- Email: [email protected]
Ammonia synthesis over cobalt catalysts doped with cerium and barium. Effect of the ceria loading
Applied Catalysis A: General 445–446 (2012) 280–286
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Ammonia synthesis over cobalt catalysts doped with cerium and barium. Effect of the ceria loading c ˙ ˛ nski ´ Magdalena Karolewska a , Elzbieta Truszkiewicz a , Bogusław Mierzwa b , Leszek Kepi , a,∗ Wioletta Raróg-Pilecka a
Warsaw University of Technology, Faculty of Chemistry, Noakowskiego 3, 00-664 Warsaw, Poland Institute of Physical Chemistry of the Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland c Institute of Low Temperature and Structure Research of the Polish Academy of Sciences, Okólna 2, 50-950 Wrocław, Poland b
a r t i c l e
i n f o
Article history: Received 28 May 2012 Received in revised form 6 August 2012 Accepted 25 August 2012 Available online 31 August 2012 Keywords: Ammonia synthesis Cobalt catalyst Barium and cerium promoters Ceria loading
a b s t r a c t The effect of ceria loading on the activity and thermal stability of a cobalt catalyst for ammonia synthesis has been studied. The concentration of CeO2 in the Co3 O4 + CeO2 mixtures ranged from 0 to 29 wt%, whereas the barium content was constant and equaled 13 wt% in relation to Co3 O4 . Kinetic measurements in NH3 synthesis were carried out in a flow differential reactor operating under standard conditions (p = 6.3 MPa, T = 400 ◦ C, H2 :N2 = 3:1). The kinetic measurements of NH3 synthesis were supplemented with the characterization studies of the obtained materials – N2 physisorption, XRPD, H2 chemisorption, TPR-TG/DTG-MS and SEM-EDS. It has been found that ceria plays the role of a structural promoter; it hinders the sintering of Co3 O4 during calcination and stabilizes the surface area of Co under reaction conditions. Moreover, it seems that ceria stabilizes the hcp phase of metallic cobalt, which results in a higher activity of the cobalt catalysts. The co-promoted cobalt catalyst containing 11.5 wt% of CeO2 (symbol Co/Ce(11.5)/Ba) was shown to be more active in NH3 synthesis than the conventional fused iron catalyst (KM I, H. Topsoe). Moreover, the catalyst Co/Ce(11.5)/Ba is very resistant to overheating. Heat treatment at 600 ◦ C for 160 h (3H2 :N2 = 30.0 l/h, p = 0.1 MPa) results in a decrease of its activity in NH3 synthesis by only a few percent. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Ammonia synthesis is one of the most important and the largest, in terms of the production scale, catalytic processes in the chemical industry today. The global production of NH3 was estimated to be 124 million tons in 2006 and continues to exhibit a steady increase [1]. Ammonia is used in the production of fertilizers, explosive materials, and as a refrigerant. Since the hydrogen used in fuel cells has to be pure, i.e. not contaminated with carbon oxides, ammonia is thought to become the best source of hydrogen suitable for this application in the near future [2–7]. The catalytic reaction of hydrogen and nitrogen, which leads to the formation of ammonia, is of significant importance in the development of catalysis. Two Nobel Prizes have been awarded for research on this topic [8]. Moreover, studies on ammonia synthesis have led to many significant improvements and innovations. The basic terms used in catalysis, such as promoter or poison, were first used with regard to ammonia synthesis. Furthermore, new catalytic concepts, e.g. structure
∗ Corresponding author. Tel.: +48 22 234 57 66; fax: +48 22628 27 41. E-mail address: [email protected] (W. Raróg-Pilecka). 0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apcata.2012.08.028
sensitivity of a reaction, have been developed during studies of ammonia synthesis catalysts [9]. For the above reasons, the researchers from industrial centres and academia are still active in the field of NH3 synthesis and they try to work out a completely new catalyst [10–18] or to improve the formula and properties of the conventional iron one [19–25]. Among several new catalytic systems investigated in the last 40 years, only ruthenium supported on high surface area graphite (HSAG) has been implemented in the industrial practice so far. A combination of both the conventional iron (first catalytic bed) and the modern Ru/C catalyst (last three beds) in the so called Kellog Advanced Ammonia Process (KAAP) has been shown to be very advantageous [26,27], i.e. the pressure in ammonia loops could be significantly reduced (to 9.0 MPa), thus resulting in a lower energy consumption. However, the high price of ruthenium and its tendency to catalyzes the depletion of the support, i.e. methanation [28–31], are the main reasons why the ruthenium catalyst did not revolutionize the process of NH3 synthesis. Research on the development of an entirely new catalytic system with other metals which exhibit catalytic activity in this reaction [32] is still ongoing. Cobalt-based catalysts are among the most frequently investigated systems [15,33–35]. The latest research of Raróg-Pilecka et al. has shown
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
that unsupported cobalt catalysts promoted with cerium and barium (Co/Ce/Ba) obtained by the templating method are more active than the conventional iron catalyst [36]. However, due to a rather difficult technique of catalyst preparation, further work of this team was aimed at obtaining promoted cobalt catalysts by a simpler method, namely coprecipitation [37]. The catalysts obtained by the latter method were even more active in ammonia synthesis. However, research on both types of catalysts shows that the essential components of an active cobalt catalyst are the promoters: barium and cerium. Their synergistic effect is evident in both works [36,37]. This paper also investigates the catalytic properties of Co/Ce/Ba in NH3 synthesis. However, this work focuses on the effect of ceria loading on the activity of the cobalt systems. The kinetic measurements of NH3 synthesis were supplemented with the characterisation studies of the obtained materials – N2 physisorption, XRPD, H2 chemisorption, TPR-TG/DTG-MS and SEMEDS. Moreover, thermal stability tests of the most active contact have been performed. 2. Experimental 2.1. Catalyst preparation The mixtures of cobalt and cerium carbonates with different Co/Ce ratios were prepared by co-precipitation with potassium carbonate as the precipitant. Our recent studies [38] have shown that the application of carbonates, in particular potassium carbonate, in the role of the precipitant, leads to obtaining materials with a higher surface area and more favourable catalytic properties. Appropriate amounts of a mixture of Co(NO3 )2 ·6H2 O and Ce(NO3 )3 ·6H2 O, both of analytical grade, were dissolved in distilled water and warmed to approximately 90 ◦ C. Then, a warm (90 ◦ C) solution of K2 CO3 (analytical purity) was slowly added. The obtained precipitate, i.e. mixtures of cobalt carbonate and cerium carbonate, was filtered under reduced pressure and washed with cold distilled water several times until the pH was ∼7. The materials were then dried and calcined at 500 ◦ C overnight. The resulting black powder was a mixture of Co3 O4 and CeO2 . Thus obtained materials were impregnated using an aqueous solution of barium nitrite (analytical purity). The resulting precursors were then dried at 120 ◦ C overnight. The amount of barium was constant and equal to 13 wt% in relation to Co3 O4 . The final step of preparation was crushing and sieving the solid materials in order to get the 0.2–0.63 mm fraction. The ceria loading of the resulting materials is presented in Table 1. This table also contains data for cobalt oxide promoted with barium and pure CeO2 , also obtained by precipitation and a subsequent calcination. Additional XRF measurements have shown that the potassium concentration in the co-precipitated samples is less than 0.05%, other impurities were lower than the detection limit.
281
2.2. Characterization studies The precursors of the catalysts, that is the oxidized forms, and the reduced samples were characterized by nitrogen physisorption (ASAP 2020, Micromeritics), XRPD and SEM-EDS studies. The reduced samples were also characterized by hydrogen chemisorption (TPD-H2 ). Moreover, TPR experiments were performed for all materials using a thermobalance equipped with a quadrupole mass spectrometer (TPR-TG/DTG-MS). X-ray powder diffraction (XRPD) data were collected with a Siemens D 5000 diffractometer in a Bragg–Brentano configuration using a Cu-sealed tube operating at 40 kV and 40 mA with a Ni filter. Measurements of each system were performed in the scattering angle range of 5–100◦ with a 0.02◦ step and a counting speed of 1.0 s/step. The average size of cobalt or cobalt oxide crystallites was determined from the Scherrer equation using the integral width of reflections fitted to the analytical Pearson VII functions. The morphology and homogeneity of the samples was determined with a Scanning Electron Microscope (FEI NovaNanoSEM 230) equipped with an EDX spectrometer (EDAX Genesis XM4). The H2 chemisorption measurements were carried out in a flow set-up (a fully automated PEAK 4 instrument) equipped with a TCD cell and supplied with high purity (≥99.9999 vol%) gases [39]. The details of the experimental procedure were described previously [35]. The sample (0.4 gCo3 O4 ) was reduced in a flowing H2 :Ar = 80:20 mixture (40 ml/min) at 550 ◦ C for 20 h, flushed with argon (40 ml/min) at that temperature, and cooled to 150 ◦ C in argon. Then, the sample was cooled in H2 (40 ml/min) to 0 ◦ C and the sample was flushed with argon (40 ml/min, 0 ◦ C, 15 min) to remove weakly adsorbed hydrogen. Next, the catalyst was heated in flowing Ar with a 15 ◦ C/min temperature ramp – step 1 (TPD-H2 ). The concentration of H2 in the outlet gas was monitored. Consequently, the amount of hydrogen evolved to the gas phase was determined by integrating the H2 trace. The hydrogen TPD data were used for calculating the Co surface area (SH2 ). The H:Cos = 1:1 stoichiometry [40] and the formula proposed by ´ Borodzinski and Bonarowska [41] were used for calculations. In addition, the samples were overheated twice (steps 2 and 3) in a mixture of H2 :Ar = 80:20 (40 ml/min) at 600 ◦ C for another 20 h. After each stage of overheating, hydrogen chemisorption measurements were performed according to the procedure described above. Hence the influence of overheating on the size of the total area of cobalt available for hydrogen was determined. The TPR-TG/DTG-MS measurements were performed using a NETZSCH STA 449C thermobalance equipped with a quadrupole mass spectrometer (NETZSCH QMS 403C). The experiments were conducted using samples of approximately 30 mg of catalyst powder heated up to 550 ◦ C at the constant rate of 10 ◦ C/min in a pure (≥99.999 vol%) H2 :Ar = 10:90 mixture (100 ml/min). This temperature was maintained for 3 h. The reference crucible was empty. The mass change, temperature and selected m/z signals were monitored throughout the entire experiment. In order to avoid water
Table 1 Characteristics of catalyst precursors. Sample symbol
Ceria loading (wt%)
N2 physisorption SP
Co3 O4 Co/Ce(3.7) Co/Ce(11.5) Co/Ce(17.2) Co/Ce(29.0) CeO2 b a b
0 3.7 11.5 17.2 29.0 100
a
2
(m /gCo3 O4 )
40 69 94 91 109 32
SP , VP – BET surface area and pore volume of materials in oxide form, respectively. Values with respect to mCeO2 [g].
