Effects of additives on continuous hydrate-based flue gas separation

Applied Energy 221 (2018) 374–385

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Effects of additives on continuous hydrate-based flue gas separation☆ ⁎

T

Mingjun Yang, Hang Zhou, Pengfei Wang, Yongchen Song

Key Laboratory of Ocean Energy Utilization and Energy Conservation of Ministry of Education, Dalian University of Technology, Dalian, Liaoning 116024, China

H I GH L IG H T S

stage continuous hydrate-based gas separation process was proposed. • Multiple movement characters was observed and analyzed for experimental cycles. • Solution saturation for 5% TBAB + 5% THF was higher than that of 19% THF. • Hydrate • 80 min is the appropriate time for rapid hydrate formation process.

A R T I C LE I N FO

A B S T R A C T

Keywords: Hydrate-based gas separation Carbon dioxide hydrate Additives Continuous experimental process Solution movement

CO2 capture from fossil fuel power plants is the main method of CO2 storage. Hydrate-based gas separation is regarded as a potential method for CO2 capture from flue gas. In this study, hydrate-based gas separation (HBGS) was used to capture CO2 from flue gas (19.96 mol% CO2 and 80.04 mol% N2), and the continuous experimental process was monitored using magnetic resonance imaging (MRI). The effects of two additives (5 wt% TBAB + 5 wt% THF and 19 wt% THF), two gas injection methods (constant pressure and constant flow rate processes), and of different pressures and flow rates on the hydrate saturation and solution movement were investigated. The results show that both additives effectively promote hydrate formation. The constant pressure process was superior to the constant flow rate process for hydrate formation. Furthermore, the flow rate had little influence on the hydrate saturation. The process was most efficient when a hydrate formation stage time of approximately 80 min was used. The solution movement resulting from the continuous multiple cycles tended to decrease during subsequent cycles. Moreover, the addition of 19 wt% THF had a more obvious effect on the solution movement than 5 wt% TBAB + 5 wt% THF. Solution concentration phenomena were observed in the presence of 19 wt% THF at 285.15 K; these phenomena may have been affected by the formation and dissociation of hydrates. Due to solution movement during the continuous industrial hydrate-based gas separation process, the solution might need to be replenished. Finally, in terms of the resulting the hydrate saturation, the use of 5 wt% TBAB + 5 wt% THF was found to be more suitable in this study, while the 19 wt% THF was more suitable for a higher experimental temperature.

1. Introduction Fossil fuel combustion is the main sources of CO2 emissions. A high concentration of CO2 is present in the flue gas produced by fossil fuel power plants. CO2 capture is an effective way to reduce the emission of CO2 into the atmosphere [1–4]. Thus, it is necessary to capture CO2 from flue gas, and the CO2 separation process is the key step of CO2 capture. Many methods for CO2 separation have been reported, including chemical and physical absorption, solid physical adsorption, cryogenic distillation, and membrane gas separation [5]. CO2 separation is an energy-consumption process [6]. Although some of the above

☆ ⁎

methods yield satisfactory results for CO2 separation, their energy cost is still too great for commercial use. Furthermore, the cost of hydratebased gas separation (HBGS) is believed to be lower than that of chemical absorption [7]. Unlike cryogenic distillation, the operating temperature of HBGS is above the freezing point, and thus less energy is used for refrigeration [8,9]. Furthermore, the water used in HBGS can be recycled, which reducing the raw material costs. The feasibility of HBGS has been proven in the laboratory [10–14]. In HBGS, N2 and CO2 are the main components in the flue gas. Because the hydrate phase equilibrium pressure of N2 is significantly higher than that of CO2 at the same temperature, CO2 preferentially enters the hydrate cavities

The short version of the paper was presented at ICAE2017, Aug 21–24, Cardiff, UK. This paper is a substantial extension of the short version of the conference paper. Corresponding author. E-mail address: [email protected] (Y. Song).

https://doi.org/10.1016/j.apenergy.2018.03.187 Received 4 January 2018; Received in revised form 17 March 2018; Accepted 31 March 2018 0306-2619/ © 2018 Elsevier Ltd. All rights reserved.

Applied Energy 221 (2018) 374–385

M. Yang et al.

making separation the flue gas via HBGS possible [15]. The key limitations of the HBGS process are the hydrate formation rate and the demanding conditions required to reach phase equilibrium. Therefore, researchers have attempted to determine these values through a variety of different methods [16–18]. Enhancing the gas–water contact and the heat and mass transfer are believed to increase the hydrate formation rate. At present, the main methods to enhance gas-water contact and heat and mass transfer are mixing, spraying, bubbling, or porous media; other physical methods and the use of chemical additives have also been considered [19]. The use of porous media in particular is an important method to promote hydrate formation in hydrate applications. The water in porous media could improve the gas-water contact, which could promote hydrate formation by increasing the hydrate formation rate [20]. Babu et al. found that the addition of porous media resulted in 54% conversion of water into hydrate within two hours. In addition, higher gas consumption and a decreased induction time were observed [21]. Park et al. reported increased gas consumption and water conversion when silica gels were employed [22]. Adeyemo et al. observed that the gas consumption, CO2 recovery, separation factor and the conversion of water to hydrate increased with larger pore and particle size [23]. In our previous work, we found that less solution concentration occurred BZ-01, which contained smaller pores, than in BZ-02 and BZ-04, which contained larger pores [7]. Zhong et al. found that stirring was beneficial for hydrate formation [24]. In addition to physical methods, the use of chemical additives to improve the HBGS process has been widely studied. These chemical additives are divided into two classes: kinetic and thermodynamic additives. Kinetic additives include sodium dodecyl sulfate (SDS), sodium dodecyl benzene sulfonate (SDBS), and alkylpolyglucoside (APG) [25,26]. Thermodynamic additives such as tetrahydrofuran (THF), tetrabutylammonium bromide (TBAB), tetrabutylammonium chloride (TBAC), tetrabutylammonium fluoride (TBAF), and cyclopentane (CP) not only accelerate the hydrate formation rate, but can also ease the hydrate phase equilibrium conditions [8,27–33], TBAB and THF are currently in wide use. The principle behind the use of TBAB as an additive is as follows. During hydrate formation, Br- can displace water molecules in the crystal structure of the hydrate to form hydrate cages. Subsequently, TBA+ can occupy the large cages of the hydrate, while CO2, N2, H2S, and other small molecules fill the remaining small cages. TBAB is involved in the formation of hydrate cages, stabilizing the hydrate structure and resulting in a substantial reduction in the pressure of hydrate dissociation [34]. Several studies have verified that TBAB can allow phase equilibrium to be achieved under less demanding conditions. Li et al. measured the phase equilibrium of CO2 hydrate in the presence of TBAB. Their results showed that TBAB significantly decreased in the phase equilibrium pressure of CO2 hydrate [35,36]. Furthermore, the TBAB hydrate is believed to have an obvious effect on CO2 separation from mixed gases. Kim et al. investigated the thermal stability of hydrates of mixed gases (20% CO2 + 80% N2) in the presence of TBAB, TBAC, and TBAF, and found that the concentration for CO2 in the hydrate was as high as about 60%, and that the selectivity of CO2 did not depend on the type of quaternary ammonium salt used [34]. Hashimoto et al. studied hydrate-based mixed gas (CO2 + N2) separation in the presence of TBAB, and found that the CO2 selectivity was highest at a pressure of 1 MPa and a TBAB concentration of 32 wt %. They also found more CO2 was captured without gas flow than with gas flow [37]. In addition to these experimental results, models have also been proposed [38,39]. A model for the phase equilibrium conditions for the formation of hydrates of CO2, CH4, and N2 in TBAB or pure water has been proposed, in which the temperature and the TBAB concentration are considered [40,41]. The principle behind the use of THF as a promoter is as follows. When the CO2/N2 gas mixture and water form a hydrate in the presence of THF, THF molecules occupy the large cages of the hydrate, while CO2 fills the small cages to form the S-II THF-CO2 hydrate. This changes the

