3.1. Calorimetry
The objective of the MCDSC experiments was to identify the best candidates to form binary hydrogen hydrates. Three different aqueous solutions of THF (2.78, 5.56 and 8.34 mol %) and two of TBAB (1.38 and 3.59 mol %) were compared at H
2 pressures of 6.89, 10.37 13.79 MPa (gauge). In these tests, during the last stage of the programmed cooling and heating process shown in
Figure 8, the calorimeter sample temperature was slowly ramped from 263 K to 293 K at a constant rate of 0.1 K/min. Triplicate experiments were performed for each concentration of THF and TBAB.
Calorimeter thermograms of the final hydrate (and ice) decomposition (heating) step, for 2.78, 5.56 and 8.34 mol % THF, are shown in
Figure 10,
Figure 11 and
Figure 12, respectively. While the calorimeter measures the rate of heat flow (µJ/s) to or from a sample, we have divided these data by the measured mass of the initial liquid solution in each cell, to account for small variations of this quantity between samples, and also the user-selected temperature ramping rate (0.1 K/min). This results in a parameter with the same units as specific heat capacity (J/kg-K) but which, in this case, includes both the sensible and latent heat components of the material comprising the sample. This parameter is plotted as a function of temperature, where endothermic processes correspond to positive values.
An initial test was conducted at atmospheric pressure (without H
2 addition) as a baseline. At 2.78 mol % THF, the solution contains about half the THF required to form pure hydrate [
12].
Figure 10 shows a large endothermic peak with a small shoulder to the right. For the atmospheric pressure case, the apex of the large peak occurs at approximately 273.5 K and corresponds to ice melting. The small shoulder to the right of the ice peak is associated with the THF hydrate and its apex occurs at 277 K. As H
2 pressure in the cell increases, the binary hydrate peak decouples from the ice peak and moves to higher temperatures. As expected, the ice melting peak shifts to lower temperatures with increasing pressure.
Results for 5.56 mol % THF are shown in
Figure 11. All thermograms exhibit a large endothermic peak near 273 K, indicating that ice continues to form even though the solution contains sufficient THF to ensure no excess water. The second peak between 277 K and 280 K occurs for the no-H
2 case (i.e., 0 MPa gauge pressure) and is believed to represent decomposition of pure THF hydrate. As the sample is pressurized with H
2, a small shoulder appears at higher temperatures to the right of the THF peak, possibly due to the dissociation of hydrogen hydrate. To test this, a calorimetry experiment was performed where a 5.56 mol % THF solution was pressurized to 10.34 MPa with pure N
2 gas. The measured thermogram is compared with the thermogram of the 5.56 mol % THF solution pressurized with H
2 in
Figure 13. In the absence of H
2, the third peak to the right of the THF peak does not appear. There is, however, another new peak, this time to the left of the THF hydrate peak which may reflect the presence of N
2 in the solid phase.
When THF concentration in the solution was increased to 8.34 mol %, the ice melting peak in the thermograms disappears, as seen in
Figure 12. All of the solution is converted to hydrate. The two remaining endothermic events (indicated by the two peaks) correlate well with thermograms obtained at 5.56 mol % THF.
The fraction of the solution that forms hydrate at different pressures can be estimated by deconvolving the ice and hydrate peaks, and calculating the area under the ice peak and above a baseline that accounts for the sensible heat component of the sample. Different algorithms to determine the sensible heat baseline from the thermograms were tested and are discussed by Weissman [
20]. The area under the ice peak can be multiplied by the sample mass to obtain the total energy required to melt the ice component. Dividing this by literature values of the latent heat of pure water ice, ΔH
f, yields the mass of the ice component, which is then subtracted from the total mass of the sample. Following this approach, about 35% of the 5.56 mol % THF solution is converted into hydrate at atmospheric pressure. Pressurizing the sample with H
2 gas increases this percentage to 75%, 76%, and 78% at 6.89, 10.34, and 13.79 MPa, respectively.
Calorimetry results for binary hydrate formed from aqueous solutions of TBAB are shown in
Figure 14 and
Figure 15. TBAB has been proposed to reduce the formation pressure of H
2 hydrate [
6,
21,
22]. TBAB is known to form semi-clathrate hydrate in which the water cage is broken to accommodate the large TBAB molecule [
23]. The ratio of TBAB:water in the hydrate has been reported to range from 1:2 to 1:36. 20% and 40% by mass solutions of TBAB were investigated which correspond to 1.38 mol % (~1:72) and 3.59 mol % (~1:27), respectively.
