, 10]. The high H2S removal capacities of Fe/MgO could be explained by the finely dispersed iron oxide on MgO support. These
(c) Fe 3+
Intensity (a.u.)
Fe 2+
(b) Fe 3+ Fe 2+
(a) Fe 3+ Fe 2+ 718
716
714
712
710
708
706
704
Binding Energy /eV
123
138
catalysts prepared by a dissolution-precipitation method resulted in well dispersion of Fe on MgO support. The easy reduction of Fe3+ to Fe2+ in Fe/MgO by H2S was attributed to the small particle size of Fe component in Fe/MgO. Thus, it should be emphasized to disperse Fe particles well on MgO support in H2S wet oxidation. The amounts of paramagnetic Fe3+ cations on Fe/MgO were correlated with the H2S removal capacities of Fe/MgO catalysts from Mo¨ssbauer experiments [10]. The active sites were attributed to these Fe 3+ cations on Fe/MgO. The Fe/MgO catalysts prepared by the suspension impregnation were more effective than those prepared by the incipient impregnation for H2S removal in catalytic wet oxidation [18]. In the case of the Fe/MgO catalysts prepared with iron nitrate, it was suggested that well dispersed Fe2O3 can be active phase and crystalline a-Fe2O3 should be avoided for H2S oxidation. The influence of the iron precursors on the structure and activity of Fe/MgO catalysts for catalytic wet oxidation of H2S to elemental sulfur is shown in Fig. 5. Figure 5 shows the H2S removal capacity of Fe(N)/ MgO, Fe(S)/MgO, and Fe(Cl)/MgO catalysts with different iron loadings. The H2S removal capacity was obtained by calculating the total amount of H2S removed up to 50% of the H2S removal efficiency. The removal capacity in H2S oxidation increases up to 15 wt% Fe, and then decreases with further increase in Fe loading. The removal capacity of 15 wt% Fe/MgO is 3.75 gsulfur/gcat for iron nitrate precursor, 2.8 gsulfur/gcat for iron sulfate precursor, and 2.4 gsulfur/gcat for iron chloride precursor, respectively. H2S removal capacity of Fe(N)/MgO catalyst is the highest in all of the ranges of iron loadings, indicating that the
E.-K. Lee et al. Table 2 Physical properties of Fe/MgO prepared from different iron precursors Sample
Iron precursor
SBET (m2/g)
Pore volume (cm3/g)
Average pore ˚) diameter (A
Fe–N
Iron nitrate
32.2
0.18
223.4
Fe–S
Iron sulfate
25.7
0.07
109.2
Iron chloride
18.8
0.07
147.2
Fe–Cl
activity of the prepared catalysts is dependent on the iron precursor. Table 2 shows the physical properties of 6 wt% Fe/MgO samples prepared from different precursors. BET surface areas of Fe/MgO samples are in order of Fe(N)/MgO > Fe(S)/MgO > Fe(Cl)/MgO, which is correlated with H2S removal capacity. Therefore, BET surface area can be responsible for the highest H2S removal of Fe(N)/MgO sample since Fe(N)/MgO samples have the highest BET surface area. To understand the decomposition patterns of the precursor impregnated on MgO during the calcinations, the TG-DSC profiles of Fe–N, Fe–S, and Fe–Cl impregnated on MgO for preparing 6 wt% Fe/MgO were analyzed. Figure 6 and 7 show the TG and DSC profiles of Fe–N, Fe–S, and Fe–Cl samples, respectively. As shown in Fig. 6, two stages of weight loss can be observed on the TGA curve of the Fe–N sample. The first stage of weight loss of about 2–3% is observed between 0 C and 300 C and the second stage of weight loss of about 20% is observed above 300 C. This corresponds to the two endothermic peaks on the DSC profiles (Fig. 7), demonstrating that the decomposition proceeded in two steps [19, 20]:
Weight loss (%)
