Iron-manganese mixed oxide catalysts for ethanol combustion. Influence of the fe:mn atomic ratio and the calcination temperature

F. G. Durán, L. E. Cadús and B. P. Barbero

Instituto de Investigaciones en Tecnología Química (INTEQUI), UNSL - CONICET, Chacabuco 917, D5700BWS San Luis, Argentina. bbarbero@unsl.edu.ar

Abstract— A Fe-Mn mixed oxide catalyst for ethanol combustion was optimized using different Fe:Mn atomic ratios and varying the calcination temperature (500°C, 650°C and 800°C). The results indicated that the increase of the calcination temperature caused a decrease of the catalytic activity due to a strong decrease of the specific surface area. Then, among the catalysts calcined at 500°C, the most active one was the catalyst with lower amount of iron. The results of X-ray diffraction (XRD) and temperature programmed reduction (TPR) indicated that a small amount of iron in the Fe-Mn mixed catalyst favored the formation of dispersed surface amorphous MnOx species. These species are easily reducible and very active in total oxidation reactions.

Keywords— Total Oxidation; Ethanol; Catalyst Preparation.

I. INTRODUCTION

In the last decade, ethanol has been considered a promising alternative fuel. It is used in cars as an octane enhancer and an oxygenated additive to gasoline, and even, as a replacement for gasoline (Wang et al., 2000; Goldemberg et al., 2008). Although exhaust emissions of ethanol are less toxic than those of gasoline, partially oxidized products such as acetaldehyde, CO and acetic ester may cause severe environmental and health problems. For these reasons, it is important to eliminate the pollution of ethanol and their by products, being the catalytic combustion an attractive method. Therefore, it is necessary the development of efficient catalysts, which would directly convert ethanol to CO₂ and H₂O at relatively low temperature.

Several groups have investigated the reactivity of noble metal-based catalysts and transition metal oxide-based catalysts in oxidation reactions. Generally speaking, the noble metal catalysts are more active than other metal oxide catalysts, but they are more expensive and easily poisoned as well. Particularly, manganese oxide-based catalysts have received great attention because they exhibit an interesting behavior in combustion reactions and they are environmentally friendly materials (Wahlberg et al., 1999; Gallardo-Amores et al., 1999). Their catalytic properties are attributed to a high capacity to store oxygen and to form manganese oxides in different oxidation states, for the reason that manganese is highly efficient in reduction-oxidation cycles. The redox abilities frequently increase if it is combined with other elements (Lin et al., 2001). For instance, Mn-Cu mixed oxides show better catalytic behaviour than Mn₂O₃ and CuO pure oxides in the combustion of propane and ethanol (Morales et al., 2006).

In a previous work (Morales et al., 2007), iron-manganese mixed oxide based catalysts, obtained by coprecipitation showed a higher catalytic activity in ethanol combustion than Mn₂O₃ and Fe₂O₃ pure oxides. Mishra et al. (2008) prepared manganese-iron mixed oxide pillared clays with varying the manganese to iron ratio and they observed that the catalyst with higher Mn/Fe ratio was the most active in the acetone decomposition while the catalytic activity in trichloroethylene decomposition increased with the increase in iron content up to certain value (10.14%). These results encouraged us to attempt the optimization of the iron-manganese mixed oxide catalysts prepared by coprecipitation with varying the Fe:Mn atomic ratio and the calcination temperature. The catalytic behavior was evaluated in the ethanol combustion.

II. EXPERIMENTAL SECTION

A. Preparation of the catalysts

Aqueous solutions of Fe(NO₃).9H₂O, 0.48M (Riedel-de Haën), Mn(NO₃).4H₂O, 0.48M (Fluka), and NH₄OH 1N (used as precipitant agent) were prepared (Wang et al., 2005). The nitrate solutions were mixed in adequate proportions for obtaining Fe:Mn atomic ratios equal to 1:1, 3:1, and 1:3. Then, the mixed solution was added drop by drop at the ammonium hydroxide solution kept under agitation and at pH = 11.5 - 12.5. The pH was regulated adding ammonium hydroxide at 28 vol% when it was necessary. Meanwhile, the precipitate was aging in the mother solution for 24 hours. Then, it was separated from the solution and dried slowly in a vacuum oven increasing the temperature from 50°C to 120°C at 5°C/h. The resulting solid was milled and calcined in air at 500°C, 650°C or 800°C for 3 hours. The catalysts were denoted as FexMny-T, where the ratio x:y = 1:1, 3:1, and 1:3 represents the Fe:Mn atomic ratio and T is the calcination temperature in Celsius degrees. Also, Fe₂O₃ and Mn₂O₃ were prepared by the same synthesis method and calcined at 500°C, for comparison.

