12710
J. Phys. Chem. B 2005, 109, 12710-12717
Effect of Expanded Graphite Lattice in Exfoliated Graphite Nanofibers on Hydrogen Storage Angela D. Lueking,*,†,‡ Ling Pan,§ Deepa L. Narayanan,†,‡ and Caroline E. B. Clifford‡ Department of Energy and Geo-EnVironmental Engineering, The Energy Institute, and The Materials Research Institute, The PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802 ReceiVed: March 9, 2005; In Final Form: May 5, 2005
A graphite exfoliation technique, using intercalation of a concentrated sulfuric/nitric acid mixture followed by a thermal shock, has successfully exfoliated a herringbone graphite nanofiber (GNF). The exfoliated GNF retains the overall nanosized dimensions of the original GNF, with the exfoliation temperature determining the degree of induced defects, lattice expansion, and resulting microstructure. High-resolution transmission electron microscopy indicated that the fibers treated at an intermediate temperature of 700 °C for 2 min had dislocations in the graphitic structure and a 4% increase in graphitic lattice spacing to 3.5 Å. The fibers treated at 1000 °C for 36 h were expanded along the fiber axis, with regular intervals of graphitic and amorphous regions ranging from 0.5 to >50 nm in width. The surface area of the starting material was increased from 47 m2/g to 67 m2/g for the 700-°C treatment and to 555 m2/g for the 1000-°C treatment. Hydrogen uptake measurements at 20 bar indicate that the overall hydrogen uptake and operative adsorption temperature are sensitive to the structural variations and graphitic spacing. The increased surface area after the 1000-°C treatment led to a 1.2% hydrogen uptake at 77 K and 20 bar, a 3-fold increase in hydrogen physisorption of the starting material. The uptake of the 700-°C-treated material had a 0.29% uptake at 300 K and 20 bar; although low, this was a 14-fold uptake over the starting material and higher than other commonly used pretreatment methods that were tested in parallel. These results suggest that selective exfoliation of a nanofiber is a means by which to control the relative binding energy of the hydrogen interaction with the carbon structure and thus vary the operative adsorption temperature.
Introduction The use of hydrogen as an energy carrier will require a means to transport and store hydrogen, yet no existing technology meets the performance requirements established by the U. S. Department of Energy. Early hydrogen storage claims in carbon nanomaterials have been highly disputed due to artifacts such as water contamination,1,2 metal contamination,3 and/or temperature effects.4 Early reports of exceptionally large hydrogen storage in graphite nanofibers (GNFs) at ambient temperature and 120 bar5,6 have not been reproduced by other laboratories.4,7-9 Nonetheless, storage reports continue to emerge showing hydrogen uptake ranging from 6.5% at 120 bar10 to 10-17% at 120 bar11-13 for GNFs and 1.97% at 40 bar14 to 6.3% at 148 bar15 for structurally similar multiwalled carbon nanotubes (MWNTs). Measurements of hydrogen storage in GNFs produced in two separate laboratories suggest that the uptake is dependent upon synthesis catalyst and pretreatment conditions.9 Variations in synthesis conditions and thermal processing in these reports suggest that defects within the bulk graphitic structure or at terminal carbons lead to enhanced hydrogen storage and the graphitic layering in both GNFs and MWNTs perhaps make them more amenable to these structural deformations. The first reports of hydrogen storage in GNFs suggested expansion of the graphitic lattice upon hydrogen adsorption,5 * Author to whom correspondence should be addressed. E-mail:
[email protected]. † Department of Energy and Geo-Environmental Engineering. ‡ The Energy Institute § The Materials Research Institute.
