Subscriber access provided by UNIV AUTONOMA MADRID

Article

Activation of Waste Tire Char upon Cyclic Oxygen Chemisorption#Desorption

Francisco Heras, Noelia Alonso, Miguel A#. Gilarranz, and Juan J. Rodriguez
Ind. Eng. Chem. Res., 2009, 48 (10), 4664-4670 Publication Date (Web): 31 March 2009 Downloaded from http://pubs.acs.org on May 13, 2009

More About This Article

Additional resources and features associated with this article are available within the HTML version: Supporting Information Access to high resolution figures Links to articles and content related to this article Copyright permission to reproduce figures and/or text from this article

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036

4664

Ind. Eng. Chem. Res. 2009, 48, 46644670

APPLIED CHEMISTRY Activation of Waste Tire Char upon Cyclic Oxygen Chemisorption-Desorption
Francisco Heras,* Noelia Alonso, Miguel A. Gilarranz, and Juan J. Rodriguez
S.D. Ingeniera Qumica, UniVersidad Autonoma de Madrid, Ctra. Colmenar Viejo, km15. 28049, Madrid, Spain

Activation of waste tire char upon cyclic oxygen chemisorption-desorption permits a controlled development of porosity versus burnoff using air as feed gas for the activation process. A slow but monotonical increase of BET surface area is obtained from cycle to cycle. Initially the process led to the development of mesoporosity without generating micropores and then the micropore volume is increased whereas a decrease of narrow mesopore (20-80 nm) volume is observed, probably as a consequence of mesopore widening. Although the SBET reaches relatively low values (below 250 m2/g) even after 15 cycles, this surface development is associated with low burnoff values (around 22% for indicated BET surface area) and corresponds in an important percentage (up to about 50%) to external (nonmicropore) area. Temperatures around 210 and 550 C for chemisorption and desorption, respectively, have been found as optimum for the purpose of preparing granular mesoporous carbons which can be interesting candidates as catalytic supports for liquid phase applications.
Introduction Waste tires represent an important environmental problem nowadays due to the high generation, more than 5 million of tons per year in Europe, United States, and Japan, and potential polluting power.1 In addition to physical recycling by retreading and as additive in asphalt formulation or other applications, several thermal methods based on pyrolysis and/or gasication of waste tire rubber have been studied. The typical composition of pyrolysis products is 0-30% gases, 20-55% oil, and 30-40% char.2,3 The chemical composition of oil and gas fractions has been extensively studied to recover valuable chemicals as aromatics, or to be used as fuel because of their high heating power: about 35 MJ/Nm3 and 40 MJ/kg for gas and oil fraction, respectively.2,4 Important attention has also been paid to the valorization of char. The chars from waste tires pyrolysis usually do not have adequate characteristics to be used directly in commercial applications. For this reason several methods have been studied to modify the properties of these chars to be applied as carbon black or activated carbons. The activated carbons obtained from waste tires have been tested successfully as adsorbents in the decolorization of wastewater, in terms of methylene blue adsorption capacity,5 and the removal of a variety of organic compounds and metals from aqueous solutions.2,6 The preparation of activated carbons is the most studied alternative for the valorization of waste tires char. In general, the chars obtained from pyrolysis of waste tires show low surface area values, around 50-100 m2/g of BET surface area (SBET).2,7-9 Hence, activation by different methods, mainly using steam and CO2, is needed.1 In this way it is possible to produce activated carbons with a wide range of pore volume and pore size distributions to be used in different applications. Helleur et al.2 reported SBET values of 240-275 m2/g after the activation with steam and carbon dioxide at temperatures around 900 C at burnoff values of 10-40%. Substantially higher values (200-1100 m2/g) have been reported by San Miguel et al.6 for the activation with steam and CO2 at 925-1100 C, but with
* To whom correspondence should be addressed. E-mail: [email protected].