100
90
(c) (b) 80
(a) 70
Fig. 5 H2S removal capacities in H2S catalytic wet oxidation over Fe(N)/MgO (n), Fe(S)/MgO (d), and Fe(Cl)/MgO (m), and catalysts. H2S flow rate: 5 ml/min, O2 flow rate: 100 ml/min
123
0
200
400
600
800
1000
Temperature (°C)
Fig. 6 TGA curves of Fe–N (a), Fe–Cl (b), and Fe–S (c) samples
Catalytic Systems for the H2S Wet Oxidation
139
sample. The first stage of weight loss can be primarily due to water loss. However, the second weight loss can be due to evolution of chloride in reduced (HCl) form. Temperature regions can be divided into two steps [20]:
(c)
Heat flow (W/g)
(b)
Region I ð\320 CÞ
(a)
FeCl3 6H2 O ! FeCl3 þ 6H2 O Region II ð > 320 CÞ 2FeCl3 + 3H2 O ! a - Fe2 O3 + 6HCl 0
100
200
300
400
500
600
Temperature (°C)
Fig. 7 DSC curves of Fe–N (a), Fe–S (b), and Fe–Cl (c) samples
Region I ð\300 CÞ 2FeðNO3 Þ3 9H2 O ! 2FeðNO3 Þ3 þ 9H2 O Region II ð > 300 CÞ 2FeðNO3 Þ3 þ 3=2O2 ! a Fe2 O3 þ 6NO3 Three stages of weight loss are identified for the Fe–S sample. The first stage of weight loss can corresponds to the evolution of water, the second and the third stage of weight loss can be attributed to the evolution of gases. These reactions can occur [20]: Region I ð\300 CÞ FeSO4 7H2 O ! FeSO4 H2 O + 6H2 O Region II ð300 480 CÞ 6FeSO4 H2 O + 3/2O2 ! a - Fe2 O3 + 2Fe2 (SO4 )3 + 6H2 O
Region b ð > 480 CÞ Fe2 (SO4 )3 ! a - Fe2 O3 + 3SO3 Two stages of weight losses of Fe–Cl sample are observed with similarity to the TGA profile of the Fe–N
Based on these results, the Fe(S)/MgO sample can partially contain Fe2(SO4)3 phase after it is calcined at 460 C for 5 h, while Fe(N)/MgO and Fe(Cl)/MgO samples not have each precursor phase. The calcination temperature of 460 C was optimized for preparing well dispersed Fe/MgO catalyst, since the higher calcination temperature than 460 C resulted in the lower H2S removal capacity of Fe/MgO catalyst by the agglomeration of a-Fe2O3. Figure 8 shows XRD patterns for the 6 wt% Fe/MgO catalysts after calcinations at 460 C. The sharp diffraction peaks at 2h = 36.9, 42.9, 62.3, 74.6, and 78.6 are ascribed to MgO support (JCPDS 4-829). All of Fe/MgO catalysts show the peaks attributable to MgO and no evidence for Fe2O3. It indicates that well dispersed Fe particles can be present on Fe/MgO catalysts, regardless of the kind of Fe precursor. Figure 9 shows the TPR profiles of 6 wt% Fe/MgO catalysts prepared from different iron precursors. TPR patterns of these catalysts exhibit quite different peaks. Fe(N)/MgO sample shows the first peak at 570 C, whereas for Fe(S)/MgO and Fe(Cl)/MgO samples the first peak occurs at 650 C and 680 C, respectively. The first TPR peak corresponds to the reduction of the highly dispersed Fe2O3 species, which are not detected by the XRD. The second peak can be ascribed to the reduction of Mg2Fe2O4 species, which was previously confirmed by Mo¨ssbauer spectroscopy [10]. These results can be related to the Fe2O3 particle size in Fe/MgO catalysts. Hence, It indicates that Fe(N)/MgO shows the highest iron dispersion. The binding energies (BE) of some characteristic core levels of Fe and Mg in the samples and atomic surface ratios are shown in Table 3 [21, 22]. The presence of sulfate ion on the Fe(S)/MgO sample surface was detected by the binding energy at around 168.7 eV [23]. It indicates that the catalyst prepared from iron sulfate has sulfur species on the catalyst surface. In all cases, the binding
123
140
E.-K. Lee et al.
