10884
Ind. Eng. Chem. Res. 2009, 48, 10884–10891
Adsorption/Reduction of NO2 on Graphite Oxide/Iron Composites Svetlana Bashkova and Teresa J. Bandosz* Department of Chemistry, The City College of New York, 160 ConVent AVenue, New York, New York 10031
Adsorption of NO2 and the retention of NO (the product of NO2 reduction by carbon) on the materials prepared from iron acetato complex and its mixture with graphite oxide were studied. The surface of the materials was characterized using adsorption of nitrogen, X-ray diffraction (XRD), scanning electron microscopy (SEM), Fourier transform infrared (FTIR), and thermogravimetric analysis (TGA). The results showed that the development of texture and porosity along with formation of some active surface species on the graphite oxide/iron composite materials leads to an increase in NO2 adsorption. An immediate reduction of NO2 to NO by carbon was noticed for all samples except for the one that contained a smallest amount of carbon in a mostly oxidized form. The retention of NO on the surface of the materials does not depend on their structural characteristics. Better performance is found for the materials that contain active iron species, such as γ-FeOOH and R-Fe2O3, which likely react with NO molecules forming surface nitrates. 1. Introduction Reduction of nitrogen oxides (NOx) emissions, originating both from the mobile exhausts and stationary power generating sources has become a global issue of concern. Applications of various carbonaceous materials for the removal of NOx have been widely investigated.1-6 The results of these investigations are very promising and show that carbons have capacities for the removal of NOx. The surface modification of activated carbons with different metals further improves their catalytic properties.7-12 Nevertheless, the problem still exists in an almost simultaneous adsorption of NO2 and its reduction to NO by a carbon surface with the formation of CO and CO2.2,3,5,6 Thus, Kaneko7 reported the adsorption of NO on the activated carbon fiber with highly dispersed γ-FeOOH at room temperature. He proposed the mechanism, where NO molecules are first chemisorbed on γ-FeOOH and deposited near the entrances of the micropores, and then, they move into the micropore to form a dimer. Illan-Gomez and co-workers studied the reduction of NO by activated carbons impregnated with transition metals,8,9 potassium,10 and calcium.12 They found that, among transition metals, iron, cobalt, and nickel were the most effective catalysts for NO being able to chemisorb it dissociatively at low temperatures.8 The same group of scientists also found potassium and iron to be the most effective catalysts in the reaction of NO with carbon, by undergoing an oxidation-reduction cycle.9,10 Carabineiro and co-workers investigated the role of transition metal oxides and their binary mixtures, supported on the activated carbons, on the conversion of NO and N2O.13-17 They found that even at low temperatures NO can be reduced to N2 and N2O on the carbon surfaces. For this process to take place, NO has to be dissociatively adsorbed, provided that the surface has the sites to retain oxygen atoms. Such sites can be related to the presence of metals species. Enhanced adsorption of NO was also found on activated carbon with reduced surfaces on which NO was hypothesized to be adsorbed on dangling carbons and edge sites.18 Other types of materials successfully applied for NOx removal are mineral oxide particles19-21 and zeolites.22-24 A group of materials which has potential for application in electrochemical devices,25 catalysis,26 and adsorption27 is based * To whom correspondence should be addressed. Tel.: 1 212 650 6017. Fax: 212 650 6107. E-mail address:
[email protected].
on graphite oxide (GO). The versatility of GO modifications is linked to its layered structure and relatively well-defined surface chemistry.28,29 The modification process is made possible due to (i) the hydrophobic character of GO, (ii) its high dispersion in water, and (iii) its delamination in alkaline media or alcohols. The restacking of graphite layers is easily achieved and depends on their degree of orientation and on the drying method. A very popular way of GO modification is an introduction of amines or conduction polymers within the interlayer space for the production of nanometer-scale structures, devices, and cathode materials.25 Graphite oxides and their composite materials, such as GO/bentonite,30 GO/aluminum polycation, and GO/ zirconium-aluminum polyoxycation,31 have been shown to be efficient adsorbents of ammonia. Although graphite oxide appeared to be inactive for the adsorption of NO2, its modification with polyaniline,32 and more so, with iron oxide,26 had a positive effect on NO2 adsorption. In the case of iron oxide, the obtained materials were R-Fe2O3- and Fe3O4-pillard graphite. The TEM observations showed that for the latter material the layers of Fe3O4 were sandwiched by the thin graphite sheets and that the mesopores of this material had a hydrophobic nature, which was suggested as a factor contributing to the adsorption activity for NO and NO2 at low pressures.26 The objective of this paper is to study the adsorption of NO2 and the retention of NO, the product of NO2 reduction by carbon, on the graphite oxide-iron composites at ambient conditions. We focus on the investigation of the effects of the chemical nature of these composites, their surface and structural characteristics, and the conditions of their preparation on the performance as adsorbents. 