- Email: [email protected]
liquid system: An experimental study using methane and micelle-forming surfactants
Chemical Engineering Science 73 (2012) 79–85
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Science journal homepage: www.elsevier.com/locate/ces
Surfactant effects on hydrate formation in an unstirred gas/liquid system: An experimental study using methane and micelle-forming surfactants Naoki Ando, Yui Kuwabara 1, Yasuhiko H. Mori n Department of Mechanical Engineering, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan
a r t i c l e i n f o
abstract
Article history: Received 15 November 2011 Received in revised form 16 January 2012 Accepted 18 January 2012 Available online 28 January 2012
This paper reports an experimental study in which we intended to obtain a better understanding of the possible role of surfactant micelles on the formation of a clathrate hydrate in a quiescent methane/ liquid-water system. The experiments were performed using a laboratory-scale, isobaric hydrateforming reactor, which was initially composed of a 300-cm3 aqueous phase and a 640-cm3 methanegas phase, then successively provided with methane such that the system pressure was held constant at 3.9 or 4.0 MPa. The surfactants used in this study were lithium dodecyl sulfate (LDS), dodecylbenzene sulfonic acid (DBSA), and sodium oleate (SO), which have sufficiently low Krafft points and hence, unlike sodium alkyl sulfates (sodium dodecyl sulfate and its homologues), allow the micelle formation under such typical hydrate-forming conditions as those used in this study ( 275 K in temperature and 3.9 4.0 MPa in pressure). Significant increases in the rate of hydrate formation and the final water-tohydrate conversion ratio were simultaneously observed by the addition of LDS to the aqueous phase up to concentrations in the range from 0.6 to 1.6 times the relevant critical micelle concentration (CMC). Neither the rate of hydrate formation nor the final water-to-hydrate conversion ratio exhibited any appreciable change in the above concentration range. Similar observations were obtained by the addition of DBSA to the aqueous phase up to the concentrations that ranged from 0.5 to 2.9 times the relevant CMC. Based on these observations, we have concluded that micelles of LDS and DBSA have no practical effect on hydrate formation. No substantial promotion of hydrate formation was detected by the addition of SO to the aqueous phase up to concentrations that ranged from 0.8 to 4.2 times the relevant CMC. & 2012 Elsevier Ltd. All rights reserved.
Keywords: Clathrate hydrate Gas hydrate Crystallization Solutions Surfactant Energy
1. Introduction This study is concerned with the effects of surfactant additives on the formation of a clathrate hydrate (abbreviated hydrate, hereafter) in a system containing an aqueous phase and a methane gas phase in mutual contact. Many studies, mostly experimental studies, have been reported so far about the effects of various surfactant additives on the formation of hydrates, crystalline solid compounds formed from water and various guest substances including light hydrocarbons, carbon dioxide and some fluorocarbons. The fact that the hydrate formation is substantially promoted by the addition of some types of surfactants is of potential importance for the industrial applications of such hydrates, for example, the storage and transport of natural gas or hydrogen, the separation of carbon dioxide from flue gas at coal-fired power plants, the recovery of clean water from the waste water generated at paper-making mills, and the cool
n
Corresponding author. Tel.: þ81 45 566 1522; fax: þ 81 45 566 1495. E-mail address: [email protected] (Y.H. Mori). 1 Present address: Daihatsu Motor Co. Ltd., Ikeda-shi, Osaka 563-8651, Japan.
0009-2509/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2012.01.038
storage for residential air conditioning. The point of particular interest in the hydrate formation in surfactant-containing systems is that, even in the absence of any mechanical means for mixing the aqueous and the guest-gas phases inside a hydrate-forming reactor, high-rate hydrate formation continues, generating porous hydrate layers climbing on the wall of the reactor from the level of the horizontal interface between the two fluid phases (Kutergin et al., 1992; Mel’nikov et al., 1998; Zhong and Rogers, 2000; Sun et al., 2003a, 2003b, 2003c; Watanabe et al., 2005a; Pang et al., 2007; Okutani et al., 2007, 2008). If the water (or the aqueous liquid) inside the reactor is not replenished during the hydrate-forming process while the guest gas is continuously replenished (just as in ordinary isobaric, semi-batch hydrate-forming operations), the highrate hydrate formation lasts until the aqueous phase reduces to 5 20% of its initial volume (Okutani et al., 2007, 2008). These findings indicate the possibility of developing an economical, highperformance industrial hydrate-forming technology utilizing appropriate surfactant additives. One of the fundamental and, at the same time, practically important question still left for a better understanding about the hydrate formation in surfactant-containing systems is the effect
80
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
of surfactant micelles formed in the aqueous phase on the hydrate-formation kinetics. As surveyed by, for example, Okutani et al. (2008), this issue was once a subject of controversy. Although the controversy is apparently over at present, the issue is still not yet completely settled, which is briefly discussed below. Zhong and Rogers (2000) were probably the first to raise this issue. For interpreting their experimental results about the effect of ethane hydrate formation in the presence of an anionic surfactant, sodium dodecyl sulfate (SDS), they suggested that SDS micelles once formed would work as the carriers of ethane molecules to the bulk of the aqueous phase and thereby promoting the hydrate formation. However, this hypothesis has been disputed and denied by several research groups on the basis of different arguments (Di Profio et al., 2005, 2007; Watanabe et al., 2005a, 2005b; Gayet et al., 2005; Pang et al., 2007; Zhang et al., 2007a, 2007b, 2007c). The most straightforward argument against this hypothesis was probably that micelles could not be formed by many surfactants including SDS at temperatures used in ordinary hydrate-forming operations, i.e., typically less