Promoted disappearance of CO2 hydrate self-preservation effect by surfactant SDS

Capture, storage and utilization of CO 2 gas through the hydrate technology is deemed as a promising approach to solve the problem on global warming. After ecient transportation of CO 2 gas with the self-preservation effect of hydrate, the dissociations of hydrate are required in some practical scenarios. In order to investigate the initial dissociation properties of hydrates while the self-preservation effect was removed, CO 2 hydrates were formed in different experimental conditions and media and then frozen. After that, the self-preservation effect of hydrates was slowly removed through a uniform heating method. The measurement results exhibit that comparing with the pure water (PW), the initial dissociation temperatures (IDT) are signicantly lowered by the silica gel powder (SG) and sodium dodecyl sulphate (SDS) solution (SS). Similarly, the initial dissociation rates (IDR) of hydrates are also reduced. These patterns are opposite to the traditional elucidating mechanism of the self-preservation effect of hydrate. Hence, the theory of ice shell was supplemented furthermore. In addition, it is found that the nal duration time of the entire dissociation process of hydrate after losing the preservation effect presents obvious dependence on the initial dissociation rate (IDR). These ndings are expected to provide some references for the future transportation of CO 2 gas with the self-preservation effect of hydrate.


Introduction
Gas hydrates are one kind of crystalline compounds, in which some cavities are composed of water molecules and guest molecules are trapped inside.1 Under suitable temperature and pressure conditions, many gases can form hydrates, e.g. carbon dioxide, methane, hydrogen and other similar size gases, as well as several low molecular hydrocarbons, e.g. neo-hexane (NH), tetrahydrofuran (THF) and other hydrocarbons of similar size.2 Owing to these different gas components, hydrates present different applicable contributions.3 -456 Due to the favourable formation conditions, CO 2 seabed sequestration in the form of hydrate is considered as one of the prospective ways to mitigate global warming.7 ,8 According to the signi cant difference between the thermal stabilities of CO 2 and CH 4 hydrates, it also possesses a great practical potential for using the formation of CO 2 hydrate to replace CH 4 stored in the CH 4 hydrate formed on the seabed.9 ,10 As estimated, at least 17% of CO 2 emitted cumulatively from power plants can be captured and sequestrated through the hydrate technology. This means that the rise in global temperatures can be limited to less than 2 °C in a long term. 11 In recent years, other applications related with gas hydrates, e.g. seawater desalination, 12,13 gas puri cation  and cool storage, 17,18 have also caused wide attention. Before these applications, CO 2 gas capture from the power plants and then transport to the intended destination on land or in sea is the rst step. In order to solve the problem on e cient transportation of formed hydrate, the application of self-preservation effect was proposed.
So-called self-preservation effect of hydrate is a phenomenon that in some non-equilibrium temperature regions below the ice point, gas hydrates can still maintain stable even at atmospheric pressure.  Owing to the extremely slow dissociation rate under this state, gas hydrates can be transported at atmospheric pressure, safely and e ciently. Hence, the dissociation properties of hydrates under the selfpreservation effect have been studied detailedly, under different experiment conditions. Because this effect occurs only below the ice point, most researchers think that the effect is mainly cause by the ice shell wrapped outside the frozen gas hydrate. The dissociation rate of "self-preserving" methane hydrate and the rate at which the Ih ice transforms to hexagonal ice successively were measured in different temperature ranges by Takeya et al. And it was pointed out that at the initial dissociation stage, a layer of ice forms on the surface of hydrate and thus prevents the further dissociation. 22,23 Subsequently, the in uence of hydrate particle size and the ice structure on the self-preservation effect was investigated and further pointed out that the generation of the effect depends on the action of ice shell and interaction between host and guest molecules.  And then, the dissociation rates of different types of "selfpreserving" hydrates were measured. in different temperature ranges. The dissociation processes of structure type I (sI) and II (sII) hydrates at 193-290K 27 and 190K-273K 28 were studied by Stern et al., using successive temperature delaying and rapid decompression method. According to the dissociation rates of methane hydrate measured between -7.5 to 0 ℃ at normal pressure, Shirota et al pointed out that the dissociation rate at -5 ℃ was the lowest. 29 In addition, the in uences of surface structure of hydrate and ice "defects", 30 ice particle size, 31,32 thermodynamic conditions and the phase composition, 33 and supercooled water 34 on the self-preservation effect of hydrates have also been considered by other researchers.
