Influence of Surface Chemistry on the Adsorption of Oxygenated

Nov 30, 2010 - Hydrocarbons on Activated Carbons. Camelia Matei Ghimbeu,* Roger Gadiou, Joseph Dentzer, Dominique Schwartz, and. Cathie Vix-Guterl...
0 downloads 0 Views 4MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Influence of Surface Chemistry on the Adsorption of Oxygenated Hydrocarbons on Activated Carbons Camelia Matei Ghimbeu,* Roger Gadiou, Joseph Dentzer, Dominique Schwartz, and Cathie Vix-Guterl Institut de Science des Mat eriaux de Mulhouse, (CNRS LRC 7228), 15 rue Jean Starcky, 68057 Mulhouse, France Received August 26, 2010. Revised Manuscript Received November 8, 2010 The objective of this work was to study the adsorption of different oxygenated hydrocarbons (methanol, ethanol, 1 and 2-butanol, methyl acetate) on activated carbons from organic mixtures with cyclohexane. Three activated carbons prepared by thermal and chemical treatments of a commercial carbon were employed for this purpose. Their textural properties were found to be similar, whereas their surface chemistries were modified, as shown by temperatureprogrammed desorption coupled to mass spectrometry (TPD-MS) and X-ray photoelectron spectroscopy (XPS). The adsorption isotherms were obtained by depletion method, and the analysis of adsorbed species was evaluated by TPDMS to obtain new insight into the interactions between the different hydrocarbons and the carbon surface. Ethanol leads to a high-energy interaction between its hydroxyl function and the oxygenated surface groups and also to a lower energy interaction between the aliphatic part of the molecule and the carbon material. The desorption activation energy for this hydrophilic interaction is high (50 to 105 kJ/mol), and it is related to the nature of the carbon surface groups. The relative importance of these two interactions depend on the size of the alcohol/methanol is similar to ethanol, whereas butanols lead to more dispersive interactions. Methyl-acetate cannot undergo this kind of strong interaction and behaves like cyclohexane, having desorption activation energies ranging between 25 and 45 kJ/mol no matter the molecule and the carbon surface chemistry.

1. Introduction Carbon/carbon composites are widely used in brushes and collectors in electrical motors, ensuring the electrical current to flow from a rotating part to a stationary part.1,2 In applications such as fuel pumps in cars, these motors are in permanent contact with hydrocarbons. Recently, because of the increase in crude oil prices and limited resources of fossil oil, great attention has been paid to the development of biofuels (bioethanol and biodiesel).3 Besides the advantages of using such fuels (biodegradable, nontoxic, better engine performance, higher octane number, environmental and health aspects related to decrease in the level of pollutants and carcinogens),4-6 there are also short- and longterm problems caused by their use (cold weather engine starting, excessive engine wear, etc.).7 Concerning the brushes and the collectors, an unexpected wear was observed in some cases when using biofuels. The differences between the biofuels and the classical fuels can be of physical nature (electrical conductivity, density, viscosity, etc.) or chemical nature (presence of alcohols, esters, etc.). Moreover, there are several different phenomena (electrical, tribological, and chemical)8-11 that can play an *To whom correspondence should be addressed. E-mail: camelia.ghimbeu@ uha.fr. Tel: 33 389608719. (1) Wilk, A.; Moson, I. Wear 2002, 253, 935–945. (2) Zhongliang, H.; Zhenhua, C.; Jintong, X.; Guoyun, D. Wear 2008, 265, 336– 340. (3) Shay, E. G. Biomass Bioenergy 1993, 4, 227–242. (4) Krawczyk, T. Inform 1996, 7, 801–828. (5) Ma, F.; Hanna, M. A. Biores. Technol. 1999, 70, 1–15. (6) Rocha, M. S.; Simoes-Moreira, J. R. Fuel 2005, 84, 447–452. (7) Harwood, H. J. J. Am. Oil Chem. Soc. 1984, 61, 315–324. (8) Yoichi, M.; Satoshi, K.; Shozo, I.; Takeshi, M.; Tetsushi, U. DENSO Tech. Rev. 2002, 7, 84–88. (9) Yasar, I.; Canakci, A.; Arslan, A. Tribol. Intern. 2007, 40, 1381–1386. (10) Hershberger, J.; Woodford, J. B.; Erdemir, A.; Frenske, G. R. Surf. Coat. Technol. 2004, 183, 111–117. (11) McKee, D. W.; Savage, R. H.; Gunnoe, G. Wear 1972, 22, 193–214.

