The injection of CO2 into CH4 hydrate-bearing sediments causes the release of CH4 and the formation of CO2 hydrate within the CH4 hydrate stability field. CH4-CO2 replacement allows for the recovery of an energy source, CH4, while trapping CO2. In this study, we monitor pore-scale changes in electrical resistance and relative stiffness during CH4 hydrate formation, CH4-CO2 replacement, and hydrate dissociation; experiments are also observed using high-resolution time-lapsed photography. Results show that CH4-CO2 replacement occurs locally and gradually so that the overall hydrate mass remains solid and no stiffness loss should be expected at the sediment scale. Other experimental results confirm the slow diffusion of CH4 through the hydrate shell that forms between water and gas; this may allow for the coexistence of gas-hydrate-water phases for long periods of time.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The extent of the CH4-CO2 replacement is affected by multiple factors and coexisting processes, such as pressure- and temperature-dependent relative viscosity, permeability, density and solubilities among water, CH4 and CO2 [Jung et al., 2010]. Previous studies have observed no apparent dissociation during replacement [Stevens et al., 2008], and have monitored replacement ratios and rates which show that the CH4-CO2 replacement rate increases near the CH4 hydrate phase boundary and with increasing CO2 gas pressure, reaching a constant value when the CO2 liquefies [McGrail et al., 2007; Ota et al., 2005, 2007]. The replacement ratio increases when a mixture of CO2 and N2 is used for replacement because the smaller N2 molecule facilitates the replacement of CH4 from the small cage in structure I hydrate [Park et al., 2006].
The stability of hydrate-bearing sediments during CH4-CO2 replacement is not yet well understood. In this study, we monitor pore-scale changes in electrical resistance and stiffness to gain an in-depth view of ongoing process. We choose these measurements because of the pronounced sensitivity of underlying physical parameters to phase changes. In particular, the electrical resistivities of water, hydrate, liquid CO2 and CH4 gas are ordered as ρH2O < ρhyd < ρCO2-liquid < ρCH4gas from ρ∼0.2 Ω m for seawater to ρ∼∞ for gas. On the other hand, stiffness ranks as follows BCH4gas < BCO2-liquid < BH2O < Bhyd (note that the bulk modulus of liquid CO2 is almost one order of magnitude lower than that of water). These observations guide the design of the device and test methodology used in this study.
2. Experimental Study
The experimental device is designed to explore hydrate formation and CH4-CO2 replacement at a small scale, such as at the water meniscus that forms between particles in a partially water-saturated sediment.
The test consists of a thin cylindrical water layer (8.8 mm diameter, 0.9 mm in height; and 55 mg water mass) retained by surface tension between two conductive aluminum disks (Figure 1a). These disks are bonded onto corresponding piezocrystals. The device is housed in a high-pressure chamber within a temperature controlled environment (Figure 1b). The water droplet is recorded using time-lapse photography to confirm phase changes and to observe volume changes (resolution: 1 pixel∼10 μm). Pressure and temperature are measured with a pressure transducer and a thermocouple, respectively, and values are recorded every 2 s using a data logger.
Figure 1c shows the electrical circuit and peripheral electronics used to measure electrical resistance and relative stiffness. Electrical resistance is determined at 50 kHz to avoid electrode polarization effects. The resistance of the medium R is a function of measured voltages V1 and V2, and the known resistance of the series resistor R* = 4700 Ω,
The source piezocrystal is connected to a sinusoidal signal generator operated at ∼60 kHz. The signal amplitude produced by the output piezocrystal is measured using an oscilloscope.
2.2. Experimental Procedure
Multistage P-T trajectories are imposed in three different experiments. For clarity, a single, complete test sequence is reported in this manuscript. Similar results were obtained in all other tests. The P-T trajectory during this experiment consists of three stages (Figure 2): (1) ice formation and melting followed by CH4 hydrate formation, (2) CH4-CO2 replacement, and (3) hydrate dissociation. Details for each stage follow.
2.2.1. Transient Ice Formation
A droplet of deaired water (ρw = 231 Ω m) is placed between the two aluminum substrates, creating a cylindrically shaped meniscus (see Figure 2a). The chamber is briefly vacuumed, then pressurized with CH4 gas to 8.1 MPa and kept at a temperature ∼277°K for 11 h under quiescent conditions. The pressure and temperature are rapidly decreased to 3.7 MPa and 250°K to form ice (some hydrate may form as well).