phase equilibrium conditions of the original hydrate, reducing the induction time of the process significantly [42]. Several studies have verified that THF can allow phase equilibrium to be achieved under less demanding conditions. Wang et al. investigated the CO2 hydrate phase equilibrium conditions in the presence of THF using a stepwise heating method; their results demonstrated that the equilibrium pressure of the CO2 hydrate was significantly reduced in the presence of THF, and that as the THF concentration was increased, the equilibrium pressure was reduced [43,44]. Moreover, THF hydrate is believed to have an obvious effect on gas separation from mixed gases. Cai et al. studied the hydrate-based CO2 separation process in the presence of THF, using realtime Raman spectroscopy [45]. Park et al. found that CO2 could be enriched to 90% in the hydrate phase when THF and mixed gas (60% H2 + 40% CO2) were used for hydrate formation [22]. Another study verified that more than 99 mol% CO2 could be recovered from flue gas by HBGS when the hydrate promoter THF was used [46]. Sun et al. used a CO2 hydrate with THF for refrigeration applications, and proposed a kinetic model of the CO2-THF hydrate to predict hydrate decomposition [47]. The effect of various mixed additives has also been studied. Our group has investigated the effect of the mixed additive THF + TBAB on the separation of flue gas and found similar phase equilibrium pressure using 5% THF + 5% TBAB and 5% THF + 10% TBAB. Furthermore, as the THF or TBAB concentration was increased, the gas consumption also increased [17]. Lirio et al. measured the storage capacity, induction time, and induction temperature of CO2 hydrates in pure water, SDS solution, THF solution, and a mixed solution of THF and SDS. Their results showed that while SDS increased the CO2 hydrate storage capacity, its effect on the CH4 hydrate was less significant. However, because of the coupling effect, a mixture of SDS and THF reached 121 ( ± 12)% of the maximum theoretical value, and the induction temperature and the induction time were reduced [48]. Tzirakis et al. studied the phase equilibrium of the hydrate of a CO2 + N2 mixture with TBAB, CP, and TBAB + CP as promoters, respectively, and found that the mixture of TBAB and CP had a greater effect on reducing the equilibrium pressure of hydrate than pure TBAB [49]. The results were similar for a TBAF + CP mixture. Moreover, they found that the TBAF + CP mixture had a greater effect on reducing the equilibrium pressure of hydrate than pure TBAF [50]. Xia et al. found that CO2 and H2S could be captured simultaneously from biogas or natural gas in the presence of a physical gas solvent (TMS) and TBAB [51]. Moreover, mixtures of many additives, including TBAB, SDS, THF, dimethylsulfoxide (DMSO), and CP have been investigated, and mixed additives have been shown to have a significant effect on gas separation, leading to over 90 mol% CO2 in the hydrate phase and separation factors as high as 17.82 [52–57]. Therefore, HBGS is a feasible potential method for CO2 capture, and further investigation the factors affecting industrial CO2 capture is necessary. In a review of the literature, many works about additives have been reported, while most of them focused on the hydrate phase equilibrium conditions, separate factors and CO2 concentrations after the gas separation, and their works on separate factor have been studied in great detailed. However, we found that there have been few simulations of continuous industrial hydrate-based gas separation processing in the presence of additives, which is very important for hydrate-based technology. Therefore, the effects of different pressures and reaction times on hydrate formation, as well as the effects of different gas injection methods and additives, and the use of continuous multiple cycles on the continuous industrial hydrate-based gas separation process have been studied in this work. 2. Experimental material and methods 2.1. Apparatus The experimental apparatus consisted of several parts (see Fig. 1): a 375

Applied Energy 221 (2018) 374–385

M. Yang et al.

Fig. 1. Experimental system.

completely when the ratio of THF to water is 1:17. Therefore, two additives were used in this work. In addition, the experimental conditions (pressure, temperature and flowrate) chosen in this work were based on our previous works [7,17].

Table 1 Properties and suppliers of the materials. Material

Purity/composition

Particle size/mm

Supplier

CO2/N2

–

BZ-02 THF

80.6 mol%CO2, 19.4 mol%N2 Soda glass ≥99.0%

TBAB

≥99.0%

–

Dalian Guangming special Co., Ltd., China As-one Co., Ltd., Japan Kermel Chemical Reagent Co., Ltd., China Kermel Chemical Reagent Co., Ltd., China

0.177–0.250 –

2.3. Calculations The amount of solution in the field of view (FOV) was monitored using the MRI system. The protons (H1) in liquid water can be detected by MRI, while the protons (H1) in the hydrate will not be detected by MRI. That is, the amount and distribution of water in the FOV can be obtained from the mean intensity (MI). Thus, in the absence of other factors that could affect the MI, a decrease in the MI can be interpreted as water being converted into hydrate [7]. The solution saturation (volume fraction) and hydrate saturation (volume fraction) were calculated from the mean intensity (MI) of the solution. The discussions of the MI data correspond to the FOV. It has been reported that 1 m3 fresh water can form 1.25 m3 hydrate at standard temperature and pressure (STP) [58]. Therefore, the solution saturation and hydrate saturation can be calculated as follows:

high-pressure cylindrical vessel with a jacket filled with FC-40 (3M Company, USA), two thermostatic baths, an MRI system (operating at 400 MHz, Varian, Inc., USA), three high-precision syringe pumps (260D, Teledyne Isco, Inc., USA), and pressure transducers (Nagano Co., Ltd., Japan), which have been described previously [7]. In this study, BZ-02 glass beads (As-one Co., Ltd., Japan) were used as the porous media, and 19 wt% THF and 5 wt% TBAB + 5 wt% THF were used as the hydrate promoters. The solution was diluted with deionized water produced in the lab, and flue gas was simulated using a gas mixture of 19.96 mol% CO2 and 80.04 mol% N2. More details are listed in Table 1 and reported in the previous work [7].

