Dissolution Process Observation of Methane Bubbles in the Deep Ocean Simulator Facility

To investigate the temperature dependency of the methane bubble dissolution rate, buoyant single methane bubbles were held stationary in a countercurrent water flow at a pressure of 6.9 MPa and temperatures ranging from 288 K to 303 K. The 1 to 3 mm diameter bubbles were analyzed by observation through the pressure chamber viewport using a bi-telecentric CCD camera. The dissolution rate in artificial seawater was approximately two times smaller than that in pure water. Furthermore, it was observed that the methane bubble dissolution rate increased with temperature, suggesting that bubble dissolution is a thermal activation process (the activation energy is estimated to be 9.0 kJ/mol). The results were different from the expected values calculated using the governing equation for methane dissolution in water. The dissolution modeling of methane bubbles in the mid-to-shallow depth of seawater was revised based on the current results.


Introduction
When methane (CH 4 ) and other gaseous hydrocarbons are released into the ocean from natural bottom seeps or oil and gas industry activities such as the development of offshore production wells or during deep oil spills, they can lead to significant environmental consequences. These hydrocarbons can dissolve or may escape through the air-ocean interface. Microbial metabolism is recognized as a major sink of hydrocarbon contaminants in the ocean, but the mechanisms and rates of biological degradation and sedimentation are not well understood.
Natural gas hydrates (NGHs) observed in offshore sediments are formed under high pressure and low temperatures. NGHs contain water and a large amount of natural gases, including CH 4 generated by biogenic and thermogenic processes. NGHs have been reported to exist in the seabed in various parts of the world, and they have attracted attention as unconventional natural gas resources (e.g., [1,2]) and, in the geological record [3], as one of the contributing factors to climate change (e.g., [4,5]) and large-scale continental slope instability [6].
Investigations of the distribution of NGH in the offshore seabed are in progress. In areas where biogenic natural gas is dominant, hydrates are observed in the pore spaces of unconsolidated clastic Recently, as part of offshore CH 4 -hydrate research and development activities, the "Methane Gases from Deepwater (MEGADEEP)" model [34] and "MEGADEEP-Eco" model [35] were created to simulate the diffusion behavior of CH 4 gas released from the seabed. These models were applied to CH 4 -hydrate activities in the Nankai trough [36] by the MH21 project in Japan. They were also used to simulate CH 4 emission from offshore CH 4 -hydrate reservoirs in the East China Sea [37] and in the Japan Sea [38].
Previous research has provided significant insight into the dissolution process of CH 4 bubbles in seawater; however, the majority of laboratory experiments have been performed using pure water and field studies have been conducted in seawater. The effect of salts on the dissolution rate of CH 4 bubbles is estimated only indirectly, and studies that have made direct comparisons are limited. Furthermore, laboratory experiments have mainly focused on the effects of the hydrate film formation on the dissolving bubbles, and only a very small number of measurements have been made of the bubble dissolution rate under non-hydrate forming conditions. In consideration of these deficiencies, we have attempted to measure the dissolution rate of CH 4 bubbles outside of the hydrate film-forming regime. We determined the dissolution rate of single CH 4 bubbles in a high-pressure vessel in pure water and artificial seawater and also evaluated the temperature dependency of the dissolution rates. These experimental results were compared with predicted values computed by models [29].

Materials and Methods
Experiments were performed in a facility that can simulate conditions in the ocean from the surface down to 1000 m depth. The facility allows the continuous monitoring of buoyant gas bubbles for extended periods. A downward flow of water over the rising bubble generates a downward drag force. The flow rate of water over the bubble can be adjusted so that buoyancy and drag forces are balanced and the bubble is held stationary in space. In the frame of reference of the bubble, this simulates buoyant rise through the water column. In this study, the dissolution rates of single CH 4 bubbles under dynamic (i.e., advective), non-hydrate-forming conditions were inferred from data on bubble shrinkage over time.
The University of Hawai'i Deep Ocean Simulator (DOS) was employed in this study. The DOS has an internal volume of about 100 L and can operate safely at internal hydrostatic pressures up to 10.4 MPa. The DOS consists of a pressure vessel equipped with observation windows, a high-pressure water circulation system, a temperature controller, a gas supply system, and an imaging system ( Figure 1). Previous studies [11,12] have confirmed that this system is capable of monitoring the behavior of single and multiple buoyant droplets and bubbles over long time periods.
