engleski

Combustion of active carbon as model carbon material : comparison of non-catalytic and catalytic oxidation

Kinetics of non-catalytic and Pt-catalysed oxidation of active carbon, selected as a model carbon material, was investigated using thermogravimetric analysis (TGA). Investigations were performed in the temperature range from 40 º ; C to 1000 º ; C at different heating rates (5-25 º ; C min-1). The influence of Mn- and Pt-based catalyst on the combustion kinetics was examined as well. Values of the kinetic parameters, such as activation energy, Ea and Arrhenius pre-exponential factor, A were determined using isoconversional method proposed by Kissinger-Akahira-Sunose. The obtained values were in good agreement with the literature data. Linear relationship between the activation energy and pre-exponential factor was studied and defined by the compensation effect. 1. Introduction The use of Diesel engines has increased recently due to their reliability, durability and efficiency and low consumption of fuels [1]. However, they are the main source of emission of soot particles into atmosphere [2, 3]. It is well known that these particles are very harmful to human health, due to their carcinogenic effect [4]. One of the available technologies for the control of diesel particulate emissions employs the filters, which capture soot particles from the exhaust stream. However, the filters must be regenerated periodically or continuously by combustion/oxidation of the trapped soot to prevent the back-pressure build-up. Soot oxidation takes place at high temperatures (> 600˚ C) while the temperature of diesel exhaust gases for small engines may be as low as 200˚ C. Therefore, the catalyst is used to prevent accumulation of soot on the monolithic filter. The catalyst is introduced by addition of the oxidation catalyst precursors as fuel additives or by impregnation of the filter walls with the oxidation catalyst. In this study we used the powder catalyst mixed with soot sample to enhance the oxidation rate. Our primary aim was to investigate the influence of the two applied catalysts on oxidation rate of the model soot particles. Thermo-oxidative degradation and catalytic oxidation were investigated applying dynamic thermogravimetric analysis. Kinetic parameters, activation energy and pre-exponential factor were evaluated by the Kissinger-Akahira-Sunose isoconversional method [5, 6]. The second aim was to determine the strengths and weaknesses of isoconversional method. 2. Experimental Active carbon, supplied by Kemika, was selected as a model carbon material to simulate diesel soot combustion. Chloride, sulphate, heavy metals, zinc, iron and carbon content of the manufacturer’ s sample were 0.001 ; 0.01 ; 0.005 ; 0.0005 and 91.969 wt. % respectively. Moisture content after heating at 120 º ; C was 8 %. A model sample was used for non-catalytic oxidation without further modifications. A commercial, Pt-based catalyst (code F 105 R/W ; 5 wt. % of Pt), supplied by Degussa-Huls AG, (particle diameter 23  m) was thermally pre-treated at 60 º ; C for 24 hours. A Mn-based catalyst was prepared by wet impregnation of Mn(SO4) 4 H2O solution on the alumina support. Thermo-oxidative degradation and catalytic oxidation of active carbon were investigated by dynamic termogravimetric analysis under the excess air and employment of PerkinElmer Thermobalance TGS-2. A 7-10 mg sample was used for thermogravimetric analysis. The appropriate amount of active carbon or a mixture of active carbon and catalyst (1:1) was loaded in the crucible and heated from 40  C to 1000  C. The experiments were carried out at constant heating rates (5, 10, 15, 20 and 25 ˚ C min-1). Mass loss and sample temperature were monitored by a computerised data system. The oxidant was dry air (30 and 150 cm3min-1) flowing downward the cylindrical sample holder. 3. Results and discussion 3.1. Catalytic and non-catalytic oxidation Figure 1 shows comparison of the results obtained from non-catalytic and catalytic oxidation of the soot at constant heating rate of 10˚ C min-1. The characteristic temperature of soot oxidation moved to low temperature region in the presence of the Pt-based catalyst. Maximum temperature for catalytic oxidation of the soot-Pt catalyst mixture was reduced by approximately 180-200˚ C compared to thermo-oxidative degradation in the absence of a catalyst. Conversely, the Mn-based catalyst had no influence on combustion rate of the active carbon sample. Surprisingly, because some researchers observed good activity of the Mn-based catalyst in soot oxidation [7]. Probably, wet impregnation is not appropriate for preparation of the Mn-based catalyst active in soot oxidation. Further ongoing work is aimed at investigating the influence of other preparation methods and conditions on activity of the Mn-catalyst. 