Shi = 1.25 ×

(I0−Ii ) × Ss0 × 100%, I0

Ii × Ss0 × 100%, I0

(1)

2.2. Experimental procedures

Ssi =

The glass beads were packed into the vessel tightly and uniformly, and the vessel was then placed in the center of the MRI magnet. Once the vessel was under vacuum, the solution was injected into the vessel slowly until the pressure in the vessel reached 6 MPa, saturating the porous medium with the solution. Subsequently, the vessel was allowed to rest for 15 min before part of the solution in the porous medium was displaced by the gas mixture. At this point, the porous medium was partially saturated with the solution. The gas mixture was injected slowly to the set pressure, and the constant pressure process or constant flow rate process was begun. The pressure in the vessel was controlled by a gas injection pump during the constant pressure process, while the pressure in the vessel was controlled by the back pressure control pump and the flow rate of the gas mixture was controlled by the gas injection pump. After the hydrate formation process had finished, the hydrate dissociation process began. This process lasted more than 30 min, and the pressure in the vessel was controlled by the back pressure control pump during the hydrate dissociation process. After the entire cycle had completed, a new cycle was started. By the way, we found that the induction time of hydrate formation was the shortest in the presence of 5 wt% TBAB + 5 wt% THF, and the hydrate could be formed

where Shi and Ssi are the hydrate saturation (volume fraction) and solution saturation (volume fraction) at time i (min), Ss0 is the initial solution saturation at 0 min, and I0 and Ii are the MIs of the solution at 0 min and i min, respectively.

(2)

3. Results and discussion Five experiments were carried out in this study. Different gas mixture flow rates, different pressures, two additives, and two injection methods (a constant pressure process and a constant flow rate process) were considered. Hydrate formation and dissociation were repeated several times for each case to simulate the continuous industrial hydrate-based gas separation process. Further details of the experimental results and parameters are listed in Table 2. 3.1. A typical experimental process Fig. 2 shows the values of the experimental parameters for Case 1. A typical experimental process is discussed in detail below to explain the experimental process clearly. Case 1 includes five cycles at different 376

Applied Energy 221 (2018) 374–385

M. Yang et al.

Table 2 Experimental parameters and resultsa. Case

Cycle

1

1 2 3 4 5 1 2 3 1 2 3 4 1 2 3 1 2 3

2

3

4

5

T

Add

flow

K 279.15

5%TBAB + 5%THF

no

285.15

19%THF

no

yes

yes

p

V

SH

V1

t

Cs

MPa 6.5 5.5 5 4 3 4 4 5 4 3 4 3 4 4 4 3–4 3 –

mL/min – – – – – – – – – – – – 1 2 0.5 2–0 2 –

% 18.9 27.6 29.2 29.4 – 24.3 23.2 39.3 2.38 12.62 16.27 7.21 8.85 9.73 8.89 12.76 5.13 –

%/min

min 10 0 0 0 – 12 0 0 – – – – – – – – – –

% 2.08 −6.14 10.53 −16.58 – −5.49 6.51 −9.47 10.32 −25.98 3.56 −4.70 −17.13 −0.60 −1.43 −17.34 −4.06 –

0.284 0.268

0.283 0.289 – – – – – – – – – –

a The symbols in this table are defined as follows: T is the temperature during the experiments (K); Add is the additive used in the experiments; p is the vessel pressure during the hydrate formation (MPa); V is the flow rate of gas mixture injection during the constant flow rate process (mL/min); SH is the residual hydrate saturation (volume fraction); V1 is the rate of hydrate formation (%/min); t is the induction time for hydrate formation (min); and Cs is the change in the solution saturation (volume fraction).

flow through the porous medium. As soon as Case 1-1 was finished, the gas mixture was injected into the vessel at a rate of 10 mL/min, beginning Case 1-2. As shown in Fig. 2, the MI decreased during the gas injection; this might have been caused by gas displacement, hydrate formation, or both. The MI obviously changed during the gas injection. One possible explanation for this behavior is that hydrate formation can occur during gas injection after the first cycle. This phenomenon might be caused by the memory effect of the water. After the hydrate dissociation, water memory may have remained in the solution, which would be helpful for hydrate formation. Thus, the hydrate formation induction time could be reduced. It always needs a period of time to observe the onset of hydrate formation in a system brought out of equilibrium. And the “period of time” was named “induction time” [59]. The pressure in the vessel was then maintained at 5.5 MPa; during this time, the MI fell rapidly to 616 and then decreased slowly to 433. The hydrate saturation in Case 1-2 was 27.6%, which was higher than

pressures. A high-pressure gas mixture was injected into the vessel at a rate of 10 mL/min. The MI changed only slightly during the gas injection, which indicated that no hydrate was generated during this step. The pressure in the vessel was maintained at 6.5 MPa after the gas injection. The MI then fell rapidly to 774 and then underwent a further slow decreased to 683, and a significant amount of gas was consumed. This behavior indicates that most of the hydrate was generated during the initial rapid decrease in MI. After the MI had stabilized, the vessel was depressurized gradually to 0.5 MPa. The hydrate saturation in Case 1-1 was 18.9%. During depressurization, the MI rose rapidly, and gas was produced, indicating that hydrate dissociation had begun. The MI increased to about 1300, which was higher than the initial MI, meaning that the solution in the FOV had moved. This “solution movement” was the change in the distribution of the solution in the porous medium during the experimental process, which might be caused by the hydrate formation and dissociation or by the injection of the gas mixture and its

Fig. 2. Changes in the experimental parameters for Case 1. 377

Applied Energy 221 (2018) 374–385

M. Yang et al.

solution movement, a darker image indicates greater hydrate. Fig. 3(a), (f), (k), (p), and (u) correspond to the initial states of the five cycles. The five images are slightly different because of the solution movement caused by hydrate formation and dissociation. Fig. 3(b), (g), (l), (q), and (v) correspond to the states after gas injection of during the five cycles, respectively. We found that these states differed slightly from the states before gas injection, in agreement with the MI curve in Fig. 2. Fig. 3(c), (h), (m), (r), and (w) correspond to the hydrate formation stage of the five cycles. The images show obvious darkening due to hydrate formation. The brightness of each image is different, and Fig. 3(h), (m), and (r) are darker than Fig. 3(c) and (w), which indicates that the hydrate saturation in Case 1-1 and Case 1-5 was lower than in Case 1-2, Case 1-3, and Case 1-4. Fig. 3(d), (i), (n), (s), and (x) correspond to the hydrate dissociation stage of the five cycles. These images are brighter because of the hydrate dissociation. Fig. 3(e), (j), (o), (t), and (y) correspond to the state after hydrate dissociation of five cycles, respectively. These images are similar to the following description. Hydrates were first generated at the bottom of the FOV, and, subsequently, the hydrate formation was observed in the throughout the FOV. This phenomenon was mainly caused by the contact between the gas and solution. The gas and solution initially made contact at the bottom of the FOV, and hydrate was simultaneously generated there. After the gas mixture had diffused throughout the FOV, hydrate formation was observed in the over the entire FOV. We found that no hydrate formation occurred during gas injection in Case 1-1, but hydrate was generated when the gas mixture was injected into the vessel in later cycles. This was mainly due to the memory effect of the water produced by hydrate dissociation. A slight darkening can be seen in Fig. 3(w), unlike in the other cycles. This was caused mainly by solution movement rather than hydrate formation. This phenomenon confirmed that there was no hydrate formation in Case 1-5.