burnoff levels up to 80%. Literature shows as a general trend that to reach SBET higher than 300-400 m2/g through physical activation it is necessary to operate at high burnoff implying low nal yields. Most of the works in literature are devoted to physical activation with steam and CO2, whereas little attention has been paid to oxygen as activating agent. Oxygen exhibits a very high reactivity with chars and thus provokes a poor development of porosity, centered mostly in macropores and associated to high burn off values.2 Py et al.10 reported an activation method based on oxygen chemisorption/desorption cycles to control the high reactivity of this activating agent. In this method, the rst step of the cycle is chemisorption of oxygen at moderate temperatures using air or oxygen-containing mixtures. Afterward, the chemisorpted oxygen is desorbed at high temperature in inert atmosphere. In these conditions the oxygen is desorbed as CO and CO2 leaving new reactive carbon sites. This method is also reported as effective for pore size widening. Activation by cyclic chemisorption/desorption requires the optimum temperature and time for each step of the cycle to be conveniently established. When the char is heated in the presence of oxygen, chemisorption and desorption are competitive phenomena. At low temperatures chemisorption prevails, while desorption is the dominant process at high temperature. Desorption temperature is of importance since it inuences not only the amount of carbon that is evolved as CO or CO2, but also the type of oxygen-carbon associations that decompose giving rise to pore development. In this work air has been used as activating agent for surface area development by cyclic chemisorption-desorption from a char obtained from pyrolysis of waste tires. In a rst approach the temperature and time for chemisorption and desorption steps
Table 1. Composition of Waste Tire Char (% Weight Basis) C 89.3
a

H 0.5

N 0.3

S 2.1

othersa 7.8

Includes ash (Si, Zn, Ca, and other minor elements) and oxygen.

10.1021/ie801764x CCC: $40.75 2009 American Chemical Society Published on Web 03/31/2009

Ind. Eng. Chem. Res., Vol. 48, No. 10, 2009

4665

Figure 1. Experimental setup for the pyrolysis (a) and activation (b) runs.

Figure 2. Relative mass variation upon oxygen chemisorption at constant temperature.

are optimized. Then the development of porous structure is studied with a focus on surface area and pore volume. Materials and Methods Experimental Setup. The char was obtained from pyrolysis of waste tires in a quartz vertical reactor placed inside of a sandwich-type electrical furnace. The reactor was 50 mm in diameter and 500 mm in length. Figure 1a shows the experimental setup. A downward nitrogen ow was maintained through the reactor by means of a mass ow controller. A sample cup was held in the reactor by means of a rod introduced from the top. The rod pemits the displacement of the cup with the rubber sample from the top of the reactor, where an inert atmosphere can be maintained due to the downward nitrogen ow, to the center of furnace where the reaction takes place. After the programmed reaction time, the cup is raised again to

the top to be cooled by the incoming nitrogen ow. The gases produced upon pyrolysis are cooled, washed, and evacuated. The activation cycles were carried out in a quartz xed bed reactor heated by an electrical furnace (Figure 1b). The bulb where the sample is placed is 20 mm 50 mm. The gas ow was introduced by a mass ow controller through the bottom of the reactor after preheating in the lower part of the furnace. The sample was supported on a sintered porous quartz plate to avoid channelling. The temperature was controlled thanks to two thermocouples, one of them inserted in the bed and the other one placed close to the furnace wall. Waste gases were cooled, H2O-scrubbed, and evacuated. Preparation of Char. Waste rubber was separated mechanically from the tread strip of Pirelli P2000 waste tires. After cryogenic grinding in liquid nitrogen and sieving, a fraction of 1-2 mm particle size was selected as starting material to produce char. The char was obtained by isothermal fast pyrolysis

4666

Ind. Eng. Chem. Res., Vol. 48, No. 10, 2009

Figure 4. TGA desorption patterns at 550 and 750 C. Figure 3. TPD curves of char after oxygen chemisorption.

