Development of Fe-based Catalysts for Purification of Coke Oven Gases

Fe-based catalysts of different geometry are developed for the purification of coke oven gases: bulk, supported on alumina and supported on alumina silicate monoliths. Adsorption and decomposition of H2S on the catalysts developed are studied. Influence of active component content, type of support material and modification by Mn and Mo on the catalyst activity in de-H2S process is elucidated. Supported monolith catalysts show superior activity over bulk and supported spherical catalysts in H2S decomposition reaction and demonstrate stable operation in ammonia decomposition process during 2 hours at 900°C giving 100% ammonia conversion.


Introduction
Main impurities of coke oven gases (COG) are H 2 S and NH 3 . More than 99% of these impurities, which form harmful products upon oxidation or processing of COG, have to be removed taking into account environmental and technological aspects. On the other hand, COG containing up to 60% of hydrogen is considered as essential source for production of pure H 2 . Therefore, an efficient and low cost process for cleaning of COG is highly desirable.
According to literature data, many metal sulfides M x S y , where M = Na, K, Li, Fe, Co, Ni, Cr, Mo, V and metal oxides M x O y , where M = Fe, Cr, Cu, Mn, V, Ti Mo, Al, Zn are active in H 2 S decomposition reaction [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]. When metal oxides are used, the fist step in the process of H 2 S decomposition is the transformation of metal oxide into metal sulfide. Metal sulfide formed acts as a catalyst for the process of H 2 S decomposition. It was found that metal sulfide made in result of treatment of metal oxide by H 2 S shows superior activity compared with the readymade metal sulfide in reaction conditions [5].
Different sulfides form the following order accord-ing to their activity in H 2 S decomposition reaction: CoS 2 [18]<MoS 2 [19]<Co-Mo-S [20]<Ru-Mo-S [21]. It is necessary to note that besides Mo-and V-based catalysts, which are commonly studied in de-H 2 S reaction, Fe 2 O 3 catalyst also deserves attention as it exhibits high level of activity not only in H 2 S decomposition process, but in ammonia decomposition reaction as well [11,13,14], making this catalyst very promising for the use in the complex cleaning of COG. In [17]  Phase composition of the catalysts was registered by XRD using HZG-4 (Cu K α radiation) diffractometer with copper radiation in the 2θ range of 10-70 degrees. Specific surface area of the samples was measured by BET method. Quantitative analysis of elements in supported spherical and monolith catalysts was performed by the use of method of atomabsorption spectroscopy.
Prepared catalysts were tested in a 1% H 2 S-99% Ar flow. Catalyst was placed into quartz reactor, free volume of reactor was filled by sieved quartz (1-2 mm). Catalyst was treated in N 2 first at 150°C during 30 min and then at 600°C during 30 min. Activity test included activation of the catalysts by reaction mixture at 600°C till establishment of a steady outlet concentration of H 2 S and subsequent study of H 2 S decomposition at the temperatures of 600→ 700→800→900→800→700→600°C at residence time equal to 1 s. Duration of catalyst activation was characterized by value ρ representing total amount of H 2 S (mmole) fed into the reactor divided by the weight of a catalyst sample at which a steady level of H 2 S conversion is achieved.
Activity test in de-NH 3 reaction was performed as follows: catalyst was placed into quartz reactor, at first was treated in He at 150°C during 30 min, then at 600°C during 1 hour and catalyst activity was measured within temperature interval of 600-900°C. Both fresh catalyst and catalyst after activity test in H 2 S decomposition reaction were tested in NH 3

