12360
J. Phys. Chem. B 2006, 110, 12360-12364
Development of Porosity in a Char during Reaction with Steam or Supercritical Water Miguel Molina-Sabio,† M. Jesu´ s Sa´ nchez-Montero,‡ Juan M. Juarez-Galan,† Francisco Salvador,‡ Francisco Rodrı´guez-Reinoso,*,† and Aurelio Salvador‡ Laboratorio de Materiales AVanzados, Departamento de Quı´mica Inorga´ nica, UniVersidad de Alicante, Apartado 99, E-03080, Alicante, Spain, and Departamento de Quı´mica Fı´sica, Facultad de Quı´mica, UniVersidad de Salamanca, 37008 Salamanca, Spain ReceiVed: March 8, 2006; In Final Form: May 3, 2006
Two series of activated carbon have been prepared by reaction of a char (from olive stones) with supercritical water (SCW) with the objective of studying the effect of temperature and residence time on the development of porosity. The results have been compared with those obtained using the same char but with classical activation with steam. Both procedures develop porosity, but (i) the reaction rate is critical in the development of porosity for steam but not for SCW activation, and (ii) SCW activation produces a larger development of microporosity at low degrees of burnoff, whereas steam produces more meso- and macroporosity. The differences have been explained by assuming that the mechanism for the carbon-water reaction is common but the transport properties of water in the supercritical state are more favorable to facilitate the access of water to the interior of the char particles. In contrast, when steam is used for the activation of the char, the diffusion of the molecules cannot keep up with the chemical rate and, consequently, the reaction is preferentially taking place at the most accessible surface sites, thus facilitating the development of larger pores and the widening of microporosity.
1. Introduction Steam activation is the most common method used to manufacture activated carbon from a given char. The reaction takes place at 800-1000 °C, and the objective is the development of the maximum porosity at a minimum weight loss. For this reason the conditions have to be selected to minimize the external burning of the particles.1 Above 374 °C and 22 MPa water is in a supercritical state, and as such it exhibits properties that are far from those of the liquid state. For instance, it can mix well with nonpolar gases, it can solve many organic compounds (even with some having large molecular dimensions) and it may oxidize organic matter. Since the transport properties are favorablesit exhibits low viscosity and high diffusion coefficientssupercritical water (SCW) can extract components from solids and/or react with them.2 For this reason there is interest in studying the use of SCW for the regeneration and even preparation of activated carbon as initially proposed by Salvador.3 In the case of regeneration research is focused in finding the conditions under which the adsorbed species are desorbed from the spent carbon without loss of carbon and, consequently, without modification of the porosity.4 In the case of manufacture it is important to study the possible positive effect of SCW activation on the development of porosity in case the process can be scaled to industrial production. Although the number of publications concerning the activation of chars with SCW is very small, there is some significant information. Thus, in a study of the gasification of coconut char with SCW it is suggested that the mechanism for the C-H2O * Corresponding author. Phone: +34 96 5903544. Fax: +34 96 5903454. E-mail:
[email protected]. † Universidad de Alicante. ‡ Universidad de Salamanca.
reaction may be the same for steam and for SCW.5 On the other hand, there is a report with a reduced number of data indicating that an increase in pressure of water above the critical point has no effect on the development of porosity of phenolic resin char.6 The influence of activation temperature, particle size, and water flow rate on the development of porosity during activation of carbonized oak wood with SCW has also been recently described.7 However, the literature search did not find any single publication describing a systematic analysis of the development of porosity upon gasification with SCW and the comparison with the activation with steam of the same char. This is the main objective of this work. Additionally, such comparison will contribute to gaining a more in-depth knowledge of the overall reaction scheme for the heterogeneous gas-solid reaction in the particular case of water with a char particle. 2. Experimental Section Crushed and sieved olive stones (particle size 1.0-3.5 mm) were carbonized under a flow of nitrogen (80 mL/min) at 850 °C for 2 h. The char was then activated with either steam or supercritical water. Activation with SCW has been carried out in a continuous flow reactor inside a high-temperature furnace. An HPLC pump (Shimadzu, LC-10AS) was used to circulate water through the reaction chamber. Pressure in the reactor was controlled by a high-pressure valve. More details of the experimental setup can be found in ref 3. For all experiments the flow of water was fixed at 4 mL/min and the pressure was fixed at 29 MPa. Using a fixed temperature of 655 °C, the residence time was changed (1-11 h) to produce a series of activated carbons (series B) with burnoff ranging from 10 to 70%; the actual value was included in the nomenclature of the carbon. For instance, carbon B-10 was prepared by activation of the char up to 10% burnoff.