B. Characterization of the catalysts

The specific surface areas were measured in a Micromeritics Gemini V equipment by the nitrogen adsorption method (BET) after to degas the samples at 250°C under N₂ flow using a FlowPrep 060 - Sample Degas System - from Micromeritics. The crystalline phases were identified by X-ray diffraction (XRD) using a Rigaku diffractometer equipped with a radiation source of V filtered - Cr Kα (λ = 0.22897 nm) operated at 35 kV and 25 mA and the PDF-ICDD files as references. The composition was determined by X-ray fluorescence (XRF) with a Philips PW 1400 apparatus where the calibration standards were prepared with known amounts of Fe₂O₃ and Mn₂O₃. The study of temperature programmed reduction (TPR) was carried out using a U-type quartz tubular reactor and a thermal conductivity detector (TCD) for recording the hydrogen consumption. A 25mg-sample and a reducing gas mixture (5 vol% H₂/N₂) at a flow rate of 30 cm³/min were used. The temperature was increased at a heating rate of 10°C/min from room temperature to 700°C, and kept at that temperature until the hydrogen consumption signal returned to the initial value.

C. Catalytic evaluation

The catalysts were evaluated in the total oxidation of ethanol using a glass fixed bed reactor operated at atmospheric pressure. The powdered catalyst was pelletized to particles of 0.5 - 0.8 mm diameter. The sample (300 mg) was mixed with glass particles of the same size in a weight ratio of catalyst:glass = 1:5. The reactive mixture was C₂H₅OH:O₂:He = 1:20:79 in volume with a total flow rate of 100 cm³/min. The reaction temperature, measured with a thermocouple placed in the centre of the catalytic bed, was increased from 80°C until total conversion of ethanol. The reagent and products were analyzed by on-line gas chromatography using a Buck Scientific Mod 910 chromatograph equipped with a Porapak T columm and thermal conductivity detector.

III. RESULTS AND DISCUSSION

In a previous work (Morales et al., 2007), it was observed that an iron-manganese mixed oxide catalyst was more active than the pure oxides, Fe₂O₃ and Mn₂O₃, in ethanol combustion. That catalyst was prepared with a Fe:Mn atomic ratio equal to 1:1 calcining the precipitate at 500°C. This work attempts to optimize the Fe-Mn mixed oxide catalyst by using different calcination temperatures and Fe:Mn atomic ratios.

The catalyst composition (Fe:Mn atomic ratio) was determined by means of XRF and the results are presented in Table 1. As it can be observed, the experimental values are significantly different from the nominal ones. The Fe:Mn atomic ratio for Fe1Mn1-T series was 1.70±0.04 being 1 the theoretical value; for Fe1Mn3-T series, it was 0.82±0.07 (theoretical value = 0.33), and for Fe3Mn1-T it was 5.71±0.10 (theoretical value = 3). The higher iron proportion detected in all of the catalysts could be due to a difficulty in precipitating the manganese during the preparation stage. In this work, the iron and manganese ions from the nitrate solutions were precipitated as hydroxides using NH₄OH as precipitant agent. The solution pH was kept between 11.5 and 12.5 adding concentrated NH₄OH when it was necessary. However, this practice was not enough to achieve the complete precipitation of the manganese that partially remained in solution.

Table 1. Results of chemical composition determined by X-ray fluorescence (XRF), specific surface area measured by the BET method and catalytic activity in ethanol combustion.
T₁₀, T₅₀, and T₈₀: reaction temperatures corresponding to 10%, 50%, and 80% conversion of ethanol, respectively.
X₂₀₀: Ethanol conversion at 200°C.

Compositions closed to the nominal ones were obtained by means of the citrate method (Durán et al., 2009) but this method presents a great difficulty during the period of calcination, where small amounts of sample and low heating rate must be used to avoid an explosion. Therefore, the large-scale catalyst production by the citrate method would be not viable.

Although the nominal Fe:Mn atomic ratios were not achieved in this work, the study was continued since the effect of the composition on the catalytic activity could be evaluated regardless of these values.