2. Experimental Section 2.1. Materials. The starting material, GO, was synthesized from the commercial synthetic graphite (Sigma-Aldrich) by the Hummers method.33 Trinuclear acetato-hydroxo iron(III) nitrate, [Fe3(OCOCH3)7OH · 2H2O]NO3, was prepared according to the method described by Yamanaka.34 The preparation procedure was the following: Fe(NO3)3 · 9H2O (40.4 g) was added to 25 mL of ethyl alcohol and then 70 mL of acetic anhydride were added in small portions. When the reaction started with an evolution of heat, the mixture was cooled in the ice bath. After all acetic anhydride reacted, the solution was kept in the ice
10.1021/ie901054p CCC: $40.75 2009 American Chemical Society Published on Web 10/09/2009
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
bath until a precipitate was formed. Then, the precipitate was separated by filtration and dried in air. To prepare GO-iron composites, 2 g of GO were dispersed in 100 mL of deionized water and sonicated for 40 min. The GO suspension was made basic (pH ) 9-10) by an addition of 0.1 M NaOH. Two solutions of trinuclear acetato complex of different concentrations were prepared, made basic to pH 9-10 by an addition of 0.1 M NaOH, and mixed with the GO suspensions. Resulting mixtures were stirred for 3 days, centrifuged, and washed with deionized water until the supernatant was clear. Finally, GO-iron gels were dried in air at 50 °C. The first sample was designated “GOFe1”, and it contains 10 mmol of iron complex per 1 g of GO. The second sample was designated “GOFe2”, and it contains 15 mmol of iron complex per 1 g of GO. The trinuclear acetato complex was also prepared without mixing with GO and was designated “Fe”. The GO-iron composites and the trinuclear acetato complex were further heat-treated in air or in nitrogen by raising the temperature (at 10 °C/min) to 300 °C and holding the materials at this temperature for 3 h. The heated mixtures were designated “GOFe1-a”, “GOFe2-a”, and “Fe-a”, where “a” stands for the heat treatment in air, and “GOFe1-n”, “GOFe2-n”, and “Fe-n”, where “n” stands for the heat treatment in nitrogen. After adsorption of NO2 letter “E” was added to the names of the samples. 2.2. Methods. 2.2.1. Evaluation of NO2 Adsorption Capacity. Evaluation of NO2 adsorption capacity was conducted in a laboratory-scale, fixed-bed reactor system. The GO-iron composite materials and iron complex materials were loaded into the glass column, 370 mm long by 9 mm diameter, such that the volume of the carbon bed was about 2 cm3. A dry air with 0.1% of NO2 was passed through the column of an adsorbent. The flow rates were controlled by the flow meters (Cole Palmer). The breakthrough concentration of NO2 and the concentration of NO were monitored using a multiple gas monitor with the electrochemical sensors (RAE Systems, MultiRAE Plus PGM50/5P). The tests were conducted until the concentrations of NO2 and NO reached the sensors’ upper limit values of 20 and 200 ppm, respectively. The experiments were conducted at room temperature and atmospheric pressure with a total gas flow rate of 0.45 L/min. 2.2.2. Surface Area and Pore Size Distribution Measurements. The nitrogen adsorption isotherms on the materials studied were determined at -196 °C using an ASAP 2010 instrument (Micromeritics). The samples were outgassed at 120 °C to a constant vacuum (10-4 torr) immediately prior to the isotherm measurements. The specific surface area (SBET) was calculated from the isotherm data using the Brunauer, Emmet, and Teller (BET) model.35 The total pore volume (Vtot), the mesopore volume (Vmes), and the micropore volume (Vmic) were calculated using the density functional theory approach (DFT).36 2.2.3. Characterization of Carbon Surface Chemistry. The pH of the adsorbent surface was obtained by mixing a sample (0.4 g) with deionized water (20 mL), stirring the suspension overnight to equilibrate, and then measuring the pH of this suspension. A Fourier transform infrared spectroscopy (FTIR) procedure was carried out on a spectrometer (Thermo Electron Corporation, Nicolet 380) using a smart diffuse reflectance mode. The spectrum was collected 16 times and corrected for the background noise. The surface chemistry of the carbons was also analyzed by thermogravimetric analysis (TGA) using TA Instruments thermal analyzer (SDT 2960) with a nitrogen flow rate of 100 cm3/min
10885