with variations in pore size distribution and X-ray diffraction patterns after hydrogen exposure to support this claim.6 Subsequent neutron diffraction studies of graphite ball-milled in hydrogen show evidence of hydrogen intercalation within the graphitic structure,16,17 and transmission electron microscopy (TEM) micrographs in recent reports of hydrogen storage in GNFs indicate expansion of the lattice after exposure to highpressure hydrogen.11-13 Hydrogen uptake for exfoliated carbon materials has been reported previously to result in a 2.5-fold increase in the hydrogen uptake at 10.5 MPa and 295 K,18 but it is not clear what type of starting material resulted in this 2.5fold uptake as few details were given for the exfoliated material. Theoretical simulations of hydrogen adsorption in rigid GNFs cannot reproduce hydrogen uptakes in excess of a few weight percent without significantly modifying the interaction potentials even with varying lattice spacing,19 nor do theoretical simulations find lattice expansion due to hydrogen adsorption thermodynamically feasible.19 In light of the continued reports of hydrogen uptake in the slit-pore-type structures of GNFs or conical MWNTs, we have adapted a method used to exfoliate graphite to expand the slit pores of GNFs a priori to test the effect on the resulting carbon structure and the hydrogen adsorption. Our exfoliation method relies on well-developed methods for graphite exfoliation, in which vaporization of a concentrated acid intercalant expands graphite along its c-axis, with evidence that up to a several 100fold expansion is possible.20 Graphite exfoliation has been described as a phase transition and may be either reversible or irreversible, depending upon the temperature used to provide
10.1021/jp0512199 CCC: $30.25 © 2005 American Chemical Society Published on Web 06/09/2005
Exfoliated GNFs for Hydrogen Storage the thermal shock to vaporize the intercalant.20 Pitch-based or activated-carbon-based carbon microfibers (denoting diameters on the order of 10-200 µm) require elevated exfoliation temperatures due to a decreased orientation and graphitic order within the microfibers. Exfoliation of carbon microfibers results in a modified shape, morphology, texture, and generally a very nonuniform structure, as evidenced by typical scanning electron microscopy (SEM) micrographs.21 The term exfoliation has also been applied to single-walled carbon nanotube (SWNT) bundles when using a much milder technique to separate the bundles into individual tubes. In a sense, the term “exfoliation” has come to have a meaning that is dependent upon the original structure of the material. To be clear, here we use the term exfoliation in the original sense of graphite exfoliation, or expansion along the graphitic c-axis. In general, the fractional expansion of graphite increases sharply with increasing degree of graphitic order or crystallite orientation, and exfoliation requires a good alignment of the carbon layers.20 The degree of expansion upon exfoliation is dependent on the activity of the intercalant, its concentration, and the disruption of the van der Waals force between graphite layers by the intercalant rather than the inherent number of defects or dislocations within the graphite structure; in fact, preexisting or induced defects within the graphitic structure inhibit graphite exfoliation or expansion along the c-axis.20 Wellstructured, highly ordered, herringbone graphite nanofibers (GNFs) provide an interesting candidate for carbon exfoliation, with their slit-pore geometry, nanoscale dimensions, high aspect ratio, and graphitic layers that terminate along the fiber axis. Methods Materials. Ultra-high-purity hydrogen (99.999%) was used in adsorption experiments, and ultra-high-purity helium (99.999%) was used to determine the helium density of the adsorbents for buoyancy corrections. Prior to adsorption, both hydrogen and helium were passed through an in-line 3-Å molecular sieve zeolite column to eliminate possible moisture contamination which can be problematic for gravimetric measurements. Highly ordered, herringbone graphite nanofibers were purchased from Catalytic Materials, with a metal content of less than 1% as reported by the manufacturer. Acids used for exfoliation were purchased from Sigma-Aldrich and used in concentrated form, sulfuric acid (95-98%) and nitric acid (70%). Exfoliation. Methods to exfoliate graphite are well-known20 and were extended to GNFs. On the basis of previous work22 and preliminary studies with graphite, the primary exfoliation method used in this study was a mixture of nitric and sulfuric acids followed by thermal shocking at 700 °C.20 GNFs were mixed with 50/50 mixtures of sulfuric/nitric acids for 2 h at room temperature. The mixture was filtered, transferred to crucibles, and then rapidly heated to 700 °C for about 2 min. A portion of this sample was subjected to an additional hightemperature treatment by heating the sample under flowing argon (ultra-high-purity) at 1000 °C for 36 h; the sample was exposed to air upon transfer to characterization equipment, then subjected to the 4-h 150-°C degas after transfer. These samples will be referred to as GNF1, EGNF-700, and EGNF-1000, for the raw, intermediate heat treatment at 700 °C, and the hightemperature heat treatment at 1000 °C, respectively. A control experiment, E2-GNF, was a variation of the preparation used for EGNF, with extended rinsing of the intercalated acid prior to thermal shock. A portion of the raw GNF1 was heated in argon at 1000 °C for 36 h without exfoliation (GNF-1000), and a separate portion of the raw GNF1 was subjected to high-