2.2.2. CH4 Hydrate Formation
Within 2 min after partial depressurization, pressure and temperature are increased back to 7.6 MPa and 277°K, to melt the ice within the CH4 hydrate stability field (see Figure 2a). These P-T values are maintained constant for 23 h to allow for CH4 hydrate growth.
2.2.3. Injection of CO2
CH4 gas is allowed to leak out of the chamber, and P-T condition is maintained inside of the CH4 hydrate stability field, while CO2 is injected into the chamber (see Figure 2b). Eventually the hydrate mass is submerged in liquid CO2. Pressure and temperature are kept at P = 7 MPa and T = 276 °K for 19 h.
2.2.4. Hydrate Dissociation
Depressurization is conducted in three steps: from liquid CO2 to gas CO2 (points c0 to c1 in Figure 2c), between CH4 and CO2 phase boundaries (points c2 to c3 in Figure 2c), and out of the CO2 hydrate stability field (points c3 to c4 in Figure 2c).
3. Experimental Results
Similar results were obtained in all three multistage tests. For clarity, a data set from a single complete test is reported here. Pressure, temperature, electrical resistance R and relative stiffness K measured during the tests are summarized in Figure 3. All parameters are plotted versus time. Note that time is zeroed at the center of the main process under consideration in each column, and plotted using a cubic root scale to show short-time effects in high resolution together with long-time changes. The evolution of the water droplet photographed through the sapphire window is documented in Figure 4 (for clarity, we show traces of the original photographs).
3.1. Transient Ice Formation
A pronounced increase in resistance and stiffness accompany ice formation (Figure 3a). There is only a minor volume change (Figure 4b).
3.2. CH4 Hydrate Formation
Ice melts and CH4 hydrate starts forming upon repressurization back inside CH4 hydrate stability field. The electrical resistance R and relative stiffness K decrease fast as the ice melts (between points a2 and a3 in Figure 3a). Therefore, there is virtually no hydrate formation during ice melting even though P-T conditions are within the hydrate stability field. This suggests that thermal diffusion-limited ice melting is much faster than gas diffusion-controlled hydrate formation.
Any hydrate that may have formed dissociates between points a3 and a4 (Figures 2 and 3a), then both resistance R and stiffness K begin to gradually increase during CH4 hydrate formation (after point a4 in Figure 3, duration 23 h), however, neither resistance nor stiffness reach the values attained during ice formation. Volume expansion during hydrate growth causes water to flow out of the meniscus, and some hydrate forms on the aluminum surface (Figure 4c).
3.3. Injection of CO2
Minor changes in electrical resistance R and relative stiffness K are observed during the injection of CO2 gas (Figure 3, points b1 to b2 (note that this is confirmed in all our tests)). However, resistance R and stiffness K increase fast as soon as liquid CO2 conditions are exceeded [see also Ota et al., 2007]. Both K and R reach values higher than during CH4 hydrate formation (point b2 in Figure 3). The mixed CH4-CO2 gas leads to a modified G-L CO2 boundary, and liquid CO2 forms above the liquid-gas P-T condition for pure CO2 (Figure 2, point b2 (see related data given by Donnelly and Katz ).
3.4. Hydrate Dissociation
Depressurization from liquid CO2 to gas CO2, and out of the CH4 phase boundary, causes no observable change in the electrical resistance R and relative stiffness K. Therefore, we infer that CO2 hydrate fills the meniscus (Figure 3, points c1 and c2). Finally, hydrate dissociates at the CO2 hydrate phase boundary (Figures 2 and 3, point c3). As hydrate dissociates, resistance R and stiffness K return to the initial values measured for the water droplet at the beginning of the test. The water loss from the beginning to the end of the test is estimated to be ∼15% based on the photographic record.
4. Analyses and Discussion
4.1. Volume Expansion
There is pronounced volume expansion during CH4 hydrate formation; a theoretical estimate shows that VCH4hyd/Vw = 1.23 for a hydration number n = 6. Volume expansion causes water to flow out of the meniscus, readily forming hydrate on the sides of the aluminum block (Figure 4c). CH4-CO2 replacement and additional CO2 hydrate formation of any remaining free water inside the meniscus can cause additional volume expansion as seen in Figure 4d (VCO2hyd/Vw = 1.28). Note that the volume of CO2 hydrate is slightly larger than for CH4 hydrate (VCO2hyd/VCH4hyd = 1%–6% [Jung et al., 2010]).