and a heating rate of 10 °C/min. The ash content related to iron oxide content was determined by burning the samples in oxygen up to 1000 °C. Scanning electron microscopy (SEM) images were obtained using a Zeiss Supra 55 VP with an accelerating voltage of 15.00 kV. Scanning was performed in situ on a powder sample. SEM images with energy-dispersive X-ray (EDAX) analysis were done at 10 K magnification. X-ray diffraction (XRD) measurements were conducted using a standard powder diffraction procedure. Adsorbents were ground with methanol in a small agate mortar. The mixture was smear-mounted onto the zero-background quartz window of a Phillips specimen holder and allowed to dry in air. Samples were analyzed by Cu KR radiation generated in a Phillips XRG 300 X-ray diffractometer. A quartz standard slide was run to check for instrument wander and to obtain accurate location of 2θ peaks. 3. Results and Discussion The morphologies of the materials studied were investigated using SEM images (Figures 1 and 2). The morphology of GO is well-preserved for the GOFe1 sample with some fragments of iron phase on the surface of the layers. The flakes of GO are better visible for the GOFe1 samples than for the GOFe2 samples and are better preserved for the samples heated in nitrogen than for the samples heated in air. The SEM images of Fe-a and Fe-n (Figure 2E and F) suggest that iron species are distributed in the nonporous glassy carbon phase. The morphologies of these samples remain the same as for GOFe2, where relatively large quantity of iron compound was used. The results of the EDAX analysis along with the pH values of the materials are presented in Table 1. They suggest that iron in the composite materials is present mostly in the form of iron oxides. The most oxidized form of iron (from O/Fe ratio) was found for GOFe2-n, and the most reduced form of iron was found for GOFe2 and GOFe2-a. It is an interesting result, taking into account the oxidizing atmosphere in the case of the latter material. Besides, the latter two samples have a higher Fe/C ratio than the rest of the samples. Regarding the iron samples, Fe-a has a much higher O/Fe ratio than Fe-n. Since the total content of an inorganic matter is over 80% and based on the composition of our materials that ash must consist of iron oxides, only frabrication of iron is seen at the external surface. The content of ash inrceases for the GOFe samples after heat treatment. This increase is the highest for GOFe2-a and is likely related to gasification of GO. The surface pH values indicate that the materials are basic in their nature. The basicity of the materials could be explained by their preparation procedure in which NaOH was added. The XRD diffraction patterns for the materials studied are presented in Figure 3. In general, the materials look quite amorphous, with only two peaks revealed at about 33 and 37°. These peaks represent the crystallographic phase of RFe2O3.26,37-40 The crystallite size of R-Fe2O3 in Fe-a, estimated from the line broadening of the main XRD peak (at 33°), using the Scherrer equation,41 is about 8 nm. The intensities of the diffraction peaks were higher for the GOFe2 sample series than for the GOFe1 sample series. However, the most pronounced R-Fe2O3 phase was found for Fe-a. Additional peaks at about 17 and 27° found only for Fe-n indicate the presence of ferrihydrites, particularly γ-FeOOH.39,40 The FTIR spectra of the composite materials and iron samples are presented in Figure 4. Several bands are revealed on these spectra. The band at about 1570 cm-1 for GOFe composites
10886
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
Figure 1. SEM micrographs for GO (A), GOFe1 (B), and GOFe2 (C).
represents either unoxidized aromatic regions of GO or asymmetric C-O stretch from oxygen surface compounds, like cyclic ethers on GO.28,42 The combination of the bands at 1640-1550 and at 1430-1300 cm-1 found in the spectra of the materials could be assigned to the CdO vibration in the carboxylate ion42-44 and stretching vibrations of the NO2 group in nitro compounds and nitrates44 of iron complex species. The peaks of a different intensity in a range of 1060-1150 cm-1 are present for all materials except for Fe-a. These peaks indicate the presence of γ-FeOOH-lepidocrocite39 and are strongly pronounced for GOFe1 and GOFe1-a. The 795 and 895 cm-1 bands are clearly visible only for Fe-a, Fe-n, and GOFe-2 and are commonly assigned to the in-plane and outof-the plane OH- vibrations in γ-FeOOH.39,45 Low frequency bands, below 700 cm-1, are well-defined only for Fe-a and Fe-n samples and are assigned to the Fe-O stretch in iron oxides.45 The differential thermogravimetric (DTG) curves in nitrogen are presented in Figure 5. The peak at about 100 °C corresponds to the physically adsorbed water, a sharp peak between 150-250 °C for GO represents decomposition of epoxides and surface carboxylic groups,31 and the peaks in the range of 100 to 400 °C likely come from the decomposition of the iron complex. The higher temperature peaks, in the range of 600-800 °C, are possibly due to the reduction of iron oxides by carbon.37,46 The peak at about 380 °C, which is well pronounced for Fe-n, may correspond to water elimination from iron hydroxides and their conversion to oxides.37,39,40 Interestingly, the intensities of the higher temperature peaks were significantly reduced for the samples heated in air. For GOFe1-a only the intensity of the 800 °C peak was reduced, but for GOFe2-a the intensities of all peaks detected at T > 500 °C were significantly reduced. This could be explained by the fact that during the heat-treatment in air the carbon surface was oxidized and less carbon became available for the reduction of the iron species. More so, GOFe2