J. Phys. Chem. B, Vol. 109, No. 26, 2005 12711 pressure hydrogen at 120 bar (12 MPa) for 16 h (GNF-HP), analogous to other studies that suggest that exfoliation may occur with simple exposure to high-pressure hydrogen. All samples were pretreated at 150 °C and ∼10-8 bar prior to characterization and adsorption measurements. Characterization. Materials were characterized using standard Brunauer-Emmett-Teller (BET) methods with nitrogen at 77 K (Quantachrome Autosorb I) and helium densitometry measurements (Hiden IGA-003), both after a 150-°C outgas for 4 h. Scanning electron microscopy (Philips XL20) and TEM (JEOL 2010 and JEOL 2010F) were used to characterize the microstructure of the material. Total ash content was determined by temperature programmed oxidation on a low-pressure PerkinElmer thermogravimetric analyzer 7. As the high carbon content rendered CHN analysis inconclusive (Leco CHN 600), the effects of exfoliation on the elemental composition were analyzed semiquantitatively using X-ray energy-dispersive spectroscopy (EDS) on an SEM (Hitachi S 3500N) with ZAF corrections. Hydrogen Uptake. A high-pressure gravimetric analyzer (Hiden Isochema IGA-003) was used to evaluate hydrogen uptake at pressures up to 20 bar. The IGA provided a highly sensitive ((1 µg) measurement with precise temperature and pressure control for automated measurements of adsorption and desorption isotherms. All samples were subjected to an in situ degas at 150 °C, unless otherwise stated. Hydrogen uptake measurements were normalized to sample mass after pretreatment, with buoyancy corrections determined from density measurements with helium. With a typical sample size of 50 mg, the error in the hydrogen due to instrumental limitations is (0.02 wt % absolute. Selected samples were chosen for quality checks to ensure the reproducibility of the measurements. Gases used in the IGA were pretreated in a 3-Å molecular sieve to remove impurities; although the outlet gas was not analyzed, we have found moisture-sensitive samples tend to show an irreversible adsorption, and this was not observed in the present work. Results and Discussion Resulting Structures. Raw Graphite Nanofibers (GNF1). The starting material GNF1 has a herringbone structure with a high degree of graphitic order and was specified by the manufacturer to be less than 1% ash content by weight, which was similar to that measured with a bulk ash analysis (Table 1). The BET surface area of GNF1 was 47.0 m2/g, and the helium density was 1.4 g/cm3 (Table 1). TEM images of GNF1 show the presence of both straight and curly GNFs, with diameters ranging from 30 to ∼300 nm (Figure 1a). The graphitic planes are clearly evident in the high-resolution TEM (HRTEM) images, as shown in Figure 1b, at an angle relative to the fiber axis, consistent with the herringbone structure. The inset in Figure 1b is the corresponding Fourier transform, which is used to determine the spacing of the graphitic layers by measuring the distance from the center spot to the diffracted spot. For GNF1, this spacing is determined to be 3.34 + 0.04 Å. Graphite Nanofibers after 1000-°C Heat Treatment (GNF1000). Several previous reports have shown that annealing a nanocarbon sample in argon at 1000 °C increases the hydrogen storage. GNF1 was treated in argon at 1000 °C for 36 h (GNF1000). This argon treatment for GNF-1000 resulted in a slight increase in surface area to 66 m2/g and an increase in helium density to 1.78 g/cm3 (Table 1). The increase in helium density is consistent with partial graphitization of the fibers. No structural or microstructural changes were observed in the TEM or HRTEM micrographs (data not shown).
12712 J. Phys. Chem. B, Vol. 109, No. 26, 2005
Lueking et al.
TABLE 1: Characterization of Graphite Nanofibers
treatment (m2/g)
BET surface area helium density (g/cm3) hydrogen uptake, 77 K, 20 bar (wt %) hydrogen uptake, 77 K, 20 bar (wt %/surface area) hydrogen uptake, 300 K, 20 bar (wt %) ash (wt %)
GNF1
GNF-1000
GNF-350-Aira
EGNF-700
EGNF-1000
none
argon, 1000 C°, 36 h, after refs 10-13
air, 350 °C for 12 h, after ref 9
HNO3/H2SO4 intercalation, 700 °C for 2 min
EGNF-700 + argon, 1000 °C, 36 h
47.0 1.38 0.34%
66 1.78 0.18%
7.2 × 10-3
2.7 × 10-3
0.02%
0.11%
1.5%
2.2%
82.4 1.47
67 1.04 0.39%
554.7 1.36 1.23%
5.8 × 10-3
2.2 × 10-3
0.15%
0.29%
0.03%
2.7%
0.61%
1.1%
a The GNF sample oxidized in air was used to determine if oxygen functional groups might give rise to the increased adsorption of EGNF-700 at 300 K; thus data for GNF-350-Air was collected only at 300 K.
Figure 1. (a) Low-magnification and (b) high-resolution TEM images of untreated GNF1. The inset to Figure 1b is the corresponding Fourier transform, by which the spacing between the graphite sheets is calculated to be 3.34 Å.