than 283 K (Di Profio et al., 2005; Watanabe et al., 2005a, 2005b; Zhang et al., 2007b, 2007c). That is, the lowest micelle-forming temperature, known as the Krafft point, for each of such surfactants is generally higher than 283 K (Watanabe et al., 2005a). Thus, we can safely claim that the promotion of hydrate formation observed in experimental systems containing SDS or its homologues has no relation to any surfactant micelles. However, the following question still remains unanswered: if micelles were actually formed in a hydrate-forming system containing a low Krafft-point surfactant, how will the hydrate formation be affected? The only previous study concerned with this point was reported by Di Profio et al. (2007). Using an electricalconductometric technique, they confirmed that three anionic surfactants, dodecylbenzene sulfonic acid (DBSA), sodium oleate (SO) and cetyltripropylammonium bromide (CTPABr), formed micelles in the aqueous phase in contact with methane at a pressure of p¼ 4 MPa and temperature of T¼275 K as far as their initial concentrations in the aqueous phase, c, were in excess of the relevant critical micelle concentrations (CMCs). Unexpectedly, Di Profio et al. (2007) found that, as for DBSA and SO, the hydrate formation in a magnetically stirred reactor was retarded when c4CMC as compared to that in the range of c oCMC. The reduction of hydrate-formation rate in the range of c 4CMC was moderate for DBSA, but significant for SO. The effect of CTPABr on the hydrate formation was so low that we could hardly discern any substantial difference between the rates of hydrate formation at c 4CMC and coCMC. Based on such observations, Di Profio et al. (2007) concluded that surfactant micelles do not promote, but inhibit the hydrate formation from methane. Considering that these observations were limited to a specific stirred system and that the mechanism of inhibition has not yet been clarified, we planned this study to observe methane-hydrate formation in an unstirred system containing a micelle-forming surfactant. Except for the selection of the surfactants to use, we performed this study based on the procedure of our previous study of methanehydrate formation in a surfactant-containing system (Okutani et al., 2008). Our observations obtained in such an unstirred system were significantly different, regarding the effect of surfactant micelles, from those reported by Di Profio et al. (2007).
dodecylbenzene sulfonic acid (DBSA), and sodium oleate (SO). DBSA and SO were two of the three surfactants used in the previous study by Di Profio et al. (2007). We selected them for the purpose of comparing the results of the experiments using the same surfactants and different setup/procedures. The selection of LDS was due to its structural similarity to SDS, the surfactant the most extensively used in previous hydrate studies, and its sufficiently low Krafft point (Smejkal et al., 2003). We expected that the effect of the counterions (Li þ or Na þ ) of the dodecyl sulfate surfactants could be realized by comparing the hydrate formation observed in the presence of LDS to that we previously observed in the presence of SDS using the same experimental apparatus and procedure (Okutani et al., 2008). The three surfactants were used as received from the suppliers. They were LDS (C12H25LiO4S) having a certified purity of 0.99 in mass fraction (supplied by Nacalai Tesque, Inc., Kyoto), SO (C18H33NaO2) having a certified purity of 0.98 in mass fraction (supplied by Nacalai Tesque, Inc., Kyoto), and DBSA (C18H30O3S) having a certified purity of 0.90 in mass fraction (supplied by Tokyo Chemical Industry Co., Ltd., Tokyo). Each of these chemicals was weighed on an electronic balance (A&D model ER-180A) with a 0.1 mg readability and dissolved in a known volume of deionized and distilled water to prepare each solution sample for the hydrate-forming experiments. The methane used in the experiments was a research-grade gas with a purity of 0.999 in mass fraction supplied by Toyoko Kagaku Co., Tokyo.
2.2. Surfactant concentrations For each of the three surfactants (LDS, DBSA and SO), we intended to vary the surfactant concentration c in the aqueous solution to be used in each hydrate-forming experiment from run to run over a range extending from the sub-CMC regime to the super-CMC regime. For this purpose, we should know, even roughly, the CMC for each surfactant under the thermodynamic condition to be adjusted in the relevant hydrate-forming experiments, i.e., the condition in which an aqueous phase and a methane-gas phase are in mutual contact at p¼ 3.9 or 4.0 MPa and T¼275 K. Di Profio et al. (2007) reported the CMC values for DBSA and SO determined by electrical conductivity measurements of the solutions in contact with methane gas at p ¼4.0 MPa and T¼275 K. We recently estimated the CMC values for LDS, DBSA and SO based on our own surface tension measurements using pendant drops suspended in a methane-gas phase adjusted at p¼3.9 MPa (for LDS) or 4.0 MPa (for DBSA and SO) and T¼275 K (Ando et al., 2012). The CMC values due to the above two sources are listed in Table 1. Unexpectedly, we find significant disagreements in the CMC values for DBSA and SO between the two sources. Although it is empirically known that surface tensiometry is apt to provide lower CMC values than electrical conductometry (Jana and Moulik, 1991; Das and Das, 2008), the differences between the two sources, particularly that for SO, seem to be too large to be simply interpreted by the above nature of surface tensiometry in comparison to electrical conductometry. This issue has not yet been clarified (Ando et al., 2012). We thus extended the c range of our hydrate-forming experiments for each surfactant such that it safely included, except for SO, the relevant CMC value whichever source we may rely on.2
2. Description of experiments 2.1. Materials We selected three commercially available anionic surfactants for use in this study. They were lithium dodecyl sulfate (LDS),
2 As for SO, the c range extended up to 100 ppm on the lower side which was not definitely lower than the relevant CMC value due to our surface-tensionmetrybased estimate, 120 7 30 ppm (Ando et al., 2012). This is because we had planned the hydrate-forming experiments with SO after the CMC value reported by Di Profio et al. (2007) and had finished the experiments before obtaining the relevant surface-tension data by ourselves.