in Case 1-1. Subsequently, the MI remained stable, and the vessel was depressurized gradually. The MI increased to 1088 during hydrate dissociation, which was lower than the initial MI. This result indicated that the solution might have been displaced by the gas mixture during the gas injection. Next, Case 1-3 was investigated. A high-pressure gas mixture was injected into the vessel, and the pressure in the vessel was maintained at 5.0 MPa. Similar to the previous case, the MI dropped quickly at first, and then became stable at about 400. The hydrate saturation in Case 1-3 was found to be 29.2%. The pressure in the vessel was then reduced to induce hydrate dissociation. The MI increased to 1365 due to hydrate dissociation and solution movement; the solution movement might have been caused by the hydrate dissociation. Case 14 followed Case 1-3. The gas mixture was injected into the vessel until the pressure in the vessel reached 4.0 MPa; simultaneously, a constant pressure process was employed. The MI decreased to about 380 during the constant pressure process. The hydrate saturation in Case 1-4 was 29.4%, which was similar to that of Case 1-3. Hydrate dissociation was then induced by depressurization, and the MI increased to 1060. At about 560 min, gas collection was carried out, causing the pressure to drop to 0. A sudden increase in the MI was observed due to the injection of the gas mixture into the vessel. After this injection, Case 1-5 was carried out. The gas mixture was injected into the vessel to a pressure of 3.0 MPa, and the MI decreased during the constant pressure process. However, the rate of reduction in the MI was lower than in other cycles, and the MI did not return to the original value after depressurization. The amount of solution in the FOV would be decreased and the images would be dark when the hydrate was generated. Water would be produced from the hydrate and the image would turn bright, when the hydrate was dissociated. The phenomenon could be shown in Fig. 3(a)–(e). However, the image did not turn bright after the depressurization in Case 1-5. This indicates that no hydrate formation occurred in Case 1-5. Based on our previous work, the hydrate phase equilibrium temperature is 286 K at 3.0 MPa in the presence of 5 wt% TBAB + 5 wt% THF, but the temperature in the vessel was 279.15 K. in other words, This phenomenon (no hydrate formation) might be caused by the volatilization of additives during the depressurization; thus, the hydrate could not be generated without additives at 3.0 MPa [17]. Similarly, Sun et al. carried out the experiments on the separation of gas mixture via hydrate-based technology in the presence of THF [60]. They found that the volatilization of THF from liquid phase was considerable during the experiments, especially during the hydrate dissociation process. we found that the phase equilibrium pressure of CO2 /N2 hydrate will increase as the additives are changed from 5% TBAB + 5% THF (284.65 K, 1.90 MPa) into 5% TBAB + 1% THF (284.55 K, 4.56 MPa) [17,61].Besides the volatilization of additives, we considered that the pressure in the vessel was lower than other cycles, and the lower pressure meant a lower driving force of hydrate formation, which might cause a longer induction time of hydrate formation. Therefore, we did not observe the hydrate formation in Case 1-5. In general, the hydrate formation induction time in the first cycle was longer than in later cycles. In addition, the MI values after hydrate formation in Case 1-2, Case 1-3, and Case 1-4 were similar: approximately 400, lower than the MI in Case 1-1. This indicates that the hydrate saturation in the first cycle was lower than in later cycles. This was due to the memory effect of the water produced by hydrate dissociation. The memory effect was more significant than the pressure. Additionally, solution movement was observed. Furthermore, these might be helpful for the analysis of the distribution of solution in actual industrial processes. Fig. 3 shows the distribution of the solution in the FOV for Case 1. The five images in each row show the whole process over one cycle, and the individual images correspond to the inflection points in Fig. 2. The bright areas in the image represent areas with a high concentration of solution, and a darkening of the image indicates a decrease in a solution concentration. The reduction of the solution concentration is caused by solution movement or hydrate formation. Thus, in the absence of

3.2. The effect of the continuous process on hydrate formation One of the main effects of the continuous process on hydrate formation is the memory effect of the water. The memory effect of water allows hydrate to be formed more easily in water that has previously undergone hydrate dissociation [62]. The formation of hydrates was accelerated by the memory effect. As listed in Table 2, the hydrate formation induction times of Case 1-1 and Case 2-1 were about 10 min, while the hydrate formation induction times of later cycles in Case 1 and Case 2 were close to 0 min. The hydrate saturation increased as the number of cycles increased, even though the pressure in the vessel was lower. This phenomenon might be due to the memory effect of water, which had a greater influence than the pressure, on hydrate formation. A similar phenomenon was observed in Case 3. The hydrate saturation in Case 3-2 was 10.2% higher than in Case 3-1, although the pressure in Case 3-2 was 1.0 MPa lower. However, the water memory had little effect on the hydrate formation in the constant flow rate process. The hydrate saturation in Case 4-2 was only 0.9% higher than that in Case 4-1. In general, water memory had a significant effect on hydrate formation in the constant pressure process; however, the effect of the water memory on hydrate formation in the constant flow rate process was not obvious. Fig. 4 shows the changes in solution saturation for all the cases. Fig. 4(a) shows the changes in solution saturation after each cycle in the constant pressure process. The change in the solution saturation (Cs) is the difference between the solution saturation before hydrate formation and after hydrate dissociation in a given cycle. The “sum” is defined as the accumulation of Cs with increasing cycle number. We found that the value of Cs alternated between positive and negative values in the constant pressure process. This indicated that the amount of solution in the FOV fluctuated, although the amount of the solution tended to decrease. The Cs values ranged from -16.58% to 10.53% in the constant pressure process in the presence of 5 wt% TBAB + 5 wt% THF, and from −25.98% to 10.32% in the constant pressure process in the 378

Applied Energy 221 (2018) 374–385

M. Yang et al.

Fig. 3. Distribution of solution in the FOV for Case 1.