Exposure of GNF to High-Pressure Hydrogen (GNF-HP and EGNF-HP). It has been suggested that high-pressure hydrogen exposure at 120 bar leads to an expanded graphite lattice.6,12,13 Therefore, GNF1 was exposed to 120 bar hydrogen (after 150-
°C degas). This hydrogen exposure led to a slight reduction in the measured BET surface area and helium density, decreasing to 35 m2/g and 1.6 g/cm3, respectively. However, this highpressure hydrogen exposure led to no real structural changes in the GNF-HP, as shown by TEM (data not shown). Similarly, EGNF-700 exposed to high-pressure hydrogen did not have any observable changes in the microstructure observable by TEM. Graphite Nanofibers after Acid Intercalation and a 700-°C Thermal Shock (EGNF-700). The acid intercalation followed by an intermediate temperature (700 °C) thermal shock led to a slight increase in surface area to 67 m2/g and a 25% decrease in helium density to 1.0 g/cm3 (Table 1). An HRTEM image of the EGNF-700 (Figure 2b) showed a significant number of dislocations present in the graphitic structure, as indicated by the arrows in Figure 2b. The spacing between the graphitic planes was calculated from the Fourier transform of the realspace HRTEM image (inset in Figure 2b) to be around 3.47 + 0.04 Å, a 4% increase from the untreated sample GNF1. There was no clear evidence of edge attack by the exfoliation treatment in the TEM micrographs. EDS data indicates the presence of residual sulfur to be high (∼10-20%) after intercalation but significantly reduced (∼1-2%) after vacuum degas at 150 °C; no nitrogen was detected (data not shown). It is not absolutely clear whether the dislocations arose during acid intercalation or were due to the thermal shock or whether the dislocations might be due to the presence of residual acid intercalants. Residual intercalant is a common phenomenon in exfoliation of graphite, as the thermal shock during exfoliation includes both vaporization and desorption from the intercalated graphite.20 Similar dislocations have been described as “rogue bonds” in the exfoliation literature, indicating linkages between carbon layers. Intercalation of graphite is known to be a staging phenomenon, meaning that intercalated layers are periodically arranged in the matrix of graphite layers.23 The graphitic dislocations observed in EGNF-700 have been depicted in cartoons of staging transitions for graphite intercalation compounds.23 However, as subsequent evidence (i.e., HRTEM of EGNF-1000, discussed below) suggests a low staging in the intercalation compounds, the resulting TEM structure of the EGNF-700 is not likely due to a staging transition. Previous exfoliation studies suggest that both acid intercalation and thermal shock were necessary for the resulting dislocations. Both HNO3 and H2SO4 are “acceptor” intercalants that draw electrons away from adjacent graphitic planes; they are expected to contract the overall graphite structure23 and can therefore not account for the observed increase in average lattice spacing. Graphite Nanofibers after an Additional 1000-°C Heat Treatment (EGNF-1000). Annealing the exfoliated sample in argon at 1000 °C increased the BET surface area to 555 m2/g,
Exfoliated GNFs for Hydrogen Storage
Figure 2. (a) Low-magnification and (b) high-resolution TEM images of untreated EGNF-700, intercalated with concentrated nitric/sulfuric acid followed by a thermal shock at 700 °C. The inset to Figure 2b is the corresponding Fourier transform, by which the spacing between the graphite sheets is calculated to be 3.47 Å, a 4% increase over GNF1.
over a 10-fold increase over the raw material (Table 1). The helium density of the sample was measured to be 1.36 g/cm3, an increase over the intermediate temperature treatment but approximately equal to the untreated GNF1. EDS analysis of EGNF-1000 showed no indication of oxygen or sulfur species (data not shown). TEM images (Figure 3) of EGNF-1000 showed a significant modification of the GNF microstructure, exhibiting alternating high- and low-contrast V-shaped regions along the fiber axis. Close examination of the HRTEM images (as shown in Figure 3b) indicate that the dark regions are graphite layers and the light regions are not pores but rather amorphous carbon. The dark and light contrast follows from the high and low mass density in the ordered and amorphous regions, respectively. The V-shape is the result of the graphitic plane orientation of the herringbone structure. The widths of the dark and light regions vary for different fibers, ranging from 5 to >50 nm. The regularity of these repeating regions likely arises from staging of the original intercalation compound. Each
J. Phys. Chem. B, Vol. 109, No. 26, 2005 12713
Figure 3. (a) Low-magnification and (b) high-resolution TEM images of EGNF-1000, the resulting structure after subjecting EGNF-700 to an additional thermal treatment in 1000-°C argon for 36 h.
intercalant layer tends to be completely filled prior to starting a new intercalant layer, as the energy for inserting a molecule between graphite is less for an existing layer when compared to a new layer.23 The staging phenomenon in graphite is shown by the repeating structure of the EGNF-1000, and the dimensions of the graphitic dark regions (i.e., 5-50 nm) suggest the staging that arose during the HNO3/H2SO4 intercalation was low prior to the high-temperature argon treatment. Structural and Chemical Changes of the Fibers Upon Exfoliation. SEM micrographs provide evidence of modifications in the different structures. SEM of GNF1 shows a relatively smooth surface (Figure 4a); SEM of EGNF-700 indicates the appearance of interesting structural features along the termination of the fibers (Figure 4b) as well as disklike structures (inset of Figure 4b) that suggest that some of the GNFs were destroyed by the intermediate temperature exfoliation process. SEM of EGNF-1000 indicates several regions with structural defects along the surface (Figure 4c).
12714 J. Phys. Chem. B, Vol. 109, No. 26, 2005
Lueking et al.
Figure 5. Hydrogen uptake for the various GNFs at 77 K and up to 20 bar. Hydrogen uptake is normalized per mass of carbon remaining after each pretreatment.
Figure 4. SEM micrographs indicate the changes in surface structure upon various stages in the exfoliation process. (a) GNF1 has relatively smooth surfaces; (b) EGNF-700 shows roughened surfaces, variations in fiber termination (arrows), and the appearance of disklike carbon structures indicating partial destruction of the GNFs; (c) EGNF-1000 shows roughened surfaces, surface defects (arrows), and expanded regions.