Energies 2020, 13, x FOR PEER REVIEW 3 of 11 used to simulate CH4 emission from offshore CH4-hydrate reservoirs in the East China Sea [37] and in the Japan Sea [38]. Previous research has provided significant insight into the dissolution process of CH4 bubbles in seawater; however, the majority of laboratory experiments have been performed using pure water and field studies have been conducted in seawater. The effect of salts on the dissolution rate of CH4 bubbles is estimated only indirectly, and studies that have made direct comparisons are limited. Furthermore, laboratory experiments have mainly focused on the effects of the hydrate film formation on the dissolving bubbles, and only a very small number of measurements have been made of the bubble dissolution rate under non-hydrate forming conditions. In consideration of these deficiencies, we have attempted to measure the dissolution rate of CH4 bubbles outside of the hydrate film-forming regime. We determined the dissolution rate of single CH4 bubbles in a high-pressure vessel in pure water and artificial seawater and also evaluated the temperature dependency of the dissolution rates. These experimental results were compared with predicted values computed by models [29].

Materials and Methods
Experiments were performed in a facility that can simulate conditions in the ocean from the surface down to 1000 m depth. The facility allows the continuous monitoring of buoyant gas bubbles for extended periods. A downward flow of water over the rising bubble generates a downward drag force. The flow rate of water over the bubble can be adjusted so that buoyancy and drag forces are balanced and the bubble is held stationary in space. In the frame of reference of the bubble, this simulates buoyant rise through the water column. In this study, the dissolution rates of single CH4 bubbles under dynamic (i.e., advective), non-hydrate-forming conditions were inferred from data on bubble shrinkage over time.
The University of Hawai'i Deep Ocean Simulator (DOS) was employed in this study. The DOS has an internal volume of about 100 L and can operate safely at internal hydrostatic pressures up to 10.4 MPa. The DOS consists of a pressure vessel equipped with observation windows, a high-pressure water circulation system, a temperature controller, a gas supply system, and an imaging system ( Figure 1). Previous studies [11,12] have confirmed that this system is capable of monitoring the behavior of single and multiple buoyant droplets and bubbles over long time periods.   The system is filled with about 100 L of either tap water (pure water) or synthetic seawater. While natural seawater was available for the present experiments, complications associated with the variability of its properties-even when collected at the same sampling site but at different times-were encountered in earlier investigations by the authors. To minimize this problem, and to ensure that results could be replicated by others, we elected to use artificial seawater with known (and fixed) composition and minimal biological loading. We believed that this would facilitate comparisons between the data and model results for the gas dissolution rate. Moreover, the use of artificial seawater in laboratory marine science experiments is a common practice.
Synthetic seawater is prepared by mixing Instant Ocean with tap water to obtain a salinity of 3.5 ± 0.2 wt% NaCl. The Instant Ocean aquarium mix contains naturally occurring salts including sodium chloride, magnesium chloride, sodium sulfate, calcium chloride, and potassium chloride [39]. We selected Instant Ocean for preparing the artificial seawater since it is widely available and has been previously characterized and employed in a number of scientific studies (e.g., [40]). The pressurized liquid in the system is circulated with a variable speed, magnetically coupled gear pump (Micropump model 223/56 C; Leeson Micro Series AC Inverter, maximum discharge rate: 12 L/min). Two bath chillers (Thermo Scientific model ThermoChill III and model IsoTemp 6200) are employed to maintain the water temperature in the range extending from 278 to 308 K, with an accuracy of ±1 K. The water flow rate in the water tunnel can be adjusted between 5 and 50 cm/s. For a dissolving CH 4 bubble with a diameter between 10 and 1 mm, the flow rate in the water tunnel test section was adjusted over a range of approximately 20 to 16 cm/s. The water tunnel consists of a clear acrylic inlet contraction and constant area viewing section cemented to a polycarbonate diffuser. The diffuser downstream of the viewing section ensures that gas bubbles under observation are not swept out of the water tunnel as they dissolve, since the downward fluid velocity decreases rapidly by a factor of about 20 from the inlet to exit in the diffuser. This allows us to periodically adjust (i.e., reduce) the fluid flow rate and reposition the dissolving bubbles back in the viewing section for continued observation. The inlet contraction of the plastic water tunnel is employed to minimize boundary layers and reduce turbulence levels in the 25.4 mm inner diameter x 102 mm length, cylindrical test (viewing) section.
To investigate the temperature dependency of a rising bubble from the deep sea to the surface layer under the condition where CH 4 hydrate does not form, similar tests were conducted at 288, 293, 298, and 303 K (each ±1 K). The pressure was fixed at 6.9 MPa ± 0.14 MPa.