10.1021/jp0614289 CCC: $33.50 © 2006 American Chemical Society Published on Web 06/03/2006
Steam vs Supercritical Water Activation
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12361
Figure 2. Cumulative pore volume for meso- and macropores (mercury porosimetry) for carbons of series B.
Figure 1. Adsorption isotherms of N2 at -196 °C for carbons of series B.
Four additional activated carbons were prepared by combining the temperature (585-655 °C) and residence time (40-5.2 h) to reach a common burnoff (around 36%). Activation of the same char with a flow of 100 mL/min steam at 800 °C for different periods of time led to series W of activated carbons. The details of their preparation and the evolution of porosity upon activation can be found in refs 8 and 9. The porosity of the activated carbon was evaluated by adsorption of N2 at -196 °C, by adsorption of CO2 at 0 °C (Coulter Omnisorp 610), and by mercury porosimetry (Micromeritics, Autopore 9510). Experimental results were used to calculate the following volumes: (i) The volumes of micropores Vmi(N2) and Vmi(CO2) were calculated by application of the Dubinin-Radushkevich equation to the corresponding adsorption isotherms. A simple comparison of both sets of values will provide information about the homogeneity of the micropore size, since Vmi(N2) corresponds to the volume of all micropores (up to 2 nm width) whereas Vmi(CO2) corresponds to the volume of only narrow micropores (up to 0.7 nm width).1,10 (ii) The volume of mesopores Vme was calculated by subtracting the value for Vmi(N2) from the volume corresponding to a relative pressure in the isotherm of 0.95. (iii) The volume of pores measured by mercury porosimetry, VHg, corresponding to a 7.2 nm-15 µm pore size range (equivalent to 200 MPa), was calculated. (iv) The volume of macropores Vma, the volume of mercury intruded into the carbon up to 30 MPa, equivalent to 50 nm width, was calculated. 3. Results and Discussion 3.1. Evolution of Porosity along Activation with Supercritical Water. Figure 1 shows the adsorption isotherms of N2 at -196 °C for the carbons of series B, prepared with SCW at 655 °C. Although the uptake increases with burnoff, the increase observed up to B-25 is mainly concentrated at low relative pressures (where the micropore filling is taking place), but from B-43 the increase occurs at progressively increasing relative
Figure 3. Evolution of different sets of pore volume along activation with SCW at 655 °C (series B).
pressures (where the adsorption in mesopores takes place by capillary condensation). The pore size distributions for meso- and macropores obtained by mercury porosimetry are given in Figure 2. The plots show a similarity for carbons with low burnoff, thus indicating the important role of the initial porosity of the char. The volume of pores increases with activation, mainly because of the development of pores with increasing dimension. These data, together with those of Figure 1, suggest that the progress of activation produces an increasing widening of the initial micropores into larger size pores. Figure 3 shows the evolution of the volume corresponding to the different pore sizes during activation with SCW. There is an increase in the volume of micropores up to around 25% burnoff; the increase is slower thereafter for the total microporosity and is almost nil for narrow microporosity. At the same time the volumes of mesopores Vme and macropores Vma increase in the whole range of burnoff, but increase faster for medium-to-high degrees of activation. Since Vmi(CO2) and Vmi(N2) are coincident up to 16% burnoff, the data can be interpreted as an initial increase in narrow microporosity, followed by a widening to wider micropores and finally a transformation into larger pores. This information is confirmed by the plots of Figure 2 (mercury porosimetry) since activation
12362 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Molina-Sabio et al.