For the transportation of CO 2 hydrate, the utilization of the self-preservation effect is undoubtedly a promising approach. As mentioned above, many research conclusions on the formation mechanism of this effect and the dissociation rules of "self-preserving" hydrates have been provided. However, in the practical scenarios, CO 2 hydrate should be required to dissociate into gas again for the further utilization after being transported to the intended destinations with the self-preservation effect. Hence, the disappearance patterns of this effect and the control mechanism behind them should also be considered, which has been rarely reported in the literatures. Therefore, with a method of increasing temperature uniformly, the self-preservation effects of CO 2 hydrates formed within different reaction systems, including pure water, silica gel powder and SDS solution, were made disappear slowly and the disappearance patterns were investigated in detail. Then, a novel mechanism controlling the selfpreservation effects of hydrates were proposed.

Results
Formation of CO 2 hydration and disappearance of its selfpreservation effect Fig. 1 shows the changes in temperature and pressure during the entire formation process of CO 2 hydrate. Following the temperature reduction, hydrate begins nucleating in PW and release heat acutely. After that, hydrate crystals begin growing and consume gas continuously. Hence, the CO 2 gas was repeatedly replenished into the crystallizer through the proportional-integral-derivative (PID) pneumatic valve, which showed as several sawtooth wave shapes on the pressure curve (e.g. AB in Fig. 1). After being frozen thoroughly, the formed hydrate will be dissociated through a continuous temperature raising method. Fig. 2 shows one complete dissociation process consists of three successive stages. During stage I, along with the release of free gas, the measured ow rate of gas rises abruptly and temperature decreases ercely. After that, the ow rate restores to the original 0 mmol/min and temperature to the preset value -6 ℃ in the stage II, at this stage, the hydrate is in a self-preseving state under normal pressure. In the stage III, while being heated continuously, the frozen hydrate will lose its self-preservation effect gradually and begins dissociating. Because the dissociation of hydrate is an endothermic process, the change curves for temperature are deviated from the original trends. These obvious deviating positions were de ned as the initial dissociation points of temperature (IDT), as shown in Fig. 2. These points in fact characterizes the speci c temperature conditions at which the self-preservation effect is removed and the frozen hydrate begins dissociating. Because the hydrates were formed under different experimental conditions, the speci c locations of IDT and the change curves for gas ow rate present obvious differences. As shown in Fig. 2, the initial dissociation rates of hydrates are all 0 mmol/min. The ow rates were then plotted against the cumulative amounts of the released gas. As shown in Fig. 3, while the cumulative amounts are less than 40 mmol, the relations between ow rate and cumulative amount all exhibit superior linearity. Hence, the initial dissociation rate (IDR) of hydrate was de ned as the measured instantaneous ow rate while the cumulative amount of gas reached 1 mmol, in this study. The speci c values of IDP and IDR were then statisticized and analyzed to investigate the disappearance patterns of the self-preservation effect of hydrate and the control mechanism behind.
In uence of medium and experimental conditions on IDT and IDR

Relationship between IDR and total dissociation time / ice content
After the self-preservation effect disappears, hydrate begins dissociating slowly. And after experiencing a certain period of time, hydrate nally releases all the gas contained inside and the entire experiment nishes. Therefore, besides the initial dissociation properties, the total durations of the dissociation processes were also measured. As shown in  Fig. 7 shows the relation between the dissociation time ( Fig. 6) and IDR (Fig. 5). As shown, the dissociation time presents a signi cant dependence on the IDR. This superior inverse proportional correlation, R 2 =0.70, means that along with the disappearance of self-preservation effect, the higher the initial dissociation rate (IDR) is, the shorter the duration of dissociation is.