contains less carbon than GOFe1 due to its higher amount of iron (Table 1). The structural parameters of the samples are collected in Table 2. Both GO31 and Fe complex are nonporous but a combination of them together results in the formation of some pores and in an increase in surface area. GOFe1 has a slightly higher surface area than that of GOFe2. The heat treatment of the former sample, both in air and in nitrogen, also resulted in the higher surface areas than the heat treatment of the latter one. The pore size distributions for the samples studied are collected in Figure 6. They show that the majority of the composite pores are larger than 20 Å; which is also confirmed by the data in Table 2. However, it is noticed from Figure 6 that the GOFe1 and GOFe2 samples have higher volumes of micropores than their heattreated counterparts. It is possible that heating of the materials results in a deflagration and exfoliation of graphene layers and formation of iron oxide particles, some of which may be formed in between the defective graphene sheets.26 Moreover, a partial gasification of graphene layers during heat treatment also takes place, which is confirmed by an increase in the ash content of the samples after the heat treatment (Table 1). As a result of these modifications, the mesoporosity of the materials increases (Table 2). More so, the heating of GO at 350 °C results in a decomposition of carboxylic- and epoxy-type of groups.29,31 These groups are located along the GO layer planes and on the edges of the layers,28,29 and their removal may also increase the mesoporosity of the materials. However, heating of GO in air may result in reoxidation of surface radicals formed by the decomposition of the above-mentioned groups and, as a result, in formation of smaller pores than those formed after heating of the same material in nitrogen (Figure 6). These suggestions could explain the higher surface areas for the materials heated in air than for the materials heated in nitrogen.
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
10887
Figure 2. SEM micrographs for GOFe1-a (A), GOFe2-a (B), GOFe1-n (C), GOFe2-n (D), Fe-a (E), and Fe-n (F). Table 1. Results of EDAX Analysis, the Surface pH Values, and the Ash Content for the Materials Studied sample
pH
ash wt %
C at %
O at %
Fe at %
O/Fe
Fe/C
GOFe1 GOFe1-a GOFe1-n Fe-a GOFe2 GOFe2-a GOFe2-n Fe-n
8.3 9.7 9.9 10.8 9.3 10.3 9.9 11.4
65 82 83
56 51 47 32 48 34 32 61
25 25 27 31 18 22 30 14
17 22 23 7 32 39 14 23
1.5 1.1 1.2 4.4 0.6 0.6 2.1 0.6
0.3 0.4 0.5
70 91 83
0.7 1.2 0.4
The breakthrough curves of NO2 and the concentration curves of NO obtained in dry air and ambient conditions for the samples studied are collected in Figure 7. For the GO-Fe composite materials, the breakthrough time of NO2 was longer on the GOFe1 sample than on the GOFe2 sample, but the retention time of NO on GOFe1 was shorter than that on GOFe2. After the heat treatment in air, the breakthrough of NO2 is observed much earlier for both samples. On the other hand, this modification increased the retention times of NO. Especially
long retention of NO was found on GOFe2-a, where the breakthrough test had to be stopped at the concentration of NO being at 100 ppm due to the fact that the concentration of NO2 at that point well exceeded the sensor limit of 20 ppm (the NO2 sensor was turned off when the concentration reached 20 ppm) and was approaching very high concentration levels. Heating in nitrogen decreased the breakthrough time of the GOFe1 sample even more than the heating of this sample in air. On the contrary, heating the GOFe2 sample in nitrogen resulted in its better NO2 performance than the heating of this sample in air. As for the retention time of NO, the heating of GOFe1 in nitrogen resulted in the same retention time of NO (at 200 ppm) as on the unheated GOFe1 sample and on the sample heated in air. However, the shapes of the concentration curves for these samples are quite different. The retention time of NO on GOFe2-n is shorter than that on GOFe2. For the samples derived from the iron organic complex, NO2 was detected immediately after the breakthrough tests started and reached the limiting concentration of 20 ppm within the
10888
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
Figure 3. XRD patterns for the Fe complex and GOFe composite initial materials.
Figure 5. DTG curves in nitrogen for the initial GO-Fe complex and GOFe composite materials. Table 2. Surface Area and Pore Volumes for the Materials Studied
Figure 4. FTIR spectra of the initial: (A) GOFe1, GOFe1-a, GOFe1-n, (B) GOFe2, GOFe2-a, GOFe2-n, (C) Fe-a, Fe-n materials.