18824 DOI: 10.1021/la103405j

important role in the degradation of the C/C materials. Although there are works about engines working in air,12,13 the studies concerning engines immersed in organic liquids are almost inexistent.14 In the aim of understanding the specific interaction between carbon materials and biofuels, the adsorption of mixtures of oxygenated hydrocarbons like ethanol and methyl acetate in cyclohexane on carbon was performed in this work. A key point is also that the study of the interactions between oxygenated hydrocarbons and carbon material surface are generally performed in aqueous solution,15,16 and there is a lack of information concerning pure organic systems. Taking into account the complex composition of the C/C composites (different type of graphites, binders, lubricants, additives, etc.),1,17 in the view of highlighting some adsorption trends, a model carbon material (activated carbon) was used in this study. To follow the influence of the carbon surface chemistry on the adsorption of oxygenated hydrocarbons, the activated carbon (Norit Cgran) was modified by a thermal treatment (to decrease the amount of surface oxygenated groups) or by chemical treatment (to increase the amount of surface oxygenated groups). To be representative of different types of biofuels, a model mixture of ethanol or methyl acetate in cyclohexane was used with a concentration varying between 0 and 15 vol %. Considering the complex fuel composition containing many types of hydrocarbons and additives, the cyclohexane organic medium (12) Robert, F.; Csapo, E.; Zaidi, H.; Paulmier, D. Int. J. Mach. Tools Manuf. 1995, 35, 259–262. (13) Hu, Z. L.; Chen, Z. H.; Xia, J. T. Wear 2008, 264, 11–17. (14) Yamamoto, T.; Bekki, K.; Sawa, K. In Proceedings on the Forty-First IEEE Holm Conference on Electrical Contacts; IEEE: New York, 1995; pp 323-329 (15) Namasivayam, C.; Kavitha, D. Microchem. J. 2006, 82, 43–48. (16) Ayranci, E.; Duman, O. J. Hazard. Mater. 2006, 136, 542–552. (17) Xia, J.-T.; Hu, Z.-L.; Chen, Z.-H.; Ding, G. Y. Trans. Nonferrous Met. Soc. China 2009, 17, 1379–1384.

Published on Web 11/30/2010

Langmuir 2010, 26(24), 18824–18833

Ghimbeu et al.

was selected in accordance with fuel composition (which contains 30-50% of saturated hydrocarbons).18 The adsorption of oxygenated hydrocarbons (carboxylic acids, esters, alcohols) from organic solvents was extensively studied between 1950 and 1960.19-21 However, the literature on aliphatic alcohol adsorption from organic media is scarce, with most studies being devoted to the separation of alcohols present in water. Studies on methanol or water have shown that the adsorption proceeds at first on specific active sites; then, the adsorption leads to cluster formation on the sorbent surface by specific interactions between adsorbed molecules.22 This is also supported by the fact that the adsorption differential enthalpy is similar to the energy of condensation (ΔHcond = -42.2 kJ/mol for ethanol).23 Much less work has been devoted to the influence of the aliphatic part of the molecule on the adsorption.24 This influence may proceed through two phenomena: dispersive interactions with the carbon surface and specific effects on the formation of clusters of molecules (e.g., steric effects). Moreover, these phenomena can be greatly influenced by the sorbent textural and surface chemistry properties and also by the sorbate physical properties (molecular size, polarity, structure, etc.).25 In this context, the adsorption of several molecules (methanol, ethanol, 1- and 2-butanol, and methyl acetate) from a cyclohexane binary mixture on activated carbon with different surface chemistries was performed in this work. Therefore, two types of phenomena should be taken into account: physical adsorption of the alcohol or the ester into the carbon porosity and specific chemical interactions between the carbon surface functional groups and the alcohol or the ester. In this work, to investigate these different influences, we have studied the chemical interactions between model carbon materials and esters or alcohols by temperatureprogrammed desorption coupled to mass spectrometry analysis (TPD-MS). This method has been successfully used to analyze molecules adsorbed on carbon surfaces.26 In addition, we determined the total quantity of adsorbed species by analyzing the decrease in the concentration in the solution by gas chromatography (GC).