The thermograms in
Figure 14, obtained using 1.38 mol % TBAB solution, exhibit two endothermic peaks, similar to the 2.78 mol % THF case. It can be surmised that the strong peak to the left corresponds to the melting of ice, indicating that the solution is undersaturated with TBAB such that significant excess water exists after the hydrate component forms as the sample is cooled. When the sample is pressurized with H
2 gas, the hydrate dissociation peak begins to broaden and shift to higher temperature. At 13.79 MPa, there is some evidence of a second hydrate decomposition peak. The area of the ice peaks does not appear to change significantly, suggesting that the partitioning of the sample between hydrate and ice is not affected much by increasing pressure.
No ice peak is detected in the thermograms shown in
Figure 15, indicating that at 3.59 mol % TBAB (i.e., TBAB:water ratio of 1:27), there is no excess water. At this concentration, TBAB has been reported to form two types of hydrates: Type A and Type B [
7]. The thermograms, however, exhibit three peaks, possibly suggesting three hydrate structures. Since these three peaks can be detected in the atmospheric pressure case where there is no H
2 gas, it is unclear to what extent hydrogen is being incorporated into the hydrate matrix. The general shape of the curves does not change significantly with increasing pressure, although there appears to be a subtle shift in the relative strengths of the second and third peaks, as well as the expected shift to higher melting temperatures.
The calorimetry experiments suggest that pressurization of THF solution with H2 gas during the hydrate formation process results in significant changes in the dissociation thermograms. At the two higher concentrations of THF, a second hydrate peak appears at temperatures between around 280 K and 285 K that increases in strength with increasing pressure. This suggests an additional hydrate structure, possibly due to the incorporation of H2 into the crystal matrix. On the other hand, thermograms of the TBAB solutions did not exhibit conclusive evidence of changes in structure when pressurized with H2. While this does not preclude the possibility of binary TBAB + H2 hydrate formation, it was decided that THF would be a better candidate for our scale-up tests. One advantage that TBAB has over THF is the slightly higher melting point of the hydrate (about 4 K) at the same pressure.
3.2. Raman Spectroscopy
Calorimetry data indicated that an additional hydrate decomposition peak appears when the 5.56 mol % and 8.34 mol % THF solutions were pressurized with H2 gas. It was posited that this peak is associated with H2 in the hydrate phase. Raman spectroscopy experiments were performed to test this hypothesis by confirming the existence of H2 binary hydrate under conditions where the thermogram peak was detected.
Following procedures described above, Raman spectra were taken at various stages of the gas hydrate dissociation process. Fine ice crystals of 5.56 mol % THF were slowly pressurized with H
2 gas to 6.89 MPa at 263 K and an initial spectrum was collected before significant uptake of hydrogen by the solid phase occurred. This spectrum is shown in
Figure 16, which plots light intensity measured by the CCD detector, in arbitrary units (a.u.), as a function of the Raman shift (cm
−1). In the gas phase overlying the ice and (pure) THF hydrate, four peaks are detected at 4128, 4146, 4158, and 4165 cm
−1 which corresponds to the H-H stretching modes of the H
2 molecules. Hydrogen hydrate was then allowed to form.
Figure 17 presents a spectrum of the resulting hydrate at 274 K. The broad peak around 4133 cm
−1 was observed previously by Lee et al. [
8] and is due to the entrapment of H
2 in small cages of the binary hydrate. The sample was depressurized and vented and a final spectrum was taken of the 5.56 mol % THF solution at 298 K for comparison. This spectrum is shown in
Figure 18. These results appear to confirm the formation of the binary hydrogen hydrate.
3.3. Scale-Up
Scale-up experiments were conducted to estimate the amount of H
2 stored in THF binary hydrates. Quantifying H
2 content is not possible with the MCDSC system.
Table 1 summarizes the H
2 storage capacity of 2.78, 5.56, and 8.34 mol % THF hydrate determined from GC analysis of gas samples collected from replicate experiments. In these experiments, the samples were pressurized with pre-cooled H
2 gas to 12 MPa, which falls between the two highest pressures tested with the MCDSC. Due to the pressure ratings of certain components in the scale-up facility flowtrain, a safe operating limit of approximately 2000 psig (13.9 MPa absolute) was established for the system. 12 MPa (gauge) was selected to accommodate the possibility of pressure excursions within this safety limit. As described previously, gas samples collected at the end of the N
2 purge process at 274 K are analyzed to determine the number of moles of any residual H
2 in the gas phase, based on the measured concentration, pressure, temperature, and known volume of gas in the lines and reactor head space. This is subtracted from the number of moles of H
2 collected after the hydrate fully dissociates (determined from the GC data) to estimate the amount of hydrogen that was stored in the hydrate.