Increasing the calcination temperature, structural modifications may be obtained favoring the solid state reaction of the elements which constitute the system. The phase diagram of the Fe-Mn-O system establishes that, under conditions similar to those used in this work for the synthesis of the catalysts, the Fe₂O₃ and Mn₂O₃ phases coexist when the calcination temperature is below 600°C. At higher temperatures, the formation of solid solutions by substitution has been observed (Kedesdy and Tauber, 1956). However, the results of XRD of our catalysts do not show modifications of the lattice parameters, not even for those catalysts calcined at 800°C.

In fact, only the diffraction lines of Fe₂O₃ (PDF 24-72) and Mn₂O₃ (PDF 24-508) are observed in the X-ray diffractograms (Fig. 1). The relative intensities of each phase are different, depending mainly on the Fe:Mn atomic ratio and, in lesser extent, on the calcination temperature. The Fe₂O₃ phase is detected in all catalysts while Mn₂O₃ is not detected in Fe1Mn1-500, Fe3Mn1-500, and Fe3Mn1-650 catalysts. Then, in these catalysts, manganese could be forming amorphous MnOx species dispersed on the surface, or microcrystalline Mn₂O₃ (crystals smaller than 40Å). The results of temperature programmed reduction (Fig. 2) are in agreement with this assumption. The reduction curve of Fe1Mn1-500 catalyst shows a very broad and weak signal beginning at approximately 200°C which may be assigned to the reduction of amorphous MnOx species. The curve of Fe3Mn1-500 catalyst also presents a deviation from the baseline beginning at 200°C and, although this is not defined as a reduction signal, it couldbe considered as an insight into the existence of amorphous MnOx species in a smaller amount than in the Fe1Mn1-500 catalyst. The increase of calcination temperature at 650°C for Fe3Mn1-T catalysts could be favoring the crystalline order of MnOx species towards the formation of microcrystalline Mn₂O₃, which reduction occurs at higher temperature than 400°C. A subsequent increase of the calcination temperature at 800°C induces the growth of Mn₂O₃ crystals which are detected by XRD.

Figure 1. X-ray diffractograms. (o) Mn₂O₃; ()Fe₂O₃.

Figure 2. Curves of temperature programmed reduction.

The Fe₂O₃ phase detected by XRD in all catalysts is also affected by the calcination temperature. Applying the Scherrer's equation, it was corroborated that the increase of calcination temperature increases the size of Fe₂O₃ crystals. For instance, in the case of Fe3Mn1-T catalysts, crystal sizes are 259, 432, and 642 Å when the catalysts are calcined at 500, 650, and 800°C, respectively.

The increase of the crystal size caused by the increase of calcination temperature is in agreement with the decrease of the specific surface area. As it is observed in Table 1, the increase of calcination temperature from 500°C to 650°C causes a decrease of around 80% of the specific surface area. The subsequent increase of calcination temperature up to 800°C causes a decrease in the area resulting 6 - 8% of the value obtained at 500°C. These results show clearly that the increase of calcinations temperature provokes sintering of particles and breakdown of porous structure.

The specific surface area is a very important characteristic of the solid to be used as catalyst. Therefore, from the catalytic viewpoint, the effect of the calcination temperature on the specific surface area could be more relevant than the effect on the structure of the active phase. Effectively, a decrease of the catalytic activity with an increase of the calcination temperature is observed when comparing T₁₀, T₅₀, T₈₀ (reaction temperatures corresponding to 10%, 50%, and 80% conversion, respectively), or conversion at 200°C (Table 1).

These results lead to the partial conclusion that the increase of the calcination temperature is not convenient to optimize the mixed FeMn catalyst since the decrease of the specific surface area has more influence on the catalytic activity than the formation of a solid solution or other structural change of the active phase. In consequence, in order to reveal which is the most convenient Fe:Mn atomic ratio, only the catalysts calcined to the lowest temperature (500°C) are considered.

To compare the catalytic activity, the values of T₅₀, T₈₀ or conversion at 200°C (presented in Table 1) are taken into account. Although the differences are small, there is a clear tendency. The activity of the catalysts calcined a 500°C decreases in the following order: Fe1Mn3-500 > Fe1Mn1-500 > Fe3Mn1-500. Furthermore, it is worth to note that the specific surface area decreases with the decrease of the iron content (Fe3Mn1-500 > Fe1Mn1-500 > Fe1Mn3-500). Therefore, considering the intrinsic activity (conversion per square meter of surface area), it is deduced that the mixed iron-manganese oxide catalyst with low proportion of iron (Fe1Mn3-500) is the most active one for ethanol combustion.