XRPD VP
a
0.16 0.29 0.42 0.35 0.45 0.11
3
(cm /gCo3 O4 )
dCo3 O4 (nm) 20 12 10 9 8 –
SCo3 O4 (m2 /gCo3 O4 ) 49 82 98 109 123 –
282
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
physisorption all the necessary apparatus parts were kept heated to 280 ◦ C. 2.3. Activity measurements The activity of the catalysts in ammonia synthesis was measured in a flow tubular reactor supplied with a very pure (≥99.99995 vol%) H2 :N2 = 3:1 mixture of controlled ammonia content (x1 ; 0–8%) [42]. The activities of various samples were tested under steady-state conditions (6.3 MPa, 400 ◦ C, gas flow rate = 70.0 l/h) and for a fixed catalyst mass (0.4 gCo3 O4 ). The concentration of ammonia in the outlet gas (x2 ) was monitored and the productivity – average reaction rate, was calculated (integral measurement; x1 = 0). Prior to kinetic tests, the samples were reduced in a H2 :N2 stream at 470 ◦ C for 45 h and then at 550 ◦ C for 40 h. The subsequent experimental procedure involved the following steps: determination of the initial reaction rate (step 1), i.e. the rate after stabilization at 470 ◦ C (40 h) + 550 ◦ C (45 h); short-term (24 h) overheating in a 3H2 + N2 mixture at 600 ◦ C, and another determination of the reaction rate under the same conditions as in the initial rate measurement (step 2). Additionally, another step (step 3) was performed, which was analogous to step 2. Both, the activation (reduction) and overheating were performed under atmospheric pressure and gas flow rate = 30.0 l/h. In a more complex experiment, both outlet (x2 ) and inlet (x1 ) streams were analysed (differential measurement; x2 − x1 x1 ) to determine the reaction rate corresponding to the mean ammonia concentration (x = (x1 + x2 )/2) in a catalyst layer; the details of such examinations can be found elsewhere [42]. The relationship between the NH3 synthesis rate and conversion was established for the best catalyst according to this procedure. 3. Results and discussion Fig. 1 presents a sequence of the diffraction patterns of the catalyst precursors (Co3 O4 + CeO2 ) which additionally contain barium. As seen, all patterns display a set of maxima characteristic for Co3 O4 (JCPDS 42-1467) and barium carbonate (JCPDS 05-0378) phases. Moreover, the diffraction patterns of samples Co/Ce(17.2)/Ba and Co/Ce(29.0)/Ba exhibit signals typical for CeO2 (JCPDS 34-0394). In the case of the other three systems (that is Co/Ba, Co/Ce(3.7)/Ba and Co/Ce(11.5)/Ba) ceria does not have a noticeable contribution to the XRD patterns. This means that ceria species are finely dispersed on the cobalt oxide surface, or they are poorly organized, nearly amorphous.
Fig. 1. XRPD patterns of the fresh catalysts (prior to reduction).
Fig. 2. Pore size distribution curves of the catalyst precursors.
In Table 1 and Fig. 2 N2 physisorption data of the samples of mixed cobalt–cerium oxides (precursors of the catalysts) are presented. In the case of both Table 1 and Fig. 1 the data for pure Co3 O4 and CeO2 , obtained by precipitation and subsequent calcination, are also shown. As can be seen (Table 1) pure oxides exhibit the worst textural parameters, i.e. they have the lowest SP and VP values. For mixed oxides, these parameters are higher. The highest values of the total surface area and pore volume were obtained for the sample Co/Ce(11.5). Moreover, the area of this sample is more than twice as high as the sum of the components (i.e. Co3 O4 and CeO2 ), calculated with accordance to their content in the double oxide system. The distribution of the pores of the samples in the oxide forms was determined from the nitrogen adsorption isotherm at 77 K using the Barrett–Joyner–Halenda model [43]. The spectrum of CeO2 has one maximum at the average pore width of ∼40 nm, whereas the spectrum of Co3 O4 has two characteristic peaks at ∼40 and 90 nm. Samples containing both oxides (Co3 O4 and CeO2 ) have a pore distribution similar to that of pure cobalt oxide. However, the total volume of the pores is much higher than that of the Co3 O4 sample. The results of the physisorption studies show that addition of cerium prevents the sintering of these materials during calcination. Moreover, the larger the cerium content, the higher the value of the overall surface area (SP ). It should be emphasized that the contribution of a separate Ce-containing phase (SBET = 32 m2 /g) to the overall surface area of the Co + Ce materials would be lower than 9%, even if the entire content of cerium in the catalyst was taken into account in the calculations. This indicates the structural nature of this promoter in cerium–cobalt systems. The physisorption data (SP ) are in good agreement with the XRPD results (SCo3 O4 – Table 1), for which the increase of the Co3 O4 surface area is also observed as a result of CeO2 addition. Nevertheless, the SEM studies for these materials do not show visible differences between them. The sample images for Co/Ba and Co/Ce(17.5)/Ba catalyst precursors are shown in Fig. 3(a) and (c). The XRPD results for the promoted catalysts after activity measurements are presented in Fig. 4. In all of these patterns metallic cobalt is the only visible cobalt-containing phase. As expected, Co3 O4 is reduced to metallic Co (JCPDS 15-0806 and 05-0727) as a result of activation. Signals coming from Co correspond to both the fcc phase (44.4◦ , 51.6◦ , 75.9◦ and 92.3◦ ) and the hcp phase (e.g. 41.7◦ , 47.6◦ ). However, the asymmetry and broadening of the signal at the angle of 47.6◦ and the absence of signals at angles 84.2◦ and 94.7◦ indicate strong disorder of the hcp phase in the [1 0 1] direction. In contrast, the cubic phase is well crystallized. Promoters i.e. barium and cerium are present in the diffraction patterns in the form of Ba(CeO3 ) (JCPDS 82-2425). The pattern obtained for
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
283
Fig. 3. SEM micrographs of the catalyst precursors (a and c) and the post-reaction catalysts after exposure to air (b and d); (a, b) Co/Ba sample and (c, d) Co/Ce(17.2)/Ba sample.
the sample with the highest concentration of ceria (Co/Ce(29.0)/Ba) contains additional signals from cerium oxide. Based on SEM-EDS studies it was found that Co/Ba is not homogeneous. Barium creates separate particles mixed with cobalt. In contrast, all samples containing both barium and cerium exhibit high homogeneity. The EDS analysis of several points of the analysed area shows that all major elements of all samples, i.e. Co, Ba and Ce, are evenly distributed with a Ba/Ce atomic ratio close to 1. Moreover, a difference between the sizes of cobalt particles in
Co/Ba and Co/Ce/Ba samples was observed, which is clearly visible on presented images in Fig. 3(b) and (d). The cobalt particles in Co/Ba are bigger than those in Co/Ce(17.2)/Ba. This indicates that the addition of Ce inhibited the growth of Co particles. The TPR-TG/DTG-MS data obtained for the all catalyst precursors are depicted in Figs. 5 and 6. The first derivative of the mass change (DTG curve) observed when a sample is heated in a flowing mixture of H2 /Ar is depicted in Fig. 5. The other figure shows the changes in the amount of evolved water vapour, which is the main reaction product, in time. As described previously [37,44,45]
Fig. 4. XRPD patterns of promoted cobalt catalysts used in ammonia synthesis.
Fig. 5. TPR-TG/DTG-MS studies of the cobalt catalysts. DTG curves of samples.
284
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286 Table 2 XRPD and N2 physisorption characteristics of catalysts used in ammonia synthesis. Catalyst symbol
XRPD
Co/Ba Co/Ce(3.7)/Ba Co/Ce(11.5)/Ba Co/Ce(17.2)/Ba Co/Ce(29.0)/Ba a
Fig. 6. TPR-TG/DTG-MS studies of the cobalt catalysts. Water vapor evolution signals (MSH2 O ) of samples.
during the reduction of pure Co3 O4 two clear signals are observed on each curve – see Figs. 5 and 6. They correspond to the reduction of Co3 O4 to CoO (the first signal at the lower temperature), and a subsequent reduction of CoO to Co (the second signal at the higher temperature). The addition of CeO2 to Co3 O4 causes a shift of the reduction signals towards higher temperatures (Figs. 5 and 6). As a result the process of complete reduction of Co3 O4 → Co in samples with a CeO2 content above 10 wt% ends in the isothermal segment of the experiment. Only the sample with a low CeO2 content behaves similarly to pure Co3 O4 . It should also be noted that under the conditions of the TPR-TG/DTG-MS measurements, the mass of CeO2 decreases only very slightly in the temperature range: 400–550 ◦ C, which, according to literature [46], should be attributed to the reduction of surface oxygen species. The introduction of barium to the Co + Ce systems neutralizes the negative impact of ceria on the reduction of Co3 O4 , as shown in Fig. 7 for the sample richest in cerium, i.e. Co/Ce(29.0)/Ba. For comparison, Fig. 7 shows the DTG curves for the corresponding measurements for pure Co3 O4 and Co3 O4 doped only with CeO2 , i.e. Co/Ce(29.0). In our studies a slight mass loss was also observed at a temperature close to 100 ◦ C (Fig. 5). This effect is associated with the emissions of water (see Fig. 6), which is either adsorbed on the surface of these materials or derives from the initial Co3 O4 reduction. In conclusion, the TPR-TG/DTG-MS studies revealed that ceria delays the reduction of Co3 O4 to metallic Co, probably because it consumes H2 [44,47–50]. Ceria is reduced more easily, that is at lower temperatures, as a result of hydrogen spillover from the transition metal (Co) to ceria. Hence, ceria uptakes additional hydrogen. This is why more hydrogen is consumed and a longer reduction time is required to fully reduce the cobalt oxide in ceria-containing
N2 physisorption
dCo (nm)
SCo (m2 /gCo )
SR
20 25 22 20 20
33.9 27.1 30.8 33.9 33.9
9.8 25.5 23.0 21.2 11.5
a
(m2 /gCo )
BET surface area of the catalysts.