sample
SBET (m2/g)
Vtot (cm3/g)
Vmes (cm3/g)
Vmic (cm3/g)
GO GOFe1 GOFe1-a GOFe1-n GOFe2 GOFe2-a GOFe2-n Fe Fe-a Fe-n
6 248 125 78 221 80 26 0 6 6
0.01 0.15 0.20 0.16 0.15 0.16 0.11 0.00 0.03 0.02
0.01 0.09 0.19 0.16 0.10 0.16 0.11 0.00 0.03 0.02
0.00 0.06 0.01 0.00 0.05 0.00 0.00 0.00 0.00 0.00
first minute of adsorption. For this reason, the breakthrough curves for these samples are not shown in Figure 7. The situation was very different for the retention of NO on these materials. All iron samples were able to retain NO on the surface for some time, but the best performance was found for Fe-a. For this
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
10889
Figure 7. Outlet concentrations of NO2 and NO as a function of time for the materials studied. Table 3. Surface pH Values of the Exhausted Samples, the Breakthrough Capacities for NO2, and the Retention Times of NO at 200 ppm
Figure 6. Pore size distributions for the materials studied.
sample, the outlet concentration of NO increased fast during the first 10 min of NO2 adsorption, reaching 57 ppm, and then slowed down, reaching a plateau at about 100 ppm. The breakthrough capacities of NO2 were calculated per volume and per mass of a sample by integration of the area above the NO2 breakthrough curve, taking into account the initial concentration of NO2, total flow rate, breakthrough time, and mass of the sample. The breakthrough time used to calculate the NO2 capacity is defined here as the elapsed time from the beginning of NO2 challenge until the concentration of NO2 in the effluent gas had reached 20 ppm. The breakthrough capacities of NO2 along with the retention times of NO, defined here as the elapsed time from the beginning of NO2 challenge until the concentration of NO had reached 200 ppm, and pH values are presented in Table 3. The highest capacities were found for the initial GOFe samples, with that of GOFe1 having a higher capacity than GOFe2. Among the samples heat-treated in air, GOFe1-a had a much higher capacity of NO2 than GOFe2-a. On the other hand, for the samples heat-treated in nitrogen, GOFe2-n had a higher capacity for NO2 than GOFe1n. The iron samples did not retain any NO2. The longest retention times of NO, as indicated above in the analysis of the
sample
pHE
NO2 breakthrough capacity, mg/cm3
GOFe1 GOFe1-a GOFe1-n GOFe2 GOFe2-a GOFe2-n Fe Fe-a Fe-n
7.2 9.2 9.1 8.7 9.8 9.8 11.3 10.8 11.1
45 19 5 27 1 15 0 0 0
a
NO2 breakthrough capacity, mg/g
NO retention time, min
39 23 4 24 1 12 0 0 0
38 40 37 60 58a 23 9 58a 17
At 100 ppm.
concentration curves, were found on the Fe-a, GOFe2-a, and GOFe2 samples. The small changes in pH of the samples after the adsorption of NO2 showed that the surface products are not of an acidic nature. On the basis of the results presented, it is suggested that the development of texture and the creation of porosity in the process of building GOFe composites is an important factor for the effective adsorption of NO2. Two materials with the highest surface areas, GOFe1 and GOFe2, had the best adsorption capacities for NO2. The same capacity of GOFe1-a (mg/g) as for GOFe2, despite a smaller surface area of the former one, could be explained by the differences in the forms of iron present in these samples. Thus, from the FTIR spectra (Figure 4A and B), it is seen that GOFe1-a has a band in the range of 1060-1150 cm-1, characteristic for γ-FeOOH, that GOFe2 does not have. It is possible that this form of iron participated in the adsorption process of NO2/NO. In fact, the FTIR spectrum of GOFe1-a after the NO2 adsorption (Figure 8A) showed some decrease in the intensity of a 1060-1150 cm-1 band and the
10890
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
Figure 8. FTIR spectra of the exhausted (A) GOFe1, GOFe1-a, GOFe1-n, (B) GOFe2, GOFe2-a, GOFe2-n, (C) Fe-a, and Fe-n materials.