After the intermediate temperature thermal shock at 700 °C, EGNF-700 was observed to have a unique thermal instability. When the sample was heated to the standard 150-°C degas temperature under vacuum, the sample lost ∼70% of its mass. The mass loss began at approximately 100 °C and peaked at ∼150 °C, with the degradation complete by ∼160 °C. This mass loss likely corresponds to loss of some residual acid intercalant, as the EDS of the sample prior to the 150-°C treatment showed 10-20% residual sulfur in the sample. The temperature at which this mass loss occurs indicates that this material was strongly bound to the surface and/or there were highly reactive carbon structures that were decomposing. SEM micrographs of the EGNF-700 prior to the 150-°C degas indicate the presence of disklike structures (Figure 4b, inset), which we believe to be carbonaceous byproducts from carbon exfoliation. These disklike structures were only observed in SEM micrographs of the
EGNF-700 prior to heating in vacuo to 150 °C. Subsequent studies on the EGNF-700 suggest that this thermal decomposition activates the material for hydrogen adsorption and that this activation must be conducted in situ to avoid exposure to air, which will deactivate the material for hydrogen uptake. No oxygen or sulfur was detectable in EGNF-1000, suggesting complete desorption of the intercalant with the extended heat treatment. The exact nature of the functional groups incorporated into the EGNF-700 surface is not known, and this will be explored in future work. We believe that the treatment at 700 °C only partially exfoliated the GNF structure, whereas 1000 °C was needed to fully exfoliate and expand the GNF. The graphite exfoliation literature differentiates between vaporization of the intercalant to exfoliate the graphitic structure and complete desorption of the intercalant, stating that it is common to desorb as little as 1/ th of the intercalant upon exfoliation.20 These results are 8 consistent with similar studies showing the effect of exfoliation temperature upon H2SO4-intercalated graphite; an increasing exfoliation temperature leads to increasing surface area and pore volume. However, in these previous studies, variations in the starting graphite have equal influence on these properties, with temperature effects being most pronounced for materials showing relatively little expansion.24 Hydrogen Uptake. At 77 K, physisorption is expected to dominate hydrogen-carbon interactions. The EGNF-700 did not have a significant variation in hydrogen uptake at 77 K compared to the untreated sample, and the variation for GNF1000 actually decreased (Figure 5). However, the 10-fold increase in surface area of EGNF-1000 was reflected in hydrogen uptake at 77 K, with a hydrogen uptake of 1.2% at 20 bar. A monolayer surface coverage of hydrogen on a graphene sheet has been calculated to be H/C ) 0.18, corresponding to 2.28 × 10-3 wt % uptake for each square meter of surface area.25 This analysis requires a flat graphene sheet and ignores the contributions of microporosity. Surface areas accessible to nitrogen may not necessarily be accessible to the smaller hydrogen molecule, and thus hydrogen uptake does not always correlate to BET surface area, particularly for nanocarbons.7-10,26 It is interesting to note that the normalized hydrogen uptake of the EGNF-1000 corresponds to that
Exfoliated GNFs for Hydrogen Storage
Figure 6. Hydrogen uptake for the various GNFs at 300 K and up to 20 bar. Hydrogen uptake is normalized per mass of carbon remaining after each pretreatment.
predicted by monolayer surface coverage of hydrogen on carbon: 2.2 × 10-3 wt % per unit surface area. This implies that the increased hydrogen uptake for EGNF-1000 is solely due to surface area and that nitrogen has access to the same pores accessible by hydrogen. GNF-1000 has a similar normalized uptake, 2.7 × 10-3 wt % per unit surface area, whereas untreated GNF1 and EGNF-700 have a normalized uptake of (6-7) × 10-3 wt % per unit surface area. Hydrogen uptake at 300 K implies a stronger carbonhydrogen interaction. When measured at 300 K and 20 bar, the baseline uptake of untreated GNF1 was significantly reduced to 0.02% with saturation at pressures less than 1 bar (Figure 6). Argon annealing, shown previously to increase hydrogen uptake, increased the uptake only to 0.1% uptake at 20 bar. The high-surface-area EGNF-1000 has a low hydrogen uptake at 300 K, only 0.03% at 20 bar and little functional pressure dependence. The hydrogen uptake for EGNF-700, although still relatively low at 0.29% at 20 bar and 300 K, exhibited a 14-fold increase over the hydrogen uptake of the raw material and did not appear to be saturated at 20 bar. Previous reports of exfoliated carbon (unspecified structure) showed a 2.5-fold increase in hydrogen uptake at 295 K after exfoliation.18 As stated above, EGNF700 had significant mass loss during the standard 150-°C pretreatment. The material was pretreated in vaccuo at degassing temperatures of 80, 100, and 150 °C, temperatures corresponding to no, partial, and complete thermal degradation, with hydrogen uptake measured after each temperature treatment. As the thermal degradation proceeded, not only did the normalized hydrogen uptake increase, but there was a significant change in the shape of the isotherm, suggesting that this thermal degradation led to an activated material for hydrogen storage (Figure 7). We do not precisely know the chemical functional groups present on the surface after the various stages of thermal degradation but suspect that these were removed with the significant mass loss that accompanied the 150-°C treatment. Other studies of oxidized GNF did not have the uptake of EGNF-700. GNF-Air, GNF1 oxidized in air at 350 °C, had an uptake at 300 K and 20 bar of 0.15% (Table 1). E2-GNF-700, a modification of the exfoliation treatment to wash the acidintercalated sample prior to thermal shock, exhibited a hydrogen uptake at 300 K and 20 bar of 0.03% and was virtually
J. Phys. Chem. B, Vol. 109, No. 26, 2005 12715
Figure 7. Highly reactive EGNF-700 showed significant mass loss upon heating to the standard degas temperature of 150 °C. Normalized hydrogen uptake at various stages in this treatment shows that the material becomes activated for hydrogen uptake as this decomposition proceeds. The hydrogen uptake of the E2-GNF-700 control (see text) indicates that any oxidation during the acid intercalation step was not responsible for the increased hydrogen uptake. Adsorption isotherms only are shown.