samples [48–50]. In contrast, barium neutralizes the negative impact of CeO2 . The effect of Ba may result from interactions of the two promoters and the formation of a common phase, i.e. Ba(CeO3 ) (see Fig. 4). However, further studies are needed to resolve this issue. Data for catalysts unloaded from the NH3 reactor are presented in Table 2. The values concern the results of calculations of the surface area of cobalt based on the XRPD measurements (XRPD column) and BET surface area (N2 physisorption column). The physisorption data shows that samples containing cerium and barium besides cobalt have higher specific surface areas than the sample containing barium only. However, Co/Ce(3.7)/Ba, Co/Ce(11.5)/Ba and Co/Ce(17.2)/Ba catalysts exhibit the highest SR values. In the case of the XRPD results (SCo ) no meaningful differences between samples are detected. Moreover, the surface areas of cobalt estimated from the XRPD are higher than those obtained from N2 physisorption measurements. This may be due to the fact that XRPD estimates average crystallite diameters, whereas nitrogen sorption occurs on the whole available surface. If Co crystallites create bigger clusters discrepancies between the obtained values are observed. It should be also remembered that both types of measurements were performed ex situ, that is after exposure of samples to air. Hence, the surface of the metallic phase was oxidized before the measurements and as a consequence the obtained values can be overestimated. This has already been observed in previous studies [36]. According to our prior research [37] reliable values of the surface of the active phase can be obtained with hydrogen chemisorption. Table 3 presents chemisorption data (SH2 ), i.e. the surface of cobalt available for H2 determined from TPD-H2 measurements, of the cobalt catalyst promoted only with barium and for a series of cobalt catalysts promoted with both ceria and barium. For comparison the results for CeO2 and Ba(CeO3 ) are also shown. After reduction at 550 ◦ C (step 1) the barium promoted catalyst with the highest Ce content exhibits the lowest SH2 value. This area is over 30% smaller than the SH2 of the catalyst promoted with barium alone (Co/Ba) – Table 3. Other Co–Ce–Ba catalysts have much higher SH2 values than the one obtained for the Co/Ba catalyst. The SH2 value noted for Co/Ce(11.5)/Ba is more than twice that of Co/Ba. It is noteworthy that the contribution of cerium-containing Table 3 Hydrogen chemisorption characteristics (SH2 ) of catalysts. Catalyst symbol
Fig. 7. TPR-TG/DTG-MS studies of the cobalt catalysts. DTG curves for the following samples: Co3 O4 , Co/Ce(29.0) and Co/Ce(29.0)/Ba.
Co/Ba Co/Ce(3.7)/Ba Co/Ce(11.5)/Ba Co/Ce(17.2)/Ba Co/Ce(29.0)/Ba CeO2 a Ba(CeO3 )a a
SH2 (m2 /gCo ) After reduction at 550 ◦ C (step 1)
After overheating at 600 ◦ C (step 2)
After successive overheating at 600 ◦ C (step 3)
4.1 7.1 9.8 5.8 2.7 2.3 2.6
3.3 5.8 6.3 5.3 2.8 2.9 –
2.6 5.4 6.0 4.8 4.4 3.4 –
Values with respect to mCeO2 [g].
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286 Table 4 Activity of the promoted cobalt catalysts, T = 400 ◦ C, p = 6.3 MPa, gas (3H2 + N2 ) flow rate = 70 dm3 [STP]/h. Catalyst symbol
Co/Ba Co/Ce(3.7)/Ba Co/Ce(11.5)/Ba Co/Ce(17.2)/Ba Co/Ce(29.0)/Ba
Productivity (gNH3 /(gCo3 O4 · h)) After reduction at 550 ◦ C (step 1)
After overheating at 600 ◦ C (step 2)
After successive overheating at 600 ◦ C (step 3)
0.61 2.20 3.33 2.81 2.17
0.53 1.60 3.05 2.31 2.81
– 1.60 3.00 2.20 2.81
phases into the SH2 value does not exceed 7%, except for the case of Co/Ce(29.0)/Ba, in which it equals approximately 20%, as evidenced by combination of the Ce contents in the catalytic materials and the SH2 storage capacity (2.3–3.4 m2 /gCeO2 ) determined in a separate chemisorption experiment for ceria. For ceria the increase of SH2 as a result of the overheating at 600 ◦ C was observed, which can be due to the reduction surface of oxygen species for this material. The consequence is the increase of the amount of chemisorbed hydrogen and also in the surface area calculated on this basis. Overheating samples to 600 ◦ C causes the cobalt to sinter. This results in a decrease of SH2 values. When the samples are overheated to 600 ◦ C twice, the decrease of the total area of cobalt available for H2 ranges from ∼20 to ∼40%. A different behaviour is noted only for the system with the highest CeO2 loading, i.e. (Co/Ce(29.0)/Ba), for which no significant change in the SH2 value is observed after the first stage of heating to 600 ◦ C (step 2). However, after the second stage of overheating this sample (step 3), an increase of the SH2 value by approximately 40% is noted. The observed phenomenon is probably associated with a much slower reduction of surface Co atoms in the presence of large amounts of CeO2 . This effect had been observed during TPR-TG/DTG-MS measurements and explained above. The results of activity of cobalt catalysts in the synthesis of NH3 are shown in Table 4. These results include the values of productivity for catalysts reduced at 550 ◦ C (step 1) and then overheated to 600 ◦ C (steps 2 and 3). As seen in Table 3 the productivity values of all Co/Ce/Ba systems are significantly higher than that of Co/Ba. After reduction at 550 ◦ C the activity of the doubly promoted catalysts is about 3 to more than 5 times higher than for those with only one promoter. The most active of the prepared systems is Co/Ce(11.5)/Ba. As a result of overheating to 600 ◦ C, the activity of cobalt catalysts promoted only with barium or barium and CeO2 (CeO2 content: 3.7, 11.5 and 17.2 wt%) decreases. This decrease in both stages of overheating ranges from 10 to ∼30%. These data correlate well with the results of the TPD-H2 experiments (Table 3), for which a decrease in the surface area of cobalt (SH2 ) is observed as a result of overheating to 600 ◦ C. In contrast, for the sample with the highest content of ceria, i.e. (Co/Ce(29.0)/Ba) an increase of activity by approximately 23% is observed as a result of overheating to 600 ◦ C (after step 3). A different behaviour of this catalyst than of the other systems has already been observed during the chemisorption studies (Table 3). These results confirm that CeO2 significantly delays the reduction of the catalysts, especially in those with a high ceria loading. Therefore, the Co/Ce(29.0)/Ba system requires a much longer time and/or higher temperatures to achieve the maximum level of SH2 values and catalytic activity. The second point which is worth noting is the lack of a perfect correlation between the chemisorption and NH3 activity results. In fact, the only good agreement between these results was obtained for the Co/Ce(11.5)/Ba sample. This sample has the highest value of the SH2 and the largest productivity (see Tables 3 and 4). For the other systems the situation is more complicated. Both Co/Ba and
285
Co/Ce(3.7)/Ba have SH2 values greater than or comparable to those obtained for Co/Ce(17.2)/Ba and Co/Ce(29.0)/Ba (see Table 3). The catalytic activity is reversed, i.e. samples with a higher CeO2 content (17.2 wt% and 29.0 wt%) almost always have a higher activity. For example, the Co/Ba has a SH2 value comparable with the value obtained for Co/Ce(29.0)/Ba, whereas the productivity is almost three times lower (Tables 3 and 4). It seems that the observed dependence may be the effect of several things. It can result from a relative content and dispersion of CeO2 and Ba(CeO3 ) specimen upon the variation of ceria loading, overheating treatment, as well as the presence or absence of the hcp phase of metallic Co. According to [51,52] this latter issue is responsible for the high activity of Co in ammonia synthesis. Metallic Co is said to undergo an allotropic transformation from hexagonal close packed (hcp) to face centred cubic (fcc) at approximately 420 ◦ C [51,53]. The temperature depends on many factors, such as the content of dopants, thermal and mechanical history, i.e. whether the sample is cooled or heated, grain size, etc. It is known from literature [52,54] that cobalt-based catalysts undergo an allotropic transformation while catalysing several reactions. The transformation takes place when the temperature of the catalysed reaction and the onset temperature of the transformation are relatively close. The phase transformation of Co strongly influences the catalytic activity of cobalt in ammonia synthesis. Since hydrogen and nitrogen chemisorb differently on hcp and fcc cobalt, the reaction rates on these two allotropic forms differ. It has been found that the former exhibits a higher turnover frequency. The dependence of the reaction rate on the allotropic form means that ammonia synthesis on cobalt is a structure-sensitive reaction [51]. Another example of a reaction in which the cobalt hcp form is more reactive than the fcc form, is the catalytic methanation of carbon dioxide [54]. There is a large temperature gap in the formation and transformation of the hcp phase in cobalt-containing samples upon heating and cooling. When a sample is cooled, the fcc phase changes to the hcp phase at 380–340 ◦ C, whereas during heating of a sample the formation of this phase occurs at 460–500 ◦ C [51]. In our research, the samples were cooled from ∼600 to 400 ◦ C. In this temperature region, cobalt in the catalyst without ceria is in the form of the fcc phase. In the presence of ceria, it is in the form of the hcp phase, which is the more beneficial one. Similar results have been observed by [52]. According to the authors [52] it is possible to stabilize the hcp cobalt phase at temperatures below 600 ◦ C by promoting the cobalt catalysts with CeO2 . Ceria might play a similar role in our systems, i.e. stabilizes the hcp phase, but our XRPD studies (Fig. 4) show that in described catalysts occurs well-crystallized fcc phase after reduction. However, it should be remembered that this measurement was carried out at room temperature and in accordance with [51] a large amount of cubic phase remained at this temperature and cubic Co does not ever disappear completely even at low temperatures. It seems that further research is required, especially in situ XRPD to gain a broader view of the observed phenomenon. Based on the research carried out within this work, it can be stated that there is an optimum amount of CeO2 in a cobalt-based system, which provides a high surface area and a high catalytic activity in the synthesis of NH3 . Moreover, the most active of the studied systems, i.e. Co/Ce(11.5)/Ba, also shows high thermal stability. Keeping it for over 160 h at 600 ◦ C (3H2 :N2 = 30.0 l/h, p = 0.1 MPa) resulted in only a 10% drop in activity. The dependence of the reaction rate on the partial pressure of ammonia for the co-promoted catalyst (Co/Ce(11.5)/Ba) is illustrated in Fig. 8. For comparison, an analogous relationship derived for the fused iron catalyst (KMI, H. Topsoe) has also been shown in Fig. 8. Both catalysts were activated (reduced) in such a way that their optimum activity was obtained. In both cases (Co and Fe), the reaction rate decreases with the increase in the ammonia content
286
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
Fig. 8. Influence of ammonia partial pressure on the NH3 synthesis rates (r) over the Co/Ce(11.5)/Ba catalyst and triply promoted fused iron catalyst (KM I) at 400 ◦ C (6.3 MPa).