Intensity (a.u.)
Fig. 8 X-ray diffraction patterns of Fe(N)/MgO (a), Fe(S)/MgO (b), and Fe(Cl)/ MgO (c) catalysts
(c)
(b)
(a) 20
30
40
50
60
70
80
2θ
TCD signal (a.u.)
number of active sites on the surface, resulting in activity differences of H2S wet oxidation.
300
4 Effect of Support
(a) (b)
400
500
(c)
600
700
800
Temperature (°C )
Fig. 9 TPR profiles of Fe(N)/MgO (a), Fe(S)/MgO (b), and Fe(Cl)/ MgO (c) catalysts
energies are essentially constant for the catalyst, regardless of the kind of the salt precursor [24]. It is concluded that Fe/MgO samples are all in a similar valence state, indicating no electronic effect of Fe precursor by means of XPS regardless of precursors. The peak intensity ratios of Fe/MgO were calculated using atomic-sensitive factors. Fe atomic ratio to Mg on the surface of catalysts increase in the order: chloride < sulfate < nitrate samples. It indicates that the Fe sites are lower on Fe–S and Fe–Cl catalysts. So it is concluded that the difference in precursor affected the
123
In the present study, the supported iron oxide catalysts were characterized using Mo¨ssbauer spectroscopy, XRD, and TPR techniques. These catalysts were investigated to obtain fundamental knowledge of catalytic behavior for the catalytic wet oxidation of H2S. It is found that different kinds of Fe3+ species are formed on the surface of the supports with the same Fe loading. The highly dispersed species of iron oxide are proposed to be responsible for the high removal capacity of H2S in catalytic wet oxidation. The support effects on H2S oxidation at room temperature are examined with Fe/MgO, Fe/Al2O3, Fe/SiO2, and Fe/ZrO2 as shown in Fig. 10. The Fe/MgO catalyst shows a much higher removal capacity (2.6 gsulfur/gcat) in wet oxidation of H2S. The high removal capacity of Fe/MgO cannot be ascribed to MgO, since MgO alone also shows low H2S removal capacity [9]. Interestingly, it was reported that the activity of H2S oxidation above 450 K was high with the following order: Fe/SiO2 > Fe/ZrO2 > Fe/ TiO2 > Fe/Al2O3 > Fe/MgO [25]. Especially, Fe2O3 supported on a SiO2 catalyst showed the highest activity and best selectivity above 177 C. The reaction at room temperature was shown to change the activity order of the catalysts. The Fe/MgO catalyst with the lowest activity above 177 C shows the highest activity for the wet
Catalytic Systems for the H2S Wet Oxidation
141
Table 3 XPS binding energies and surface atomic ratios of Fe/MgO catalysts determined by XPS peak areas Fe precursor
Binding energy (eV)
Atomic ratio
Fe3/2p
Fe/Mg · 10–2
Mg2p
NO3–/Fe · 10–3
–3 SO2– 4 /Fe · 10
Cl–/Fe · 10–3
Nitrate
710.9 ± 0.1
49.8 ± 0.1
4.63
–
–
–
Sulfate
710.8 ± 0.2
49.6 ± 0.3
3.75
–
1.2
–
Chloride
710.9 ± 0.2
49.7 ± 0.1
2.49
–
–
–
100
(d) 60
(c)
40
Intensity (a.u.)
