D

10 nm

1 nm

5-50 nm

50-200 nm

2.1. Hydrogen Storage

Measurement Techniques

The main methods used for determining the hydrogen storage capacity are temperature-programmed desorption (TPD), volumetric analysis, thermogravimetric analysis (TGA), and the electrochemical method.

The temperature-programmed desorption method consists of using a mass spectrometer to measure the hydrogen desorbed from a carbon sample during controlled heating [26]. Thermal desorption spectroscopy is highly sensitive, allowing the study of samples with masses below 1 mg. The sensitivity can be even improved by using deuterium-loaded samples, and avoiding any disturbing background from water or other hydrogen-containing adsorbates [27]. For calibration, the hydrogen desorption of a well-known metal hydride or an alloy of known hydrogen content has to be measured [28]. The activation energy of desorption can be directly measured from the temperature at which hydrogen appears in the spectrometer. As different adsorption mechanisms and sites will have different activation energies, each peak in the thermogram will indicate a different site or mechanism. It will be shown that carbon adsorbents have a desorption peak around 130 K, and nanotubes can have one (at ~300 K) or more additional peaks. The amount of hydrogen desorbed requires integration of the signal, and may have significant errors [26]. In volumetric analysis, the adsorption of hydrogen is measured by a pressure change in a fixed volume, with pressure variations attributed to adsorption or desorption. A complication with this technique is the presence of thermal effects during the filling of the sample cell. Hydrogen has an appreciable temperature-dependent compressibility that needs to be taken into account as the temperature changes due to compressing the gas during filling and the exothermic adsorption. Unless this problem is specifically addressed, the change in pressure caused by temperature fluctuations will cause overestimation of the amount of hydrogen adsorbed [18]. The calibration should be done by measuring a well-known metal hydride.

The thermogravimetric analysis method consists of measuring the weight of a sample as the temperature is varied under constant pressure. The technique is well known, with the difficulty being the sensitivity of the instrumentation on the relatively small sample sizes used in these studies. The gravimetric analysis is capable of measuring very low sample masses of about 10 mg in specially designed devices, but it is a nonselective analysis. The high specific surface area of carbon nanotubes increases the possibility of the adsorption of residual gases. Therefore, the apparatus has to be extremely clean, and high-purity hydrogen has to be used.

Hydrogen can be incorporated into the sample due to the electrochemical technique. To prepare the carbon-containing working electrode, the carbon material has to be mixed with conductive powder, for example, nickel or gold and compacted [29-31]. The counterelectrode is, for example, metallic nickel. Both electrodes are placed in a KOH solution which provides the hydrogen atoms, and they are separated by a polymer separator. During the charging process, water dissociates at the negative working electrode, and atomic hydrogen may intercalate into the carbon material. The following discharge process results in the recombin-ing of water. By maintaining a constant current, the voltage is measured across the two electrodes during the charging and discharging. The amount of desorbed hydrogen is determined by measuring the electric charge in a galvanostatic setup.

A comprehensive study which involved a broad range of carbon materials was conducted recently by Tibbetts et al. [32], whose experimental results cast serious doubts on many claims so far for room-temperature hydrogen sorption in carbon materials larger than 1 wt%. The authors analyzed the possible pitfalls in the gaseous hydrogen adsorption process using volumetric techniques [32], and proposed techniques to handle all of those problems [33]. The chief areas where hydrogen adsorption experiments may be in error are as follows.

The first is that leaks can be difficult to distinguish from hydrogen storage. Hydrogen molecules disappear from a pressurized reservoir. If hydrogen molecules leak from a system of volume V at a nearly constant rate L = dP/dt over the period of the experiment t, an observer could incorrectly attribute the pressure decrease to a wt% hydrogen stored in a sample of mass m proportional to VL/m. For small sample masses, the consequences of this can be particularly treacherous.

A second source of error is that, in a high-pressure experiment, pressure changes arising from ambient temperature variations can be mistakenly interpreted as substantial sorption. At 10 MPa and 300 K, the pressure change with temperature is 33.3 kPa/K. If this were interpreted as sorption by a 1 g sample in a 1 L volume, the result would be 2.6 wt% per degree temperature drop. Most reports in the literature have samples smaller than 1 g.

A more significant, but less obvious source of temperature variation stems from the thermodynamic principle that a chamber being pressurized experiences a temperature rise. The enthalpy of high-pressure gas is higher than that of low-pressure gas, and so the flux of high-enthalpy molecules from a high-pressure chamber to a low-pressure chamber raises the temperature of the gas in the low-pressure chamber. This temperature is not easily measured if the temperature sensor is not in good thermal contact to the gas or if the walls of the chamber have a large heat capacity. For an ideal gas adiabatic ally bled into an evacuated chamber, the temperature rise is (Cp/Cv - 1)/nitial, corresponding to a 41% increase in absolute temperature for hydrogen [34]. The authors believed that many high-pressure adsorption experimental results should be carefully corrected to exclude the pressure drop brought by return to thermal equilibrium after pressurization. Figure 3 shows the effect of thermal equilibrium after pressurization on the pressure change in the sample chamber [32].