in the gas mixture. However, the co-promoted cobalt catalyst is more active than the commercial iron catalyst in the whole range of conversion values. Moreover, the cobalt catalyst is kinetically less sensitive to the presence of ammonia (Fig. 8): an increase in the NH3 content from 1 to 8% leads to a eleven-fold decrease of the reaction rate over Co + Ce + Ba, whereas the reaction rate over the magnetite-based KMI material decreases by a factor of more than 15. Due to the kinetic differences between Co and Fe, the differences in the reaction rates under industrial conditions (15–20% NH3 conversions) will be significantly larger. 4. Conclusions In the investigated systems cerium oxide plays the role of a structural promoter, because it prevents the Co3 O4 and Co crystallites from sintering under calcination and reaction conditions, respectively. Moreover, CeO2 can stabilize the hcp phase of metallic cobalt under the reaction conditions. This phase is considered to be more active in ammonia synthesis [51], so the catalysts promoted with ceria and barium are more active in ammonia synthesis than the contacts promoted only with barium. The optimum content of ceria is between approximately 10 and 13 wt%. The comparison of the most active catalytic system – contact Co/Ce(11.5)/Ba(13), with the industrial iron catalyst KM I has shown that cobalt systems are more active than the conventional iron catalyst in the entire investigated range of ammonia concentrations. The superior catalytic activity of the cobalt catalyst is even more pronounced with the increase of ammonia content in the gas mixture. Furthermore, Co/Ce(11.5)/Ba is very resistant to overheating. The heat treatment at 600 ◦ C for 160 h (3H2 :N2 = 30.0 l/h, p = 0.1 MPa) results in only a slight decrease of activity in NH3 synthesis. Acknowledgement This work has been carried out within the Research Project N N209119737, sponsored by the Ministry of Science and Higher Education. References [1] U.S. Geological Survey, http://minerals.usgs.gov/minerals/pubs/commodity/ nitrogen, data from 31.03.08.
[2] R.Z. Sørensen, A. Klerke, U. Quaade, S. Jensen, O. Hansen, C.J.H. Christensen, Catal. Lett. 112 (2006) 77–81. [3] P.F. Ng, L. Li, S.B. Wang, Z.H. Zhu, G.Q. Lu, Z.F. Yan, Environ. Sci. Technol. 41 (2007) 3758–3762. [4] L. Li, Z.H. Zhu, Z.F. Yan, G.Q. Lu, L. Rintoul, Appl. Catal. A 320 (2007) 166–172. [5] L. Li, Z.H. Zhu, G.Q. Lu, Z.F. Yan, S.Z. Qiao, Carbon 45 (2007) 11–20. [6] S.F. Yin, B.Q. Xu, S.J. Wang, C.T. Au, Appl. Catal. A 301 (2006) 202–210. [7] X.K. Li, W.J. Ji, J. Zhao, S.J. Wang, C.T. Au, J. Catal. 236 (2005) 181–189. [8] http://nobelprize.org/nobel prizes/chemistry/laureates/index.html, data from 26.10.2007. [9] K. Tamaru, in: J.R. Jennings (Ed.), Catalytic Ammonia Synthesis: Fundamentals and Practice, Plenum Press, New York, 1991, pp. 1–18. [10] C.J.H. Jacobsen, Chem. Commun. 12 (2000) 1057–1058. [11] R. Kojima, K. Aika, Appl. Catal. A 215 (2001) 149–160. [12] R. Kojima, K. Aika, Appl. Catal. A 218 (2001) 121–128. [13] R. Kojima, K. Aika, Appl. Catal. A 219 (2001) 157–170. [14] S. Hagen, R. Barfod, R. Fehrmann, C.J.H. Jacobsen, H.T. Teunissen, I. Chorkendorff, J. Catal. 214 (2003) 327–335. [15] S. Hagen, R. Barfod, R. Fehrmann, C.J.H. Jacobsen, H.T. Teunissen, K. Stahl, I. Chorkendorff, Chem. Commun. 11 (2002) 1206–1207. ´ ˙ [16] A. Jedynak, D. Szmigiel, W. Raróg, J. Zielinski, J. Pielaszek, P. Dłuzewski, Z. Kowalczyk, Cattal. Lett. 81 (2002) 213–218. ´ Pol. J. Chem. 75 (2001) [17] A. Jedynak, Z. Kowalczyk, D. Szmigiel, J. Zielinski, 1801–1805. [18] R. Kojima, H. Enomoto, M. Muhler, K. Aika, Appl. Catal. A 246 (2003) 311–322. [19] R. Schlögl, Angew. Chem. Int. Ed. 42 (2003) 2004–2008. ´ [20] I. Jasinska, K. Lubkowski, W. Arabczyk, Przem. Chem. 82 (2003) 230–233. ´ S. Lenart, Appl. Catal. A 266 [21] M.J. Figurski, W. Arabczyk, Z. Lendzion-Bielun, (2004) 11–20. ´ R.J. Kalenczuk, ´ S. Lenart, Appl. [22] M.J. Figurski, W. Arabczyk, Z. Lendzion-Bielun, Catal. A 247 (2003) 9–15. [23] N. Pernicone, E. Ferrero, I. Rossetti, L. Forni, P. Canton, P. Riello, G. Fagherazzi, M. Signoretto, F. Pinna, Appl. Catal. A 251 (2003) 121–129. [24] Y.F. Zheng, H.Z. Liu, Z.J. Liu, X.N. Li, J. Solid State Chem. 182 (2009) 2385– 2391. [25] H.Z. Liu, X.N. Li, Z.N. Hu, Appl. Catal. A 142 (1996) 209–222. [26] R.B. Strait, Nitrogen Methanol 238 (1999) 37–43. [27] R.B. Strait, International Conference Exhibition, Caracas, 28 February–2 March, 1999, pp. 43–59. [28] Z. Kowalczyk, W. Raróg-Pilecka, Przem. Chem 85/2 (2006) 124–129. ´ J. Pielaszek, Appl. Catal. A 173 [29] Z. Kowalczyk, S. Jodzis, W. Raróg, J. Zielinski, (1998) 153–160. [30] I. Rossetti, N. Pernicone, L. Forni, Appl. Catal. A 208 (2001) 271–278. [31] I. Rossetti, N. Pernicone, L. Forni, Catal. Today 102-103 (2005) 219–224. [32] C.J.H. Jacobsen, S. Dahl, B.S. Clausen, S. Bahn, A. Logadottir, J.K. Norskov, J. Am. Chem. Soc. 123 (2001) 8404–8405. [33] O.V. Shcheglov, G.K. Boreskov, M.G. Slinko, Dokl. Akad. Nauk SSSR 105 (1955) 123–133. ´ [34] R.J. Kalenczuk, Appl. Catal. A 112 (1994) 149–160. ˛ nski, ´ Z. Kaszkur, K. Kielar, Z. Kowalczyk, [35] W. Raróg-Pilecka, E. Mi´skiewicz, L. Kepi J. Catal. 249 (2007) 24–33. [36] W. Raróg-Pilecka, E. Mi´skiewicz, Z. Kowalczyk, Catal. Commun. 9 (2008) 870–873. [37] W. Raróg-Pilecka, M. Karolewska, E. Truszkiewicz, E. Iwanek, B. Mierzwa, Catal. Lett. 141 (2011) 678–684. [38] M. Karolewska, O. Nalewajko, E. Truszkiewicz, W. Raróg-Pilecka, Przem. Chem. 91/6 (2012) 1224–1228. [39] W. Raróg-Pilecka, E. Mi´skiewicz, S. Jodzis, J. Petryk, D. Łomot, Z. Kaszkur, Z. ´ Z. Kowalczyk, J. Catal. 239 (2006) 313–325. Karpinski, [40] R.C. Reuel, C.H. Bartholomew, J. Catal. 85 (1984) 63–77. ´ M. Bonarowska, Langmuir 13 (1997) 5613–5620. [41] A. Borodzinski, [42] Z. Kowalczyk, Catal. Lett. 37 (1996) 173–179. [43] E.P. Barrett, L.G. Joyner, P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380. [44] L. Xue, Ch Zhang, H. He, Y. Teraoka, Appl. Catal. B 75 (2007) 167–174. [45] H. Wang, J.L. Ye, Y. Liu, Y.D. Li, Y.N. Qin, Catal. Today 129 (2007) 305–312. [46] H.C. Yao, Y.F.Y. Yao, J. Catal. 86 (1984) 254–265. [47] M. Kang, M.W. Song, Ch H. Lee, Appl. Catal. A 251 (2003) 143–156. [48] B. Ernst, L. Hilaire, A. Kiennemann, Catal. Today 50 (1999) 413–427. [49] L.F. Liotta, G. Di Carlo, G. Pantaleo, A.M. Venezia, G. Deganello, Appl. Catal. B 66 (2006) 217–227. [50] J.Y. Luo, M. Meng, X. Li, X.G. Li, Y.Q. Zha, T.D. Hu, Y.N. Xie, J. Zhang, J. Catal. 254 (2008) 310–324. [51] G. Rambeau, A. Jorti, H. Amariglio, J. Catal. 94 (1985) 155–165. [52] S.S.Y. Lin, D.H. Kim, S.Y. Ha, Appl. Catal. A 355 (2009) 69–77. [53] W.O. Soboyejo, T.S. Srivatsan, Advanced Structural Materials: Properties, Design, Optimization and Applications, CRC Press, Boca Raton, FL, 2007. [54] A. Amariglio, M. Eddouasse, H. Amariglio, Surf. Sci. 162 (1985) 375–381.