4.2. Relative Stiffness
Relative stiffness measurements can be analyzed assuming a mechanical system made of three springs in series held between fixed boundaries: the two end springs represent the two piezocrystals, and the central spring corresponds to the meniscus (either water, ice or hydrate). Infinite stiffness connectors between the springs represent the two aluminum disks. The relative amplitude between the input Vi and output Vo voltages is a function of the displacement δi and δo in both input and output piezocrystals, which depends on the meniscus response δm = −δo− δi through a function that combines the stiffness of piezocrystals kpiezo, the meniscus height (Lm = 0.9 mm), the medium Young's modulus Em, and the area of the meniscus Am = 60.8 mm2,
where α is the ratio between the mechanoelectric and electromechanical piezocrystal effects. Parameters α and kpiezo are inferred by assuming known condition at 100% ice and 100% CO2 hydrate (α = 1.39 and kpiezo = 2.62 × 109 N/m assuming Eice = 9.5 GPa and Ehyd = 8.4 GPa). Equation (2) shows that the voltage ratio Vo/Vi is indeed a measure of meniscus stiffness EmAm/Lm relative to the stiffness of piezocrystals kpiezo. The CH4 hydrate mass obtained using the measured voltage ratio (Vo/Vi)CH4hyd = 0.129 is 47% of the meniscus volume.
4.3. Electrical Resistance
Electrical resistance R is a function of resistivity ρ, meniscus length Lm, area Am, and a shape factor β,
When an annular CH4 hydrate shell forms, the measured resistance reflects the contributions of water and hydrate in parallel disregarding ion exclusion.
where the final approximation applies to a shape factor β = 1 for a short cylinder and a ratio of resistivities ρice/ρwater≈ρhyd/ρwater ≪ 1.0. For an initial water resistivity ρwater = 231 Ω m measured before CH4 hydrate formation, a lower bound estimated (disregarding ion exclusion) of the CH4 hydrate volume is 48% of the total meniscus volume. We conclude that (1) a significant part of the meniscus remains as free water 23 h after the initiation of CH4 hydrate formation and (2) the computed CH4 hydrate growth rate confirms that CH4 hydrate formation is CH4 diffusion-limited through the annular hydrate shell (CH4 gas diffusivity through CH4 hydrate 7.6 × 10−13 m2/s [Davies et al., 2008]).
Both relative stiffness and electrical resistance increase at all times during replacement. Therefore, while the transformation requires the opening of the hydrate cage to release the CH4 and entrap the CO2 molecule [Jung et al., 2010], this solid-liquid-solid exchange takes place locally at the reaction front, while the rest of the hydrate mass remains solid. Therefore, no stiffness loss should be expected at the sediment scale.
The CH4-CO2 exchange rate is faster than the rate of CH4 hydrate formation (data in Figures 3b and 3c), and there is additional volume expansion (compare pictures traced in Figures 4c and 4d). Both observations point toward the formation of a porous and pervious CO2 hydrate shell, probably due to the liberation and expansion of CH4 gas.
Pore-scale electrical resistance and relative stiffness measurements provide unique insight into hydrate formation, CH4-CO2 replacement, and hydrate dissociation.
In the absence of fluid flow, CH4 hydrate formation is diffusion-controlled initially through the water phase until hydrate forms. Thereafter, CH4 must diffuse through the hydrate mass to reach any isolated free water that is surrounded by hydrate. Consequently, free water can remain in an excess CH4 gas system for a relatively long time.
Hydrate formation is much slower than thermal diffusion limited ice melting (at mm scale). Therefore, hydrate formation is not concurrent with ice melting within hydrate stability field conditions in most laboratory situations.
Both CH4 hydrate formation and CH4-CO2 replacement cause pronounced volume expansion. During replacement, the newly formed CO2 hydrate shell must be fractured or porous in order to allow for the high exchange rates observed in this study.
While CH4-CO2 replacement requires the opening of the hydrate cage (i.e., a solid-liquid-solid transformation), both electrical resistance and relative stiffness measurement suggest that CH4-CO2 replacement occurs locally and gradually so that the overall hydrate mass remains solid and no stiffness loss should be expected at the sediment scale.
Ratio between the mechanoelectric and electromechanical piezocrystal effects.
Electrical resistivity (Ωm).
Bulk modulus (Pa).
Young's modulus (Pa).
Support for this research was provided by U.S. Department of Energy. Additional funding was provided by the Goizueta Foundation. We are grateful to Connor Barrett for proofreading the manuscript.