The CH 4 gas prepared by Airgas, Inc. that was used in the experiments had 99.99% purity. The gas was injected into the water from a needle positioned at the base of the pressure vessel.
The clear, constant area viewing section of the water tunnel was aligned horizontally with the two large viewports. Gas bubbles stabilized with the downward flow of water were illuminated with an LED panel (Metaphase Technologies model FR-BL) mounted in front of one viewport and were monitored with a telecentric CCD camera (Prosilica GT1920 with Thorlabs 0.128 Bi-telecentric lens) through the opposite viewport. The telecentric camera lens provided constant magnification across the span of the 25.4 mm inner diameter viewing section, to avoid errors in estimating the changes in size over time of the dissolving bubbles as they wandered laterally toward and away from the camera. Images were recorded at intervals of up to 0.5 s. Images were analyzed every 3 to 30 s, depending on the size change rate. Image analysis was done post-experiment with LabView's Vision software. Due to minor bubble movement during image capture, the spatial resolution of the image measurements was estimated to be 0.05 mm.
The shape of the bubble was assumed to be an oblate spheroid, and the major axis (a) and the minor axis (b) were estimated from the 2D image data. The time change of the spherical diameter (d e ) of the equivalent volume (v), calculated by Equation (1) The Heywood circularity factor (H) also was calculated to assess the bubble-shape change. Three or more replicates were performed at each condition; the number of samples analyzed for each data point reported below was n = 3 to 6. Figure 2 shows a typical image of a CH 4 bubble observed in the high-pressure vessel (a sample of the bubble tracking animation is provided in Supplementary Materials Video S1).

Results and Discussion
Based on the (time) sequence of the bubble images, the temporal evolution of the bubble diameter was determined from Equation (1). As shown in Figure 3, the equivalent diameter d e appears to decrease linearly with time. The dissolution rate of a CH 4 bubble, V, is estimated to be the slope dd e /dt of this linear decay. software. Due to minor bubble movement during image capture, the spatial resolution of the image measurements was estimated to be 0.05 mm. The shape of the bubble was assumed to be an oblate spheroid, and the major axis (a) and the minor axis (b) were estimated from the 2D image data. The time change of the spherical diameter (de) of the equivalent volume (v), calculated by Equation (1) The Heywood circularity factor (H) also was calculated to assess the bubble-shape change. Three or more replicates were performed at each condition; the number of samples analyzed for each data point reported below was n = 3 to 6. Figure 2 shows a typical image of a CH4 bubble observed in the high-pressure vessel (a sample of the bubble tracking animation is provided in Supplementary Materials Video S1).

Results and Discussion
Based on the (time) sequence of the bubble images, the temporal evolution of the bubble diameter was determined from Equation (1). As shown in Figure 3, the equivalent diameter de appears to decrease linearly with time. The dissolution rate of a CH4 bubble, V, is estimated to be the slope dde/dt of this linear decay.  The temperature dependence of V at 6.9 MPa is shown in Figure 4. In pure water, V was about 0.46 mm/min (7.6 µm/s) at a temperature of 298 K. V in seawater was observed to be significantly smaller than in pure water. At 298 K, the data indicate a value of V of about 0.21 ± 0.1 mm/min (3.3-3.7 µm/s). The data also appear to show a very slight increase in V with temperature over the range tested.
The present experiments indicate that the V of a CH 4 bubble rising in pure water is larger than in seawater under very similar conditions (i.e., pressure, temperature, bubble size, and ambient flow conditions). This is the first direct comparison to determine the effect of salts on CH 4 dissolution. This result is consistent with the prediction by Vascanselos et al. [41] that the dissolution rates of bubbles Energies 2020, 13, 3938 6 of 11 in seawater are smaller than in pure water because contaminants in seawater tend to accumulate on the bubble surface, which inhibits mass transfer from bubble to liquid. Energies 2020, 13, x FOR PEER REVIEW 6 of 11 The temperature dependence of V at 6.9 MPa is shown in Figure 4. In pure water, V was about 0.46 mm/min (7.6 μm/s) at a temperature of 298 K. V in seawater was observed to be significantly smaller than in pure water. At 298 K, the data indicate a value of V of about 0.21 ± 0.1 mm/min (3.3 ~ 3.7 μm/s). The data also appear to show a very slight increase in V with temperature over the range tested.