In order to investigate the key in uencing factor of the IDR, the nal conversion rates of liquid water during formations of hydrate were calculated. Considering the fact that the formed hydrates were all frozen thoroughly at negative temperatures, the calculation results were then subtracted from 100% to characterize the ice content contained in the frozen hydrates. Fig. 8 shows the relation between the calculated ice content and IDR. As shown, the IDR presents an obvious dependence on the ice content in the frozen hydrate, with a R 2 =0.67 of the tted linear equation. And this positive proportional correlation means that the higher the IDR is, the greater the ice content in the frozen hydrate is.

Discussion
As con rmed by the X-ray diffraction measurements, the self-preservation effect of hydrate is mainly caused by the ice shell formed outside the solid hydrate crystals. 21,22 The surfactant SDS utilized in this study is only a kind of dynamic additive for the formation reactions of hydrate, but does not change the thermodynamic equilibrium conditions of hydrates.  This promotion action is carried out by improving the activities of water molecules and then dissolution capabilities of gas. 38 Hence, the freezing point of water is also not changed by SDS. Logically, the dissociation properties of hydrate with selfpreservation effect should not be in uenced by the addition of SDS. However, comparing with in the PW, the IDT in the other three media were signi cantly reduced (Fig. 4). This means that comparing the hydrate with the self-preserving effect in PW, the dissociation of hydrate can be brought forward by SDS, signi cantly. In addition, according to the theory of ice shell explaining the self-preservation effect, the IDR while the effect begins disappearing should present a reverse proportional correlation with the ice content contained in the frozen hydrates. However, our experimental results show that the assumption is invalid (Fig. 8). Base on these two obvious defects on explaining the formation reason of selfpreservation effect with the theory of ice shell, the formation mechanism was further supplemented in this study.
As shown in Fig. 9, we think the frozen hydrate is preserved not thoroughly through the ice shells ( Fig. 9 B) but some nonenclathrated liquid water 39 and some gas molecules dissolved inside (Fig. 9 C). As known, SDS is one kind of macromolecular organic matter and can improve the activities of water molecules and then the dissolution capabilities of gas molecules. 40 During the formation processes of hydrate, SDS can not enter into the clathrate structures and is only dissolved in the nonenclathrated liquid water outside the solid hydrate crystals (Fig. 9 B). Along with the continuous consumption of liquid water caused by the formation of clathrate structures, the concentration of SDS dissolved in the nonenclathrated liquid water will be improved continuously. As a result, once the clathrate structures of hydrate begin losing stability during the heating process, the trapped gas molecules will be snatched away by the adjacent SDS molecules (Fig. 9 C). As a result, the IDT in the SS and SG/SS can be brought forward signi cantly, comparing with that in PW (Fig. 4). Because there are enormous capillaries inside the SG, the activities of water molecules are signi cantly inhibited by the capillary force. Hence, while the SG media are frozen, there is still enormous unfrozen liquid water inside. Because the dissolution capability of gas can be enhanced by low temperature, the inhibition in the activities of water from capillary force is then weakened. As a result, the lower the frozen temperature is, the lower the IDT are (Fig. 4). And the SG performs a similar function in promoting the disappearance of self-preservation effect (Fig. 4). With the combination of SDS and capillary force, the disappearance of self-preservation effect can be further brought forward (Fig. 4).