at low solution saturation. The solution movement was more significant in the constant pressure process than the constant flow rate process in one cycle in the presence of 19 wt% THF. However, the accumulation of solution movement was more significant in the constant flow rate process than in the constant pressure process as the number of the cycles increased in the presence of 19 wt% THF. As mentioned above, no hydrate was generated in Case 1-5, and possibly due to the volatilization of the additives. Similar phenomena were observed in Case 3-4 and Case 5-3. Therefore, additional solution should be added after several cycles in the continuous gas separation process, and further experiments are required to study how and when to replenish the solution. Fig. 5(a) and (b) shows the ratio of the area in the image with a solution saturation above a given threshold (10%, 20%, 30%, 40%, and 50% solution saturation) and the total area of the FOV for Case 3 and Case 4, respectively. For example, the 10% curve represents the ratio between the area with solution saturation over 10% and the total area

presence of 19 wt% THF. The sum in the presence of 5 wt% TBAB + 5 wt% THF was lower than in the presence of 19 wt% THF. This indicated that the 19 wt% THF additive had a greater effect on the solution movement than 5 wt% TBAB + 5 wt% THF during the constant pressure process. The solution concentration phenomenon was observed in the presence of 19 wt% THF, but not when using 5 wt% TBAB + 5 wt% THF. The more obvious solution movement might be caused by the solution concentration phenomenon. Fig. 4(b) shows the changes in the solution saturation after each cycle in the constant flow rate process. All Cs values were negative, indicating that the amount of solution in the FOV tended to decrease in the constant flow rate process. The Cs values ranged from −17.34% to −0.6%. The Cs was about −17.0% in Case 4-1 and Case 5-1, while the Cs was higher than −5.0% in later cycles. The reason for this might be that the solution movement was more significant when a greater amount of the solution was present in the constant flow rate process, and less solution movement occurred 379

Applied Energy 221 (2018) 374–385

M. Yang et al.

Fig. 4. Changes in the solution saturation for all cases.

main factors affecting the continuous process of gas separation. The memory effect of water reduced the hydrate formation induction time in later cycles, and had a more significant effect on the hydrate formation in the constant pressure process than in the constant flow rate process in the presence of 19 wt% THF. In terms of solution movement, the amount of solution tended to decrease with increasing cycle number. The addition of 19 wt% THF affected solution movement more than 5 wt% TBAB + 5 wt% THF in the constant pressure process. The solution movement was more significant in the constant flow rate process than the constant pressure process in the presence of 19 wt% THF, and solution concentration occurred in the presence of 19 wt% THF at 285.15 K. Thus, it is necessary to replenish the solution during

of the FOV. We found that the 20%, 30%, 40%, and 50% curves decreased more rapidly than the 10% curve during hydrate formation. This indicates that, initially, hydrate was generated where the solution was most abundant in the presence of additives; then, the hydrate was generated where less solution was present. Furthermore, according to Fig. 5(a), the 10%curve decreased during hydrate formation in Case 31, whereas the 20%, 30%, 40%, and 50% curves increased. This indicates that solution concentration occurred in the presence of 19 wt% THF at 285.15 K. The 10% curve was the first react during the hydrate dissociation, indicating that hydrate dissociation occurred gradually rather than instantaneously. Overall, the water memory effect and solution movement were the 380

Applied Energy 221 (2018) 374–385

M. Yang et al.

Fig. 5. Variation in the area ratios for different solution saturations in Case 3 (a) and Case 4 (b).

TBAB + 5 wt% THF were set to be longer than 76 min. Similarly, the reaction time for hydrate formation was extended to 640 min in Case 32. The hydride formation could also be divided into two parts in this case; the first part (rapid hydrate formation stage) occurred from 298 to 380 min, and the second part (slow hydrate formation stage) took place from 380 to 938 min. The hydrate formation rate decreased to approximately 0.009%/min in the second part (slow hydrate formation stage), which was about one-twentieth of the rate of hydrate formation in the first part. In addition, the duration of hydrate formation was about 80 min (1066–1148 min) in Case 3-3, and the hydrate formation rate was about 0.2% per min. This was consistent with the hydrate formation rate in Case 3-2′s rapid hydrate formation stage. Moreover, the duration was about 80 min, which was also corresponding to the mentioned conclusions. The reaction time of the rapid hydrate formation stage was 82 min in the presence of 19 wt% THF. Thus, the reaction time of hydrate formation should be longer than 82 min. In general, 5 wt% TBAB + 5 wt% THF and 19 wt% THF had similar effect on the reaction time of hydrate formation, and the optimal reaction time of hydrate formation was found to be about 80 min, and this could provide theoretical support for improving efficiency in industrial processes. .

the continuous industrial hydrate-based gas separation process because of the solution movement caused by hydrate formation and dissociation. 3.3. The effect of the constant pressure duration on hydrate saturation Fig. 6 shows the changes in the experimental parameters during Case 1-1 and Case 2-1. Efficiency is an important factor in the continuous industrial hydrate-based gas separation process, and the hydrate formation reaction time is the key to the efficiency. Complete generation of the hydrate takes a long time. A large amount of hydrate is generated initially, and then the rate of hydrate formation decreases. Hence, it is necessary to identify a suitable reaction time for hydrate formation. The reaction time of hydrate formation was extended to 514 min in Case 2-1 in the presence of 5 wt% TBAB + 5 wt% THF, which was several times greater than for the other tested conditions. The hydrate formation process can be divided into two parts. The first part (rapid hydrate formation stage) occurred from 34 to 110 min, and the second part took place from 110 to 536 min. During the rapid hydrate formation stage, with a hydrate formation rate of approximately 0.2% per min. During the slow hydrate formation stage, the hydrate formation rate dropped to about one-tenth of the initial rate (0.02% per min). The duration of the first part was 76 min; thus, both the hydrate saturation and the efficiency of the gas separation could be studied using a reaction time of slightly longer than 76 min. Thus, all the reaction times for hydrate formation in the presence of 5 wt%

3.4. The effect of pressure on hydrate formation Fig. 7 shows the hydrate saturation under different pressure conditions during the constant pressure process. The pressures in the vessel

Fig. 6. Changes in the experimental parameters for Case 1-1 and Case 2-1.

Fig. 7. Hydrate saturation under different pressure conditions. 381

Applied Energy 221 (2018) 374–385

M. Yang et al.

Fig. 8(m)–(u) are images of the experimental process for Case 4. Fig. 8(a), (d), (g), (j), (m), (p), and (s) correspond to the initial states of Case 3-1, Case 3-2, Case 3-3, Case 3-4, Case 4-1, Case 4-2, and Case 4-3, respectively. Fig. 8(c), (f), (i), (l), (o), (r), and (u) correspond to the hydrate formation states of the seven cycles. We found that the solution movement was more significant in the constant pressure process than in the constant flow rate process in the presence of 19 wt% THF. Solution concentration was observed in both the constant pressure and constant flow rate process. The distribution of the solution concentration was different in the constant pressure process and the constant flow rate process. Solution concentration was observed over the whole FOV in the constant pressure process, while in the constant flow rate process this phenomenon mainly appeared at the midline of the FOV and near the tube wall. The solution concentration area was greater in the constant pressure process than in the constant flow rate process. Fig. 9 shows the area ratios for different solution saturation thresholds (10%, 20%, 30%, and 40%). As shown in Fig. 9(a), the 10% curve dropped by approximately 50% after the first cycle in Case 4, whereas it decreased by approximately 10% in Case 3. Furthermore, the 20%, 30%, and 40% curves showed a decreasing trend in Case 4, unlike in Case 3. Moreover, the 10% and 20% curves in Case 3 showed higher values than in Case 4. These phenomena also confirmed that the solution concentration was more significant in the constant pressure process than that in the constant flow rate process. Three cycles were carried out in Case 4. A pressure of 4.0 MPa was used for all cycles; however, their flow rates were 1.0, 2.0, and 0.5 mL/ min. The hydrate saturation in each cycle was about 9.0%. This indicates that the flow rate had little effect on the hydrate formation in the presence of 19 wt% THF. However, a hydrate saturation value of 16.27% was observed in Case 3-3 at 4.0 MPa in the constant pressure