Bulk catalysts
XRD showed formation of α-Fe 2 O 3 after calcination at 600°C in all samples, except sample F-6, synthesized via thermal decomposition of iron (III) citrate hydrate, where formation of additional phase Fe 3 O 4 was registered, Table 1. Specific surface area of the samples varies within interval of 2.9-13.0 m 2 /g.
Use of oxide catalyst for H 2 S decomposition reaction requires preliminary activation of catalyst by means of H 2 S adsorption in order to transform metal oxide into metal sulfide. Detailed study of H 2 S adsorption on the catalysts developed was performed in order to clarify the process of sulfidation of metal oxide. These experiments were fulfilled at the tem-perature of 600°C in order to decrease the contribution of the H 2 S decomposition reaction in the experiments. Main results on the H 2 S adsorption on bulk Fe 2 O 3 catalysts are given in Fig. 1 showing dependence of H 2 S conversion (X H 2 S , %, left axis) and quantity of hydrogen produced due to H 2 S decomposition reaction (C H 2 , vol.%, right axis) on the magnitude ρ. Value ρ at which a steady level of H 2 S conversion is achieved, can characterize the duration of catalyst activation. At the beginning of the experiments H 2 S conversion attained value near 100%. Meantime, hydrogen was not registered in the reaction mixture. Thus, it can be concluded that almost all hydrogen sulfide is quantitatively adsorbed and interacted with the catalyst. After certain period of time, different for each catalyst, the increase of H 2 S concentration in reaction mixture with simultaneous   Table 2 Physical chemical properties of supported spherical catalysts increase of hydrogen concentration is observed. It is obvious that chemical transformation of a catalyst proceeds during H 2 S adsorption by formation of a new phase which is active in H 2 S decomposition reaction. Indeed, XRD registered sulfidation of the catalysts with formation of iron sulfide after reaction (Table 1, Fig. 2). It is interesting to note that duration of catalyst activation for the samples studied is different. The longest activation period was registered for the sample F-7 prepared via thermal decomposition of ammonium iron(III) dihydrocitrate citrate, while the shortest -for the sample F-4 prepared via thermal decomposition of iron(III) oxalate pentahydrate.   Table 1). ρ represents amount of H 2 S supplied (mmole) related to the 1 g of the sample. Fig. 3. Histogram of reaction rate of H 2 formation related to the 1 g of the sample (mole H 2 /s·g) demonstrates that Fe 2 O 3 catalyst synthesized via thermal decomposition of iron(III) nitrate nanohydrate (F-1) is the most effective. Essential decrease of specific surface area of the samples after activity test (down to ~ 0 for F-6 and F-7) is observed ( Table 1). The sample F-5 synthesized via thermal decomposition of ammonium iron (III) oxalate trihydrate demonstrates the highest thermal stability in respect to specific surface area. Although the catalysts F-1, synthesized via thermal decomposition of iron (III) nitrate nanohydrate, exhibited the highest activity in H 2 S decomposition reaction, the ammonium iron (III) oxalate trihydrate was chosen for the preparation of supported catalysts in order to exclude emission of toxic nitrogen oxides which are formed during iron (III) nitrate nanohydrate decomposition. Moreover, the sample F-5 synthesized by the use of Fe(NH 4 ) 3 (C 2 O 4 ) 3 ·3H 2 O shows sufficiently high activity in H 2 S decomposition reaction and demonstrates the highest thermal stability in respect to specific surface area.

Calculated reaction rates of H 2 formation according to the decomposition reaction H 2 S → H 2 + S on bulk Fe 2 O 3 catalysts are given in
Activity of the catalysts supported on γ-Al 2 O 3 and α-Al 2 O 3 depends significantly on the content of active component, Fig. 4. Reaction rate of H 2 formation decreases with increase of Fe content in the sample for catalysts supported on γ-Al 2 O 3 . Highest activity demonstrates catalyst F-8 with Fe content equal to 2.9%. On the contrary, increase of Fe content from 3.1 to 4.1% in the catalyst F-9 leads to increase of the catalyst activity.