appearance of a 1260 cm-1 band, which may indicate the formation of some surface nitrates/nitrites.19 An increase in the intensity of a 1350 cm-1 band for GOFe1 (Figure 8A), and to some extent for GOFe2 (Figure 8B), could be related to the oxidation of a carbon part of the composite by NO2. Interestingly, the composite material with the lowest surface area, GOFe2-n, has a higher capacity for NO2 than two materials with higher surface areas, GOFe2-a and GOFe1-n. The reason for this could be in the presence of carbonate species in GOFe2-n marked by a 1430 cm-1 band (Figure 4B), characteristic only for this composite’s FTIR spectrum. These species can be formed in reaction of CO2 evolved during the reduction of NO2 by carbon surface with iron nitrates/nitrites. Regarding the retention of NO on the surface of the materials studied, two of them, Fe-a and GOFe2-a, have better capabilities to retain NO than the rest of the samples. It is interesting that the forms of iron in these two samples seem to be quite different. The high O/Fe ratio for Fe-a and the low O/Fe ratio for GOFe2a (Table 1) suggest that in the former one iron is mostly oxidized and in the latter one it is mostly reduced. For GOFe2-a, the presence of GO can contribute to the reduction of iron. A different shape of NO concentration curve for GOFe2-a and a high retention time of NO on this sample (Figure 7) could be attributed to such factors as a high Fe/C ratio (Table 1) and a high oxidation state of carbon. The latter feature is confirmed by the reduced intensity of a 1570 cm-1 band, attributable to the unoxidized aromatic regions of carbon, in the FTIR spectrum of GOFe2-a (Figure 4B) and by the reduced intensities of 600-800 °C DTG peaks (Figure 5), related to the reduction of iron oxides by carbon.37,46 The fact that the outlet concentration of NO for GOFe2-a increased very slowly is related to a low
degree of NO2 reduction to NO by oxidized carbon. The later increase in NO concentration for GOFe2-a is likely due to a gradual occupation of the active iron centers with NO and their further unavailability for the NO molecules. When the iron samples without GO are considered, the oxidized form of iron seems to be preferred by NO molecules over the reduced form of iron. Although there is no carbon from GO, it is the carbon from the heat-treated iron complex that makes the reduction of NO2 to NO on both Fe-a and Fe-n almost immediate (Figure 7). The observed difference between the NO concentration curves for Fe-n and Fe-a is in a different rate of NO release from the surface. Thus, the release of NO from Fe-a was much slower than that from Fe-n, and the concentration of NO for Fe-a reached an equilibrium after about 10 min of NO2 adsorption. It is possible that such species in Fe-a as γ-FeOOH and R-Fe2O3 (confirmed by FTIR and XRD), where iron is present in an oxidized form, play a significant role in the retention of NO. Thus, a 1260 cm-1 band in combination with a 1640 cm-1 band, as in the spectrum of Fe-aE (Figure 8C), could be related to the presence of bridging nitrates.19 More so, the combination of a 1260 cm-1 band with a 1330 cm-1 band, for the same sample, may also indicate the presence of bridging nitrites.19 On the other hand, increase in the intensity of a 1595 cm-1 band for Fe-nE could be related to NO bonding to Fe2O3;47 a higher intensity band at 650 cm-1 for Fe-n than for the other samples indicates the presence of Fe2O3.45 It was also found by Kaneko7 that the presence of γ-FeOOH in porous carbons assists the micropore filling of NO through their chemisorptive action. That could be a possible reason for a good retention of NO on GOFe2; this composite material exhibited higher porosity than other samples and is the only composite, which has clearly pronounced γ-FeOOH absorption bands in its FTIR spectrum (Figure 4B). 4. Conclusions The results presented in this paper demonstrate the importance of the porosity and the form of iron species for the effective removal of NO2 on the GOFe composites. The composite materials with the highest surface areas have the best adsorption capacities for NO2. However, the lack of porosity in a material could, to some extent, be compensated by the presence of active iron species. The FTIR analysis of the samples showed that these active species are lepidocrocite and carbonate species, which participate in the process of NO2 adsorption with the formation of surface nitrates/nitrites. The obtained data also showed that for most of the samples an immediate reduction of NO2 to NO by carbon has occurred, which resulted in the release of NO from the surface. The only sample that exhibited a marked delay in the release of NO and, for which the outlet concentration of NO increased very slowly, was GOFe2a. Such behavior of this composite material could likely be explained by its high Fe/C ratio and the oxidized form of carbon, which significantly reduced a degree of NO2 reduction to NO. The materials derived from the iron complex are nonporous, have a zero capacity for NO2, but are able to retain NO on the surface. Despite an immediate detection of NO in the outlet gas, the retention of NO on the surface of the iron sample heated in air was significantly better than on the surface of the sample heated in nitrogen. This can be explained the presence of such iron species as γ-FeOOH and R-Fe2O3, which participate in the interaction with NO molecules and formation of bridging nitrates detected by FTIR.