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
81
Table 1 Critical micelle concentrations (CMCs) for the three surfactants each contained in an aqueous phase in equilibrium with a methane-gas phase at a temperature of 275 K. The pressure is 3.9 MPa for the LDS-containing system and 4.0 MPa for the other two systems. The surface tension g measured at and above the CMC for each system is indicated in the right-hand column. Sufactant
Data source
Measurement method
CMC (g/kg ¼ 103 ppm)
g (mN/m) at c ZCMC
LDS
Ando et al. (2012)
Tensiometry
3.1 70.3
36.8 70.6
DBSA
Ando et al. (2012) Di Profio et al. (2007)
Tensiometry Conductometry
0.39 70.01 0.574a
31.9 70.5 –
SO
Ando et al. (2012) Di Profio et al. (2007)
Tensiometry Conductometry
0.12 70.03 0.435b
20.27 0.3 –
a b
Converted from original CMC value given as 1.76 mM. Converted from original CMC value given as 1.43 mM.
2.3. Apparatus The experimental apparatus used in this study was just the same as that used in our previous study of methane-hydrate formation in the presence of sodium alkyl sulfate surfactants, i.e., SDS and its homologues having longer alkyl chains, and detailed elsewhere (Okutani et al., 2008). Thus, we only briefly describe this apparatus. The apparatus consists of a cylindrical stainlesssteel test chamber, 100 mm ID and 120 mm height, used as the hydrate-forming reactor and auxiliary machinery and instruments. Connected to the top lid of the test chamber is the tubing for evacuating the chamber with the aid of a vacuum pump, supplying an aqueous surfactant solution before each hydrateforming experiment, and supplying methane gas before and during the experiment. Two sheathed type-T thermocouples were inserted into the chamber through the top lid to measure the temperatures in the gas and aqueous phases. The chamber was immersed in a thermostated water bath. The gas supply line connecting a high-pressure methane cylinder to the test chamber is equipped with a pressure regulator (Tescom model 44-5266), a mass-flow meter (Brooks model 5860E or Oval model F-121), and a digital pressure gauge (Valcom model VPMC-A4-4). This allowed us to measure the instantaneous rate of gas supply into the test chamber within an uncertainty of 75 cm3/min NTP [converted to the volume flow rate at NTP (273.15 K and 101.3 kPa) condition] and the pressure inside the chamber within a 75-kPa uncertainty. 2.4. Experimental procedure As in our previous studies in this series (Watanabe et al., 2005a; Okutani et al., 2007, 2008), we operated the apparatus such that the hydrate formation during each experimental run was completed under a nearly constant pressure, i.e., within a narrow range of fluctuation about a prescribed pressure level. This was accomplished by allowing the methane gas from the high-pressure cylinder to flow through the pressure regulator to the test chamber, thereby replenishing the loss in the methane gas inside the chamber due to hydrate formation. The prescribed pressure level for the experiments using LDS was 3.9 MPa, the same as in our previous study using SDS and its longer-chain homologues, but it was raised to 4.0 MPa for the experiments using DBSA or SO to make the experimental conditions coincide with those imposed in the study by Di Profio et al. (2007) for the sake of comparing the results of both studies. Following the way of our previous studies (Watanabe et al., 2005a; Okutani et al., 2007, 2008), we focused our attention on the macroscopic behavior of hydrate formation following the nucleation which could be recognized by the visual observation, through the sight window of the test chamber, of the onset of hydrate formation and/or by the apparent start of methane flow into the chamber
responding to the methane uptake into the hydrate. The procedure of each hydrate-forming experiment as well as the preparation of the apparatus in advance of the experiment were almost the same as those in our previous study (Okutani et al., 2008), hence we only briefly describe the procedure. The chamber was first charged with a 300-cm3 surfactant solution and cooled to the target temperature of 275 K. After evacuating the chamber as well as the methane-supply tubing connected to the chamber, methane was supplied to the chamber to pressurize its interior to 3.9 MPa (when using LDS) or 4.0 MPa (when using DBSA or SO). Because the phase-equilibrium temperatures corresponding to the pressures of 3.9 MPa and 4.0 MPa are predicted by CSMGem, a phase-equilibrium calculation program (Sloan and Koh, 2008), to be 277.2 K and 277.5 K, respectively, the temperature driving force for hydrate formation was estimated to be 2.2 K for the experiments using LDS and 2.5 K for the experiments using DBSA and SO. The onset of hydrate formation was detected by both the visual observations through the sight windows of the test chamber and monitoring of V 0g , the volumetric rate (at NTP condition) of the methane inflow into the test chamber. Throughout the experiment, V 0g as well as the pressure p and temperature T inside the test chamber was continuously recorded by a data logger (Eto Denki Co., Cadac 21). Once the hydrate formation started inside the chamber, p generally tended to decrease while T tended to increase. The magnitudes of the variations in p and T during the hydrateforming period in each experimental run were 0.01 0.02 MPa and 0.4 2.7 K, respectively. The variations in p and T throughout the hydrate-forming periods for all the experimental runs were within 3.9070.05 MPa and 275.0 ( þ2.7/ 0.1) K, respectively. 2.5. Data processing The quantitative data that we directly obtained during each experimental run were those of V 0g ðtÞ, the instantaneous rate of methane supply [converted to the volume flow rate at NTP (273.15 K and 0.1013 MPa) condition]3 to the test chamber, or Vg(t), the cumulative volume (NTP) of methane supplied to the chamber after the inception of hydrate formation, where t denotes the time lapse after the inception of hydrate formation. As we pointed out earlier (Mori and Komae, 2008; Okutani et al., 2008), V 0g ðtÞ and Vg(t) do not necessarily approximate V 0g,uptake ðtÞ, the instantaneous volumetric rate (NTP) of methane uptake into the hydrate, and Vg,uptake(t), the cumulative volume (NTP) of 3 Both of the two types of thermal mass-flow meters used in this study directly indicate the measured quantities in terms of V 0g ðtÞ or Vg(t). These volumetric values can readily be converted to the mass- or mole-base values by multiplying them by rg, the mass density of methane (NTP), or by dividing them by v^ g , the molar volume of methane (NTP). According to an NIST database (Lemmon et al., 2002), rg ¼ 0.71746 kg/m3 and v^ g ¼ 22.361 m3/kmol.