In addition, owing to the promotion function of SDS in gas dissolution and the similar function of capillary force at low temperatures, it is very di cult for these dissolved gas molecules to desorb from the solution. 40 As a result, comparing with in the PW, the IDR are signi cantly lowered in the other experimental media (Fig. 5). And the higher the conversion rate of liquid water to hydrate is, the higher the concentration of SDS dissolved in the nonenclathrated liquid water is. Hence, the IDR of the selfpreserving hydrates with the highest conversion rates (Table S1) will be the lowest, as shown by the 6.5 ℃ and 3.5 ℃ in Fig. 5. Different from the IDT (Fig. 4), the frozen temperatures do not exhibit dramatic in uence on the IDR (Fig. 5). The reverse proportional correlation between the IDR and total dissociation time exhibits a signi cant dependence of subsequent dissociation on the initial dissociation properties.
Therefore, for the transportation of CO 2 gas through the self-preservation effect of hydrate, it is very important for the investigation on the initial dissociation properties while the effect begins disappearing.

Experimental apparatus and materials
As shown in Fig. 10, the core of the utilized experimental system is a crystallizer, made of 316 stainless steel, with a magnetic stirrer inside and observation windows on both sides, of which the diameter and the height are respectively 45 mm and 130 mm. The crystallizer is wrapped by a water vest connecting to a low-temperature circulating cooling bath. Then, the temperature can be controlled accurately, in the range of -50~150 ℃ ± 0.1 ℃. The pressure and temperature inside the crystallizer are measured by a digital pressure meter (DPM) and temperature meter (DTM), respectively. The CO 2 gas with high pressure is stored in a gas cylinder also made of 316 stainless steel. There is a proportional-integral-derivative (PID) pneumatic valve connected between the crystallizer and the cylinder. Then, the crystallizer can be pressurized through the PID valve. There is a vent on the bottom of the crystallizer, which connects to a digital gas ow meter (DGM). The measurement range of DGM is 0~5L/min, the resolution 0.001L/min, and the precision ±0.12%. The DGM, DPM and DTM are controlled through a data acquisition system (DAS). During the experiments, all of the measured parameters were recorded and saved by the DAS at an interval of 5 s. CO 2 with a purity 99.99% was chosen to form hydrate. During the formation experiments of CO 2 hydrates, sodium dodecyl sulphate (SDS) has been widely applied as a promoter.  Hence, besides in the pure water, CO 2 hydrates were also formed in SDS solution. The transfer rate of gas to the aqueous phase can be signi cantly enhanced by SDS 44 and a 1 wt% is an effective concentration for enhancing the formation rate of hydrate. 43 Therefore, an SDS solution with 1 wt% concentration was prepared with analytical grade SDS and deionized water with a resistivity higher than18.00 MΩ·cm, which is self-made in lab. In order to ensure the comprehensiveness of the research results, silica gel powder (Fig. S1) was also chosen as a porous medium in which CO 2 hydrates were formed. The density of the medium is 0.35 g/cm 3 , porosity 77.44% and average particle size between 25 and 58 μm. At saturated state, the water content of medium is W water /W media = 2.2:1. During all formations, the utilized experimental media are all non-saturated with a xed water content of W water /W media = 1.5:1. This water content leads to a large gas-water contact area and enormous interconnected pore spaces facilitating the formation of hydrate. 45 As a result, high water conversion ratios of hydrate formation can be achieved through this water content. 46,47 Experimental procedure CO 2 hydrates were formed in two liquid phases, i.e. pure water and SDS solution (abbreviated as PW and SS) and one porous medium containing two different liquids, i.e. silica gel powder mixed with pure water and SDS solution (abbreviated as SG/PW and SG/SS). Four temperatures 8.5 ℃, 6.5 ℃, 3.5 ℃, 0.5 ℃ were designed for the formation experiments (Table S1). The corresponding pressure values of phase equilibrium of CO 2 hydrate were calculated as 3.67 MPa, 2.77 MPa, 1.68 MPa and 1.32 MPa (Table S1), with the CSMGem software (Natural Gas Hydrate Center, Colorado School of Mines).