Ind. Eng. Chem. Res., Vol. 48, No. 24, 2009
Acknowledgment This work was supported by ARO Grant W911NF-05-1-0537 and NSF collaborative grant 0754945/0754979. The authors are grateful to Dr. Mykola Seredych for the help in GO synthesis and Dr. Jorge Morales for the help with SEM/EDAX analysis. Literature Cited (1) Zhang, W. J.; Rabiei, S.; Bagreev, A.; Zhuang, M. S.; Rasouli, F. Study of NO adsorption on activated carbons. Appl. Catal., B 2008, 83, 63. (2) Zhang, W. J.; Bagreev, A.; Rasouli, F. Reaction of NO2 with activated carbon at ambient temperature. Ind. Eng. Chem. Res. 2008, 47, 4358. (3) Pietrzak, R.; Bandosz, T. J. Activated carbons modified with sewage sludge derived phase and their application in the process of NO2 removal. Carbon 2007, 45, 2537. (4) Teng, H.; Suuberg, E. M. Chemisorption of nitric oxide on char. 2. Irreversible carbon oxide formation. Ind. Eng. Chem. Res. 1993, 32, 416. (5) Shirahama, N.; Moon, S. H.; Choi, K. H.; Enjoji, T.; Kawano, S.; Korai, Y.; Tanoura, M.; Mochida, I. Mechanistic study on adsorption and reduction of NO2 over activated carbon fibers. Carbon 2002, 40, 2605. (6) Jeguirim, M.; Tschamber, V.; Brilhac, J. F.; Ehrburger, P. Interaction mechanism of NO2 with carbon black: Effect of surface oxygen complexes. J. Anal. Appl. Pyrolysis 2004, 72, 171. (7) Kaneko, K. Anomalous micropore filling of NO on γ-FeOOHdispersed activated carbon fibers. Langmuir 1987, 3, 357. (8) Illan-Gomez, M. J.; Solano, A. L.; Salinas-Martinez de Lecea, C. NO reduction by activated carbons. 6. Catalysis by transition metals. Energy Fuels 1995, 9, 976. (9) Illan-Gomez, M. J.; Solano, A. L.; Radovic, L. R.; Salinas-Martinez de Lecea, C. NO reduction by activated carbons. 5. Catalytic effect of iron. Energy Fuels 1995, 9, 540. (10) Illan-Gomez, M. J.; Solano, A. L.; Radovic, L. R.; Salinas-Martinez de Lecea, C. NO reduction by activated carbons. 2. Catalytic effect of potassium. Energy Fuels 1995, 9, 97. (11) Lee, Y. W.; Choi, D. K.; Park, J. W. Performance of fixed-bed KOH impregnated activated carbon adsorber for NO and NO2 removal in the presence of oxygen. Carbon 2002, 40, 1409. (12) Illan-Gomez, M. J.; Solano, A. L.; Radovic, L. R.; Salinas-Martinez de Lecea, C. NO reduction by activated carbons. 4. Catalysis by calcium. Energy Fuels 1995, 9, 112. (13) Carabineiro, S. A.; Bras Fernandes, F.; Vital, J. S.; Ramos, A. M.; Silva, I. F. Uncatalyzed and catalyzed NO and N2O reactions using various catalyst and binary barium mixtures supported on activated carbon. Catal. Today. 1999, 52, 559. (14) Carabineiro, S. A.; Bras Fernandes, F.; Vital, J. S.; Ramos, A. M.; Fonseca, I. M. NO conversion using binary vanadium mixtures supported on activated carbon. Appl. Catal., B 2003, 44, 227. (15) Carabineiro, S. A.; Bras Fernandes, F.; Vital, J. S.; Ramos, A. M.; Fonseca, I. M. N2O cnversion using manganese binary mixtures supported on activated carbon. Appl. Catal., B 2005, 59, 181. (16) Carabineiro, S. A.; Bras Fernandes, F.; Silva, R. J. C.; Vital, J. S.; Ramos, A. M.; Fonseca, I. M. N2O reduction by activated carbon over iron bimetallic catalysts. Catal. Today B 2008, 1330135, 441. (17) Carabineiro, S. A.; Bras Fernandes, F.; Vital, J. S.; Ramos, A. M.; Silva, I. F. Vandium as a catalyst fro NO, N2O and CO2 reaction with activated carbon. Catal. Today 2000, 57, 305. (18) Xia, B.; Phillips, J.; Chen, C. H.-K.; Radovic, L. R.; Silva, I. F.; Menedez, J. A. Impact of pretreatments on the selectivity of carbon for NOx asorption/reduction. Energy Fuels 1999, 13, 903. (19) Underwood, G. M.; Miller, T. M.; Grassian, V. H. Trasmission FT-IR and Knudsen cell study of the heterogeneous reactivity of gaseous nitrogen dioxide on mineral oxide particles. J. Phys. Chem. A 1999, 103, 6184. (20) Toops, T. J.; Smith, D. B.; Partridge, W. P. NOx adsorption on Pt/K/Al2O3. Catal. Today 2006, 114, 112. (21) Zhu, R.; Guo, M.; Ci, X.; Ouyang, F. An exploratory study on simultaneous removal of soot and NOx over Ir/γ-Al2O3 catalyst in the presence of O2. Catal. Commun. 2008, 9, 1184. (22) Chen, H. Y.; Wang, X.; Sachtler, M. H. Reduction of NOx over various iron/zeolite catalysts. Appl. Catal., A 2000, 194, 159.