82
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
methane taken into the hydrate, with sufficient accuracy because of the change in the gas-phase volume inside the chamber with the progress of the liquid-to-hydrate conversion. As detailed elsewhere (Mori and Komae, 2008), V 0g ðtÞ and Vg(t) during an isobaric hydrate-formation process may be converted to V 0g,uptake ðtÞ and Vg,uptake(t), respectively, by multiplying them by the following factor:
z
1 , ð1ðrG =rH ÞÞ þnw ðMw =M g Þ=ððrG =rL ÞðrG =rH ÞÞ
ð1Þ
where rG, rL and rH are the mass densities of the gas, liquid and hydrate phases, respectively, inside the test chamber; nw is the hydration number of the hydrate formed in the chamber; and Mg and Mw are the molar masses of the guest gas (i.e., methane in case of this study) and water, respectively. Thus, we deduced V 0g,uptake ðtÞ and Vg,uptake(t) from the V 0g ðtÞ and Vg(t) data for each experimental run using Eq. (1). For the hydrate-forming conditions set in this study, V 0g,uptake ðtÞ and Vg,uptake(t) should be higher than the corresponding V 0g ðtÞ and Vg(t) by 5 6%.
3. Results and discussion 3.1. Qualitative observations In general, the behavior of macroscopic hydrate-phase growth observed in the experiments using LDS or DBSA was almost the same as what we observed in our previous experiments using SDS, STS or SHS (Okutani et al., 2008) and also in qualitative agreement with the descriptions given by Kutergin et al. (1992), Mel’nikov et al. (1998), Zhong and Rogers (2000), Watanabe et al. (2005a), Gayet et al. (2005) and Pang et al. (2007) of their observations with gaseous guests (propane, ethane, HFC-32, and methane) and SDS. That is, in the LDS- or DBSA-containing system, hydrate crystals first formed discrete porous layers on the surface of the aqueous phase, then generated porous layers on the chamber wall above the aqueous phase. The latter layers laterally grew, while thickening at the same time, and integrated into a single riftless layer spreading over the entire wall of the chamber. Such an observation indicates that, as we already pointed out based on similar observations with SDS, STS and SHS (Okutani et al., 2008), the major mechanism of the hydrateformation enhancement due to the addition of LDS or DBSA is the capillary-driven supply of water into the porous hydrate layers growing on the chamber wall. In contrast, we observed only the formation of a thin hydrate film lying on the surface of the aqueous phase in the SO-containing system. Apparently, the addition of SO to any concentration c up to almost 4.2 times the surface-tensiometry-based CMC exhibited no appreciable change in the hydrate-forming behavior. This fact was unexpected in view of the intensive function of SO for decreasing the surface tension of the aqueous solution (see Table 1). 3.2. Quantitative evaluation of hydrate formation Fig. 1 shows the methane-uptake data obtained in the three experimental runs performed with aqueous LDS solutions of different initial concentrations. Each curve represents the time evolution of Vg,uptake, the cumulative volume (NTP) of methane taken into the hydrate formed in the test chamber, during each run after the instant t ¼0 at which the hydrate formation was first detected. In general, the Vg,uptake versus t curve initially exhibits a gentle increase, then a much steeper slope, and finally a plateau. This pattern of the Vg,uptake evolution is essentially the same as those observed in our previous experiments (Okutani et al., 2008)
Fig. 1. Time evolution of Vg,uptake, the cumulative volume (NTP) of methane uptake into the hydrate, in each of the three experimental runs performed with different LDS concentrations. The pressure p and the liquid-phase temperature T inside the test chamber were held at 3.907 0.01 MPa and 275.0 (þ2.0/ 0.1) K, respectively, throughout the experiments. The CMC was estimated based on surface tension measurements (Ando et al., 2012) to be 3100 ppm. The Vg,uptake(t) curve for no surfactant addition (c¼ 0 ppm) is not plotted here because it almost overlaps with the abscissa.
in which three sodium alkyl sulfate surfactants (SDS, STS and SHS) were separately used. Following the method used in a previous study (Okutani et al., 2008) to quantitatively evaluate the hydrate-formation rate during each experimental run, we defined the effective hydrateformation time as the duration from the inception of hydrate formation (t ¼0) to a specific moment, t ¼ tt, at which the rate of methane uptake V 0g,uptake ð ¼ dV g,uptake =dtÞ had decreased to 20% of its maximum value recorded in the same run. We also defined Vgt,uptake, the total amount of methane uptake, as the Vg,uptake value at t ¼ tt. The average rate of methane uptake during the effective hydrate-formation time, V_ g,uptake , is deduced by dividing Vgt,uptake by tt. Besides, Vgt,uptake may be normalized by dividing it by Vgt,uptake,max, the amount of methane that would be consumed if the water inside the test chamber was completely converted into the structure-I (sI) hydrate. The ratio of Vgt,uptake to Vgt,uptake,max indicates the fraction of the water that had been converted from the liquid state into the hydrate inside the test chamber during the effective hydrate-formation time, tt. For calculating Vgt,uptake,max, we estimated the hydration number for the sI methane hydrate at the pressure relevant to the experiments (3.90 MPa for the case of experiments using LDS) on the basis of CSMGem (Sloan and Koh, 2008). Figs. 2 and 3 plot the data of V_ g,uptake and Vgt,uptake/Vgt,uptake,max, respectively, versus c, the initial surfactant concentration in the aqueous phase. Here we note that neither V_ g,uptake nor Vgt,uptake/Vgt,uptake,max appreciably change with the variation in c from 2000 ppm to 5000 ppm crossing the CMC (E3100 ppm). The level of V_ g,uptake observed here is almost comparable to that observed with SDS in the c range from 500 to 3000 ppm, while the level of Vgt,uptake/Vgt,uptake,max observed here is slightly less than that for SDS in this c range (Okutani et al., 2008). The experimental results for DBSA are shown in Figs. 4–6 correspondingly to those for LDS shown in Figs. 1–3. The experiments were performed at seven different DBSA concentrations to closely examine the possible variation in the hydrate-formation rate with an increase in c from the sub-CMC to the super-CMC regimes with reference to either of the two estimates of the CMC, i.e., about 390 ppm due to the surface tensiometry (Ando et al., 2012)
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
Fig. 2. Rate of methane uptake averaged over the effective hydrate-formation time, V_ g,uptake , in each experimental run versus the LDS concentration c.