2. Experimental Section 2.1. Modification of Surface Chemistry of Carbon Materials. For the adsorption studies, the activated carbon Norit CGran (Norit, Amersfoort, The Netherlands) was used as the pristine carbon material. It is a granular carbon with particle size ranging from 0.5 to 2 mm that is prepared from wood by chemical activation process with phosphoric acid. To study the effect of the amount of surface functional groups in the adsorption process, the carbon surface was oxidized (CGran-ox) and reduced (Cgranred): (1) For oxidation process, the liquid phase oxidation with H2O2 was selected because it allowed us to add oxygenated functions while maintaining the carbon textural properties nearly unchanged.27 For a typical batch, 10 g of carbon was treated at room temperature with 50 mL of H2O2 (30 vol %) under stirring for 3 h. This step was followed by washing the carbon with distillate water until the pH became neutral. Finally, the carbon (18) Perry, R.; Gee, I. L. Sci. Total Environ. 1995, 168, 149–156. (19) Kipling, J. J. Adsorption from solutions of non-electrolytes; Academic Press: New York, 1965. (20) Gasser, G. G.; Kipling, J. J. J. Phys. Chem. 1960, 64, 710–715. (21) Kipling, J. J.; Wright, E. H. M. J .Phys. Chem. 1963, 67, 1789–1793. (22) Gun’ko, V.; Bandosz, T. J. Phys. Chem. Chem. Phys. 2003, 5, 2096–2103. (23) Naono, H.; Hakuman, M.; Shimoda, M.; Nakai, K.; Kondo, S. J. Colloid Interface Sci. 1996, 182, 230–238. (24) Wright, E. H. M. J. Colloids Interface Sci. 1967, 24, 180–184. (25) Wang, C.-M.; Huang, C.-M.; Wu, H. J. Chem. Eng. Data 2005, 50, 811–816. (26) Gadiou, R.; dos santos, E.; Vijayaraj, M.; Anselme, K.; Dentzer, J.; Soares, G.; Vix-Guterl, C. J. Colloids Surf. B 2009, 73, 168–174. (27) Figueiredo, J. L.; Pereira, M. F. R.; Freitas, M. M. A.; Orfao, J. J. M. Carbon 1999, 37, 1379–1389.