at 800 C under 100 NmL/min of nitrogen ow and it was held at that temperature for 20 min. In each pyrolysis run, 1-2 g of waste tires rubber was placed in the cup. These conditions were selected according to the results obtained in previous works focused on the optimization of waste tires carbonization.11 The yield was 41% and the elemental composition of the char obtained is shown in Table 1. Activation of Char. In each experiment, 0.5 g of char was placed into the xed bed reactor and heated under 100 NmL/ min nitrogen ow up to 210 C; the heating rate up to this temperature was 15 C/min. Then the nitrogen ow was switched to an 80 NmL/min air ow which was maintained during 3 h to enable oxygen chemisorption. Afterward, the ow was switched again to nitrogen and the sample was heated, using the same heating rate, to the desired desorption temperature (550 or 750 C) and maintained for 2 h. After completion of each cycle, the sample was weighted and its porous structure analyzed. Characterization of Samples. Oxygen chemisorption was studied by temperature programmed desorption (TPD) and thermogravimetric analysis (TGA). TPD was carried out by heating the sample up to 900 C in He ow at a heating rate of 10 C/min and analyzing the CO and CO2 evolved with a nondispersive infrared absorption analyzer (Siemens Ultramat 22). The TGA system consisted of a quartz tube where a quartz sample pot was suspended from a CI Instruments electronic head balance. The SBET and the pore size distribution or the samples were determined from nitrogen adsorption-desorption at 77 K using a Quantachrome Autosorb I apparatus. The SBET was calculated by a multipoint method while the t-method was used to calculate the micropore volume and external or nonmicropore area. Scanning electron microscopy (SEM) was performed in a Hitachi S-3000N device. Results and Discussion TGA Study of the Oxygen ChemisorptionDesorption. In a rst approach a series of chemisorption/desorption experiments were carried out to establish the most adequate temperatures for the activation cycles. Oxygen chemisorption and desorption rates are known to increase with temperature. Chemisorption can take place from temperatures as low as 200 C, whereas desorption can be already signicant at such temperature.12 The maximum net uptake of oxygen for a number of carbonaceous materials has been found to take place between 200 and 350

C.10 Figure 2 shows the TGA curves corresponding to the chemisorption of oxygen by the char between 200 and 250 C, where an increase can be observed during the rst three hours and beyond that time no signicant change can be appreciated. The stabilization of the weight of the sample suggests a balance between sorption and desorption. Therefore, the chemisorption time was set to 3 h. A weight increase of 0.32, 0.65, 0.41, and 0.23% (initial weight basis) can be observed for the runs carried out at 200, 210, 230, and 250 C, respectively. The increase in weight during the chemisorption is rather low (<1%), therefore the burnoff expected in each cycle is also low. At 210 C the balance between chemisorption and desorption leads to the maximum net weight gain, which is convenient in order to achieve a burnoff as high as possible in each activation cycle. This result is in agreement with a previous work by Py et al.10 that reported a temperature of 200 C as the most convenient to perform oxidation without combustion for several carbonaceous materials. Attending to the TPD curves (Figure 3), the amount of CO2 desorbed up to 850 C is similar in all the cases, except when the chemisorption experiment was carried out at 200 C, where the lower desorption of CO2 is consequent with the lower uptake of oxygen observed. For chemisorption temperatures of 230 and 250 C, the desorption pattern of CO2 is similar, that is, desorption becomes important beyond 300 C and becomes maximum around 550 C. In the case of chemisorption at 210 C, no signicant desorption was observed up to 400 C, showing two CO2 desorption peaks at 525 and 625 C approximately; besides, a signicantly higher amount of CO2 was evolved beyond 700 C in this case. On the basis of the TPD proles, two desorption temperatures, 550 and 750 C, were selected for the study of the activation cycles. A heat treatment of the oxidized char at 550 C would provide the evolution of most of the rst CO2 peak, whereas at 750 C it would cover almost completely both the rst and second peak. The amount of oxygen evolved as CO in TPD was found to be almost negligible below about 850 C. Figure 4 shows the TGA curves in nitrogen atmosphere corresponding to desorption step at 550 and 750 C for a char oxidized at 210 C. Although desorption is much slower at 550 C, at both temperatures most of the weight loss takes place within the rst four hours. Thus, a 4 h time was set for the desorption step. Activation by Cyclic Oxygen Chemisorption-Desorption. Figure 5 shows the evolution of burnoff with the number of cycles for the set of experiments carried out at

Ind. Eng. Chem. Res., Vol. 48, No. 10, 2009

4667

Figure 5. Burn-off evolution upon cyclic chemisorption-desorption.