in Case 1-1, Case 1-2, Case 1-3, and Case 1-4 were 6.5, 5.5, 5.0, and 4.0 MPa, respectively, and the hydrate saturation values were 18.9%, 27.6%, 29.2%, and 29.4%. The hydrate saturation increased slightly from Case 1-2 to Case 1-4, which was approximately 10% higher than in Case 1-1. This was due to the memory effect of water. The solution in the first cycle consisted of deionized water with additives, whereas the solution contained the memory water with additives in later cycles. The memory water might have had a more significant effect on hydrate formation than the pressure. Therefore, Case 1-1 and Case 2-1 were compared to analyze the effect of the pressure. As shown in Fig. 6, the pressures in Case 1-1 and Case 2-1 were 6.5 and 4.0 MPa, respectively. The hydrate saturation reached 18.9% 76 min after gas injection in Case 1-1, while the hydrate saturation 76 min after gas injection was 9.7% in Case 2-1. In Case 2-1, 104 min were required for the hydrate saturation to reach 18.9%. Furthermore, the hydrate formation induction time in Case 1-1 was shorter than in Case 2-1. This indicates that the higher pressure was beneficial for hydrate formation. Moreover, the hydrate saturation increased from 23.2% to 39.3% when the pressure was increased from 4.0 MPa in Case 2-2 to 5.0 MPa in Case 2-3. There are two main reasons for this: the effect of the water memory and the effect of the pressure. A similar phenomenon was observed in Case 3-2 and Case 3-3. Overall, the higher pressure was beneficial for hydrate formation in the constant pressure process, and the hydrate formation induction time was shorter at higher pressures. Thus, it is feasible to increase the pressure in the first cycle after solution replenishment. 3.5. Differences between the flow and constant pressure processes Fig. 8 shows the distribution of the solution during Case 3 and Case 4. Fig. 8(a)–(l) are images of the experimental process for Case 3, and

Fig. 8. Distribution of the solution in the FOV for Case 3 and Case 4. 382

Applied Energy 221 (2018) 374–385

M. Yang et al.

Fig. 9. Variation in the area ratios for different solution saturations in Case 3, Case 4, and Case 5 (a: 10%, b: 20%, c: 30%, d: 40%).

could not be generated within a short time at 285.15 K in the presence of 5 wt% TBAB + 5 wt% THF, indicating that 19 wt% THF was more suitable for use at higher temperatures, which could reduce the amount of energy consumed for cooling in industrial hydrate-based gas separation. If temperature is ignored, the hydrate saturation was found to be 29.4% in Case 1-4 in the presence of 5 wt% TBAB + 5 wt% THF, which was about 1.8 times greater than in Case 3-3 (16.27%) in the presence of 19 wt% THF. In general, the solution movement was more significant in the presence of 5 wt% TBAB + 5 wt% THF than 19 wt% THF. In addition, the solution became concentrated in the presence of 19 wt% THF at 285.15 K, whereas this phenomenon did not occur in the presence of 5 wt% TBAB + 5 wt% THF at 279.15 K. Finally, in terms of hydrate saturation, 5 wt% TBAB + 5 wt% THF was more suitable in this study, while 19 wt% THF was more suitable for a higher experimental temperature.

process. Similarly, the hydrate saturation was 12.62% in Case 3-2 and 7.21% in Case 3-4 at 3.0 MPa in the constant pressure process, both of which were higher than in Case 5-2 (5.13%). This indicated that the use of the constant pressure process was more beneficial for hydrate formation than the constant flow rate process. In summary, the constant flow rate process had a greater effect on solution movement than the constant pressure process, and the solution concentration was more significant in the constant pressure process than that in the constant flow rate process in the presence of 19 wt% THF. The hydrate saturation was higher in the constant pressure process than that in the constant flow rate process. Moreover, the flow rate had little effect on hydrate formation in this study. 3.6. Effect of additives on hydrate formation As shown in Figs. 3 and 8, the solution became concentrated in the presence of 19 wt% THF at 285.15 K; however, this phenomenon was not observed in the presence of 5 wt% TBAB + 5 wt% THF at 279.15 K. In contrast, hydrates were generated over the whole FOV uniformly in the presence of 5 wt% TBAB + 5 wt% THF at 279.15 K, while in the presence of 19 wt% THF at 285.15 K, the hydrate was mainly generated in the regions in which the solution was concentrated. Furthermore, the solution concentration was affected by hydrate formation and dissociation in the presence of 19 wt% THF. The experimental temperature was 279.15 K in Case 1 and Case 2 (5 wt% TBAB + 5 wt% THF), and 285.15 K in Case 3, Case 4, and Case 5 (19 wt% THF). The hydrate

4. Conclusions In this study, a hydrate-based continuous gas separation process was carried out. The effects of two gas injection methods (constant pressure and constant flow rate processes), two additives (5 wt% TBAB + 5 wt% THF and 19 wt% THF), different pressures, different reaction times of hydrate formation, and three different flow rates were investigated. The constant flow rate process had a greater effect on the solution movement than the constant pressure process. The solution movement was more significant in the presence of 19 wt% THF than 5 wt% 383

Applied Energy 221 (2018) 374–385

M. Yang et al.

TBAB + 5 wt% THF. Due to the solution movement during the continuous industrial hydrate-based gas separation process, the solution might need to be replenished. We found that the process was most efficient with a hydrate formation stage time of approximately 80 min. Solution concentration was observed in the presence of 19 wt% THF at 285.15 K, and may have been affected by the formation and dissociation of hydrates. Moreover, the flow rate had little effect on hydrate formation in the presence of 19 wt% THF. Finally, in terms of hydrate saturation, the 5 wt% TBAB + 5 wt% THF additive was more suitable in this study, while 19 wt% THF was more suitable at a higher experimental temperature.

[23]

[24]

[25]

[26] [27]

Acknowledgements [28]

This project was financially supported by the National Key Research and Development Plan of China (2017YFC0307300, 2016YFC0304001), and the Fundamental Research Funds for the Central Universities of China (DUT18ZD403).