Langmuir 2010, 26(24), 18824–18833

Article was dried in an oven in air stream at 110 °C for 24 h. (2) To reduce the surface chemical groups, 10 g of activated carbon was placed in a fused silica tubular reactor and heat-treated to 600 °C under H2 flow and kept 1 h at this temperature. The material was cooled to room temperature under the same atmosphere; after that, a flow of air was introduced into the reactor and kept for 1 h. 2.2. Material Characterizations. The textural properties of the materials were investigated from the adsorption isotherms of N2 at 77 K and CO2 at 273 K with an Autosorb 1-LP (Quantachrome) and a Micromeritics ASAP 2010 instruments. Prior to the analysis, the samples were outgassed overnight in vacuum at 300 °C. The total pore volume, Vp, was obtained from the amount of N2 adsorbed at a relative pressure of 0.95. The BET surface area, SBET, was calculated in the relative pressure range of 0.02 to 0.3. The Dubinin-Radushkevich equation was used on nitrogen and carbon dioxide isotherms to obtain the volumes of micropores, Vmicro, and of ultramicropores, Vultra, respectively. The mesopore volume Vmeso was obtained by Vp - Vmicro. The pore size distribution was determined using the QSDFT model for nitrogen adsorption in carbon slit pores.28 The morphology of the materials was examined by scanning electron microscopy (FEI Quanta 400) equipped with an energy dispersive spectrometer, which allows us to determine the elemental mapping of the samples. XPS spectra were obtained on a SCIENTA 200 X-ray photoelectron spectrometer equipped with a conventional hemispherical analyzer. The latter was operated at constant pass energy of 100 eV in the fixed transmission mode. The incident radiation used was a monochromatic Al KR (1486, 6 eV) operated at 420 W (14 kV; 30 mA). Analysis was performed using a takeoff angle of 90°, and the base pressure in the analysis chamber was ∼10-9 mbar. The analyzed surface area was ∼3 mm2. The spectrometer energy scale was calibrated to the Ag 3d5/2, Au 4f7/2, and Cu 2p3/ 2 core level peaks set, respectively, at a binding energy of 368.2, 84.0, and 932.7 eV. The survey and multiregions spectra were recorded for C1s, O1s and P2p photoelectron peaks. CasaXPS software was used to deconvolute the peaks using a Tougaard baseline type. The models used to decompose the C1s and P2p peaks were those proposed by Desimoni et al.29 and Puziy et al.,30 respectively. The O1s region was decomposed, as suggested by Zhou et al.;31 in addition, two components corresponding to P-O and PdO contributions were used. To evaluate the surface chemistry of the materials and to study the desorption of alcohols or esters, we used a homemade TPDMS setup. The samples were placed in a quartz tube in a furnace and heat-treated with a linear heating rate in vacuum. For surface chemistry determination, a heating rate of 2 °C/min in the temperature range 25-950 °C was selected, whereas for organic molecules desorption experiments, several heating rates were applied (0.5 to 8 °C/min) in the temperature range 25-450 °C. The amount of carbon used was ∼15 mg. The gases evolved during the heating process were continuously analyzed quantitatively by a mass spectrometer. Before experiments, the mass spectrometer was calibrated using N2 (m/z = 28), H2 (m/z = 2), CO (m/z = 28), CO2 (m/z = 44), O2 (m/z = 32), CH3-COOCH3 (m/z = 43), CH3-OH (m/z = 31), C2H5OH (m/z =31), 1-C4H9-OH (m/z = 31), 2-C4H9-OH (m/z = 45), and C6H12 (m/z = 56). The total pressure of the gas released during the heat treatment was measured as a function of the temperature using a Bayard-Alpert gauge. This total pressure was then compared with the sum of the partial pressure of the species deduced from the quantitative analysis by mass spectrometry. The total amount (28) Radovic, L.; Castilla-Moreno, C.; Rivera-Utrilla, J. Carbon Materials As Adsorbents in Aqueous Solutions. In Chemistry and Physics of Carbon; Radovic, L. R., Ed.; Marcel Dekker: New York, 2001; Vol. 27, pp 227-405. (29) Desimoni, E.; Casell, G. I.; Salvi, A. M. Carbon 1992, 30, 521–526. (30) Puziy, A. M.; Poddubnaya, O. I.; Socha, R. P.; Gurgul, J.; Wisniewski, M. Carbon 2008, 46, 2113–2123. (31) Zhou, J. H.; Sui, Z. J.; Zhu, J.; Li, P.; Chen, D.; Dai, Y. C.; Yuan, W. K. Carbon 2007, 45, 785–796.

DOI: 10.1021/la103405j

18825

Article

Ghimbeu et al.

of each gas released was computed by time integration of the TPD curves. 2.3. Adsorption Measurements. The adsorption of different oxygenated hydrocarbons from their mixtures with cyclohexane onto carbon materials was performed using a conventional bath equilibrium technique. In a first step, the adsorption of ethanol (C2H5-OH) and methyl acetate (CH3-COOCH3)/cyclohexane (C6H12) mixtures on activated carbons (CGran, CGran-red, and CGran-ox) was studied and compared. The concentration of ethanol and methyl acetate was varied between 0.1 and 15 vol %. In a second step, the influence of the adsorption of some alcohols with different hydrocarbon chain lengths from their binary mixture with cyclohexane on CGran-red was evaluated. The CGranred carbon was selected for this study considering its hydrophobic nature compared with the other carbons. The following alcohols were used: methanol (CH3-OH), ethanol (C2H5-OH), and 1- and 2-butanol (1 and 2-C4H9-OH) with concentration of 1 vol % in cyclohexane. The chemicals with high purity (99.5 to 99.9%) were purchased from Sigma-Aldrich and Fluka. Prior to the adsorption measurements, the samples were outgassed under vacuum at 150 °C for 24 h. Typically, 100 mg of activated carbon was placed in glass flasks (25 mL) in which 5 mL of binary liquid mixture of oxygenated hydrocarbon in cyclohexane with a known concentration was added. The flasks were shaken in a thermostatted bath at room-temperature for 2 days to reach the adsorption equilibrium. The hydrocarbon concentration at equilibrium was measured with a GC (Hewlett-Packard 5890 Series II Plus) having a capillary column (length 60 m, internal diameter 0.25 mm, stationary phase CP-1301, Varian) and a flame ionization detector (FID). The solutions were analyzed under isothermal conditions (200 °C), and the carrier gas used was high purity nitrogen. The amount of solute adsorbed was calculated according to the eq 132 qe ¼ V ðC 0 - C e Þ=m

ð1Þ

where V is the volume of the solution, C0 and Ce are the initial and the equilibrium concentration, and m is the weight of absorbent.