Figure 6. SBET vs cycle number for 550 and 750 C desorption temperature.

desorption temperatures of 550 and 750 C. In all cases the chemisorption step was carried out at 210 C. As it can be seen, the burnoff shows a nearly linear dependence on the number of cycles applied to the char, excluding the rst cycle which yields a higher burnoff. For the rest, a linear t provided always values of r above 0.99. From the slope a variation of 1.09 and 1.41 points of percent burnoff per cycle can be deduced at 550 and 750 C desorption temperatures, respectively. Thus, the chemisorbed oxygen evolved as CO2 at temperatures higher than 750 C in TPD (Figure 3) does not seem to build up on the char surface blocking the reactive sites. Therefore, successive activation cycles do not signicantly affect the reactivity of the char. The nal burnoff values achieved after 15 activation cycles were 17.4 and 22.3% for the samples treated at desorption temperatures of 550 and 750 C, respectively. These burn-offs are equivalent to that reported by Py et al.10 for the activation of coal and pyrolyzed tar pitch by cyclic adsorption/desorption of oxygen from air ow. Figure 6 shows the evolution of SBET with the number of cycles applied. The highest increase is observed up to the fourth or fth cycle an then a lower but monotonical increase is maintained. During the rst three cycles the samples treated at 550 and 750 C exhibit a fairly similar behavior, but beyond

that point the development of SBET area is signicantly higher for the samples treated at 750 C desorption temperature, which is in accordance with the higher burnoff achieved. After 15 cycles, values of 156 and 239 m2/g were obtained at 550 and 750 C, respectively. These results are rather low when compared with those reported by Mui et al.1 for the activation of waste tires char with steam and CO2 (up to 1200 m2/g), although fairly different conditions were used leading to much higher burnoff values (up to 80%) and consequently substantially lower nal yields. Helleur et al.2 studied the preparation of activated carbons from steam and CO2-gasication of waste tires char reporting SBET values between 200 and 300 m2/g at burnoff values of 13 to 38%, whereas negligible SBET values were obtained when using 2% O2 at 725-750 C up to 10-15% burnoff. At high temperature, the high reactivity of oxygen led to uncontrolled oxidative destruction of the char particles without signicant development of porosity. This problem can be avoided through the cyclic oxygen chemisorption-desorption process tested in this work which provides a way of controlling the activation process. Figure 7 shows the evolution of the micropore and narrow mesopore (2-8 nm) volumes with the number of cycles. In the rst cycles the development of surface area is associated almost

4668

Ind. Eng. Chem. Res., Vol. 48, No. 10, 2009

Figure 7. Micropore and narrow mesopore volumes vs cycle number.

Figure 8. Evolution of the mean pore size with the number of cycles at 550 and 750 C desorption temperatures.

exclusively to mesoporosity for the two desorption temperatures tested. A maximum development of narrow mesoporosity is observed after the fth cycle regardless the desorption temperature. Thus, a decrease of mesopore volume within this size range (2-8 nm) is noticed, most probably as a result of mesopore widening. The microporosity increases with the number of cycles applied and the values obtained after fteen cycles are 0.035 and 0.060 cm3/g for a desorption temperature of 550 and 750 C, respectively. These values are substantially lower than those reported by Py et al.10 for the activation of other materials by cyclic adsorption/desorption (0.2-0.6 cm3/ g). These differences in surface area development can be due mostly to the starting materials used by Py et al., where the initial microporosity leads to a higher oxygen chemisorption enhancing pore development. The starting char is composed by the original carbon black, the carbon generated from rubber, and some deposited pyrolytic carbon. Since these carbons may have different reactivity, the low increase observed in microporosity could be partly due to partial regeneration of the

active surface associated to the type of carbon gasied preferentially in the rst activation cycles, which deserves further study to understand pore development control. Figure 8 shows the evolution of the mean pore size with the number of cycles applied. The starting char is essentially mesoporous. As a result of chemisorption-desorption cycles a decrease up to mean pore sizes around 15 nm is observed. During the rst three cycles the mean pore size decreases dramatically, which can be attributed to the generation of narrow mesopores, since micropores only were observed to develop effectively in later cycles. Beyond this point the mean pore size decreases monotonically for a desorption temperature of 750 C due to a development of microporosity. At 550 C desorption temperature the mean pore size is stabilized up to the sixth cycle and then decreases monotonically in successive cycles. The stabilization of SBET from the fourth cycle could be explained as a balance between the beginning of creation of micropores and widening of mesopores.