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Ammonia synthesis over cobalt catalysts doped with cerium and barium. Effect of the ceria loading c ˙ ˛ nski ´ Magdalena Karolewska a , Elzbieta Truszkiewicz a , Bogusław Mierzwa b , Leszek Kepi , a,∗ Wioletta Raróg-Pilecka a
Warsaw University of Technology, Faculty of Chemistry, Noakowskiego 3, 00-664 Warsaw, Poland Institute of Physical Chemistry of the Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland c Institute of Low Temperature and Structure Research of the Polish Academy of Sciences, Okólna 2, 50-950 Wrocław, Poland b
a r t i c l e
i n f o
Article history: Received 28 May 2012 Received in revised form 6 August 2012 Accepted 25 August 2012 Available online 31 August 2012 Keywords: Ammonia synthesis Cobalt catalyst Barium and cerium promoters Ceria loading
a b s t r a c t The effect of ceria loading on the activity and thermal stability of a cobalt catalyst for ammonia synthesis has been studied. The concentration of CeO2 in the Co3 O4 + CeO2 mixtures ranged from 0 to 29 wt%, whereas the barium content was constant and equaled 13 wt% in relation to Co3 O4 . Kinetic measurements in NH3 synthesis were carried out in a flow differential reactor operating under standard conditions (p = 6.3 MPa, T = 400 ◦ C, H2 :N2 = 3:1). The kinetic measurements of NH3 synthesis were supplemented with the characterization studies of the obtained materials – N2 physisorption, XRPD, H2 chemisorption, TPR-TG/DTG-MS and SEM-EDS. It has been found that ceria plays the role of a structural promoter; it hinders the sintering of Co3 O4 during calcination and stabilizes the surface area of Co under reaction conditions. Moreover, it seems that ceria stabilizes the hcp phase of metallic cobalt, which results in a higher activity of the cobalt catalysts. The co-promoted cobalt catalyst containing 11.5 wt% of CeO2 (symbol Co/Ce(11.5)/Ba) was shown to be more active in NH3 synthesis than the conventional fused iron catalyst (KM I, H. Topsoe). Moreover, the catalyst Co/Ce(11.5)/Ba is very resistant to overheating. Heat treatment at 600 ◦ C for 160 h (3H2 :N2 = 30.0 l/h, p = 0.1 MPa) results in a decrease of its activity in NH3 synthesis by only a few percent. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Ammonia synthesis is one of the most important and the largest, in terms of the production scale, catalytic processes in the chemical industry today. The global production of NH3 was estimated to be 124 million tons in 2006 and continues to exhibit a steady increase [1]. Ammonia is used in the production of fertilizers, explosive materials, and as a refrigerant. Since the hydrogen used in fuel cells has to be pure, i.e. not contaminated with carbon oxides, ammonia is thought to become the best source of hydrogen suitable for this application in the near future [2–7]. The catalytic reaction of hydrogen and nitrogen, which leads to the formation of ammonia, is of significant importance in the development of catalysis. Two Nobel Prizes have been awarded for research on this topic [8]. Moreover, studies on ammonia synthesis have led to many significant improvements and innovations. The basic terms used in catalysis, such as promoter or poison, were first used with regard to ammonia synthesis. Furthermore, new catalytic concepts, e.g. structure
∗ Corresponding author. Tel.: +48 22 234 57 66; fax: +48 22628 27 41. E-mail address: [email protected] (W. Raróg-Pilecka). 0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apcata.2012.08.028
sensitivity of a reaction, have been developed during studies of ammonia synthesis catalysts [9]. For the above reasons, the researchers from industrial centres and academia are still active in the field of NH3 synthesis and they try to work out a completely new catalyst [10–18] or to improve the formula and properties of the conventional iron one [19–25]. Among several new catalytic systems investigated in the last 40 years, only ruthenium supported on high surface area graphite (HSAG) has been implemented in the industrial practice so far. A combination of both the conventional iron (first catalytic bed) and the modern Ru/C catalyst (last three beds) in the so called Kellog Advanced Ammonia Process (KAAP) has been shown to be very advantageous [26,27], i.e. the pressure in ammonia loops could be significantly reduced (to 9.0 MPa), thus resulting in a lower energy consumption. However, the high price of ruthenium and its tendency to catalyzes the depletion of the support, i.e. methanation [28–31], are the main reasons why the ruthenium catalyst did not revolutionize the process of NH3 synthesis. Research on the development of an entirely new catalytic system with other metals which exhibit catalytic activity in this reaction [32] is still ongoing. Cobalt-based catalysts are among the most frequently investigated systems [15,33–35]. The latest research of Raróg-Pilecka et al. has shown
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
that unsupported cobalt catalysts promoted with cerium and barium (Co/Ce/Ba) obtained by the templating method are more active than the conventional iron catalyst [36]. However, due to a rather difficult technique of catalyst preparation, further work of this team was aimed at obtaining promoted cobalt catalysts by a simpler method, namely coprecipitation [37]. The catalysts obtained by the latter method were even more active in ammonia synthesis. However, research on both types of catalysts shows that the essential components of an active cobalt catalyst are the promoters: barium and cerium. Their synergistic effect is evident in both works [36,37]. This paper also investigates the catalytic properties of Co/Ce/Ba in NH3 synthesis. However, this work focuses on the effect of ceria loading on the activity of the cobalt systems. The kinetic measurements of NH3 synthesis were supplemented with the characterisation studies of the obtained materials – N2 physisorption, XRPD, H2 chemisorption, TPR-TG/DTG-MS and SEMEDS. Moreover, thermal stability tests of the most active contact have been performed. 2. Experimental 2.1. Catalyst preparation The mixtures of cobalt and cerium carbonates with different Co/Ce ratios were prepared by co-precipitation with potassium carbonate as the precipitant. Our recent studies [38] have shown that the application of carbonates, in particular potassium carbonate, in the role of the precipitant, leads to obtaining materials with a higher surface area and more favourable catalytic properties. Appropriate amounts of a mixture of Co(NO3 )2 ·6H2 O and Ce(NO3 )3 ·6H2 O, both of analytical grade, were dissolved in distilled water and warmed to approximately 90 ◦ C. Then, a warm (90 ◦ C) solution of K2 CO3 (analytical purity) was slowly added. The obtained precipitate, i.e. mixtures of cobalt carbonate and cerium carbonate, was filtered under reduced pressure and washed with cold distilled water several times until the pH was ∼7. The materials were then dried and calcined at 500 ◦ C overnight. The resulting black powder was a mixture of Co3 O4 and CeO2 . Thus obtained materials were impregnated using an aqueous solution of barium nitrite (analytical purity). The resulting precursors were then dried at 120 ◦ C overnight. The amount of barium was constant and equal to 13 wt% in relation to Co3 O4 . The final step of preparation was crushing and sieving the solid materials in order to get the 0.2–0.63 mm fraction. The ceria loading of the resulting materials is presented in Table 1. This table also contains data for cobalt oxide promoted with barium and pure CeO2 , also obtained by precipitation and a subsequent calcination. Additional XRF measurements have shown that the potassium concentration in the co-precipitated samples is less than 0.05%, other impurities were lower than the detection limit.
281
2.2. Characterization studies The precursors of the catalysts, that is the oxidized forms, and the reduced samples were characterized by nitrogen physisorption (ASAP 2020, Micromeritics), XRPD and SEM-EDS studies. The reduced samples were also characterized by hydrogen chemisorption (TPD-H2 ). Moreover, TPR experiments were performed for all materials using a thermobalance equipped with a quadrupole mass spectrometer (TPR-TG/DTG-MS). X-ray powder diffraction (XRPD) data were collected with a Siemens D 5000 diffractometer in a Bragg–Brentano configuration using a Cu-sealed tube operating at 40 kV and 40 mA with a Ni filter. Measurements of each system were performed in the scattering angle range of 5–100◦ with a 0.02◦ step and a counting speed of 1.0 s/step. The average size of cobalt or cobalt oxide crystallites was determined from the Scherrer equation using the integral width of reflections fitted to the analytical Pearson VII functions. The morphology and homogeneity of the samples was determined with a Scanning Electron Microscope (FEI NovaNanoSEM 230) equipped with an EDX spectrometer (EDAX Genesis XM4). The H2 chemisorption measurements were carried out in a flow set-up (a fully automated PEAK 4 instrument) equipped with a TCD cell and supplied with high purity (≥99.9999 vol%) gases [39]. The details of the experimental procedure were described previously [35]. The sample (0.4 gCo3 O4 ) was reduced in a flowing H2 :Ar = 80:20 mixture (40 ml/min) at 550 ◦ C for 20 h, flushed with argon (40 ml/min) at that temperature, and cooled to 150 ◦ C in argon. Then, the sample was cooled in H2 (40 ml/min) to 0 ◦ C and the sample was flushed with argon (40 ml/min, 0 ◦ C, 15 min) to remove weakly adsorbed hydrogen. Next, the catalyst was heated in flowing Ar with a 15 ◦ C/min temperature ramp – step 1 (TPD-H2 ). The concentration of H2 in the outlet gas was monitored. Consequently, the amount of hydrogen evolved to the gas phase was determined by integrating the H2 trace. The hydrogen TPD data were used for calculating the Co surface area (SH2 ). The H:Cos = 1:1 stoichiometry [40] and the formula proposed by ´ Borodzinski and Bonarowska [41] were used for calculations. In addition, the samples were overheated twice (steps 2 and 3) in a mixture of H2 :Ar = 80:20 (40 ml/min) at 600 ◦ C for another 20 h. After each stage of overheating, hydrogen chemisorption measurements were performed according to the procedure described above. Hence the influence of overheating on the size of the total area of cobalt available for hydrogen was determined. The TPR-TG/DTG-MS measurements were performed using a NETZSCH STA 449C thermobalance equipped with a quadrupole mass spectrometer (NETZSCH QMS 403C). The experiments were conducted using samples of approximately 30 mg of catalyst powder heated up to 550 ◦ C at the constant rate of 10 ◦ C/min in a pure (≥99.999 vol%) H2 :Ar = 10:90 mixture (100 ml/min). This temperature was maintained for 3 h. The reference crucible was empty. The mass change, temperature and selected m/z signals were monitored throughout the entire experiment. In order to avoid water
Table 1 Characteristics of catalyst precursors. Sample symbol
Ceria loading (wt%)
N2 physisorption SP
Co3 O4 Co/Ce(3.7) Co/Ce(11.5) Co/Ce(17.2) Co/Ce(29.0) CeO2 b a b
0 3.7 11.5 17.2 29.0 100
a
2
(m /gCo3 O4 )
40 69 94 91 109 32
SP , VP – BET surface area and pore volume of materials in oxide form, respectively. Values with respect to mCeO2 [g].