3. Results and Discussion 3.1. Material Characterizations. The nitrogen adsorption/ desorption isotherms on carbon materials presented in Figure 1 show that they are of type IV. The hysteresis loop (P/P0 = 0.42 to 1.00) indicates the presence of mesopores. Furthermore, a significant number of micropores are also characterizing the carbon materials. Comparing the carbon textural properties gathered in Table 1, it can be seen that the CGran and CGran-ox have very similar textural properties, whereas for CGran-red, a slight increase in the microporosity and a decrease in the specific surface area can be remarked. Nevertheless, the values of the textural properties do not differ by >7%. Therefore, it can be assumed that the surface chemistry of the carbon material has been changed without any significant modification of the textural properties. The pore size distributions of the carbon materials (Figure 2) and the values of Vmicro and Vultra (Table 1) show that they exhibit a regular distribution of pore over the complete range of micropore sizes, that is, between 5 and 20 A˚. If we compare with the size of the adsorbates, then we can assume that most of these pores should be accessible to the adsorbates. The modification of the carbon structure due to the thermal and chemical treatments was studied by XRD (Phillips Xpert 2000) and Raman spectroscopy (Jobin Yvon Labram), and no significant difference between their structures has been remarked (data not shown). An important aspect of the carbons that have been activated with phosphoric acid is the residual presence of phosphor on their (32) Liu, C.; Liang, X.; Liu, X.; Wang, Q.; Teng, N.; Zhan, L.; Zhang, R.; Qiao, W.; Ling, L. Appl. Surf. Sci. 2008, 254, 1659–2665.

18826 DOI: 10.1021/la103405j

Figure 1. Adsorption-desorption isotherms of N2 at 77 K for CGran, CGran-red, and CGran-ox. (The isotherms of CGran and CGran-red have been shifted by 600 and 300 cm3/g, respectively.) Table 1. Textural Properties of Carbon Materials material

SBET (m2/g)a

Vmicro (cm3/g)b

Vp (cm3/g)c

Vmeso (cm3/g)d

Vultra (cm3/g)e

CGran 1460 0.55 1.18 0.63 0.24 CGran-red 1380 0.52 1.12 0.60 0.25 CGran-ox 1480 0.55 1.23 0.68 0.24 a SBET calculated in the relative pressure region P/P0 = 0.02 to 0.30. b Vmicro: micropore volume determined from DR equation using N2 isotherms. c Vp: total pore volume determined at P/P0 = 0.95. d Vmeso: mesopore volume obtained by subtracting the micropore volume from the total pore volume (Vp - Vmicro). e Vultra: ultra-micropore volume determined from DR equation using CO2 isotherms.

surface. The concentration of phosphor determined from EDS and XPS spectra is presented in Table 2. The differences between the quantities determined by EDS and XPS are due to the fact that XPS analyzes only the first 10 nm of the material surface, where the most of the impurities are present. It has been shown that the capacity of adsorption of carbon materials can be greatly influenced by the presence of impurities,32 which can play the role of supplementary site for the adsorption. The decrease in the concentration of phosphor for CGran-ox is due to the washing of the sample after the chemical treatment by H2O2. The sample CGran-red was exposed to only gaseous hydrogen and was not washed. The morphology of the materials (SEM images in Figure 3) shows that CGran exhibits some asperities on the surface with sizes of ∼1 μm. In the case of CGran-red, the size of these aggregates seems to increase, maybe because of the thermal treatment. On the contrary, it can be seen that these aggregates disappeared after the chemical treatment (CGran-ox). This is again related to the washing of the sample. The nature and the surface concentration of oxygenated functions were determined by TPD-MS. In such an experiment, the oxygenated surface groups decomposed and lead to the formation of CO and CO2; the nature and the release temperature give indication of their nature. CO2 arises from groups such as Langmuir 2010, 26(24), 18824–18833

Ghimbeu et al.