Figure 4. Arrhenius plots for the reaction rate of char with steam and SCW.
TABLE 1: Data for Carbons Activated with SCW up to 36% Burnoff at Different Temperatures temperature (°C)
reaction rate × 10-6 (s-1)
Vmi(CO2) (cm3/g)
Vmi(N2) (cm3/g)
V0.95(N2) (cm3/g)
585 595 600 655
2.39 3.38 6.64 17.14
0.32 0.32 0.34 0.34
0.44 0.44 0.45 0.44
0.50 0.50 0.49 0.49
develops pores with increasing dimension so that only the volume of macropores increases from 40% burnoff. Consequently, it seems that activation with SCW is taking place mainly at the interior of the char particles and the evolution of the different pore sizes suggests that the large size pores are produced by breakdown-collapse of micropores. In addition to series B, four activated carbons with a common degree of activation, 36% burnoff, have been prepared by activation with SCW at different temperatures and residence times. The results of Table 1 show the almost nil effect of the reaction rate (amount of carbon gasified per time unit per weight of solid) in the porosity for the particular case of SCW activation. This behavior is very different from the one described for the activation with steam.11 3.2. Porosity Developed by Activation with Steam and SCW. Water at 29 MPa and 585-655 °C is a supercritical fluid reacting with the char. A linear relationship between the weight loss and activation time was observed when plotting the results for the preliminary experiments carried out to define the best way to reach 36% burnoff. The slopes of the plots for the four carbons were used to calculate the reaction rates given in Table 1. Similar values were found when the same char was activated under steam, with the only difference being that the temperature required for the latter was higher (750-900 °C),11 mainly due to the low concentration (pressure) of water compared with the one used for SCW activation. Figure 4 shows that the reaction rate for both activation modes follows an Arrhenius type equation, although the increase with temperature is faster when SCW is used instead of steam. The reaction rate for the activation of the char with SCW deduced from Figure 4 is 1.31 × 105 exp(1.75 × 105/RT) (s-1). The activation energy value, 175 kJ/mol, is somewhat lower than the value found for SCW activation of oak wood char, 212 kJ/ mol.7 The difference is very probably due to differences in morphology since the values for coconut shell char in SCW5 (166 kJ/mol) and in steam12 (176 kJ/mol) are more similar to that for olive stone and lower than that for oak wood char. Furthermore, the above reaction rate for olive stone char is almost coincident with the one proposed by Long and Sykes
Figure 5. Evolution of pore volume with activation for carbons activated with steam and SCW: (a) Vmi(N2); (b) volume of pores larger than 7.5 nm from mercury porosimetry,VHg. A theoretical line has been plotted as reference in each case (see text).
for coconut char reacted with steam12 when the reaction rate reaches a limit, at very high partial pressures for water: 1.00 × 105 exp(1.76 × 105/RT) (s-1). Consequently, it is easy to deduce that (i) the fundamental mechanism of the reaction of the char with SCW cannot be very different from that proposed for the activation with steam and (ii) that the overall reaction in the supercritical state is not influenced by diffusion processes. The reaction rate for the activation of the char with steam deduced from Figure 4 is 36 exp(1.39 × 105/RT) (s-1). Both the preexponential factor and the activation energy are perceptibly lower than for the reaction of SCW with the same char. It seems then that reaction with steam is governed by a join effect of chemical control and transport phenomena. If the diffusion rate of the reactant does not reach the chemically controlled rate, a gradient of concentration of the reactant is produced at the exterior of the char particle and along the micropore, thus affecting the development of porosity. Figure 5 includes the plots for the evolution of the total volume of micropores, Vmi(N2), and the total volume of mesoand macropores detected by mercury porosimetry, VHg, as a function of the degree of activation for the two series of activated carbons, B (activated with SCW) and W (activated with steam). The plots have been limited to around 40% burnoff, since above this value the porosity is so much conditioned by the porosity of the carbon that it is being further activated. A “theoretical” line has also been included as reference. The line in Figure 5a
Steam vs Supercritical Water Activation
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12363
Figure 6. Scheme for different types of carbon atom removal along activation with water.