Similarly to Fe1Mn1-500 and Fe3Mn1-500 catalysts, the reduction of Fe1Mn3-500 catalyst begins around 200°C and a reduction peak with the maximum at approximately 325°C (Fig. 2) is detected. This signal could be assigned to the reduction of amorphous MnOx species dispersed on the surface. Then, considering that the area under the reduction curve between 200°C and 350°C is proportional to the amount of dispersed MnOx species, an increase of the catalytic activity with a higher amount of these species is clearly observed.

The dispersed surface amorphous MnOx species has been recognized as highly reactive in oxidation reactions because they are easily reducible. The close relationship between the results of catalytic activity and the amount of these species detected in the catalysts leads to propose that the ethanol oxidation on mixed Fe-Mn catalysts takes place through a redox mechanism generally known as Mars - van Krevelen. This mechanism has been also suggested on other mixed Mn-Ni (Morales et al., 2007) and type-perovskite La_1-xCa_xFeO₃ oxide catalysts (Barbero et al., 2006).

Consequently, the best catalytic activity of Fe-Mn mixed catalysts in comparison to pure Mn₂O₃ can be attributed to the formation of amorphous MnOx species favored by the presence of a small amount of iron.

In order to evaluate the level of optimization of the Fe-Mn mixed oxide catalyst achieved in this work, a comparison of the catalytic performance of Fe1Mn3-500 with other catalysts is carried out. In Table 2, T₅₀ and specific surface areas for several catalysts are presented. First of all, it is interesting to compare Fe1Mn3-500 catalyst with other Fe-Mn catalysts prepared by different methods. As it is observed, T₅₀ for Fe1Mn3-500 is lower than for FeMn-X (atomic ratio Fe:Mn = 1:1), which was prepared by coprecipitation with NaCO₃ at different aging time of the precipitated. In relation to Fe-Mn catalysts prepared by the citrate method (Durán et al., 2009), Fe1Mn3-500 is quite similar to FeMn 1:3 but less active than FeMn 1:1 catalyst. Probably, a different structure -besides of a higher specific surface area generated by the citrate method- is the cause of higher catalytic activity. The advantage of Fe1Mn3-500 catalyst is a simpler preparation method than the citrate method. Regarding to other catalyst families such as MnNi 1:1 and MnCu 1:1, prepared by coprecipitation with NaCO₃, Fe1Mn3-500 is more active. Only the optimized Mn-Cu catalyst, Mn2Cu8-500 (Morales et al., 2008), presents the same T₅₀ than Fe1Mn3-500.

Table 2. T₅₀ and specific surface areas for different catalyst series.

Comparing with other authors' results is not easy since different reaction conditions were employed. However, we included in Table 2 the results obtained with a family of manganese oxides with different structures (Santos et al., 2010), which are considered as highly active in ethanol combustion. As it can be observed, the activity of Fe1Mn3-500 is similar to Mn2 (cryptomelane) and Mn6 (cryptomelane; α-Mn₂O₃; Mn₅O₈). Only Mn5 (cryptomelane; Mn₃O₄) is more active than Fe1Mn3-500.

With regard to several perovskite-type catalyst series (LaCo_1-yFe_yO₃, La_1-xCa_xFeO₃, La_1-xCa_xMnO₃) evaluated under the same reaction conditions, Fe1Mn3-500 catalyst is considerably more active. Just an addition of Ag to La_0.6Sr_0.4MnO₃ (Wang et al., 2000) increases the catalytic activity giving a T₅₀ lower than Fe1Mn3-500 catalyst.

IV. CONCLUSIONS

The optimization of the mixed Fe-Mn oxide catalyst for ethanol combustion was achieved in this work. The results related to the catalytic activity showed that it is not convenient to increase the calcination temperature above 500°C because the specific surface area strongly decreases. On the other hand, a small amount of iron is enough to favor the formation of dispersed surface amorphous MnOx species which are easily reducible and very active in oxidation reactions.