XRPD VP
a
0.16 0.29 0.42 0.35 0.45 0.11
3
(cm /gCo3 O4 )
dCo3 O4 (nm) 20 12 10 9 8 –
SCo3 O4 (m2 /gCo3 O4 ) 49 82 98 109 123 –
282
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
physisorption all the necessary apparatus parts were kept heated to 280 ◦ C. 2.3. Activity measurements The activity of the catalysts in ammonia synthesis was measured in a flow tubular reactor supplied with a very pure (≥99.99995 vol%) H2 :N2 = 3:1 mixture of controlled ammonia content (x1 ; 0–8%) [42]. The activities of various samples were tested under steady-state conditions (6.3 MPa, 400 ◦ C, gas flow rate = 70.0 l/h) and for a fixed catalyst mass (0.4 gCo3 O4 ). The concentration of ammonia in the outlet gas (x2 ) was monitored and the productivity – average reaction rate, was calculated (integral measurement; x1 = 0). Prior to kinetic tests, the samples were reduced in a H2 :N2 stream at 470 ◦ C for 45 h and then at 550 ◦ C for 40 h. The subsequent experimental procedure involved the following steps: determination of the initial reaction rate (step 1), i.e. the rate after stabilization at 470 ◦ C (40 h) + 550 ◦ C (45 h); short-term (24 h) overheating in a 3H2 + N2 mixture at 600 ◦ C, and another determination of the reaction rate under the same conditions as in the initial rate measurement (step 2). Additionally, another step (step 3) was performed, which was analogous to step 2. Both, the activation (reduction) and overheating were performed under atmospheric pressure and gas flow rate = 30.0 l/h. In a more complex experiment, both outlet (x2 ) and inlet (x1 ) streams were analysed (differential measurement; x2 − x1 x1 ) to determine the reaction rate corresponding to the mean ammonia concentration (x = (x1 + x2 )/2) in a catalyst layer; the details of such examinations can be found elsewhere [42]. The relationship between the NH3 synthesis rate and conversion was established for the best catalyst according to this procedure. 3. Results and discussion Fig. 1 presents a sequence of the diffraction patterns of the catalyst precursors (Co3 O4 + CeO2 ) which additionally contain barium. As seen, all patterns display a set of maxima characteristic for Co3 O4 (JCPDS 42-1467) and barium carbonate (JCPDS 05-0378) phases. Moreover, the diffraction patterns of samples Co/Ce(17.2)/Ba and Co/Ce(29.0)/Ba exhibit signals typical for CeO2 (JCPDS 34-0394). In the case of the other three systems (that is Co/Ba, Co/Ce(3.7)/Ba and Co/Ce(11.5)/Ba) ceria does not have a noticeable contribution to the XRD patterns. This means that ceria species are finely dispersed on the cobalt oxide surface, or they are poorly organized, nearly amorphous.
Fig. 1. XRPD patterns of the fresh catalysts (prior to reduction).
Fig. 2. Pore size distribution curves of the catalyst precursors.
In Table 1 and Fig. 2 N2 physisorption data of the samples of mixed cobalt–cerium oxides (precursors of the catalysts) are presented. In the case of both Table 1 and Fig. 1 the data for pure Co3 O4 and CeO2 , obtained by precipitation and subsequent calcination, are also shown. As can be seen (Table 1) pure oxides exhibit the worst textural parameters, i.e. they have the lowest SP and VP values. For mixed oxides, these parameters are higher. The highest values of the total surface area and pore volume were obtained for the sample Co/Ce(11.5). Moreover, the area of this sample is more than twice as high as the sum of the components (i.e. Co3 O4 and CeO2 ), calculated with accordance to their content in the double oxide system. The distribution of the pores of the samples in the oxide forms was determined from the nitrogen adsorption isotherm at 77 K using the Barrett–Joyner–Halenda model [43]. The spectrum of CeO2 has one maximum at the average pore width of ∼40 nm, whereas the spectrum of Co3 O4 has two characteristic peaks at ∼40 and 90 nm. Samples containing both oxides (Co3 O4 and CeO2 ) have a pore distribution similar to that of pure cobalt oxide. However, the total volume of the pores is much higher than that of the Co3 O4 sample. The results of the physisorption studies show that addition of cerium prevents the sintering of these materials during calcination. Moreover, the larger the cerium content, the higher the value of the overall surface area (SP ). It should be emphasized that the contribution of a separate Ce-containing phase (SBET = 32 m2 /g) to the overall surface area of the Co + Ce materials would be lower than 9%, even if the entire content of cerium in the catalyst was taken into account in the calculations. This indicates the structural nature of this promoter in cerium–cobalt systems. The physisorption data (SP ) are in good agreement with the XRPD results (SCo3 O4 – Table 1), for which the increase of the Co3 O4 surface area is also observed as a result of CeO2 addition. Nevertheless, the SEM studies for these materials do not show visible differences between them. The sample images for Co/Ba and Co/Ce(17.5)/Ba catalyst precursors are shown in Fig. 3(a) and (c). The XRPD results for the promoted catalysts after activity measurements are presented in Fig. 4. In all of these patterns metallic cobalt is the only visible cobalt-containing phase. As expected, Co3 O4 is reduced to metallic Co (JCPDS 15-0806 and 05-0727) as a result of activation. Signals coming from Co correspond to both the fcc phase (44.4◦ , 51.6◦ , 75.9◦ and 92.3◦ ) and the hcp phase (e.g. 41.7◦ , 47.6◦ ). However, the asymmetry and broadening of the signal at the angle of 47.6◦ and the absence of signals at angles 84.2◦ and 94.7◦ indicate strong disorder of the hcp phase in the [1 0 1] direction. In contrast, the cubic phase is well crystallized. Promoters i.e. barium and cerium are present in the diffraction patterns in the form of Ba(CeO3 ) (JCPDS 82-2425). The pattern obtained for
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
283
Fig. 3. SEM micrographs of the catalyst precursors (a and c) and the post-reaction catalysts after exposure to air (b and d); (a, b) Co/Ba sample and (c, d) Co/Ce(17.2)/Ba sample.
the sample with the highest concentration of ceria (Co/Ce(29.0)/Ba) contains additional signals from cerium oxide. Based on SEM-EDS studies it was found that Co/Ba is not homogeneous. Barium creates separate particles mixed with cobalt. In contrast, all samples containing both barium and cerium exhibit high homogeneity. The EDS analysis of several points of the analysed area shows that all major elements of all samples, i.e. Co, Ba and Ce, are evenly distributed with a Ba/Ce atomic ratio close to 1. Moreover, a difference between the sizes of cobalt particles in
Co/Ba and Co/Ce/Ba samples was observed, which is clearly visible on presented images in Fig. 3(b) and (d). The cobalt particles in Co/Ba are bigger than those in Co/Ce(17.2)/Ba. This indicates that the addition of Ce inhibited the growth of Co particles. The TPR-TG/DTG-MS data obtained for the all catalyst precursors are depicted in Figs. 5 and 6. The first derivative of the mass change (DTG curve) observed when a sample is heated in a flowing mixture of H2 /Ar is depicted in Fig. 5. The other figure shows the changes in the amount of evolved water vapour, which is the main reaction product, in time. As described previously [37,44,45]
Fig. 4. XRPD patterns of promoted cobalt catalysts used in ammonia synthesis.
Fig. 5. TPR-TG/DTG-MS studies of the cobalt catalysts. DTG curves of samples.
284
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286 Table 2 XRPD and N2 physisorption characteristics of catalysts used in ammonia synthesis. Catalyst symbol
XRPD
Co/Ba Co/Ce(3.7)/Ba Co/Ce(11.5)/Ba Co/Ce(17.2)/Ba Co/Ce(29.0)/Ba a
Fig. 6. TPR-TG/DTG-MS studies of the cobalt catalysts. Water vapor evolution signals (MSH2 O ) of samples.
during the reduction of pure Co3 O4 two clear signals are observed on each curve – see Figs. 5 and 6. They correspond to the reduction of Co3 O4 to CoO (the first signal at the lower temperature), and a subsequent reduction of CoO to Co (the second signal at the higher temperature). The addition of CeO2 to Co3 O4 causes a shift of the reduction signals towards higher temperatures (Figs. 5 and 6). As a result the process of complete reduction of Co3 O4 → Co in samples with a CeO2 content above 10 wt% ends in the isothermal segment of the experiment. Only the sample with a low CeO2 content behaves similarly to pure Co3 O4 . It should also be noted that under the conditions of the TPR-TG/DTG-MS measurements, the mass of CeO2 decreases only very slightly in the temperature range: 400–550 ◦ C, which, according to literature [46], should be attributed to the reduction of surface oxygen species. The introduction of barium to the Co + Ce systems neutralizes the negative impact of ceria on the reduction of Co3 O4 , as shown in Fig. 7 for the sample richest in cerium, i.e. Co/Ce(29.0)/Ba. For comparison, Fig. 7 shows the DTG curves for the corresponding measurements for pure Co3 O4 and Co3 O4 doped only with CeO2 , i.e. Co/Ce(29.0). In our studies a slight mass loss was also observed at a temperature close to 100 ◦ C (Fig. 5). This effect is associated with the emissions of water (see Fig. 6), which is either adsorbed on the surface of these materials or derives from the initial Co3 O4 reduction. In conclusion, the TPR-TG/DTG-MS studies revealed that ceria delays the reduction of Co3 O4 to metallic Co, probably because it consumes H2 [44,47–50]. Ceria is reduced more easily, that is at lower temperatures, as a result of hydrogen spillover from the transition metal (Co) to ceria. Hence, ceria uptakes additional hydrogen. This is why more hydrogen is consumed and a longer reduction time is required to fully reduce the cobalt oxide in ceria-containing
N2 physisorption
dCo (nm)
SCo (m2 /gCo )
SR
20 25 22 20 20
33.9 27.1 30.8 33.9 33.9
9.8 25.5 23.0 21.2 11.5
a
(m2 /gCo )
BET surface area of the catalysts.