has been calculated from the volume of micropores of the char, 0.205 cm3/g as deduced from the adsorption of CO2 at 0 °C, and assuming that (i) activation only produces internal burning of the char particle and (ii) the removal of carbon atoms (a real density of 2.0 g/cm3 was taken for the char) only leads to the development of micropores. Consequently, this theoretical line indicates the maximum possible development of microporosity at each burnoff. Since no further development of meso- and macroporosity is assumed to be taking place, the theoretical line for this large porosity (Figure 5b) will show only a slight increase in pore volume accessible to mercury because the weight of the char is being reduced during activation and the data are referred to unit weight. Results for the activation with SCW at low burnoff are near those predicted by the model since the values for the volume of micropores are near the theoretical line (Figure 5a) and there is almost no development of larger pores (Figure 5b). However, activation with steam produces an initial development of large pores (Figure 5b) and a small one for the micropores (Figure 5a). Consequently, activation with SCW at low degrees of activation may be very beneficial since it produces the largest development of microporosity in a char; this is important for many industrial applications of activated carbon. Deviation of the lines for series B from the theoretical plots at degrees of activation larger than 25% burnoff is caused by the conversion/widening of micropores into meso- and macropores. The large increase observed for the volume of micropores in series W is the consequence of increasing the accessibility of the micropores to the water molecules. 3.3. Main Differences between Activation with Steam and and Activation with SCW. Previous work indicates that the fundamental mechanism for the reaction of the char with steam can be applied to a wide range of char properties (porosity and concentration of active sites) and experimental conditions (temperature, pressure, and particle size).13 It is also probable that the mechanism could be valid when water is under supercritical conditions, as suggested by Matsumura et al.5 However, there is a main difference between the conventional activation with steam and activation with SCW, with a direct effect on the development of porosity: the homogeneity of the reaction with the char particle. Supercritical water has a low viscosity and a high diffusion coefficient, which facilitates the penetration to the interior of the char particle, even into the finest pores. The presence of tar (from the pyrolysis of the precursor) causing restrictions or even blocking the spaces between micropores is not an obstacle for
the access of the reactant under supercritical conditions because it can even easily extract such tar. Because the transport properties of SCW are so favorable, the concentration of the reactant at the exterior of the char particle upon activation is not very different from that inside the pores, and consequently, the reaction can take place at the same rate everywhere at the internal and external surfaces of the char. The external surface area (including that of macropores) is very small when compared with the surface area for the micropores. Therefore, the reaction will be mainly internal, producing the loss of the molecular sieving structure14 by progressive widening of the existing pores if there is removal of layers of carbon atoms. There is also an aperture of existing microcapillaries, the development of new interconnectivity, and the creation of new microporosity. A scheme of this type of micropore development is given in Figure 6. Above 40% burnoff the destruction of walls between micropores predominates, thus leading to the formation of mesopores. Under the experimental conditions used for the activation of the char with steam, the structure of the pores will condition the diffusion rate and, consequently, the local concentration of the reactant inside the char particle. If penetration of the reactant into the char is not complete, the reaction will mainly take place at the most accessible surface, thus developing mainly larger pores. Similarly, the concentration of reactant will be larger at the entrance of the pore than inside it since the axial diffusion rate cannot keep up with the chemical rate. The result is the widening of the pores from the beginning of the activation process (Figure 6), and the development of micropores will be faster as the pores become more accessible to the reactant. On the other hand, if the diffusion and reaction rates vary with temperature in a different fashion,13 the development of porosity will be a function of activation temperature, in contrast to the behavior of SCW activation. 4. Conclusions Activation with supercritical water mainly occurs in the interior of the char particles, thus increasing the size of the pores with increasing burnoff. At low degrees of burnoff, activation with SCW leads to the largest development of microporosity in the char without a noticeable increase in the volume of larger pores. However, activation of the same char with steam increases the large porosity, with little development of narrow microporosity. This difference in behavior is mainly caused by the degree of homogeneity of the reaction at the surface (internal and external) of the char.