[29]

[30]

References [31]

[1] Raoof A, Nick HM, Wolterbeek TKT, Spiers CJ. Pore-scale modeling of reactive transport in wellbore cement under CO2 storage conditions. Int J Greenhouse Gas Control 2012;11:S67–77. [2] Krebs FC. Carbon dioxide—a themed issue. Energy Environ Sci 2012;5:7238. [3] Dou B, Song Y, Liu Y, Feng C. High temperature CO2 capture using calcium oxide sorbent in a fixed-bed reactor. J Hazardous Mater 2010;183:759–65. [4] Lee Y, Lee D, Lee J-W, Seo Y. Enclathration of CO2 as a co-guest of structure H hydrates and its implications for CO2 capture and sequestration. Appl Energy 2016;163:51–9. [5] Ma ZW, Zhang P, Bao HS, Deng S. Review of fundamental properties of CO2 hydrates and CO2 capture and separation using hydration method. Renew Sustain Energy Rev 2016;53:1273–302. [6] Zhang Y, Ji X, Lu X. Energy consumption analysis for CO2 separation from gas mixtures. Appl Energy 2014;130:237–43. [7] Yang M, Song Y, Jiang L, Zhao Y, Ruan X, Zhang Y, et al. Hydrate-based technology for CO2 capture from fossil fuel power plants. Appl Energy 2014;116:26–40. [8] Komatsu H, Hayasaka A, Ota M, Sato Y, Watanabe M, Smith RL. Measurement of pure hydrogen and pure carbon dioxide adsorption equilibria for THF clathrate hydrate and tetra-n-butyl ammonium bromide semi-clathrate hydrate. Fluid Phase Equilib 2013;357:80–5. [9] Aaron D, Tsouris C. Separation of CO2from flue gas: a review. Sep Sci Technol 2005;40:321–48. [10] Ding K, Zhong D-L, Lu Y-Y, Wang J-L. Enhanced precombustion capture of carbon dioxide by gas hydrate formation in water-in-oil emulsions. Energy Fuels 2015;29:2971–8. [11] Lee HJ, Lee JD, Linga P, Englezos P, Kim YS, Lee MS, et al. Gas hydrate formation process for pre-combustion capture of carbon dioxide. Energy 2010;35:2729–33. [12] Park S, Lee S, Lee Y, Seo Y. CO2 capture from simulated fuel gas mixtures using semiclathrate hydrates formed by quaternary ammonium salts. Environ Sci Technol 2013;47:7571–7. [13] Li S, Fan S, Wang J, Lang X, Liang D. CO2 capture from binary mixture via forming hydrate with the help of tetra-n-butyl ammonium bromide. J Nat Gas Chem 2009;18:15–20. [14] Ricaurte M, Dicharry C, Broseta D, Renaud X, Torré J-P. CO2 Removal from a CO2–CH4 gas mixture by clathrate hydrate formation using THF and SDS as watersoluble hydrate promoters. Ind Eng Chem Res 2013;52:899–910. [15] Tang J, Zeng D, Wang C, Chen Y, He L, Cai N. Study on the influence of SDS and THF on hydrate-based gas separation performance. Chem Eng Res Des 2013;91:1777–82. [16] Xu C-G, Zhang S-H, Cai J, Chen Z-Y, Li X-S. CO2 (carbon dioxide) separation from CO2–H2 (hydrogen) gas mixtures by gas hydrates in TBAB (tetra-n-butyl ammonium bromide) solution and Raman spectroscopic analysis. Energy 2013;59:719–25. [17] Yang M, Jing W, Zhao J, Ling Z, Song Y. Promotion of hydrate-based CO2 capture from flue gas by additive mixtures (THF (tetrahydrofuran) + TBAB (tetra-n-butyl ammonium bromide)). Energy 2016;106:546–53. [18] Yang M, Liu W, Song Y, Ruan X, Wang X, Zhao J, et al. Effects of additive mixture (THF/SDS) on the thermodynamic and kinetic properties of CO2/H2 hydrate in porous media. Ind Eng Chem Res 2013;52:4911–8. [19] Cai J, Xu C-G, Xia Z-M, Chen Z-Y, Li X-S. Hydrate-based methane separation from coal mine methane gas mixture by bubbling using the scale-up equipment. Appl Energy 2017;204:1526–34. [20] Seo YT, And ILM, Ripmeester JA, And JL, Lee H. Efficient recovery of CO2 from flue gas by clathrate hydrate formation in porous silica gels. Environ Sci Technol 2005;39:2315–9. [21] Babu P, Kumar R, Linga P. A new porous material to enhance the kinetics of clathrate process: application to precombustion carbon dioxide capture. Environ Sci Technol 2013;47:13191–8. [22] Park S, Lee S, Lee Y, Lee Y, Seo Y. Hydrate-based pre-combustion capture of carbon

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44] [45]

[46]

[47] [48]

[49]

[50]