The present experiments indicate that the V of a CH4 bubble rising in pure water is larger than in seawater under very similar conditions (i.e., pressure, temperature, bubble size, and ambient flow conditions). This is the first direct comparison to determine the effect of salts on CH4 dissolution. This result is consistent with the prediction by Vascanselos et al. [41] that the dissolution rates of bubbles in seawater are smaller than in pure water because contaminants in seawater tend to accumulate on the bubble surface, which inhibits mass transfer from bubble to liquid.  The temperature dependence of V at 6.9 MPa is shown in Figure 4. In pure water, V was about 0.46 mm/min (7.6 μm/s) at a temperature of 298 K. V in seawater was observed to be significantly smaller than in pure water. At 298 K, the data indicate a value of V of about 0.21 ± 0.1 mm/min (3.3 ~ 3.7 μm/s). The data also appear to show a very slight increase in V with temperature over the range tested.
The present experiments indicate that the V of a CH4 bubble rising in pure water is larger than in seawater under very similar conditions (i.e., pressure, temperature, bubble size, and ambient flow conditions). This is the first direct comparison to determine the effect of salts on CH4 dissolution. This result is consistent with the prediction by Vascanselos et al. [41] that the dissolution rates of bubbles in seawater are smaller than in pure water because contaminants in seawater tend to accumulate on the bubble surface, which inhibits mass transfer from bubble to liquid. The present data were compared with previous studies. The dissolution rates of CH4 bubbles in pure water were measured previously by Warzinski et al. [15]. They observed a dissolution rate of 0.94 ± 0.05 µm/s at a temperature of 282 ± 0.5 K and pressure of 8.5 ± 2 MPa. This value is about eight The present data were compared with previous studies. The dissolution rates of CH 4 bubbles in pure water were measured previously by Warzinski et al. [15]. They observed a dissolution rate of 0.94 ± 0.05 µm/s at a temperature of 282 ± 0.5 K and pressure of 8.5 ± 2 MPa. This value is about eight times smaller than our results. One possible explanation for this discrepancy may be differences in the experimental conditions, such as pressure and bubble size.
To date, the dissolution rates of CH 4 bubbles in seawater have generally been determined from field data. Rehder et al. [18,20] estimated the in situ dissolution rates of rising bubbles as ranging from 8.5 to 15.1 ± 3.2 µm/s. These rates are much larger than those observed in the present study. Bubble size may explain this difference. McGinnis et al. [33] showed that mass transfer coefficients changed when Energies 2020, 13, 3938 7 of 11 the diameter of a bubble fell below around 5 mm. Using their model for smaller bubbles, a dissolution rate in seawater of about 0.2 mm/min is calculated, which agrees well with data from the present study.
Equation (2) is the relationship for CH 4 dissolution in liquid that has been employed in a number of models [28][29][30]33,38]: where m is the mass of a bubble, t is time, K is the mass transfer coefficient (for small bubbles), M = 16 g/mol is the molecular weight of CH 4 , A is the surface area of the bubble, C s is the saturation concentration of CH 4 in the liquid, and C 0 is the initial gas concentration in the liquid. For comparison with the present experimental results, the ascent rate, which is needed to determine the value of K, is equivalent to the flow rate of the water in the water tunnel test section. C s was estimated from the table in Duan and Mao [42], and C 0 was assumed to equal the total mass of CH 4 injected into the 100 L of water contained in the DOS. The initial bubble diameter was set at 1.5 mm. For these parameters, Equation (2) predicts values for the dissolution rates of CH 4 bubbles presented in Figure 5.
To date, the dissolution rates of CH4 bubbles in seawater have generally been determined from field data. Rehder et al. [18,20] estimated the in situ dissolution rates of rising bubbles as ranging from 8.5 to 15.1 ± 3.2 µm/s. These rates are much larger than those observed in the present study. Bubble size may explain this difference. McGinnis et al. [33] showed that mass transfer coefficients changed when the diameter of a bubble fell below around 5 mm. Using their model for smaller bubbles, a dissolution rate in seawater of about 0.2 mm/min is calculated, which agrees well with data from the present study.
Equation (2) is the relationship for CH4 dissolution in liquid that has been employed in a number of models [28][29][30]33,38]: where m is the mass of a bubble, t is time, K is the mass transfer coefficient (for small bubbles), M = 16 g/mol is the molecular weight of CH4, A is the surface area of the bubble, Cs is the saturation concentration of CH4 in the liquid, and C0 is the initial gas concentration in the liquid. For comparison with the present experimental results, the ascent rate, which is needed to determine the value of K, is equivalent to the flow rate of the water in the water tunnel test section. Cs was estimated from the table in Duan and Mao [42], and C0 was assumed to equal the total mass of CH4 injected into the 100 L of water contained in the DOS. The initial bubble diameter was set at 1.5 mm. For these parameters, Equation (2) predicts values for the dissolution rates of CH4 bubbles presented in Figure 5. Square marks correspond to dissolution in pure water, and circles, to the seawater case. The dotted line is a fitted curve for the experimental data for seawater that is calculated using Equation (3).