It is worth noting that the metal hydride research community has devised experimental techniques to handle all of these problems [33]. First, to avoid spurious readings from gas cooling to ambient, use two calibrated volumes connected by a valve, one an initially pressurized reservoir and the other containing the sample. Then determine the precise equilibrium temperatures and pressures of both volumes before and after opening the connecting valve. Second, to unambiguously establish that there are no leaks, measure not only the gas adsorbed, but also the gas desorbed from

Figure 3. Hydrogen pressure versus time for a 307 mL chamber containing no sample. This shows pressurization to 11.97 MPa from vacuum at 25.3 h and the subsequent pressure decrease. Reprinted with permission from [32], G. G. Tibbetts et al., Carbon 39, 2291 (2001). © 2001, Elsevier Science.

Figure 3. Hydrogen pressure versus time for a 307 mL chamber containing no sample. This shows pressurization to 11.97 MPa from vacuum at 25.3 h and the subsequent pressure decrease. Reprinted with permission from [32], G. G. Tibbetts et al., Carbon 39, 2291 (2001). © 2001, Elsevier Science.

the sample. Third, present a complete pressure composition isotherm that precisely defines the pressure and composition at which the hydrogen-adsorbent bond is formed and decomposes. A complete pressure composition isotherm would show an interesting change in curvature with pressure that would be very useful in classifying the sorption, and lending credibility to the research [32].

2.2. Gaseous Hydrogen Storage in CNTs

A summary of gaseous hydrogen storage in undoped and doped carbon nanotubes from the literature is given in Table 1.

Dillon et al. [26] were the first to publish experimental data on hydrogen adsorption in nanotubes, and measured exactly the desorption of hydrogen of nonpurified SWNT samples (containing additional cobalt catalyst and amorphous carbon). Using TPD, the authors found that the hydrogen desorbed from both the nanotubes and activated carbon samples at about the same temperature (~133 K). However, a second peak appeared at higher temperature for SWNTs (~290 K), indicating that there were additional adsorption sites in SWNTs. They proposed that these sites were when hydrogen molecules had access to the nanotube cores. The authors did not see this high-temperature site in either activated carbon or in arc-generated soot produced without a catalyst. The SWNT constituent of the sample appeared to be especially effective for hydrogen adsorption, and an activation energy for hydrogen desorption was found to be 19.6 kJ • mol-1 (approximately five times that of planar graphite). In 2000, Heben's group performed inelastic neutron scattering on this material, and showed that hydrogen is physisorbed [35], but the authors made no statement about the total amount.

The hydrogen storage capacity was estimated to range between 5 and 10 wt% [26]. This figure was derived from the measured hydrogen desorption of 0.01 wt% and an SWNT content estimated at 0.2 wt%. Furthermore, it was assumed that only the SWNTs in the sample contributed to the hydrogen adsorption, so the analysis required a large correction for 99.8 wt% of material that was assumed inert. However, hydrogen adsorption in high-porosity carbon (AX-21 carbon) is as high as 5.3 wt% at a temperature of 77 K and a hydrogen pressure of 1 MPa [36].

In a more recent paper, Dillon et al. [37] presented an oxidative technique to open the nanotubes: the degassing of the samples in vacuum to 970 K, and an oxidizing in water in the 325-975 K temperature range. They measured the adsorption of hydrogen in such treated samples by their TPD method, and found a noticeable enhancement (up to a factor of 3 in the most favorable case) of the characteristic desorption peak between 250 and 300 K. They attributed this improvement of the adsorption capacity to the opening and filling of the nanotubes. Taking into account that their carbon soot samples containing only 0.05% of nanotubes can adsorb ~0.005% weight of hydrogen, Dillon and his colleagues believed that the pure adsorbed hydrogen was therefore ~10 wt%. These authors have further developed a cutting method to produce samples with a high concentration of short SWNTs with open ends that are accessible to the entry of hydrogen molecules. With these better characterized nanotube samples, they were able to achieve a more accurate determination of hydrogen adsorption, which they estimated to be at ~3.5-4.5 wt% hydrogen at room temperature and 0.07 MPa hydrogen pressure [38]. By opening and recapping the nanotubes, they concluded that most of the hydrogen is stored within the capillary, rather than in the interstitial spaces between the SWNT.

Ye et al. [39] were motivated to perform measurements on SWNT material of high purity because the previous measurements were made on dilute SWNTs which required a large correction for material that was assumed inert. Ye et al. were the first to report hydrogen-adsorption investigations on purified [40] laser-generated [41] SWNTs. They undertook their experimental measurements of hydrogen adsorption on high-purity "cut SWNTs." To cut the SWNT material and disrupt the tightly packed rope structure found in highly crystalline pristine SWNTs, a small quantity of material was sonicated in dimethylformamide, and then extracted from the solvent by vacuum filtration. TEM revealed that this treatment broadens the diameter distribution of the SWNT ropes, and also increases the number of SWNT terminations within ropes (hence the SWNTs are said to be "cut"). These differences were not evident in the X-ray diffraction patterns of the material before and after the treatment, perhaps indicating that the average rope diameter was largely unchanged.

The adsorption of hydrogen was measured by the volumetric method without correction for thermal effects. The highest gravimetric hydrogen storage capacity achieved in SWNT material treated in this manner was 8.25 wt% (an H/C atomic ratio of 1.0), at a temperature of 80 K and pressure of MPa. At a pressure of MPa, a sudden increase in the adsorption capacities of the SWNT samples was reported; the authors suspected that a structural phase transition was responsible for this effect. In their method, the ropes were split into individual tubes, thereby increasing the surface area available for physisorption. In their conclusion, the authors [39] advanced that the ability of SWNT

Table 1. Summary of the reported gaseous hydrogen storage capacity in carbon nanotubes and carbon nanofibers.

Sample

Ref.

0 0

Post a comment