indistinguishable from GNF1 (Figure 7). These results suggest some of the increased uptake may be due to oxidation of the GNFs, but the increased uptake is likely not solely due to oxidation. Future work is needed to clearly identify the functional groups present on the surface at various stages of pretreatment to assess the role of these groups in hydrogen uptake. As various oxidative treatments of GNF1 did not attain the uptake of EGNF-700, we believe the graphitic dislocations present in EGNF-700 contributed to the increased uptake. A decrease in helium density of EGNF-700 suggests that helium, at least, was able to access these graphitic dislocations. Hydrogen uptake of the exfoliated carbon sample exceeded the other commonly used pretreatments selected by a factor of 2-3 (Table 1). Park et al.14 stated oxygen functional groups blocked the access of hydrogen to GNF slit-pores and led to a decreased hydrogen storage level. Lueking et al. found edge attack to correlate with increased hydrogen uptake, whether the edge attack was induced by oxidative or hydrogasification conditions.9 Pradhan et al. attributed increased hydrogen binding energy to surface roughening upon oxidative treatment in SWNTs.27 Simple high-pressure exposure at 80-120 bar was suspected to expand the graphite lattice, leading to a reported hydrogen uptake of ∼15%.11,12 Similarly, although not the maximum storage reported, Browning et al. also reported a significant hydrogen uptake of 4.16% after pretreatment at 150 °C in vacuum.10 Parallel studies with pretreatments based on these previous reports (i.e., argon at 1000 °C, oxidation in concentrated HNO3/H2SO4, and oxidation at 350 °C in air) did not exhibit the 14-fold increase (at 300 K and 20 bar) of the exfoliation treatment. Despite this 14-fold increase in hydrogen uptake upon exfoliation, the adsorption uptake is not of the magnitude reported previously for GNF exhibiting an expanded lattice structure via TEM11-13 and/or explained by hydrogen uptake via chemisorption and hydrogen intercalation within the lattice.10 The original objective of the work, as stated above, was to use exfoliation to induce defects and increase the interlayer spacing between the graphene sheets of GNFs. We were successfully
12716 J. Phys. Chem. B, Vol. 109, No. 26, 2005 able to increase the graphite lattice by 4% as well as significantly expand the graphite lattice along the fiber axis. The thermal shock conditions used to create these two very different fibers were at opposite extremes, and future work will look to refine the exfoliation technique to create an array of GNFs with variations on lattice spacing. An optimal pore size, resulting in overlap of the carbon-hydrogen interaction potentials, is expected to maximize adsorption19 while increasing carbonhydrogen binding energy. The overall hydrogen uptake for the EGNF-700 and EGNF1000 is not close to the Department of Energy hydrogen storage targets. We are, however, encouraged by the data that suggests that exfoliation is a means by which to control the carbonhydrogen binding energy. At 77 K, the hydrogen uptake for EGNF-1000 is comparable to recent reports of 2.7 wt % uptake in a spherical nanoporous carbon with surface areas between 946 and 1646 m2/g at 77 K and 50 bar28 and 2 wt % for commercially available activated carbon fibers at 77 K and 20 bar.29 For the latter case, the surface area was not reported. Lueking has previously found the surface area of comparable commercially available activated carbon fibers to be 2000 m2/ g;30 thus we suspect the activated carbon fiber in the above report to have a surface area in the range of 1500-2000 m2/g. Residual catalysts are not expected to play a significant role in hydrogen uptake, as the ash content of the materials was low and EDS analysis showed that the acid intercalation step removed most of any surface metal. Although repeated studies have shown that residual metal is both necessary and sufficient to obtain hydrogen uptake in carbon-based materials, the low metal content for these fibers suggests that this does not likely contribute to the observed hydrogen uptake. Rather, this work focuses on optimizing the carbon sink, for which the overall uptake may be further increased by the addition of catalytic materials, a subject for future work. New Exfoliated Carbon Material. There was clear evidence of microstructure changes in the GNFs after the various exfoliation treatments. Thermal shock at ∼700 °C led to disruptions in the graphitic order and an increase of 4% in lattice spacings. The high-temperature treatment in argon led to an even more pronounced change in the microstructure, with a 10-fold increase in surface area and clear evidence via both low- and high-magnification TEM imaging of expansion of the GNF lattice. The low-magnification TEM image (Figure 3a) for our