Ind. Eng. Chem. Res., Vol. 48, No. 10, 2009

4669

Figure 9. External surface area vs cycle number.

Figure 10. SEM images of the treated char after 15 cycles for (a) 550 C and (b) 750 C.

Figure 9 shows the evolution of external or nonmicropore surface area with the number of cycles applied at the two desorption temperatures tested. The evolution shows a fairly similar behavior at both desorption temperatures that is con-

sistent with that observed for the narrow mesopore volume. After an initial development of an external surface a decline can be observed. Such decline takes place from the fourth cycle in the case of desorption at 550 C, and from the ninth cycle at 750

4670

Ind. Eng. Chem. Res., Vol. 48, No. 10, 2009

C. The maximum external surface area obtained is around 120 m2/g for the two desorption temperatures employed. This value is high in comparison with the results reported by Py et al.10 for other materials (5-70 m2/g). Although the development of porosity can be considered comparatively low, the external area achieved is above the average for most commercial activated carbons. This feature, in addition to the granular structure of the resulting activated carbons, makes them interesting candidates for some specic applications, such as catalyst supports particularly in liquidphase processes. It is well known that mesopores allow an adequate dispersion of the metal on the support which results in good catalytic activity.13-15 Gurrath et al.16 reported advantages of mesoporous activated carbons as catalyst supports in liquid-phase cyclohexene and cyclooctene hydrogenation where steric but not diffusional effects could be inuencing catalyst activity. Moreover, Calvo et al.17 found that mesopores were responsible for the good results obtained in 4-clorophenol oxidation using activated carbon supported Pd catalysts. Morphology. Figure 10 shows SEM images, at different magnication, of the samples obtained after 15 cycles at 550 and 750 C desorption temperatures. The images indicate that the morphology of the particle is maintained after that high number of cycles. It can also be seen that the particle is formed by a continuous matrix where the carbon black and ash are embedded in the carbon generated from rubber by primary pyrolysis reactions and in pyrolytic carbon. This structure can provide a higher mechanical strength than those based on agglomerated micrograins. That supports the potential interest of these activated carbons as catalytic supports for liquid-phase reactions. With this application in mind the optimum activation treatment consists of ve chemisorptions-desorption cycles. A desorption temperature of 500 C is preferred since the development of external area is achieved with lower burnoff (7% in contrast to 10% at 750 C), which can contribute to a higher strength of the activated carbon granules. Conclusion Cyclic adsorption-desorption of oxygen is revealed as a promising method for physical activation of waste tires chars under controlled conditions. This method permits a control of the high reactivity of oxygen, even at concentrations as high as those corresponding to air, to produce granular mesoporous activated carbons. With the method proposed it is possible to prepare activated carbons with BET surface area up 220 m2/g at burnoff values lower than 20%, or with external and nonmicropore area up to 120 m2/g at burnoff values as low as 6-9%. In the rst activation cycles the porosity develops mostly by the generation of narrow mesopores, and from cycles 4-6 the microporosity is developed. The different phases of porosity development could be the result of the gasication of carbon with different reactivity and capacity for the regeneration of active surface. As a result of the low burnoff values the particles morphology is maintained during activation and granular