384

dioxide in the presence of a thermodynamic promoter and porous silica gels. Int J Greenhouse Gas Control 2013;14:193–9. Adeyemo A, Kumar R, Linga P, Ripmeester J, Englezos P. Capture of carbon dioxide from flue or fuel gas mixtures by clathrate crystallization in a silica gel column. Int J Greenhouse Gas Control 2010;4:478–85. Zhong D-L, Lu Y-Y, Sun D-J, Zhao W-L, Li Z. Performance evaluation of methane separation from coal mine gas by gas hydrate formation in a stirred reactor and in a fixed bed of silica sand. Fuel 2015;143:586–94. Alberti M, Pirani F, Lagana A. Carbon dioxide clathrate hydrates: selective role of intermolecular interactions and action of the SDS catalyst. J Phys Chem A 2013;117:6991–7000. Zhang CS, Fan SS, Liang DQ, Guo KH. Effect of additives on formation of natural gas hydrate. Fuel 2004;83:2115–21. Mohammadi AH, Eslamimanesh A, Belandria V, Richon D, Naidoo P, Ramjugernath D. Phase equilibrium measurements for semi-clathrate hydrates of the (CO2+N2+tetra-n-butylammonium bromide) aqueous solution system. J Chem Thermodyn 2012;46:57–61. Fukumoto A, Dalmazzone D, Paricaud P, Fürst W. Experimental measurements and modeling of the dissociation conditions of tetrabutylammonium chloride semiclathrate hydrates in the presence of hydrogen. J Chem Eng Data 2015;60:343–50. Du J, Wang L. Equilibrium conditions for semiclathrate hydrates formed with CO2, N2, or CH4 in the Presence of Tri-n-butylphosphine Oxide. Ind Eng Chem Res 2014;53:1234–41. Babu P, Yao M, Datta S, Kumar R, Linga P. Thermodynamic and kinetic verification of tetra-n-butyl ammonium nitrate (TBANO3) as a promoter for the clathrate process applicable to precombustion carbon dioxide capture. Environ Sci Technol 2014;48:3550–8. Li X-S, Xu C-G, Chen Z-Y, Wu H-J. Hydrate-based pre-combustion carbon dioxide capture process in the system with tetra-n-butyl ammonium bromide solution in the presence of cyclopentane. Energy 2011;36:1394–403. Zhao W-L, Zhong D-L, Yang C. Prediction of phase equilibrium conditions for gas hydrates formed in the presence of cyclopentane or cyclohexane. Fluid Phase Equilib 2016;427:82–9. Zhong D-L, Wang W-C, Zou Z-L, Lu Y-Y, Yan J, Ding K. Investigation on methane recovery from low-concentration coal mine gas by tetra- n -butyl ammonium chloride semiclathrate hydrate formation. Appl Energy 2017. Kim S, Seo Y. Semiclathrate-based CO2 capture from flue gas mixtures: an experimental approach with thermodynamic and Raman spectroscopic analyses. Appl Energy 2015;154:987–94. Sfaxi IBA, Durand I, Lugo R, Mohammadi AH, Richon D. Hydrate phase equilibria of CO2+N2+aqueous solution of THF, TBAB or TBAF system. Int J Greenhouse Gas Control 2014;26:185–92. Deschamps J, Dalmazzone D. Dissociation enthalpies and phase equilibrium for TBAB semi-clathrate hydrates of N2, CO2, N2 + CO2 and CH4 + CO2. J Therm Anal Calorim 2009;98:113–8. Hashimoto H, Yamaguchi T, Kinoshita T, Muromachi S. Gas separation of flue gas by tetra- n -butylammonium bromide hydrates under moderate pressure conditions. Energy 2017;129:292–8. Fukumoto A, Sales Silva LP, Paricaud P, Dalmazzone D, Fürst W. Modeling of the dissociation conditions of H2 + CO2 semiclathrate hydrate formed with TBAB, TBAC, TBAF, TBPB, and TBNO3 salts. Application to CO2 capture from syngas. Int J Hydrogen Energy 2015;40:9254–66. Fukumoto A, Paricaud P, Dalmazzone D, Bouchafaa W, Ho TT-S, Fürst W. Modeling the dissociation conditions of Carbon Dioxide + TBAB, TBAC, TBAF, and TBPB Semiclathrate Hydrates. J Chem Eng Data 2014;59:3193–204. Zhao W-L, Yang C, Wang R-R. Phase equilibria prediction model for semiclathrate hydrates formed from CH4, CO2, and N2in the presence of tetra-n-butyl ammonium bromide. Energy Sources Part A 2016;38:621–7. Garcia M, Marriott R, Clarke MA. Modeling of the thermodynamic equilibrium conditions for the formation of TBAB and TBAC semiclathrates formed in the presence of Xe, Ar, CH4, CO2, N2, and H2. Ind Eng Chem Res 2016;55:777–87. Linga P, Adebola A, Englezos P. Medium-Pressure clathrate hydrate/membrane hybrid process for postcombustion capture of carbon dioxide. Environ Sci Technol 2007;42:315–20. Wang M, Sun Z-G, Qiu X-H, Zhu M-G, Li C-H, Zhang A-J, et al. Hydrate dissociation equilibrium conditions for Carbon Dioxide + Tetrahydrofuran. J Chem Eng Data 2017;62:812–5. Linga P, Kumar R, Englezos P. Gas hydrate formation from hydrogen/carbon dioxide and nitrogen/carbon dioxide gas mixtures. Chem Eng Sci 2007;62:4268–76. Cai J, Zhang Y, Xu C-G, Xia Z-M, Chen Z-Y, Li X-S. Raman spectroscopic studies on carbon dioxide separation from fuel gas via clathrate hydrate in the presence of tetrahydrofuran. Appl Energy 2018;214:92–102. Kang S-P, Lee H. Recovery of CO2 from flue gas using gas hydrate: thermodynamic verification through phase equilibrium measurements. Environ Sci Technol 2000;34:4397–400. Sun Q, Kim S, Kang YT. Study on dissociation characteristics of CO2 hydrate with THF for cooling application. Appl Energy 2017;190:249–56. Lirio CFdS, Pessoa FLP, Uller AMC. Storage capacity of carbon dioxide hydrates in the presence of sodium dodecyl sulfate (SDS) and tetrahydrofuran (THF). Chem Eng Sci 2013;96:118–23. Tzirakis F, Stringari P, von Solms N, Coquelet C, Kontogeorgis G. Hydrate equilibrium data for the CO2 + N2 system with the use of tetra-n-butylammonium bromide (TBAB), cyclopentane (CP) and their mixture. Fluid Phase Equilib 2016;408:240–7. Tzirakis F, Stringari P, Coquelet C, von Solms N, Kontogeorgis G. Hydrate equilibrium data for CO2+N2 system in the presence of Tetra-n-butylammonium Fluoride

Applied Energy 221 (2018) 374–385

M. Yang et al.

[51]

[52]

[53]

[54]

[55]

[56] Li X-S, Xu C-G, Chen Z-Y, Wu H-J. Tetra-n-butyl ammonium bromide semi-clathrate hydrate process for post-combustion capture of carbon dioxide in the presence of dodecyl trimethyl ammonium chloride. Energy 2010;35:3902–8. [57] Ho LC, Babu P, Kumar R, Linga P. HBGS (hydrate based gas separation) process for carbon dioxide capture employing an unstirred reactor with cyclopentane. Energy 2013;63:252–9. [58] Sloan ED. Clathrate Hydrates of Natural Gas. M.Dekker; 2007. [59] Douïeb S, Archambault S, Fradette L, Bertrand F, Haut B. Effect of the fluid shear rate on the induction time of CO2-THF hydrate formation. Can J Chem Eng. 2017;95:187–98. [60] Sun Q, Chen G, Guo X, Liu A. Experiments on the continuous separation of gas mixtures via dissolution and hydrate formation in the presence of THF. Fluid Phase Equilib 2014;361:250–6. [61] Yang M, Jing W, Wang P, Jiang L, Song Y. Effects of an additive mixture (THF +TBAB) on CO2 hydrate phase equilibrium. Fluid Phase Equilib 2015;401:27–33. [62] Takeya S, And AH, Hondoh T, Uchida T. Freezing-memory effect of water on nucleation of CO2 hydrate crystals. J Phys Chem B 2000;104:4164–8.

(TBAF) and mixture of TBAF and Cyclopentane (CP). J Chem Eng Data 2016;61:1007–11. Xia Z-m, Li X-s, Chen Z-y, Li G, Cai J, Wang Y, et al. Hydrate-based acidic gases capture for clean methane with new synergic additives. Appl Energy 2017;207:584–93. Sun Q, Guo X, Liu A, Liu B, Huo Y, Chen G. Experimental study on the separation of CH4 and N2via hydrate formation in TBAB solution. Ind Eng Chem Res 2011;50:2284–8. Mech D, Gupta P, Sangwai JS. Kinetics of methane hydrate formation in an aqueous solution of thermodynamic promoters (THF and TBAB) with and without kinetic promoter (SDS). J Nat Gas Sci Eng 2016;35:1519–34. Luo Y, Yan Z-Y, Li J, Guo G-J, Guo X-Q, Sun Q, et al. Effects of dimethyl sulfoxide on phase equilibrium conditions of CO2and IGCC fuel gas hydrate in the presence and absence of Tetra-n-butyl Ammonium Bromide. J Chem Eng Data 2017;62:188–93. Li Q, Fan S, Wang Y, Lang X, Chen J. CO2 removal from biogas based on hydrate formation with Tetra-n-Butylammonium Bromide Solution in the presence of 1Butyl-3-Methylimidazolium Tetrafluoroborate. Energy Fuels 2015;29:3143–8.

385