hightemperature exfoliated structure is similar to the low-magnification images of Gupta and Srivastava.12,13 However, this expansion effect was realized only for the complete exfoliation treatment followed by a 1000-°C argon treatment (EGNF-1000). We did not see this expansion for GNFs treated in argon without prior acid intercalation (GNF-1000) or for the sample exposed to high-pressure hydrogen (GNF-HP). Gupta et al. did not mention the acid treatment of their GNFs prior to hydrogen exposure;12,13,31 therefore we believe that the exfoliation of our fibers is a different process from what they report based on their stated methods and low-resolution TEM images. The exfoliation treatment used here is based on wellestablished methods to exfoliate graphite. Although the term exfoliation has been applied previously to carbon microfibers and single-walled carbon nanotubes to denote separation of fibrous strands or bundles into individual tubelike structures, this is fundamentally dissimilar to true graphite exfoliation, which is a phase transition and expands the graphitic spacing along the c-axis.20 It is generally found that beyond a minimum stacking height32 exfoliation is easier for materials that are “thin”
Lueking et al. with respect to the c-axis, which can be attributed to the forces needed to overcome the interplanar van der Waals forces at play between adjacent graphite layers. It was also previously reported that exfoliation of graphite was limited to particle sizes with a width along the ab-plane of graphite in excess of 75 µm.20 Particle size limitations along the ab-plane of graphite are attributed to the need for a sizable intercalate “island size” to provide the force needed to overcome the van der Waals forces at play between graphitic layers, thus limiting exfoliation of carbon fibers to those with micro- rather than nanodimensions.20 Although exfoliation of ∼10-µm carbon microfibers has been reported, these exfoliated carbon microfibers are generally highly irregular in structure, with the expansion serving to separate fibrous materials from a bundle and/or creating a flowery bundle of eye-shaped pores within the cross section of the fiber.33,34 The largest diameter observed in low-resolution images of the raw GNFs was ∼300 nm, with most fibers having a diameter less than 200 nm, giving clear indication that exfoliation of graphite layers (rather than fibrous bundles) is not limited to particles with widths of 75 µm. The structures presented here are clearly unique, in comparison to other exfoliated carbon fibers, in that the microstructure of the EGNF-1000 has repeating V-shaped regions that alternate between amorphous and graphitic character. Future work will serve to better characterize this new material. Previous work suggests that the exfoliated graphite nanofiber will have applications in catalysis, fuel cells, and as an electrode. Conclusions A new exfoliated carbon fiber with nanodimensions was synthesized by graphite exfoliation followed by high-temperature treatment. The exfoliated carbon nanofiber had a unique microstructure, with repeating units of high-density graphitic regions separated by low-density regions that were amorphous in nature. An expansion along the fiber axis was accompanied by a 10-fold increase in BET surface area. At the other end of the preparation spectrum, GNFs were prepared with mild dislocations within the graphitic structure and a 4% increased lattice spacing. Overall hydrogen uptake and operative adsorption temperature were drastically different for these two exfoliated materials, suggesting a means by which to control the carbon-hydrogen binding energy through variation of exfoliation conditions. Future work will include systematic variations on exfoliation conditions in an attempt to synthesize materials with an array of slit-pore dimensions and include subsequent chemical treatments to control the resulting surface chemistry on the exfoliated carbon structures. Acknowledgment. We thank E. Dickey for her feedback on the TEM presented in this work. This work was funded through start-up funds proved by the Institute for the Environment, the Energy Institute, and the College of Earth and Mineral Sciences, Penn State University. Funding for exfoliation of carbon for hydrogen storage was provided, in part, by the H2E Center at Penn State University. References and Notes (1) Yang, R. T. Carbon 2000, 38, 623. (2) Pinkerton, F. E.; Wicke, B. G.; Olk, C. H.; Tibbetts, G. G.; Meisner, G. P.; Meyer, M. S.; Herbst, J. F. J Phys Chem B 2000, 104, 9460. (3) Hirscher, M.; Becher, M.; Haluska, M.; Dettlaff-Weglikowska, U.; Quintel, A.; Duesberg, G. S.; Choi, Y. M.; Downes, P.; Hulman, M.; Roth, S.; Stepanek, I.; Bernier, P. Appl. Phys. A: Mater. Sci. Process. 2001, 72, 129. (4) Tibbetts, G. G.; Meisner, G. P.; Olk, C. H. Carbon 2001, 39, 2291.