Supported spherical catalysts
Phase composition of spherical catalysts depends on the content of iron introduced and type of support material used ( Table 2). Phase of α-Fe 2 O 3 was registered in the catalysts supported on α-alumina and zirconia, while active component in the catalyst F-8b (9.7%Fe/γ-Al 2 O 3 ) was recognized as the solid solution of Fe 3+ in γ-Al 2 O 3 . Fig. 4. Histogram of reaction rates of H 2 formation related to 1 g of Fe in the sample for supported spherical catalysts calcined at 600°C: F-8, F-8a, F-8b; and F-9a, F-9b, F-9 (see Table 2). Formation of iron sulfide (Fe 7 S 8 or Fe 1-x S) was registered in the samples after H 2 S decomposition reaction for all catalysts supported both on γand α-Al 2 O 3 . The catalyst F-8a (4.6%Fe/γ-Al 2 O 3 ) is the only exception. Transformation of oxide active component to iron sulfide leads to change of catalyst colour from beige (F-8) or orange (F-9) to black colour. Significant decrease of specific surface area of the samples supported on γ-Al 2 O 3 is observed.
Activity data for Fe-containing catalysts supported on various materials (γ-Al 2 O 3 , α-Al 2 O 3 , ZrO 2 , SiO 2 ) and calcined at 600°C are presented in Fig. 5. Increase of H 2 S conversion with temperature rising from 600 to 900°C is observed. Meantime, the catalysts deposited both on γ-Al 2 O 3 and α-Al 2 O 3 demonstrate the highest level of activity. Formation of iron sulfide after activity test is registered for catalysts supported on γ-Al 2 O 3 , α-Al 2 O 3 and SiO 2 materials, whereas decrease of specific surface area is observed for all catalysts (Table 2). consider this fact, reaction rates of H 2 formation reaction related to the 1 g of introduced Fe were compared. It is seen that picture is changed significantly. Supported monolith catalysts show superior activity compared with bulk and supported spherical catalysts, Fig. 7.
Results described above were devoted to the study of catalysts in H 2 S decomposition reaction. However, besides H 2 S, COG contain substantial amount of ammonia. Therefore, independent experiments on decomposition of ammonia by the use of monolith F-9/ B catalyst were performed. In order to elucidate the behavior of catalysts in the presence of H 2 S and NH 3 ,  Table 2).
Modification of the samples F-8 (2.9%Fe/γ-Al 2 O 3 ) and F-9 (4.1%Fe/α-Al 2 O 3 ) was performed to increase their activity. According to XRD analysis, modification of catalysts by manganese leads to formation of additional β-Mn 2 O 3 phase compared with phase composition of initial catalysts ( Table 2). Modification of F-8 by molybdenum does not change phase composition of the catalysts. XRD registers γ-Al 2 O 3 similarly to initial F-8 catalyst. In turn, modification of F-9 sample by molybdenum changes significantly phase composition of the catalyst: interaction of Fe with Mo occurs leading to the formation of Fe 2 MoO z or Fe 2 (MoO 4 ) 3 phases. However, activity data, presented in Fig. 6, allows to conclude that modification of the samples F-8 and F-9 by Mn and Mo results in decrease of the catalyst activity, although formation of MnS and MoS 2 is observed after activity test according to XRD data ( Table 2).

Supported monolith catalysts
To clarify the influence of catalyst geometry on its activity in H 2 S decomposition process, comparative analysis of activity data for different catalysts was made: bulk Fe 2 O 3 catalyst F-5, supported spherical F-8, F-9 and supported monolith catalysts F-8/B, F-9/B (Fig. 7). Activity of supported spherical catalysts (F-8 and F-9) exceeds activity both of bulk and supported monolith catalysts. But it is necessary to emphasize that content of introduced active component (Fe) in the catalysts compared varies within wide range: 70 wt.% for bulk Fe 2 O 3 catalyst and less than 1 wt.% for supported monolith catalysts. In order to Fig. 6. Histogram of reaction rates of H 2 formation related to 1 g of sample for supported spherical catalysts F-8 and F-9 modified by Mo and Mn, calcined at 600°C: Mo/ F-8, Mn/F-8, Mo/F-9, Mn/F-9 (see Table 2). Reaction temperature 600 o C 700 o C 800 o C 900 o C activity of fresh catalysts was compared with activity of catalyst tested preliminary in H 2 S decomposition reaction at the temperatures of 600→900→ 600°C. It was found that 100% ammonia conversion on the fresh supported monolith catalyst is observed starting from 750-800°C, whereas for catalyst after de-H 2 S test -starting from 900°C, Fig. 8. Monolith catalyst demonstrated stable operation in de-NH 3 process at 900°C for 2 hours.

Conclusions
Fe-based catalysts of different geometry are developed for the purification of coke oven gases: bulk, supported on alumina and supported on alumina silicate monoliths.
Adsorption and decomposition of H 2 S reaction were studied on bulk Fe 2 O 3 catalyst and Fe(NH 4 ) 3 (C 2 O 4 ) 3 ·3H 2 O salt was chosen as the best starting material to be used for the synthesis of supported catalysts. Fe-containing catalysts supported on alumina show highest level of activity in H 2 S decomposition process compared with Fe-containing catalysts supported on ZrO 2 and SiO 2 . It was found that modification of Fe-containing catalysts by Mn and Mo leads to the decrease of the catalyst activity.
Monolith catalysts synthesized show superior activity in de-H 2 S process over bulk and spherical catalysts when reaction rates of H 2 formation related to 1 g of Fe introduced are compared. It was revealed that monolith catalysts demonstrate stable operation in ammonia decomposition process during 2 hours at 900°C giving 100% ammonia conversion.