Fig. 3. Water-to-hydrate conversion ratio at the end of the effective hydrateformation time in each experimental run, Vgt,uptake/Vgt,uptake,max, versus the LDS concentration c.
Fig. 4. Time evolution of Vg,uptake, the cumulative volume (NTP) of methane uptake into the hydrate, in each of the three experimental runs performed with different DBSA concentrations. The pressure p and the liquid-phase temperature T inside the test chamber were held at 4.007 0.05 MPa and 275.0 (þ 2.7/ 0.1) K, respectively, throughout the experiments. The CMC was estimated to be 390 ppm based on the surface tension measurements (Ando et al., 2012) and 574 ppm based on the electrical conductivity measurements (Di Profio et al., 2007).
83
Fig. 5. Rate of methane uptake averaged over the effective hydrate-formation time, V_ g,uptake , in each experimental run versus the DBSA concentration c.
Fig. 6. Water-to-hydrate conversion ratio at the end of the effective hydrateformation time in each experimental run, Vgt,uptake/Vgt,uptake,max, versus the DBSA concentration c.
and 574 ppm due to the electrical conductometry (Di Profio et al., 2007). Comparing Fig. 4 with Fig. 1 in this paper as well as Fig. 3 in our previous paper (Okutani et al., 2008), we note a significant difference in the time evolution of Vgt,uptake in the DBSA-containing system from those in the LDS-, SDS-, STS- and SHS-containing systems. This difference may be realized in the following two points: (a) except for the result obtained at the lowest concentration, 200 ppm, Vgt,uptake(t) in the DBSA-containing system exhibited an almost linear increase from start, not accompanied by a slowly-increasing start-up period, and (b) the hydrate formation almost ceased within 4 h, a much shorter period compared to the hydrate-formation periods observed in the other systems. Such a difference in Vgt,uptake(t) leads to a substantially higher level of V_ g,uptake in the DBSA-containing system than in the other systems. In fact, the level of V_ g,uptake E200 cm3/min (NTP) as realized in Fig. 5 is about 45% higher than the level of V_ g,uptake obtained in the corresponding SDS-containing system (see Fig. 4 in Okutani et al., 2008). As for Vgt,uptake/Vgt,uptake,max, the DBSA-containing system slightly surpasses the LDS-containing system (Fig. 3 in this paper) and is slightly inferior to the SDS-containing system (Fig. 5 in Okutani et al., 2008). The other point of our concern, which we can clearly recognize in Figs. 4–6, is that the hydrate formation is insensitive to c and there is no substantial change in either V_ g,uptake or Vgt,uptake/Vgt,uptake,max over the entire c range (200 1146 ppm) covered by the present experiments. More specifically, we do not find in Figs. 4–6 any sign of suppression of the hydrate formation in the super-CMC regime as compared to the hydrate formation in the sub-CMC regime. This fact is somewhat inconsistent with the experimental observations by Di
84
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
surfactants—lithium dodecyl sulfate (LDS), dodecylbenzene sulfonic acid (DBSA) and sodium oleate (SO)—each of which could form, if added to an aqueous phase at a sufficiently high concentration, micelles even under the hydrate-forming thermodynamic condition ( 275 K in temperature and 3.9 4.0 MPa in pressure). The major findings obtained in this study may be summarized as follows:
Fig. 7. Time evolution of Vg,uptake, the cumulative volume (NTP) of methane uptake into the hydrate, in each of the three experimental runs performed with different SO concentrations. The pressure p and the liquid-phase temperature T inside the test chamber were held at 4.0070.02 MPa and 275.0 (þ 0.5/ 0.1) K, respectively, throughout the experiments. The CMC was estimated to be 120 ppm based on the surface tension measurements (Ando et al., 2012) and 435 ppm based on the electrical conductivity measurements (Di Profio et al., 2007).
Profio et al. (2007) for a stirred DBSA-containing system, in which the super-CMC regime was found to be inferior to the sub-CMC regime not only in delaying the hydrate nucleation (i.e., providing a longer induction time), but also in slightly decreasing the final water-to-hydrate conversion ratio. As for SO, we performed hydrate-forming experiments in the c range from 100 to 500 ppm that almost covered the two estimates of the CMC, i.e., about 120 ppm due to the surface tensiometry (Ando et al., 2012) and 435 ppm due to the electrical conductometry (Di Profio et al., 2007). All the Vg,uptake(t) records obtained in the experiments using SO are shown in Fig. 7. The variation in Vg,uptake(t) with c is erratic, and it is practically insignificant in view of the order of magnitude of Vg,uptake(t) during tr7 h. Note that even the highest value of Vg,uptake at t ¼7 h shown in Fig. 7 is less than 3% of the corresponding Vg,uptake values for the LDS- and DBSA-containing systems (cf. Figs. 1 and 4). It may be claimed from the above results that SO did not substantially promote the hydrate formation irrespective as to whether c is below or above the CMC. Again, this finding is apparently inconsistent with the experimental observation by Di Profio et al. (2007) in a stirred SO-containing system, in which the addition of SO to the concentration of c¼219 ppm (0.72 mM)4 in the sub-CMC regime significantly promoted the hydrate formation. At present, we cannot provide any physical interpretation about this inconsistency between the observations obtained in unstirred and stirred experimental systems.