Exfoliated GNFs for Hydrogen Storage (5) Chambers, A.; Park, C.; Baker, R. T. K.; Rodriguez, N. M. J. Phys. Chem. B 1998, 102, 4253. (6) Park, C.; Anderson, P. E.; Chambers, A.; Tan, C. D.; Hidalgo, R.; Rodriguez, N. M. J. Phys. Chem. B 1999, 103, 10572. (7) Ahn, C. C.; Ye, Y.; Ratnakumar, B. V.; Witham, C.; Bowman, R. C.; Fultz, B. Appl. Phys. Lett. 1998, 73, 3378. (8) Strobel, R.; Jorissen, L.; Schliermann, T.; Trapp, V.; Schutz, W.; Bohmhammel, K.; Wolf, G.; Garche, J. J Power Sources 1999, 84, 221. (9) Lueking, A. D.; Yang, R. T.; Rodriguez, N. M.; Baker, R. T. K. Langmuir 2004, 20, 714. (10) Browning, D. J.; Gerrard, M. L.; Lakeman, J. B.; Mellor, I. M.; Mortimer, R. J.; Turpin, M. C. Nano Lett. 2002, 2, 201. (11) Gupta, B. K.; Srivastava, O. N. Int. J. Hydrogen Energy 2000, 25, 825. (12) Gupta, B. K.; Srivastava, O. N. Int. J. Hydrogen Energy 2001, 26, 857. (13) Gupta, B. K.; Tiwari, R. S.; Srivastava, O. N. J. Alloys Compd. 2004, 381, 301-308. (14) Lee, S. M.; Park, K. S.; Choi, Y. C.; Park, Y. S.; Bok, J. M.; Bae, D. J.; Nahm, K. S.; Choi, Y. G.; Yu, S. C.; Kim, N. G.; Frauenheim, T.; Lee, Y. H. Synth. Met. 2000, 113, 209. (15) Hou, P. X.; Yang, Q. H.; Bai, S.; Xu, S. T.; Liu, M.; Cheng, H. M. J. Phys. Chem. B 2002, 106, 963. (16) Fukunaga, T.; Itoh, K.; Orimo, S.; Aoki, M.; Fujii, H. J Alloys Compd. 2001, 327, 224. (17) Fukunaga, T.; Itoh, K.; Orimo, S.; Aoki, K. Mater. Sci. Eng., B 2004, 108, 105. (18) Poirier, E.; Chahine, R.; Bose, T. K. Int. J. Hydrogen Energy 2001, 26, 831.
J. Phys. Chem. B, Vol. 109, No. 26, 2005 12717 (19) Wang, Q. Y.; Johnson, J. K. J. Phys. Chem. B 1999, 103, 277. (20) Chung, D. D. L. J. Mater. Sci. 1987, 22, 4190. (21) Yoshida, A.; Hishiyama, Y.; Inagaki, M. Carbon 1990, 28, 539. (22) Clifford, C. E. B.; Beil, A.; Boland, E.; Grove, L. Prepr. Symp.Am. Chem. Soc., DiV. Fuel Chem. 2004, 49, 633. (23) Dresselhaus, M. S.; Dresselhaus, G. AdV. Phys. 2002, 51, 1. (24) Han, J. H.; Cho, K. W.; Lee, K. H.; Kim, H. Carbon 1998, 36, 1801. (25) Zuttel, A.; Sudan, P.; Mauron, P.; Wenger, P. Appl. Phys. A: Mater. Sci. Process. 2004, 78, 941. (26) Kayiran, S. B.; Lamari, F. D.; Levesque, D. J. Phys. Chem. B 2004, 108, 15211. (27) Pradhan, B. K.; Harutyunyan, A. R.; Stojkovic, D.; Grossman, J. C.; Zhang, P.; Cole, M. W.; Crespi, V.; Goto, H.; Fujiwara, J.; Eklund, P. C. J. Mater. Res. 2002, 17, 2209. (28) Terres, E.; Panella, B.; Hayashi, T.; Kim, Y. A.; Endo, M.; Dominguez, J. M.; Hirscher, M.; Terrones, H.; Terrones, M. Chem. Phys. Lett. 2005, 403, 363. (29) Kadono, K.; Kajiura, H.; Shiraishi, M. Appl. Phys. Lett. 2003, 83, 3392. (30) Lueking, A. D. Hydrogen Storage in Carbon Nanomaterials: Investigation of Fiber Structure, Metal Content, and Hydrogen Spillover. Ph.D. Thesis, University of Michigan, Ann Arbor, MI, 2003. (31) Gupta, M. Int. J. Quantum Chem. 2000, 77, 982. (32) Dowell, M. B. In Exfoliation of Intercalated Graphites, Part I, Effect of Graphite Crystallinity, 12th Biennial Conference on Carbon, Pittsburgh, PA, July 28-Aug 1, 1975; p 31. (33) Yoshida, A.; Hishiyama, Y.; Inagaki, M. Carbon 1990, 28, 539. (34) Toyoda, M.; Inagaki, M. J. Phys. Chem. Solids 2004, 65, 109.