12364 J. Phys. Chem. B, Vol. 110, No. 25, 2006 Supercritical water, with a very low viscosity and high diffusion coefficient, exhibits a very high degree of penetration to the interior of the char particle, thus facilitating the removal of internal carbon atoms and producing narrow microporosity. However, in the case of activation with steam, the structure of narrow pores present in the char conditions the diffusion rate of water vapor and the reaction will predominantly be at the most accessible surface of the char particle, thus facilitating the formation of macropores and the widening of microporosity. Since the diffusion and reaction rate govern the overall particle reactivity, the porosity developed will be a function of temperature; this is not the case for activation with SCW. Acknowledgment. Financial support from the Ministerio de Educacio´n y Ciencia and Ministerio de Ciencia y Tecnologı´a, Spain (Project No. MAT2004-03480-C02-02 and Project No. PPQ-2000-1249, respectively), is acknowledged. Partial support from the European Network of Excellence “INSIDE POReS” (Contract No. NMP3-CT-2004-500895) is also acknowledged. References and Notes (1) Marsh, H.; Rodrı´guez-Reinoso, F. ActiVated Carbon; Elsevier Science: New York, 2006, in press.
Molina-Sabio et al. (2) Frank, E. U.; Wiegand, G.; Dahmen, N. Walter. In Ullmann’s Encyclopedia of Industrial Chemistry, 5th ed.; Elvers, B., Hawkins, S., Eds.; VCH: Weinheim, 1996; Vol. A28, pp 12-22. (3) Salvador, F. U.S. Patent 6,239,067, 1998. (4) Salvador, F.; Sanchez-Jimenez, C. Carbon 1999, 37 (4), 577-583. (5) Matsumura, Y.; Xu, X.; Antal, M. J., Jr. Carbon 1997, 35 (6), 819824. (6) Cai, Q.; Huang, Z.; Kang, F.; Yang, J. Carbon 2004, 42 (4), 775783. (7) Salvador, F.; Sanchez-Montero, M. J.; Martin-Rodriguez, M. J. To be submitted. (8) Rodriguez-Reinoso, F.; Molina-Sabio, M.; Gonzalez, M. T. Carbon 1995, 33 (1), 15-23. (9) Molina-Sabio, M.; Gonzalez, M. T.; Rodriguez-Reinoso, F.; Sepu´lveda-Escribano, A. Carbon 1996, 34 (4), 505-509. (10) Garrido, J.; Linares-Solano, A.; Martin-Martinez, J. M.; MolinaSabio, M.; Rodriguez-Reinoso, F.; Torregrosa, R. Langmuir 1987, 3, 7681. (11) Gonzalez, M. T.; Molina-Sabio, M.; Rodrı´guez-Reinoso, F. Carbon 1994, 32 (8), 1407-1413. (12) Long, F. J.; Sykes, K. W. Proc. R. Soc. London 1948, A193, 377399. (13) Laurendeau, N. M. Prog. Energy Combust. 1978, 4, 221-270. (14) Gomez-de-Salazar, C.; Sepulveda-Escribano, A.; Rodriguez-Reinoso, F. In Studies in Surface Science and Catalysis; Unger K. K., et al.; Elsevier: Amsterdam, 2000; Vol. 128, pp 303-312.