activated carbons are obtained. On the basis of this it can be postulated that the activated carbons prepared are interesting candidates for catalytic supports for liquid-phase applications. Acknowledgment The authors greatly appreciate nancial support from the Spanish Ministerio de Educacion y Ciencia (CTQ2006-13512). Literature Cited
(1) Mui, E. L. K.; Ko, D. C. K.; McKay, G. Production of ActiVe Carbons from Waste Tyres-A reView. Carbon 2004, 42, 2789. (2) Helleur, R.; Popovic, N.; Ikura, M.; Stanciulescu, M.; Liu, D. Characterization and Potential Applications of Pyrolytic Char from Ablative Pyrolysis of Used Tires. J. Anal. Appl. Pyrolysis 2001, 58-59, 813. (3) Arabiourrutia, M.; Lopez, G.; Elordi, G.; Olazar, M.; Aguado, R.; Bilbao, J. Product Distribution Obtained in the Pyrolysis of Tyres in a Conical Spouted Bed Reactor. Chem. Eng. Sci. 2007, 62, 5271. (4) Ko, D. C. K.; Mui, E. L. K.; Lau, K. S. T.; McKay, G. Production of Activated Carbons from Waste TiresProcess Design and Economical Analysis. Waste Manage. 2004, 24 (9), 875. (5) Lin, Y.-R.; Teng, H. Mesoporous Carbons from Waste Tire Char and their Application in Wastwater Discoloration. Microporous Mesoporous Mater. 2002, 54 (1-2), 167. (6) San Miguel, G.; Fowler, G. D.; Sollars, C. J. A Study of the Characteristics of Activated Carbons Produced by Steam and Carbon Dioxide Activation of Waste Tyre Rubber. Carbon 2003, 41, 1009. (7) Roy, C.; Chaala, A.; Darmstadt, H. The Vacuum Pyrolysis of Used Tires. End-Uses for Oil and Carbon Black Products. J. Anal. Appl. Pyrolysis 1999, 51, 201. (8) Cunliffe, A. M.; Williams, P. T. Inuence of Process Conditions on the Rate of Activation of Chars Derived from Pyrolysis of Used Tires. Energy Fuels 1999, 13 (1), 166. (9) Kaminsky, W.; Mennerich, C. Pyrolysis of Synthetic Tire Rubber in a Fluidised-Bed Reactor to Yield 1,3-Butadiene, Styrene and Carbon Black. J. Anal. Appl. Pyrolysis 2001, 58-59, 803. (10) Py, X.; Guillot, A.; Cagnon, B. Activated Carbon Porosity Tailoring by Cyclic Sorption/Decomposition of Molecular Oxygen. Carbon 2003, 41, 1533. (11) Alonso-Morales, N.; Gilarranz, M. A.; Heras, F.; Gonzalez, V.; Rodrguez, J. J. Inuence of Operating Variables on Carbonization of Waste Tyres; 1st European Chemistry Congress, 2006, Bulgary. (12) Figueiredo, J. L.; Pereira, M. F. R.; Freitas, M. M. A.; Orfao, J. J. M. Modication of the Surface Chemistry of Activated Carbons. Carbon 1999, 37, 1379. (13) Radovic, R. L.; Rodriguez-Reinoso, F. Chemistry and Physics of Carbon; Marcel Dekker: New York, 1997; Vol. 25. (14) Moreno-Castilla, C. Carrasco-Martn, F. Cobalt Catalists Supported on Activated Carbons: Preparation and Behavior in the Hydrogenation of Carbon Oxides. J. Chem. Soc., Faraday Trans. 1995, 91, 3519. (15) Rodrguez-Reinoso, F. The Role of Carbon Materials in Hetero geneous Catalysis. Carbon 1998, 36, 3159. (16) Gurrath, M.; Kuretzky, T.; Bohem, H. P.; Okhlpkova, L. B.; Lisitsyn, A. S.; Likholobov, V. A. Palladium Catalysts on activated carbon supports. Inuence of Reduction Temperature, Origin of the Support and Pretreatments of the carbon surface. Carbon 2000, 38, 14241. (17) Calvo, L.; Gilarranz, M. A.; Casas, J. A.; Mohedano, A.; Rodrguez, J. J. Hydrodechlorination of 4-Chlorophenol in Aqueous Phase Using Pd/ AC Catalysts Prepared With Modied Active Carbons Supports. Appl. Catal. B 2006, 67, 68.

ReceiVed for reView November 18, 2008 ReVised manuscript receiVed February 18, 2009 Accepted February 18, 2009 IE801764X