H2S removal efficiency (%)
80
20
(b) 0 0
5
10
15
20
Reaction time (h)
Fig. 10 H2S removal capacities with Fe/MgO (n), Fe/Al2O3 (s), Fe/ SiO2 (M), and Fe/ZrO2 (h) in H2S wet oxidation at room temperature: H2S flow rate: 5 ml/min, O2 flow rate: 100 ml/min, catalyst: 3 g
(a)
20
oxidation at room temperature, indicating that the mechanism can be changed with the reaction temperature. Table 4 shows that the BET surface areas and pore volumes of the supported iron oxide catalysts. The BET surface area of Fe/SiO2 is the highest of 231.3 m2/g, while that of Fe/Al2O3 is the lowest of 5.9 m2/g. The BET surface areas are in the order: Fe/SiO2 > Fe/MgO > Fe/ ZrO2 > Fe/Al2O3. As shown in Fig. 10, the H2S removal capacity is not in agreement with the order of BET surfaces of supported iron oxide catalysts. The XRD patterns of the prepared catalysts are presented in Fig. 11. The characteristic peaks in Fe/MgO are not found, indicating that a-Fe2O3 can be present in the small size. Fe2O3 characteristic peaks in other catalysts are strong, indicating that the Fe2O3 particle size of the Fe/ Table 4 B.E.T. surface area and pore volume of the supported iron oxide catalysts Surface property
Fe/MgO
Fe/SiO2
Surface area (m2/g)
74.16
231.3
Pore volume (cc/g)
0.34
0.993
Fe/ZrO2
Fe/Al2O3
8.6
5.9
0.04
0.007
30
40
50
60
70
2θ
Fig. 11 X-ray diffraction patterns of (a) Fe/MgO, (b) Fe/Al2O3, (c) Fe/SiO2, and (d) Fe/ZrO2: (d) a-Fe2O3
SiO2, Fe/ZrO2, Fe/TiO2, and Fe/Al2O3 are much larger than that of Fe/MgO [26]. It is well known that the particle below the 5 nm cannot be observed by the XRD. Figures 12 and 13 show the Mo¨ssbauer spectra of the Fe/MgO, Fe/Al2O3, Fe/SiO2, and Fe/ZrO2 catalysts at 20 C and –260 C, respectively. Table 5 shows the Mo¨ssbauer parameters of these catalysts. The doublet (two line) of the Fe/MgO catalyst at 20 C and –260 C may be due to highly dispersed Fe3+ species, that is, paramagnetic Fe3+ cations and/or super-paramagnetic Fe3+ oxide clusters [10, 27]. However, the Mo¨ssbauer spectra of other catalysts show the formation of sextets, indicating that Fe particle sizes are large enough to show the ferromagnetic properties. The hyperfine field of Fe/ MgO (H = 386.1 kOe for tetrahedral site and H = 454.0 kOe for octahedral site) are much smaller than that of pure a-Fe2O3 (ca. 525 kOe) and that of MgFe2O4 (H = 511 kOe for tetrahedral site and H = 538 kOe for octahedral site).
123
142
E.-K. Lee et al.
Fig. 12 Mo¨ssbauer spectra at 293 K of (a) Fe/MgO, (b) Fe/ Al2O3, (c) Fe/SiO2, and (d) Fe/ ZrO2
0.0
(d) 0.2
0.4
0.0
(c)
Absorption (%)
0.2
0.4 0.0 0.5
(b)
1.0
0
(a)
1 2 -12
The much lower hyperfine field of the Fe/MgO can be attributed to the small size of the MgFe2O4 particles or the iron particles [28]. The spectrum of the Fe/Al2O3 catalyst at 20 C and –260 C is only composed of a sextet, respectively. Comparing the Mo¨ssbauer parameters in Table 5 with those reported for Fe catalysts and reference compounds, the sextet is attributed to large a-Fe2O3 particles [29, 30]. It is consistent with the XRD results in Fig. 11 showing the strong peak intensity due to a-Fe2O3 crystallites. In the spectra of both Fe/SiO2 and Fe/ZrO2 catalysts at 20 C, the signals are composed of two components, a doublet and a magnetic split sextet, respectively. The doublet is attributed to highly dispersed Fe3+ species (Fe3+ oxide cluster), while the sextet is assigned to the to a-Fe2O3 crystallites [31]. The spectra of same catalysts are only composed of a sextet when measured at –260 C. These spectra at –260 C confirm the existence of super-paramagnetic Fe oxide particles (Fe3+ oxide cluster). The parameters are close to the values reported for Fe3+ oxide cluster and a-Fe2O3 crystallites, respectively. The most reliable percentages of each species (doublets and sextets) can be estimated at low temperature, since the probable differences in the recoil-free factors are minimal.