samples [48–50]. In contrast, barium neutralizes the negative impact of CeO2 . The effect of Ba may result from interactions of the two promoters and the formation of a common phase, i.e. Ba(CeO3 ) (see Fig. 4). However, further studies are needed to resolve this issue. Data for catalysts unloaded from the NH3 reactor are presented in Table 2. The values concern the results of calculations of the surface area of cobalt based on the XRPD measurements (XRPD column) and BET surface area (N2 physisorption column). The physisorption data shows that samples containing cerium and barium besides cobalt have higher specific surface areas than the sample containing barium only. However, Co/Ce(3.7)/Ba, Co/Ce(11.5)/Ba and Co/Ce(17.2)/Ba catalysts exhibit the highest SR values. In the case of the XRPD results (SCo ) no meaningful differences between samples are detected. Moreover, the surface areas of cobalt estimated from the XRPD are higher than those obtained from N2 physisorption measurements. This may be due to the fact that XRPD estimates average crystallite diameters, whereas nitrogen sorption occurs on the whole available surface. If Co crystallites create bigger clusters discrepancies between the obtained values are observed. It should be also remembered that both types of measurements were performed ex situ, that is after exposure of samples to air. Hence, the surface of the metallic phase was oxidized before the measurements and as a consequence the obtained values can be overestimated. This has already been observed in previous studies [36]. According to our prior research [37] reliable values of the surface of the active phase can be obtained with hydrogen chemisorption. Table 3 presents chemisorption data (SH2 ), i.e. the surface of cobalt available for H2 determined from TPD-H2 measurements, of the cobalt catalyst promoted only with barium and for a series of cobalt catalysts promoted with both ceria and barium. For comparison the results for CeO2 and Ba(CeO3 ) are also shown. After reduction at 550 ◦ C (step 1) the barium promoted catalyst with the highest Ce content exhibits the lowest SH2 value. This area is over 30% smaller than the SH2 of the catalyst promoted with barium alone (Co/Ba) – Table 3. Other Co–Ce–Ba catalysts have much higher SH2 values than the one obtained for the Co/Ba catalyst. The SH2 value noted for Co/Ce(11.5)/Ba is more than twice that of Co/Ba. It is noteworthy that the contribution of cerium-containing Table 3 Hydrogen chemisorption characteristics (SH2 ) of catalysts. Catalyst symbol
Fig. 7. TPR-TG/DTG-MS studies of the cobalt catalysts. DTG curves for the following samples: Co3 O4 , Co/Ce(29.0) and Co/Ce(29.0)/Ba.
Co/Ba Co/Ce(3.7)/Ba Co/Ce(11.5)/Ba Co/Ce(17.2)/Ba Co/Ce(29.0)/Ba CeO2 a Ba(CeO3 )a a
SH2 (m2 /gCo ) After reduction at 550 ◦ C (step 1)
After overheating at 600 ◦ C (step 2)
After successive overheating at 600 ◦ C (step 3)
4.1 7.1 9.8 5.8 2.7 2.3 2.6
3.3 5.8 6.3 5.3 2.8 2.9 –
2.6 5.4 6.0 4.8 4.4 3.4 –
Values with respect to mCeO2 [g].
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286 Table 4 Activity of the promoted cobalt catalysts, T = 400 ◦ C, p = 6.3 MPa, gas (3H2 + N2 ) flow rate = 70 dm3 [STP]/h. Catalyst symbol
Co/Ba Co/Ce(3.7)/Ba Co/Ce(11.5)/Ba Co/Ce(17.2)/Ba Co/Ce(29.0)/Ba
Productivity (gNH3 /(gCo3 O4 · h)) After reduction at 550 ◦ C (step 1)
After overheating at 600 ◦ C (step 2)
After successive overheating at 600 ◦ C (step 3)
0.61 2.20 3.33 2.81 2.17
0.53 1.60 3.05 2.31 2.81
– 1.60 3.00 2.20 2.81
phases into the SH2 value does not exceed 7%, except for the case of Co/Ce(29.0)/Ba, in which it equals approximately 20%, as evidenced by combination of the Ce contents in the catalytic materials and the SH2 storage capacity (2.3–3.4 m2 /gCeO2 ) determined in a separate chemisorption experiment for ceria. For ceria the increase of SH2 as a result of the overheating at 600 ◦ C was observed, which can be due to the reduction surface of oxygen species for this material. The consequence is the increase of the amount of chemisorbed hydrogen and also in the surface area calculated on this basis. Overheating samples to 600 ◦ C causes the cobalt to sinter. This results in a decrease of SH2 values. When the samples are overheated to 600 ◦ C twice, the decrease of the total area of cobalt available for H2 ranges from ∼20 to ∼40%. A different behaviour is noted only for the system with the highest CeO2 loading, i.e. (Co/Ce(29.0)/Ba), for which no significant change in the SH2 value is observed after the first stage of heating to 600 ◦ C (step 2). However, after the second stage of overheating this sample (step 3), an increase of the SH2 value by approximately 40% is noted. The observed phenomenon is probably associated with a much slower reduction of surface Co atoms in the presence of large amounts of CeO2 . This effect had been observed during TPR-TG/DTG-MS measurements and explained above. The results of activity of cobalt catalysts in the synthesis of NH3 are shown in Table 4. These results include the values of productivity for catalysts reduced at 550 ◦ C (step 1) and then overheated to 600 ◦ C (steps 2 and 3). As seen in Table 3 the productivity values of all Co/Ce/Ba systems are significantly higher than that of Co/Ba. After reduction at 550 ◦ C the activity of the doubly promoted catalysts is about 3 to more than 5 times higher than for those with only one promoter. The most active of the prepared systems is Co/Ce(11.5)/Ba. As a result of overheating to 600 ◦ C, the activity of cobalt catalysts promoted only with barium or barium and CeO2 (CeO2 content: 3.7, 11.5 and 17.2 wt%) decreases. This decrease in both stages of overheating ranges from 10 to ∼30%. These data correlate well with the results of the TPD-H2 experiments (Table 3), for which a decrease in the surface area of cobalt (SH2 ) is observed as a result of overheating to 600 ◦ C. In contrast, for the sample with the highest content of ceria, i.e. (Co/Ce(29.0)/Ba) an increase of activity by approximately 23% is observed as a result of overheating to 600 ◦ C (after step 3). A different behaviour of this catalyst than of the other systems has already been observed during the chemisorption studies (Table 3). These results confirm that CeO2 significantly delays the reduction of the catalysts, especially in those with a high ceria loading. Therefore, the Co/Ce(29.0)/Ba system requires a much longer time and/or higher temperatures to achieve the maximum level of SH2 values and catalytic activity. The second point which is worth noting is the lack of a perfect correlation between the chemisorption and NH3 activity results. In fact, the only good agreement between these results was obtained for the Co/Ce(11.5)/Ba sample. This sample has the highest value of the SH2 and the largest productivity (see Tables 3 and 4). For the other systems the situation is more complicated. Both Co/Ba and
285
Co/Ce(3.7)/Ba have SH2 values greater than or comparable to those obtained for Co/Ce(17.2)/Ba and Co/Ce(29.0)/Ba (see Table 3). The catalytic activity is reversed, i.e. samples with a higher CeO2 content (17.2 wt% and 29.0 wt%) almost always have a higher activity. For example, the Co/Ba has a SH2 value comparable with the value obtained for Co/Ce(29.0)/Ba, whereas the productivity is almost three times lower (Tables 3 and 4). It seems that the observed dependence may be the effect of several things. It can result from a relative content and dispersion of CeO2 and Ba(CeO3 ) specimen upon the variation of ceria loading, overheating treatment, as well as the presence or absence of the hcp phase of metallic Co. According to [51,52] this latter issue is responsible for the high activity of Co in ammonia synthesis. Metallic Co is said to undergo an allotropic transformation from hexagonal close packed (hcp) to face centred cubic (fcc) at approximately 420 ◦ C [51,53]. The temperature depends on many factors, such as the content of dopants, thermal and mechanical history, i.e. whether the sample is cooled or heated, grain size, etc. It is known from literature [52,54] that cobalt-based catalysts undergo an allotropic transformation while catalysing several reactions. The transformation takes place when the temperature of the catalysed reaction and the onset temperature of the transformation are relatively close. The phase transformation of Co strongly influences the catalytic activity of cobalt in ammonia synthesis. Since hydrogen and nitrogen chemisorb differently on hcp and fcc cobalt, the reaction rates on these two allotropic forms differ. It has been found that the former exhibits a higher turnover frequency. The dependence of the reaction rate on the allotropic form means that ammonia synthesis on cobalt is a structure-sensitive reaction [51]. Another example of a reaction in which the cobalt hcp form is more reactive than the fcc form, is the catalytic methanation of carbon dioxide [54]. There is a large temperature gap in the formation and transformation of the hcp phase in cobalt-containing samples upon heating and cooling. When a sample is cooled, the fcc phase changes to the hcp phase at 380–340 ◦ C, whereas during heating of a sample the formation of this phase occurs at 460–500 ◦ C [51]. In our research, the samples were cooled from ∼600 to 400 ◦ C. In this temperature region, cobalt in the catalyst without ceria is in the form of the fcc phase. In the presence of ceria, it is in the form of the hcp phase, which is the more beneficial one. Similar results have been observed by [52]. According to the authors [52] it is possible to stabilize the hcp cobalt phase at temperatures below 600 ◦ C by promoting the cobalt catalysts with CeO2 . Ceria might play a similar role in our systems, i.e. stabilizes the hcp phase, but our XRPD studies (Fig. 4) show that in described catalysts occurs well-crystallized fcc phase after reduction. However, it should be remembered that this measurement was carried out at room temperature and in accordance with [51] a large amount of cubic phase remained at this temperature and cubic Co does not ever disappear completely even at low temperatures. It seems that further research is required, especially in situ XRPD to gain a broader view of the observed phenomenon. Based on the research carried out within this work, it can be stated that there is an optimum amount of CeO2 in a cobalt-based system, which provides a high surface area and a high catalytic activity in the synthesis of NH3 . Moreover, the most active of the studied systems, i.e. Co/Ce(11.5)/Ba, also shows high thermal stability. Keeping it for over 160 h at 600 ◦ C (3H2 :N2 = 30.0 l/h, p = 0.1 MPa) resulted in only a 10% drop in activity. The dependence of the reaction rate on the partial pressure of ammonia for the co-promoted catalyst (Co/Ce(11.5)/Ba) is illustrated in Fig. 8. For comparison, an analogous relationship derived for the fused iron catalyst (KMI, H. Topsoe) has also been shown in Fig. 8. Both catalysts were activated (reduced) in such a way that their optimum activity was obtained. In both cases (Co and Fe), the reaction rate decreases with the increase in the ammonia content
286
M. Karolewska et al. / Applied Catalysis A: General 445–446 (2012) 280–286
Fig. 8. Influence of ammonia partial pressure on the NH3 synthesis rates (r) over the Co/Ce(11.5)/Ba catalyst and triply promoted fused iron catalyst (KM I) at 400 ◦ C (6.3 MPa).