4. Conclusions We have experimentally examined the effects on the hydrate formation from methane in an unstirred chamber of three anionic 4 There is a typographical error in Table 1 in the paper by Di Profio et al. (2007). The two SO concentrations, 2.86 mM and 0.72 mM, indicated in the table should be exchanged. This correction is based on a private communication (an e-mail on November 13, 2008) from Di Profio to Y.H. Mori.
(1) LDS and DBSA yield qualitatively the same change in the hydrate-formation behavior, which is in turn qualitatively the same as that we observed with the other three anionic surfactants—sodium dodecyl sulfate (SDS), sodium tetradecyl sulfate (STS) and sodium hexadecyl sulfate (SHS)—in our previous study (Okutani et al., 2008). That is, porous hydrate layers grow on the chamber wall above the level of the aqueous phase, and their growth accounts for the major portion of the total hydrate formation inside the chamber. (2) No appreciable change is detected in neither the qualitative hydrate-formation behavior nor quantitative hydrate-formation-related quantities (such as the average rate of hydrate formation, the final water-to-hydrate conversion ratio) as the result of an increase in the LDS or DBSA concentration in the aqueous phase crossing the relevant critical micelle concentration (CMC). That is, the micelle formation in the aqueous phase neither promotes nor retards the hydrate formation. (3) LDS is almost comparable to SDS as regard to its effect on the rate of hydrate formation, and is slightly inferior to SDS as regards to its effect on the final water-to-hydrate conversion ratio. (4) As regard to its effect on the rate of hydrate formation, DBSA is superior to any other surfactants tested in our previous and present studies using the same experimental apparatus/procedure, which are SDS, STS, SHS, LDS, DBSA and SO. The typical rate of hydrate formation promoted by DBSA exceeds that by SDS by about 45%. (5) The addition of SO to any concentration, either below or above the CMC, hardly promotes the hydrate formation.
Acknowledgment This study has been supported in part by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (Grant 20246040).
References Ando, N., Kuwabara, Y., Kodama, T., Mori, Y.H., 2012. Surface tensions of aqueous solutions of lithium dodecyl sulfate, sodium oleate, and dodecylbenzene sulfonic acid in contact with methane under hydrate-forming conditions. Fluid Phase Equilibria 314, 146–151. Das, C., Das, B., 2008. Effect of tetraalkylammonium salts on the micellar behavior of lithium dodecyl sulfate: a conductometric and tensiometric study. J. Mol. Liq. 137, 152–158. Di Profio, P., Arca, S., Germani, R., Savelli, G., 2005. Surfactant promoting effects on clathrate hydrate formation: are micelles really involved? Chem. Eng. Sci. 60, 4141–4145. Di Profio, P., Arca, S., Germani, R., Savelli, G., 2007. Novel nanostructured media for gas storage and transport: clathrate hydrates of methane and hydrogen. Trans. ASME—J. Fuel Cell Sci. Technol. 4, 49–55. Gayet, P., Dicharry, C., Marion, G., Graciaa, A., Lachaise, J., Nestrov, A., 2005. Experimental determination of methane hydrate dissociation curve up to 55 MPa by using a small amount of surfactant as hydrate promoter. Chem. Eng. Sci. 60, 5751–5758. Jana, P.K., Moulik, S.P., 1991. Interaction of bile salts with hexadecyltrimethylammonium bromide and sodium dodecyl sulfate. J. Phys. Chem. 95, 9525–9532. Kutergin, O.B., Mel’nikov, V.P., Nesterov, A.N., 1992. Surfactant effect on the mechanism and kinetics of gas hydrate formation. Dokl. Akad. Nauk. 323, 549–553. (in Russian).
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
Lemmon, E., McLinden, M., Huber, M., 2002. NIST Reference Fluid Thermodynamic and Transport Properties Database (REFPROP), Ver. 7.0. US Department of Commerce. Mel’nikov, V.P., Nesterov, A.N., Feklistov, V.V., 1998. Formation of gas hydrates in the presence of additives consisting of surface-active substances. Khim. Interesakh. Ustoich. Razv. 6, 97–102. (in Russian). Mori, Y.H., Komae, N., 2008. A note on the evaluation of the guest-gas uptake into a clathrate hydrate being formed in a semibatch- or batch-type reactor. Energy Convers. Manage. 49, 1056–1062. Okutani, K., Kuwabara, Y., Mori, Y.H., 2007. Surfactant effects on hydrate formation in an unstirred gas/liquid system: amendments to previous study using HFC32 and sodium dodecyl sulfate. Chem. Eng. Sci. 62, 3858–3860. Okutani, K., Kuwabara, Y., Mori, Y.H., 2008. Surfactant effects on hydrate formation in an unstirred gas/liquid system: an experimental study using methane and sodium alkyl sulfates. Chem. Eng. Sci. 63, 183–194. Pang, W.X., Chen, G.J., Dandekar, A., Sun, C.Y., Zhang, C.L., 2007. Experimental study on the scale-up of gas storage in the form of hydrate in a quiescent reactor. Chem. Eng. Sci. 62, 2198–2208. Sloan Jr., E.D., Koh, C.A., 2008. Clathrate Hydrates of Natural Gases, third ed. CRC Press, Boca Raton, FL, USA. Smejkal, G.B., Shainoff, J.R., Kottke-Marchant, K.M., 2003. Rapid high-resolution electrophoresis of multimeric von Willebrand Factor using a thermopiloted gel apparatus. Electrophoresis 24, 582–587.