3.2. Influence of gas flow rate The experiments were performed with varying flow rate of dry air (30 and 150 cm3 min-1) in order to investigate the influence of the external mass transfer limitations on the results of thermoanalytical measurements. Under experimental conditions employed in this study, the weight versus temperature curves remained unchanged within the range of experimental uncertainty. Due to a small fraction of the used soot sample, the intraphase mass transfer was also neglected. 3.3. Isoconversional kinetic analysis Isoconversional kinetic methods differ from the classical ones in that the former assume a kinetic model f(α ) and then determine activation energy (Ea) for such a process by directly calculating Ea. Their result is Ea dependence on conversion (α ). Complexity of that dependence indicates a complex reaction mechanism. Dependence analysis of Ea on α often enables identification of the scheme of a kinetic process, and can be used in modelling the reaction kinetics outside the temperature range of a particular experiment. Thermogravimetric analysis (TGA) used to determine the kinetics of composites degradation, gave the integral curves, showing dependence of weight loss on temperature. Thus, integral isoconversional methods such as Flynn-Wall-Ozawa (FWO) [8, 9] and Kissinger-Akahira-Sunose (KAS) method [5, 6] are commonly used. Differential isoconversional methods are not very suitable for analysis of TGA curves because they require numerical derivation of experimental results which contributes to experimental error. KAS method is a mathematical equivalent to FWO method but it is more accurate for a wider interval of Ea values [10]. Therefore, KAS method was chosen in this study for isoconversional calculation of Ea. Conversion, α , is calculated after the equation (1), where m0 is the initial and m∞ is final weight of a sample, ignoring the initial loss at the temperatures under 150°C attributed to evaporation of adsorbed moisture. (1) Energies of activation, Ea, were calculated from Kissinger-Akahira-Sunose (KAS) equation (2) for conversions in the range α = 0.1-0.9, where  is heating rate, T - thermodynamic temperature, A - pre-exponential factor, R - general gas constant (8.314 J mol-1 K-1), and G(α ) - integral form of the kinetic model f(α ). Figure 2 shows an example of isoconversional plots for non-catalytic oxidation of soot. Similar plots were obtained for catalytic oxidation. (2) The obtained values of Ea for non-catalytic and Pt-based oxidation were in the range from 124 -300 kJ mol-1 and 66-212 kJ mol-1 respectively (Figure 3). The pairs of activation energy and pre-exponential factor showed linear relationship defined by the compensation effect (Figure 4 and 5). Similar results were reported in the literature [11, 12]. Conclusion Thermo-oxidative degradation and catalytic oxidation of active carbon were analysed by dynamic thermogravimetric method in order to get useful information for modelling of regeneration steps in the filters used for diesel soot removal. It can be concluded that Pt catalyst considerably reduces temperature of active carbon oxidation and decreases activation energy, while the Mn-based catalyst does not influence combustion rate of active carbon. The isoconversional Kissinger-Akahira-Sunose method applied in this work appears appropriate, despite numerous approximations used in calculation of the kinetic parameters. Acknowledgement The authors highly appreciate financial support that the Ministry of Science, Education and Sport of the Republic of Croatia has given for this study. References [1] R.M. Heck, R.J. Farrauto, S.T. Gulati: Catalytic Air Pollution Control-Commercial Technology, Sec. Edit., John Wiley & Sons, Inc., New York, 2002. [2] J.P.A. Neeft, M. Makkee, J.-A. Moulijn, Fuel Process. Technol., 47 (1996) 1. [3] B.R. Stanmore, J.F. Brilhac, P. Gilot, Carbon, 39 (2001) 2247. [4] G. Saracco, N. Russo, M. Ambrogio, C. Badini. V. Specchia, Catalyst Today, 60 (2000) 33. [5] H.E. Kissinger, Anal. Chem., 29 (1957) 1702. [6] T. Akahira, T. Sunose, Res. Report Chiba Inst. Technol. 16 (1971) 1622-1631. [7] E. Saab, S. Aouad, E. Abi-Aad, E. Zhilinskaya, A. Aboukaï ; s, Catal. Today 119 (2007) 286-290. [8] T. Ozawa, Bulletin of the Chemical Society of Japan, 38 (1965) 1881-1886. [9] J.H. Flynn, L.A. Wall, J. Polym. Sci.: Polymer Letters 4 (1966) 323-328. [10] Z.M. Gao, M. Nakada, I. Amasaki, Thermochimica Acta 369 (2001) 137-142. [11] R. López-Fonseca, U. Elizundia, I. Landa, M.A. Gutiérrez-Ortiz, J.R. González-Velasco, Appl. Catal. B: Environmental, 61 (2005) 150. [12] R. López-Fonseca, I. Landa, U. Elizundia, M.A. Gutiérrez-Ortiz, J.R. González-Velasco, Combustion and Flame, 144 (2006) 398.

combustion; kinetics of oxidation; carbon material

nije evidentirano