in the gas mixture. However, the co-promoted cobalt catalyst is more active than the commercial iron catalyst in the whole range of conversion values. Moreover, the cobalt catalyst is kinetically less sensitive to the presence of ammonia (Fig. 8): an increase in the NH3 content from 1 to 8% leads to a eleven-fold decrease of the reaction rate over Co + Ce + Ba, whereas the reaction rate over the magnetite-based KMI material decreases by a factor of more than 15. Due to the kinetic differences between Co and Fe, the differences in the reaction rates under industrial conditions (15–20% NH3 conversions) will be significantly larger. 4. Conclusions In the investigated systems cerium oxide plays the role of a structural promoter, because it prevents the Co3 O4 and Co crystallites from sintering under calcination and reaction conditions, respectively. Moreover, CeO2 can stabilize the hcp phase of metallic cobalt under the reaction conditions. This phase is considered to be more active in ammonia synthesis [51], so the catalysts promoted with ceria and barium are more active in ammonia synthesis than the contacts promoted only with barium. The optimum content of ceria is between approximately 10 and 13 wt%. The comparison of the most active catalytic system – contact Co/Ce(11.5)/Ba(13), with the industrial iron catalyst KM I has shown that cobalt systems are more active than the conventional iron catalyst in the entire investigated range of ammonia concentrations. The superior catalytic activity of the cobalt catalyst is even more pronounced with the increase of ammonia content in the gas mixture. Furthermore, Co/Ce(11.5)/Ba is very resistant to overheating. The heat treatment at 600 ◦ C for 160 h (3H2 :N2 = 30.0 l/h, p = 0.1 MPa) results in only a slight decrease of activity in NH3 synthesis. Acknowledgement This work has been carried out within the Research Project N N209119737, sponsored by the Ministry of Science and Higher Education. References [1] U.S. Geological Survey, http://minerals.usgs.gov/minerals/pubs/commodity/ nitrogen, data from 31.03.08.
[2] R.Z. Sørensen, A. Klerke, U. Quaade, S. Jensen, O. Hansen, C.J.H. Christensen, Catal. Lett. 112 (2006) 77–81. [3] P.F. Ng, L. Li, S.B. Wang, Z.H. Zhu, G.Q. Lu, Z.F. Yan, Environ. Sci. Technol. 41 (2007) 3758–3762. [4] L. Li, Z.H. Zhu, Z.F. Yan, G.Q. Lu, L. Rintoul, Appl. Catal. A 320 (2007) 166–172. [5] L. Li, Z.H. Zhu, G.Q. Lu, Z.F. Yan, S.Z. Qiao, Carbon 45 (2007) 11–20. [6] S.F. Yin, B.Q. Xu, S.J. Wang, C.T. Au, Appl. Catal. A 301 (2006) 202–210. [7] X.K. Li, W.J. Ji, J. Zhao, S.J. Wang, C.T. Au, J. Catal. 236 (2005) 181–189. [8] http://nobelprize.org/nobel prizes/chemistry/laureates/index.html, data from 26.10.2007. [9] K. Tamaru, in: J.R. Jennings (Ed.), Catalytic Ammonia Synthesis: Fundamentals and Practice, Plenum Press, New York, 1991, pp. 1–18. [10] C.J.H. Jacobsen, Chem. Commun. 12 (2000) 1057–1058. [11] R. Kojima, K. Aika, Appl. Catal. A 215 (2001) 149–160. [12] R. Kojima, K. Aika, Appl. Catal. A 218 (2001) 121–128. [13] R. Kojima, K. Aika, Appl. Catal. A 219 (2001) 157–170. [14] S. Hagen, R. Barfod, R. Fehrmann, C.J.H. Jacobsen, H.T. Teunissen, I. Chorkendorff, J. Catal. 214 (2003) 327–335. [15] S. Hagen, R. Barfod, R. Fehrmann, C.J.H. Jacobsen, H.T. Teunissen, K. Stahl, I. Chorkendorff, Chem. Commun. 11 (2002) 1206–1207. ´ ˙ [16] A. Jedynak, D. Szmigiel, W. Raróg, J. Zielinski, J. Pielaszek, P. Dłuzewski, Z. Kowalczyk, Cattal. Lett. 81 (2002) 213–218. ´ Pol. J. Chem. 75 (2001) [17] A. Jedynak, Z. Kowalczyk, D. Szmigiel, J. Zielinski, 1801–1805. [18] R. Kojima, H. Enomoto, M. Muhler, K. Aika, Appl. Catal. A 246 (2003) 311–322. [19] R. Schlögl, Angew. Chem. Int. Ed. 42 (2003) 2004–2008. ´ [20] I. Jasinska, K. Lubkowski, W. Arabczyk, Przem. Chem. 82 (2003) 230–233. ´ S. Lenart, Appl. Catal. A 266 [21] M.J. Figurski, W. Arabczyk, Z. Lendzion-Bielun, (2004) 11–20. ´ R.J. Kalenczuk, ´ S. Lenart, Appl. [22] M.J. Figurski, W. Arabczyk, Z. Lendzion-Bielun, Catal. A 247 (2003) 9–15. [23] N. Pernicone, E. Ferrero, I. Rossetti, L. Forni, P. Canton, P. Riello, G. Fagherazzi, M. Signoretto, F. Pinna, Appl. Catal. A 251 (2003) 121–129. [24] Y.F. Zheng, H.Z. Liu, Z.J. Liu, X.N. Li, J. Solid State Chem. 182 (2009) 2385– 2391. [25] H.Z. Liu, X.N. Li, Z.N. Hu, Appl. Catal. A 142 (1996) 209–222. [26] R.B. Strait, Nitrogen Methanol 238 (1999) 37–43. [27] R.B. Strait, International Conference Exhibition, Caracas, 28 February–2 March, 1999, pp. 43–59. [28] Z. Kowalczyk, W. Raróg-Pilecka, Przem. Chem 85/2 (2006) 124–129. ´ J. Pielaszek, Appl. Catal. A 173 [29] Z. Kowalczyk, S. Jodzis, W. Raróg, J. Zielinski, (1998) 153–160. [30] I. Rossetti, N. Pernicone, L. Forni, Appl. Catal. A 208 (2001) 271–278. [31] I. Rossetti, N. Pernicone, L. Forni, Catal. Today 102-103 (2005) 219–224. [32] C.J.H. Jacobsen, S. Dahl, B.S. Clausen, S. Bahn, A. Logadottir, J.K. Norskov, J. Am. Chem. Soc. 123 (2001) 8404–8405. [33] O.V. Shcheglov, G.K. Boreskov, M.G. Slinko, Dokl. Akad. Nauk SSSR 105 (1955) 123–133. ´ [34] R.J. Kalenczuk, Appl. Catal. A 112 (1994) 149–160. ˛ nski, ´ Z. Kaszkur, K. Kielar, Z. Kowalczyk, [35] W. Raróg-Pilecka, E. Mi´skiewicz, L. Kepi J. Catal. 249 (2007) 24–33. [36] W. Raróg-Pilecka, E. Mi´skiewicz, Z. Kowalczyk, Catal. Commun. 9 (2008) 870–873. [37] W. Raróg-Pilecka, M. Karolewska, E. Truszkiewicz, E. Iwanek, B. Mierzwa, Catal. Lett. 141 (2011) 678–684. [38] M. Karolewska, O. Nalewajko, E. Truszkiewicz, W. Raróg-Pilecka, Przem. Chem. 91/6 (2012) 1224–1228. [39] W. Raróg-Pilecka, E. Mi´skiewicz, S. Jodzis, J. Petryk, D. Łomot, Z. Kaszkur, Z. ´ Z. Kowalczyk, J. Catal. 239 (2006) 313–325. Karpinski, [40] R.C. Reuel, C.H. Bartholomew, J. Catal. 85 (1984) 63–77. ´ M. Bonarowska, Langmuir 13 (1997) 5613–5620. [41] A. Borodzinski, [42] Z. Kowalczyk, Catal. Lett. 37 (1996) 173–179. [43] E.P. Barrett, L.G. Joyner, P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380. [44] L. Xue, Ch Zhang, H. He, Y. Teraoka, Appl. Catal. B 75 (2007) 167–174. [45] H. Wang, J.L. Ye, Y. Liu, Y.D. Li, Y.N. Qin, Catal. Today 129 (2007) 305–312. [46] H.C. Yao, Y.F.Y. Yao, J. Catal. 86 (1984) 254–265. [47] M. Kang, M.W. Song, Ch H. Lee, Appl. Catal. A 251 (2003) 143–156. [48] B. Ernst, L. Hilaire, A. Kiennemann, Catal. Today 50 (1999) 413–427. [49] L.F. Liotta, G. Di Carlo, G. Pantaleo, A.M. Venezia, G. Deganello, Appl. Catal. B 66 (2006) 217–227. [50] J.Y. Luo, M. Meng, X. Li, X.G. Li, Y.Q. Zha, T.D. Hu, Y.N. Xie, J. Zhang, J. Catal. 254 (2008) 310–324. [51] G. Rambeau, A. Jorti, H. Amariglio, J. Catal. 94 (1985) 155–165. [52] S.S.Y. Lin, D.H. Kim, S.Y. Ha, Appl. Catal. A 355 (2009) 69–77. [53] W.O. Soboyejo, T.S. Srivatsan, Advanced Structural Materials: Properties, Design, Optimization and Applications, CRC Press, Boca Raton, FL, 2007. [54] A. Amariglio, M. Eddouasse, H. Amariglio, Surf. Sci. 162 (1985) 375–381.