ACKNOWLEDGEMENTS
The authors thank the financial support from UNSL, CONICET and ANPCyT of Argentina. Also, they thank Dr. E. Perino of INQUISAL (UNSL-CONICET) for his collaboration in obtaining XRF results

REFERENCES
1. Barbero, B.P., J. Andrade Gamboa and L.E. Cadús, "Synthesis and characterisation of La_1−xCa_xFeO₃ perovskite-type oxide catalysts for total oxidation of volatile organic compounds," Appl. Catal. B, 65, 21-30 (2006).         [ Links ]
2. Durán, F.G., B.P. Barbero, L.E. Cadús, C. Rojas, M.A. Centeno and J.A. Odriozola, "Manganese and iron oxides as combustion catalysts of volatile organic compounds," Appl. Catal. B, 92, 194-201 (2009).         [ Links ]
3. Gallardo-Amores, J.M., T. Armaroli, G. Ramis, E. Finocchio and G. Busca, "A study of anatase-supported Mn oxide as catalysts for 2-propanol oxidation," Appl. Catal. B, 22, 249-259 (1999).         [ Links ]
4. Goldemberg, J., S.T. Coelho and P. Guardabassi, "The sustainability of ethanol production from sugarcane," Energy Policy, 36, 2086-2097 (2008).         [ Links ]
5. Kedesdy, H.H. and A. Tauber, "Formation of Manganese Ferrite by Solid-State Reaction," J. Am. Cer. Soc., 39, 425-431 (1956).         [ Links ]
6. Lin, R., W.P. Liu, Y.-J. Zhong and M.-F. Luo, "Catalyst characterization and activity of Ag-Mn complex oxides," Appl. Catal. A, 220, 165-171 (2001).         [ Links ]
7. Merino, N.A., B.P. Barbero, P. Ruiz and L.E. Cadús, "Synthesis, characterisation, catalytic activity and structural stability of LaCo_1−yFe_yO_{3± λ} perovskite catalysts for combustion of ethanol and propane," J. Catal., 240, 245-257 (2006).         [ Links ]
8. Mishra, T., P. Mohapatra and K.M. Parida, "Synthesis, characterisation and catalytic evaluation of iron-manganese mixed oxide pillared clay for VOC decomposition reaction," Appl. Catal. B, 79, 279-285 (2008).         [ Links ]
9. Morales, M.R., B.P. Barbero and L.E. Cadús, "Total oxidation of ethanol and propane over Mn-Cu mixed oxide catalysts," Appl. Catal. B, 67, 229-236 (2006).         [ Links ]
10. Morales, M.R., B.P. Barbero and L.E. Cadús, "Combustion of volatile organic compounds on manganese iron or nickel mixed oxide catalysts," Appl. Catal. B, 74, 1-10 (2007).         [ Links ]
11. Morales, M.R., B.P. Barbero and L.E. Cadús, "Evaluation and characterization of Mn-Cu mixed oxide catalysts for ethanol total oxidation: Influence of copper content," Fuel, 87, 1177-1186 (2008).         [ Links ]
12. Santos, V.P., M.F.R. Pereira, J.J.M. Órfão and J.L. Figueiredo, "The role of lattice oxygen on the activity of manganese oxides towards the oxidation of volatile organic compounds," Appl. Catal. B, 99, 353-363 (2010).         [ Links ]
13. Stege, W.P., L.E. Cadús and B.P. Barbero, "La_1−x Ca_xMnO₃ perovskites as catalysts for total oxidation of volatile organic compounds," Catal. Today, 172, 53-57 (2011).         [ Links ]
14. Wahlberg, A., L.J. Pettersson, K. Bruce, M. Andersson and K. Jansson, "Preparation, evaluation and characterization of copper catalysts for ethanol fuelled diesel engines," Appl. Catal. B, 23, 271-281 (1999).         [ Links ]
15. Wang, W., H.B. Zhang, G.D. Lin and Z.T. Xiong, "Study of Ag/La_0.6Sr_0.4MnO₃ catalysts for complete oxidation of methanol and ethanol at low concentrations," Appl. Catal. B, 24, 219-232 (2000).         [ Links ]
16. Wang, C., Q. Wang, X. Sun and L. Xu, "CO hydrogenation to light alkenes over Mn/Fe catalysts prepared by coprecipitation and sol-gel methods", Catal. Letters, 105, 93-101 (2005).         [ Links ]

Received: December 27, 2010.
Accepted: September 16, 2011.
Recommended by subject editor: Pedro de Alcântara Pessôa