As shown in the Figure, the agreement between measured and predicted dissolution rates is quite good for the pure water case, but slightly less so for the seawater cases. The model-estimated dissolution rates of CH4 bubbles in seawater were larger than the experimental results over the entire temperature range tested. Moreover, the predicted trend in the rates of dissolution with temperature is a mirror image of the data. Given these differences, it is worthwhile, therefore, to Square marks correspond to dissolution in pure water, and circles, to the seawater case. The dotted line is a fitted curve for the experimental data for seawater that is calculated using Equation (3). As shown in the Figure, the agreement between measured and predicted dissolution rates is quite good for the pure water case, but slightly less so for the seawater cases. The model-estimated dissolution rates of CH 4 bubbles in seawater were larger than the experimental results over the entire temperature range tested. Moreover, the predicted trend in the rates of dissolution with temperature is a mirror image of the data. Given these differences, it is worthwhile, therefore, to consider the possibility that the mechanism of dissolution of buoyant CH 4 bubbles moving through water may be different from the conventional one employed in previous studies.
It should be noted that differences between the data and the model may be associated, to some degree, with the types and amounts of components in the gas. Since natural gas includes other hydrocarbons besides CH 4 and also acid gases, the mixture solubility would be slightly different from that of the pure CH 4 used in this study. Thus, further experiments are required to investigate these effects on the gas bubble dissolution process. As shown in Figure 4, the temperature dependency of the dissolution rate exhibits a positive trend; that is, the dissolution rate increases with temperature. Equation (2) assumes that dissolution depends directly on the solubility of the gas in the liquid. Since the solubility of gases in liquid typically decreases at higher temperature, Equation (2) would not replicate the present experimental observations. As an alternative, we posit that the dissolution of gas bubbles may proceed via a thermal activation process described by: where V 0 is a pre-exponential constant, Q is the activation energy, R is the gas constant, and T is the absolute temperature of the system. If Equation (3) is fitted to the data in Figure 5 (shown as the dotted line), a value of Q of about 9.0 kJ/mol (R 2 = 0.60) is obtained. This is of the same order as the activation energy of CH 4 diffusion in pure water, approximately 11 kJ/mol [43]. It does not appear unreasonable, therefore, to propose that the rate-determining step of CH 4 dissolution from a gas bubble into liquid is the diffusion of dissolved CH 4 away from the bubble boundary layer to the bulk water. The transfer of solute molecules across the phase barrier is included in the governing equation. This transfer is expected to proceed much faster than outward (i.e., toward the water phase) diffusion from the interface. Hence, the rate-determining gas diffusion step (i.e., the dissolved gas molecules diffusing away from the gas-liquid boundary into the bulk water phase) must be included in the set of governing equations of CH 4 dissolution models.

Conclusions
Environmental impact assessments of the production of natural gas from hydrate (NGH) reservoirs in the seabed requires the development of models that can accurately simulate the behavior of CH 4 gas released into the water column. To date, many laboratory experiments and field studies have been conducted in order to investigate the dissolution process of CH 4 bubbles and to provide physical insight and data for predictive models. Most of these laboratory studies, however, have employed pure water, while field tests have been performed in the ocean. Experiments that have compared the behavior of CH 4 bubbles in pure water and seawater under identical conditions are scarce. Thus, the influence of several key factors, such as salinity, on the accuracy of model predictions remains unclear. In this study, the dissolution of single CH 4 bubbles in pure water and in artificial seawater was observed under similar pressure, temperature, and bubble size conditions using a unique Deep Ocean Simulator facility. It was found that bubbles dissolved in pure water at approximately twice the rate of bubbles in seawater and that dissolution rates in seawater appeared to increase slightly with temperature between 288 K and 303 K. These data were compared with corresponding dissolution rates calculated by models used in various NGH development projects. The measured and predicted rates in pure water were in good agreement; however, experimental data obtained with seawater yielded smaller bubble shrinkage rates than the model results. The temperature dependency of the bubble dissolution rate observed in the experiments appears to suggest that the rate of dissolution of CH 4 bubbles is limited not by the rate of transfer of CH 4 molecules across the phase interface but rather by the diffusion of dissolved CH 4 molecules from the interfacial zone to the bulk water phase. We propose that this rate-determining step be considered for inclusion in CH 4 bubble dissolution models to improve the accuracy of the model results.