85
Sun, Z.-G., Ma, R.-S., Wang, R.-Z., Guo, K.-H., Fan, S.-S., 2003a. Experimental studying of additives effects on gas storage in hydrates. Energy Fuels 17, 1180–1185. Sun, Z., Wang, R., Ma, R., Guo, K., Fan, S., 2003b. Natural gas storage in hydrates with the presence of promoters. Energy Convers. Manage. 44, 2733–2742. Sun, Z., Wang, R., Ma, R., Guo, K., Fan, S., 2003c. Effect of surfactants and liquid hydrocarbons on gas hydrate formation rate and storage capacity. Int. J. Energy Res. 27, 747–756. Watanabe, K., Imai, S., Mori, Y.H., 2005a. Surfactant effects on hydrate formation in an unstirred gas/liquid system: an experimental study using HFC-32 and sodium dodecyl sulfate. Chem. Eng. Sci. 60, 4846–4857. Watanabe, K., Niwa, S., Mori, Y.H., 2005b. Surface tensions of aqueous solutions of sodium alkyl sulfates in contact with methane under hydrate-forming conditions. J. Chem. Eng. Data 50, 1672–1676. Zhang, J.S., Lee, S.Y., Lee, J.W., 2007a. Kinetics of methane hydrate formation from SDS solution. Ind. Eng. Chem. Res. 46, 6353–6359. Zhang, J.S., Lee, S., Lee, J.W., 2007b. Does SDS micellize under methane hydrateforming conditions below the normal Kraft point? J. Colloid Interface Sci. 315, 313–318. Zhang, J.S., Lee, S.Y., Lee, J.W., 2007c. Solubility of sodium dodecyl sulfate near propane and carbon dioxide hydrate-forming conditions. J. Chem. Eng. Data 2007 (52), 2480–2483. Zhong, Y., Rogers, R.E., 2000. Surfactant effects on gas hydrate formation. Chem. Eng. Sci. 55, 4175–4187.
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Science journal homepage: www.elsevier.com/locate/ces
Surfactant effects on hydrate formation in an unstirred gas/liquid system: An experimental study using methane and micelle-forming surfactants Naoki Ando, Yui Kuwabara 1, Yasuhiko H. Mori n Department of Mechanical Engineering, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan
a r t i c l e i n f o
abstract
Article history: Received 15 November 2011 Received in revised form 16 January 2012 Accepted 18 January 2012 Available online 28 January 2012
This paper reports an experimental study in which we intended to obtain a better understanding of the possible role of surfactant micelles on the formation of a clathrate hydrate in a quiescent methane/ liquid-water system. The experiments were performed using a laboratory-scale, isobaric hydrateforming reactor, which was initially composed of a 300-cm3 aqueous phase and a 640-cm3 methanegas phase, then successively provided with methane such that the system pressure was held constant at 3.9 or 4.0 MPa. The surfactants used in this study were lithium dodecyl sulfate (LDS), dodecylbenzene sulfonic acid (DBSA), and sodium oleate (SO), which have sufficiently low Krafft points and hence, unlike sodium alkyl sulfates (sodium dodecyl sulfate and its homologues), allow the micelle formation under such typical hydrate-forming conditions as those used in this study ( 275 K in temperature and 3.9 4.0 MPa in pressure). Significant increases in the rate of hydrate formation and the final water-tohydrate conversion ratio were simultaneously observed by the addition of LDS to the aqueous phase up to concentrations in the range from 0.6 to 1.6 times the relevant critical micelle concentration (CMC). Neither the rate of hydrate formation nor the final water-to-hydrate conversion ratio exhibited any appreciable change in the above concentration range. Similar observations were obtained by the addition of DBSA to the aqueous phase up to the concentrations that ranged from 0.5 to 2.9 times the relevant CMC. Based on these observations, we have concluded that micelles of LDS and DBSA have no practical effect on hydrate formation. No substantial promotion of hydrate formation was detected by the addition of SO to the aqueous phase up to concentrations that ranged from 0.8 to 4.2 times the relevant CMC. & 2012 Elsevier Ltd. All rights reserved.
Keywords: Clathrate hydrate Gas hydrate Crystallization Solutions Surfactant Energy
1. Introduction This study is concerned with the effects of surfactant additives on the formation of a clathrate hydrate (abbreviated hydrate, hereafter) in a system containing an aqueous phase and a methane gas phase in mutual contact. Many studies, mostly experimental studies, have been reported so far about the effects of various surfactant additives on the formation of hydrates, crystalline solid compounds formed from water and various guest substances including light hydrocarbons, carbon dioxide and some fluorocarbons. The fact that the hydrate formation is substantially promoted by the addition of some types of surfactants is of potential importance for the industrial applications of such hydrates, for example, the storage and transport of natural gas or hydrogen, the separation of carbon dioxide from flue gas at coal-fired power plants, the recovery of clean water from the waste water generated at paper-making mills, and the cool
n
Corresponding author. Tel.: þ81 45 566 1522; fax: þ 81 45 566 1495. E-mail address: [email protected] (Y.H. Mori). 1 Present address: Daihatsu Motor Co. Ltd., Ikeda-shi, Osaka 563-8651, Japan.