In
Table 1, ∆n is the number of moles of hydrogen stored in the hydrate sample. The weight % given in the last column on the right is calculated by dividing the total mass of H
2 determined to have been stored in the hydrate by the mass of the fine crystals loaded into the reactor at the start of the experiment. Based on the MCDSC results, a portion of 2.78 mol % and 5.56 mol % samples probably consisted of ice, rather than hydrate. Since this fraction is unknown, however, the total sample mass was used as a basis for comparison of the results.
These results suggest that, while H2 can be stored in THF binary hydrate, the yield is low. The weight percentages fall far below the 2015 USDOE target of 5.5 weight %. It is interesting to note that the most dilute THF solution (2.78 mol %) had the highest hydrogen storage capacity. The binary gas hydrate formed from this solution released approximately 0.083 g of H2, corresponding to a 0.05 wt % storage capacity. Increasing the THF concentration to 5.56 mol % THF lowers the gas storage capacity by a factor of two. For 5.56 mol % THF, dissociation of the hydrate released approximately 0.045 g of H2, corresponding to 0.027 wt % storage capacity. The 8.34 mol % THF supersaturated solution also had a storage capacity of about 0.027 wt %, suggesting that increasing the THF concentration beyond stoichiometric proportions does not necessarily improve H2 yield.
In consideration of the unexpectedly low H
2 storage capacities determined by gas sampling and analysis, the experimental pressure records were examined to try to estimate H
2 gas uptake during the hydrate formation process to confirm these results. During the formation process, a pressure drop occurs at constant temperature due to the entrapment of H
2 in S-cages generated by (sII) THF hydrate. Since the volume occupied by the isothermal gas in the reactor is fixed, the change in the number of moles of H
2 in the gas results in a decrease in pressure. The Ideal Gas Law was employed to estimate the decrease in moles of H
2 from the measured temperature, pressure drop, and known gas volume, using an appropriate compressibility factor:
where Δn is the change in the number of moles of H
2 due to uptake by the hydrate, ΔP is the measured pressure drop, V is the volume of the gas space (98.3 cm
3) that was determined by filling the closed system with fluid and collecting and measuring the volume of that fluid, T is the measured temperature, R is the gas constant, and Z is the calculated compressibility factor, which is approximately 1.09 at the tested hydrate formation conditions.
The formation process for 2.78 mol % THF is shown as an example in
Figure 19. Measured H
2 gas pressure data are plotted as a function of temperature. At the start of the experiment, point A, H
2 pressure in the system is 11.15 MPa and sample temperature is about 274 K. As the sample temperature is cooled rapidly from 274 K to 263 K (point B), H
2 gas pressure drops slightly due to the cooling effect. From point B to point C, H
2 is drawn into and trapped in the S-cages of the THF hydrate, resulting in a reduction of gas pressure from slightly more than 11 MPa to slightly less than 10.4 MPa. Applying the above equation, this corresponds to the uptake of about 0.060 g of H
2. Note that the oscillation in the curve during this period reflects the temperature control characteristics of the freezer. The system temperature is then raised to 277 K. By comparison with
Figure 10, this should melt any ice but preserve most of the hydrate. The sample is held at this temperature for 4 h to promote the formation of more hydrate, then cooled to 274 K and held for an additional 24 h.
Figure 20 shows a second, smaller pressure drop that is recorded at 274 K. During this process, an additional 0.036 g of H
2 is consumed, for a total of 0.096 g. This value compares reasonably well with GC results of 0.083 g of H
2 released during dissociation.
Table 2 compares the results of the gas sampling and pressure drop analyses. While differences are apparent between these two methods, the results compare reasonably well and seem to confirm that, for the specific protocols and conditions examined in this exploratory study, it was not possible to produce binary THF + H
2 hydrate with significant H
2 storage capacity. The measured values in the present scale-up tests are four to ten times smaller than results reported for previous experimental studies conducted at similar THF concentrations, pressures, and temperatures [
13,
24], and which have been predicted by Grand Canonical Monte Carlo simulations performed recently by Papadimitriou et al. [
25]. Although the binary hydrate formation process was allowed to proceed for more than 40 h, kinetic effects may have contributed to the discrepancy. In addition, given the relatively large scale of the reactor, incomplete contact and associated mass transfer restrictions between the H
2 gas phase and sections of the solid ice-hydrate matrix could have limited H
2 uptake.