123
-8
-4
0 Velocity (mm/s)
4
8
12
Figure 14 shows the TPR profiles of the supported iron oxide catalysts. For the Fe/MgO sample, the TPR profile shows two peaks at 650 C and 850 C assignable to the Fe3+ cation and MgFe2O4 species, respectively. For the Fe/ SiO2 and Fe/ZrO2 samples, there are two broad peaks at relatively higher temperature compared to those of Fe/ MgO. The peak at low temperature can be attributed to Fe3+ oxide clusters and one at high temperature can be ascribed to a-Fe2O3 crystallites. In the case of Fe/Al2O3 sample, the only one peak is observed at 800 C. This can be assigned to the reduction of a-Fe2O3 crystallites. The XRD and Mo¨ssbauer spectra support that these different species are formed depending on the oxide supported metal.
5 Conclusions Magnesium oxide-supported iron metal oxide prepared from the incipient impregnation method catalyzes the wet oxidation of H2S to elemental sulfur by using oxygen as the oxidant at room temperature. Fe/MgO catalysts show the significant changes in morphology, which are explained by a dissolution-precipitation process. Fe/MgO
Catalytic Systems for the H2S Wet Oxidation
143
Fig. 13 Mo¨ssbauer spectra at 13 K of (a) Fe/MgO, (b) Fe/ Al2O3, (c) Fe/SiO2, and (d) Fe/ ZrO2
0.0
(d) 0.5
1.0 0.0
(c) Absorption (%)
0.5
0
(b) 1
2 0
(a) 1
-12
-8
-4
0 Velocity (mm/s)
4
8
12
Table 5 Mossbauer Parameter at 293 K and 13 K of Fe/MgO, Fe/Al2O3, Fe/SiO2, and Fe/ZrO2 T (C)
20
–260
Samples
Sextet
Doublet
Hhf (kOe)
EQ (mm/s)
d (mm/s)
Fe/Al2O3
517.77
–0.11
0.26
Fe/SiO2
492.90
–0.13
0.26
0.52
Fe/ZrO2
510.99
–0.11
0.25
Fe/MgO
Fe/MgO
Area (%) d (mm/s)
6 line
2 line
0.27
–
100
100
–
0.23
48
52
–0.77
0.19
70
30
0.77
0.39
24
76
EQ (mm/s) 0.65
386
–0.04
0.31
454
–0.10
0.33
Fe/Al2O3 Fe/SiO2
546.78 529.25
0.17 –0.09
0.38 0.36
100 100
– –
Fe/ZrO2
538.23
0.11
0.36
100
–
is the most effective catalyst for the wet oxidation of H2S to sulfur at room temperature among the tested catalysts. The structural analyses confirm that well-dispersed Fe species can be ascribed to the high removal capacity of H2S. Characterization of supported iron oxide catalysts by Mo¨ssbauer spectroscopy, XRD, and TPR leads us to
conclude that four kinds of Fe3+ species are formed with the various supports: both Fe3+ cation and MgFe2O4 species produced in Fe/MgO, Fe3+ oxide clusters. The high dispersion of Fe on MgO support is ascribed to the high activity in the wet oxidation of H2S to sulfur. XPS study shows that the H2S oxidation with Fe/MgO can be explained by the redox mechanism (Fe3+MFe2+)
123
E.-K. Lee et al.
TCD signal (a.u.)
144
200
(a) (b) (c) (d)
400
600
800
1000
1200
1400
Temperature (°C )
Fig. 14 TPR profiles of (a) Fe/MgO, (b) Fe/SiO2, (c) Fe/ZrO2, and (d) Fe/Al2O3
and the reduction of Fe component can be responsible for the deactivation of Fe/MgO.