Before each formation experiment in liquid phases, 104 mL PW or SS was injected into the crystallizer. For those in porous media, a certain quality of silica gel powder containing also 104 mL PW or SS was charged. The initial temperature of crystallizer was set higher than the pre-designed by 2.5 ℃. The crystallizer was then pressurized to the corresponding phase equilibrium value through the PID valve and maintained constant. While stirring is required, the magnetic stirrer was opened and the stirring speed was xed at 200 r/min. The whole system was left undisturbed over 12 h under the constant conditions to allow CO 2 to dissolve in water su ciently. After that, the temperature was uniformly reduced by 4 °C in 80 min and then retained constant, under the constant pressure condition. Finally, 1.5 °C subcooling driving force was provided for each formation experiment of CO 2 hydrate.
After hydrate was formed su ciently, the PID valve was turned off. The temperature was lowered below 0°C rapidly to freeze the formed hydrate over 10 h. Then, the free gas in the crystallizer was released slowly through the vent (Fig. 8). Even under the atmospheric pressure, the frozen hydrate dissociates very slowly and enters into the self-preserving state. As reported, the optimum temperature condition of the selfpreservation effect of methane hydrate is about -5±1℃, 27,28 then, the temperature utilized to freeze the hydrates formed in the liquids was chosen as -6 ℃. Considering the fact that there are a large number of capillaries within the silica gel power, another two lower temperatures -8 ℃ and -10 ℃ were also used to freeze the hydrates formed in the porous medium.
After the self-preserving state of hydrate was maintained over 8 h, the temperature in the crystallizer was uniformly raised to the initially set value at a rate of 3℃/h. Along with the rise in temperature, the selfpreservation effect would disappear gradually and hydrate began dissociating. DGM recorded the entire change process of the ow rate of gas released from dissociated hydrate.

Calculation method
As shown in Fig. 1, due to the instantaneous opening/closing performance of the PID valve for retaining a constant pressure, the pro le of CO 2 uptake against time during the entire formation process of hydrate consists of many separated segments and presents a sawtooth wave shape. Over each separated segment (AB in Fig. 1), the amounts of gas uptake were respectively calculated and the total amount were obtained by summing up the separated ones, with the following formulas.
where, n is the total amount of gas (with a unit mmol), n i the calculated amount of gas uptake during each separated segment (mmol), P measured pressure in the crystallizer (MPa), V the volume of headspace or residual pores of the porous medium in crystallizer (mL). T temperature (K), R 8.314 J/(mol·K). The subscripts "sta" and " n" respectively represent the starting and ending points of each separated segment. Z is the gas compressibility factor under the speci c temperature and pressure conditions, calculated by the Benedict-Webb-Rubin-Starling equation of state. 48 During the dissociation process of hydrate under atmosphere pressure and room temperature, the ow rates of released gas were measured by the DGM, with a unit in L/min. Through the gas state equation, this unit were nally converted into mmol/min.

Conclusions
Some CO 2 hydrates formed in different experimental conditions and media were frozen to generate the self-preservation effect. After that, the frozen hydrates were slowly dissociated through a uniform heating method at a xed rate. Some initial dissociation properties that were exhibited while the self-preservation effect began disappearing were investigated. The experimental results show that comparing with in pure water (PW), the initial dissociation temperatures (IDT), characterizing that at which hydrates begin lose their self-preservation effect, are reduced in SDS solution (SS), silica gel powder/pure water (SG/PW) and silica gel powder/SDS solution (SG/SS). Similarly, the initial dissociation rates (IDR) are all lowered in the above latter three media, which characterizes the instantaneous ow rate of gas while the amount of released gas reaches 1 mol. These patterns are challengeable for the traditional theory of ice shell, elucidating the generation mechanism of the self-preservation effect of hydrate. Therefore, another novel perspective was provided to consummate the generation mechanism of the effect, furthermore. Finally, we hope this study might provide some references for the e cient transportation of CO 2 gas through the self-preservation effect of hydrate.