Article

Figure 2. Pore size distributions of CGran, CGran-red, and CGran-ox materials obtained by QSDFT from nitrogen adsorption isotherm at 77 K. Table 2. Analysis of Phosphor in Pristine and Modified Carbon Materials As Determined by EDS and XPS Techniques material

CGran-red CGran CGran-ox

phosphor (mol %) EDS

XPS

0.70 0.65 0.45

1.95 0.72 0.56

carboxylic acids, anhydrides, and lactones, whereas CO derives from groups such as phenol, carbonyl, or quinone.27 Figure 4 shows the carbon desorption profiles of H2, H2O, CO, and CO2. The quantities of O2 were found to be negligible for all samples. The H2O profile exhibits two peaks; they correspond to chemisorbed water and to dehydration reactions of neighbor hydroxyl groups. This last process leads to the formation of anhydrides and lactones, which decompose at higher temperature to CO and CO2. The water and CO2 profiles show that these functions are removed by the H2 treatment on CGran-red carbon, together with the carboxyl groups. Because the thermal treatment under hydrogen was done at 600 °C, only functions with a high thermal stability are kept on CGran-red: ethers or quinones, which decompose to CO above 600 °C. Table 3 shows the total amount of oxygen surface groups evolved as CO and CO2. The oxidation treatment with H2O2 has as a result the increase in the amount of oxygenated surface groups. This increase is mainly due to functions that decompose to CO2. We observe an opposite behavior for the carbon heattreated in H2, that is, a decrease in the oxygenated groups. Furthermore, in this case, the CO2 peak is shifted to higher temperatures and the H2O peak has different desorption profile than CGran and CGran-ox because of the thermal treatment. Therefore, it is demonstrated by TPD technique that the chemical and thermal treatment is efficient to modify the surface chemistry, that is, to increase and to decrease the surface oxygenated groups, respectively. The quantities of oxygen determined by XPS are higher than those determined by TPD-MS (Table 3). The XPS technique mainly analyzes the surface of the material (10-15 nm layer thickness), whereas TPD-MS is an analysis of the bulk material. This shows that the carbon is more oxidized at the surface and that the oxygen is located in surface functional groups, in good agreement with other reported data.33 The amount of oxygenated surface groups, determined by XPS analysis of the O1s region for the three carbon materials, (33) Moreno-Castilla, C.; Lopez-Ramon, M. V.; Carrasco-Marin, F. Carbon 2000, 38, 1995–2001.

Langmuir 2010, 26(24), 18824–18833

is presented in Figure 5. These results confirm that the Cgran and Cgran-ox carbon materials have similar surface chemistries. In agreement with the TPD-MS experiments, XPS analysis also shows that Cgran-red exhibits a poor surface chemistry, but a significant amount of ether groups are present compared with the two other carbon materials. Another difference between this carbon and the two others is that the ratio of double over single P-O bond has changed. This shows that the oxidation level of the small amount of phosphor present on the surface has decreased during the hydrogen treatment. In the first part of this Article, it has been shown that the thermal and chemical treatments of the activated carbon CGran did not induce important textural, morphological, and structural modification, but on the contrary, significant changes are observed on the surface chemistry of the materials. Hence, it is expected that the adsorption capacity depend on this characteristic. 3.2. Adsorption Measurements. The adsorption isotherms of ethanol and methyl acetate adsorbed onto the surface of the carbons are presented in Figure 6. These isotherms were measured by the depletion method; then, they correspond to the total amount adsorbed. They have been normalized by the specific surface area of the materials. The error increases for the highest equilibrium concentrations because the decrease in the initial concentration due to adsorption becomes difficult to analyze quantitatively. As a general trend, CGran-ox exhibits the highest adsorption capacity for the two adsorbates, whereas CGran-red leads to the smallest adsorption capacity. Therefore, a correlation between the surface chemistry and the adsorption capacity can be remarked; that is, the higher the amount of surface oxygenate groups, the higher the adsorption capacity. This result is in good agreement with other literature reports.34-36 The adsorption capacity is also influenced by the type of adsorbent molecule. The adsorption capacity for ethanol is higher than that for methyl acetate. This behavior can be due to the different functional group (alcohol, ester) interaction with the carbon surface. Furthermore, the ethanol is a more polar molecule than methyl acetate if we consider the dielectric constant of 25 for ethanol compared with that of 6.7 for methyl acetate.37 Consequently, the better polarity of ethanol ensures a better mobility of this molecule in the carbon porosity.38 Also, the size of the molecule can be important, (34) El-Sayed, Y.; Bandosz, T. J. J. Colloid Interface Sci. 2001, 242, 44–51. (35) Mangun, C. L.; Benak, K. R.; Daley, M. A.; Economy, J. Chem. Mater. 1999, 11, 3476–3483. (36) Silvestre-Albero, A.; Silvestre-Albero, J.; Sepulveda-Escribano, A.; Rodriguez-Reinoso, F. Microporous Mesoporous Mater. 2009, 120, 62–68. (37) Shirke, R. M.; Chaudhari, A.; More, N. M.; Patil, P. B. J. Chem. Eng. Data 2000, 45, 917–919. (38) Rychlicki, G.; Terzyk, A. P. J. Therm. Anal. 1998, 54, 343–350.