10891
(23) Valyon, J.; Hall, W. K. Studies of the surface species formed from NO on copper zeolites. J. Phys. Chem. 1993, 97, 1204. (24) Raj, A.; Le, T. H. N.; Kaliaguine, S.; Auroux, A. Involvement of nitrate species in the SCR of NO by NH3 at ambient conditions over TS-1 catalysts. Appl. Catal., B 1998, 15, 259. (25) Bourlinos, A. B.; Gournis, D.; Petridis, D.; Szabo, T.; Szery, A.; Dekany, I. Graphite oxide: Chemical reduction to graphite and surface modification with primary aliphatic amines and amino acids. Langmuir 2003, 19, 6050. (26) Morishige, K.; Hamada, T. Iron oxide pillard graphite. Langmuir 2005, 21, 6277. (27) Seredych, M.; Bandosz, T. J. Removal of ammonia by graphite oxide via its intercalation and reactive adsorption. Carbon 2007, 45, 2130. (28) Szabo, T.; Berkesi, O.; Forgo, P.; Josepovits, K.; Sanakis, Y.; Petridis, D.; Dekany, I. Evolution of surface functional groups in a series of progressively oxidized graphite oxides. Chem. Mater. 2006, 18, 2740. (29) Lerf, A.; He, H.; Forster, M.; Klinowski, J. Structure of graphite oxide revisited. J. Phys. Chem. B 1998, 102, 4477. (30) Seredych, M.; Tamashausky, A. V.; Bandosz, T. J. Surface features of exfoliated graphite/bentonite composites and their importance for ammonia adsorption. Carbon 2008, 46, 1241. (31) Seredych, M.; Bandosz, T. J. Adsorption of ammonia on graphite oxide/aluminium polycation and graphite oxide/zirconium-aluminium polyoxycation composites. J. Colloid Interface Sci. 2008, 324, 25. (32) Seredych, M.; Pietrzak, R.; Bandosz, T. J. Role of graphite oxide (GO) and polyaniline (PANI) in NO2 reduction on GO-PANI composites. Ind. Eng. Chem. Res. 2007, 46, 6925. (33) Hummers, W. S.; Offeman, R. E. Preparation of graphite oxide. J. Am. Chem. Soc. 1958, 80, 1339. (34) Yamanaka, S.; Doi, T.; Sako, S.; Hattori, M. High surface area solids obtained by intercalation of iron oxide pillars in montmorillonite. Mater. Res. Bull. 1984, 19, 161. (35) Brunauer, S.; Emmet, P. H.; Teller, E. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 1938, 60, 309. (36) Lastoskie, G.; Gubbins, K. E.; Quirke, N. Pore size distribution analysis of microporous carbons: density functional theory approach. J. Phys. Chem. 1993, 97, 4786. (37) Khan, A.; Chen, P.; Boolchand, P.; Smirniotis, P. G. Modified nanocrystalline ferrites for high-temperature WGS membrane reactor applications. J. Catal. 2008, 253, 91. (38) Brebu, M.; Uddin, M. A.; Muto, A.; Sakata, Y.; Vasile, C. Catalytic degradation of acrylonitrile-butadiene-styrene into fuel oil 2. Changes in the structure and catalytic activity of iron oxides. Energy Fuels 2001, 15, 565. (39) Liu, H.; Li, P.; Zhu, M.; Wei, Y.; Sun, Y. Fe(II)-induced transformation from ferrihydrite to lepidocrocite and goethite. J. Solid State Chem. 2007, 180, 2121. (40) Chopra, G. S.; Real, C.; Alcala, M. D.; Perez-Maqueda, L. A.; Subrt, J.; Criado, J. M. Factors influencing the texture and stability of maghemite obtained from the thermal decomposition of lepidocrocite. Chem. Mater. 1999, 11, 1128. (41) Cullity, B. D.; Stock, S. R. Elements of X-Ray Diffraction, 3rd ed.; Prentice-Hall Inc.: Upper Saddle River, NJ, 2001. (42) Shen, W.; Li, Z.; Liu, Y. Surface chemical functional groups modification on porous carbon. Recent Patents Chem. Eng. 2008, 1, 27. (43) Bashkova, S.; Bandosz, T. J. The effects of urea modification and heat treatment on the process of NO2 removal by wood-based activated carbon. J. Colloid Interface Sci. 2009, 333, 97. (44) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification of organic compounds, 5th ed.; Wiley: New York, 1988. (45) Cwiertny, D. M.; Hunter, G. J.; Pettibone, J. M.; Scherer, M. M.; Grassian, V. H. Surface chemistry and dissolution of γ-FeOOH nanorods and microrods: Environmental implications of size-dependent interactions with oxalate. J. Phys. Chem. C 2009, 113, 2175. (46) West, R. C. Handbook of Chemistry and Physics, 67th ed.; CRC Press: Boca Raton, FL, 1986. (47) Kung, M. C.; Kung, H. H. IR studies of NH3, Pyridine, CO, and NO adsorbed on transition metal oxides. Catal. ReV. Sci. Eng. 1985, 27, 425.
ReceiVed for reView July 2, 2009 ReVised manuscript receiVed September 25, 2009 Accepted September 28, 2009 IE901054P