References 1. DeBerry DW (1993) Topical report to Gas Research Institute, Report No. GRI-93/0019, April 1993 2. Cantrall R (1994) GRI liquid redox sulfur conference, May 15– 17, 1994, Austin, Texas 3. Rehmat A, Goyal A, Yoshizawa J (1995) GRI liquid redox sulfur conference, September 24–27, 1995, Austin, Texas 4. Kohl AL, FC Riesenfeld (1979) Gas purification, 3rd edn. Gulf Publishing Co., Houston, pp 402–406 5. Mikhalovsky SV, Zaitsev YP (1997) Carbon 35:1367 6. Adib F, Bagreev A, Bandosz TJ (2000) Envin Sci Technol 34:686 7. Capone M (1997) Kirk-Othmer encyclopedia of chemical technology. John Wiley & Sons, Inc
123
8. Sander UHF, Fisher U, Rothe U, Kola R (1984) In: More AI (ed) Sulfur dioxide and sulfuric acid. The British Sulfur Corporation Ltd., p 153 9. Jung KD, Joo OS, Cho SH, Han SH (2001) Appl Catal 240:213 10. Jung KD, Joo OS, Kim CS (2002) Catal Letters 84:53 11. Spretz R, Marchetti SG, Ulla MA, Lombardo EA (2000) J Catal 194:167 12. Shen J, Guang B, Tu M, Chen Y (1996) Catal Today 30:77 13. Graat PCJ, Somers MA (1996) J Appl Surf Sci 100:36 14. Bukhtiyarova GA, Bukhtiyarov VI, Sakaeva NS, Kaichev VV, Zolotovskii BP (2000) J Mol Catal 158:251 15. Hardison LC (1993) AIChE spring national meeting, 2 April, New Orleans, p 124 16. Newman DW, Lynn S (1984) Am Inst Chem Eng J 30:62 17. Shin MY, Nam CM, Park DW, Chung JS (2001) Appl Catal 211:213 18. Lee EK, Jung KD, Joo OS, Shul YG (2005) Bull Kor Chem Soc 26:281 19. Shen J, Guang B, Tu M, Chen Y (1996) Catal Today 30:77 20. Cornell RM, Schwertmann U (2003) The iron oxides. WileyVCH, p 526 21. Graat PCJ, Somers MAJ (1996) Appl Surf Sci 100:36 22. Bukhtiyarova GA, Bukhtiyarov VI, Sakaeva NS, Kaichev VV, Zolotovskii BP (2000) J Mol Catal 158:251 23. Oliveira AC, Fierro JLG, Valentini A, Nobre PSS, Rangel MDC (2003) Catal Today 85:49 24. Shen W-J, Ichihashi Y, Ando H, Okumura M, Haruta M, Matsumura Y (2001) Appl Catal 217:165 25. Tero¨rde RJAM, van den Brink PJ, Visser LM, van Dillen AJ, Geus JW (1993) Catal Today 17:217 26. Okamoto Y, Kubota T, Ohto Y, Nasu S (2000) J Catal 192:412 27. Cubeiro ML, Morales H, Goldwasser MR, Pe´rez-Zurita MJ, Gonza´lez-Jime´nez F, de N CU (1999) Appl Catal 189:87 28. Bond G, Molloy KC, Stone FS (1997) Solid State Ionics 101:697 29. Huang YL, Xue DS, Zhou PH, Ma Y, Li FS (2003) Mater Sci Eng 359:332 30. Cubeiro ML, Morales H, Goldwasser MR, Pe´rez-Zurita MJ, Gonza´lez-Jime´nez F, de N CU (1999) Appl Catal 189:87 31. Julia´n Cd, Alca´zar GAP, Cebollada F, Montero MI, Gonza´lez JM, Marco JF (1999) J Magnetism Magn Mater 203:175