DOI: 10.1021/la103405j

18827

Article

Ghimbeu et al.

Figure 3. SEM images of (a) CGran, (b) CGran-red, and (c) CGran-ox.

Figure 4. TPD desorption profiles for the pristine and modified carbons. Table 3. Amounts of CO and CO2 released, Obtained by Integration of the TPD-MS Peaks for CGran, CGran-red, and CGran-ox material

CO (mmol/g)

CO2 (mmol/g)

O total (mmol/g)

TPD-MS (mol %)

XPS (mol %)

CGran-red CGran CGran-ox

1.06 1.47 2.41

0.15 0.36 0.80

1.37 2.19 4.02

1.6 2.6 4.8

6.1 11.3 12.2

ethanol being a smaller molecule (4.5 A˚)39 than methyl acetate (6 A˚),40,41 and hence the access of these molecules to the carbon ultramicropores can be blocked when their sizes are superior to that of carbon ultramicropores (100 kJ/mol, whereas the values for carboxyl groups are close to 60 kJ/mol. The peak 2 of ethanol desorption therefore seems to be characteristic of the specific interaction between the hydroxyl function of the alcohol and the surface groups of the carbon materials.

4. Conclusions The adsorption of oxygenated hydrocarbons from their binary mixture with cyclohexane on different carbons was studied in this work. The carbon materials were obtained by modification of a pristine activated carbon, and they exhibit different surface chemistries while keeping exactly the same textural properties. The adsorbed amounts were analyzed by conventional depletion method, and the isotherms obtained fitted well with a LangmuirFreundlich model. The affinity of ethanol with the surface was significantly higher than that for the other molecules, and it increased with the increase in the amount of oxygenated surface groups. The adsorption capacities were more related to the

Langmuir 2010, 26(24), 18824–18833

textural properties of the carbon materials, and they were therefore similar for the different carbon materials. In addition, direct analysis of the adsorbed molecules was assessed by temperatureprogrammed desorption with quantitative mass spectrometry analysis (TPD-MS). It was observed that ethanol undergoes two different interactions with the carbon surface: the first one is similar to the one observed for cyclohexane and methyl acetate (low activation energy: 25-45 kJ/mol) and is enhanced for hydrophobic surfaces, whereas the second interaction is specific only to alcohols (high activation energy: 50-105 kJ/mol) and is related to specific interaction of ethanol with the carbon oxygenated surface groups. Similar experiments performed with other alcohols showed that methanol has a similar behavior as ethanol, whereas 1- and 2-butanol lead to an increase in dispersive interactions and a slight decrease in hydrophilic interactions compared with the former ones. Acknowledgment. The French National Research Agency (ANR) is acknowledged for financial support of this project (ANR Materiaux et Procedes, MATETPRO06). Supporting Information Available: TPD-MS desorption profiles of ethanol and cyclohexane adsorbed on the initial and modified activated carbons and desorbed amounts of ethanol and cyclohexane from the carbon materials. This material is available free of charge via the Internet at http:// pubs.acs.org.

DOI: 10.1021/la103405j

18833