0009-2509/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2012.01.038
storage for residential air conditioning. The point of particular interest in the hydrate formation in surfactant-containing systems is that, even in the absence of any mechanical means for mixing the aqueous and the guest-gas phases inside a hydrate-forming reactor, high-rate hydrate formation continues, generating porous hydrate layers climbing on the wall of the reactor from the level of the horizontal interface between the two fluid phases (Kutergin et al., 1992; Mel’nikov et al., 1998; Zhong and Rogers, 2000; Sun et al., 2003a, 2003b, 2003c; Watanabe et al., 2005a; Pang et al., 2007; Okutani et al., 2007, 2008). If the water (or the aqueous liquid) inside the reactor is not replenished during the hydrate-forming process while the guest gas is continuously replenished (just as in ordinary isobaric, semi-batch hydrate-forming operations), the highrate hydrate formation lasts until the aqueous phase reduces to 5 20% of its initial volume (Okutani et al., 2007, 2008). These findings indicate the possibility of developing an economical, highperformance industrial hydrate-forming technology utilizing appropriate surfactant additives. One of the fundamental and, at the same time, practically important question still left for a better understanding about the hydrate formation in surfactant-containing systems is the effect
80
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
of surfactant micelles formed in the aqueous phase on the hydrate-formation kinetics. As surveyed by, for example, Okutani et al. (2008), this issue was once a subject of controversy. Although the controversy is apparently over at present, the issue is still not yet completely settled, which is briefly discussed below. Zhong and Rogers (2000) were probably the first to raise this issue. For interpreting their experimental results about the effect of ethane hydrate formation in the presence of an anionic surfactant, sodium dodecyl sulfate (SDS), they suggested that SDS micelles once formed would work as the carriers of ethane molecules to the bulk of the aqueous phase and thereby promoting the hydrate formation. However, this hypothesis has been disputed and denied by several research groups on the basis of different arguments (Di Profio et al., 2005, 2007; Watanabe et al., 2005a, 2005b; Gayet et al., 2005; Pang et al., 2007; Zhang et al., 2007a, 2007b, 2007c). The most straightforward argument against this hypothesis was probably that micelles could not be formed by many surfactants including SDS at temperatures used in ordinary hydrate-forming operations, i.e., typically less than 283 K (Di Profio et al., 2005; Watanabe et al., 2005a, 2005b; Zhang et al., 2007b, 2007c). That is, the lowest micelle-forming temperature, known as the Krafft point, for each of such surfactants is generally higher than 283 K (Watanabe et al., 2005a). Thus, we can safely claim that the promotion of hydrate formation observed in experimental systems containing SDS or its homologues has no relation to any surfactant micelles. However, the following question still remains unanswered: if micelles were actually formed in a hydrate-forming system containing a low Krafft-point surfactant, how will the hydrate formation be affected? The only previous study concerned with this point was reported by Di Profio et al. (2007). Using an electricalconductometric technique, they confirmed that three anionic surfactants, dodecylbenzene sulfonic acid (DBSA), sodium oleate (SO) and cetyltripropylammonium bromide (CTPABr), formed micelles in the aqueous phase in contact with methane at a pressure of p¼ 4 MPa and temperature of T¼275 K as far as their initial concentrations in the aqueous phase, c, were in excess of the relevant critical micelle concentrations (CMCs). Unexpectedly, Di Profio et al. (2007) found that, as for DBSA and SO, the hydrate formation in a magnetically stirred reactor was retarded when c4CMC as compared to that in the range of c oCMC. The reduction of hydrate-formation rate in the range of c 4CMC was moderate for DBSA, but significant for SO. The effect of CTPABr on the hydrate formation was so low that we could hardly discern any substantial difference between the rates of hydrate formation at c 4CMC and coCMC. Based on such observations, Di Profio et al. (2007) concluded that surfactant micelles do not promote, but inhibit the hydrate formation from methane. Considering that these observations were limited to a specific stirred system and that the mechanism of inhibition has not yet been clarified, we planned this study to observe methane-hydrate formation in an unstirred system containing a micelle-forming surfactant. Except for the selection of the surfactants to use, we performed this study based on the procedure of our previous study of methanehydrate formation in a surfactant-containing system (Okutani et al., 2008). Our observations obtained in such an unstirred system were significantly different, regarding the effect of surfactant micelles, from those reported by Di Profio et al. (2007).
dodecylbenzene sulfonic acid (DBSA), and sodium oleate (SO). DBSA and SO were two of the three surfactants used in the previous study by Di Profio et al. (2007). We selected them for the purpose of comparing the results of the experiments using the same surfactants and different setup/procedures. The selection of LDS was due to its structural similarity to SDS, the surfactant the most extensively used in previous hydrate studies, and its sufficiently low Krafft point (Smejkal et al., 2003). We expected that the effect of the counterions (Li þ or Na þ ) of the dodecyl sulfate surfactants could be realized by comparing the hydrate formation observed in the presence of LDS to that we previously observed in the presence of SDS using the same experimental apparatus and procedure (Okutani et al., 2008). The three surfactants were used as received from the suppliers. They were LDS (C12H25LiO4S) having a certified purity of 0.99 in mass fraction (supplied by Nacalai Tesque, Inc., Kyoto), SO (C18H33NaO2) having a certified purity of 0.98 in mass fraction (supplied by Nacalai Tesque, Inc., Kyoto), and DBSA (C18H30O3S) having a certified purity of 0.90 in mass fraction (supplied by Tokyo Chemical Industry Co., Ltd., Tokyo). Each of these chemicals was weighed on an electronic balance (A&D model ER-180A) with a 0.1 mg readability and dissolved in a known volume of deionized and distilled water to prepare each solution sample for the hydrate-forming experiments. The methane used in the experiments was a research-grade gas with a purity of 0.999 in mass fraction supplied by Toyoko Kagaku Co., Tokyo.
2.2. Surfactant concentrations For each of the three surfactants (LDS, DBSA and SO), we intended to vary the surfactant concentration c in the aqueous solution to be used in each hydrate-forming experiment from run to run over a range extending from the sub-CMC regime to the super-CMC regime. For this purpose, we should know, even roughly, the CMC for each surfactant under the thermodynamic condition to be adjusted in the relevant hydrate-forming experiments, i.e., the condition in which an aqueous phase and a methane-gas phase are in mutual contact at p¼ 3.9 or 4.0 MPa and T¼275 K. Di Profio et al. (2007) reported the CMC values for DBSA and SO determined by electrical conductivity measurements of the solutions in contact with methane gas at p ¼4.0 MPa and T¼275 K. We recently estimated the CMC values for LDS, DBSA and SO based on our own surface tension measurements using pendant drops suspended in a methane-gas phase adjusted at p¼3.9 MPa (for LDS) or 4.0 MPa (for DBSA and SO) and T¼275 K (Ando et al., 2012). The CMC values due to the above two sources are listed in Table 1. Unexpectedly, we find significant disagreements in the CMC values for DBSA and SO between the two sources. Although it is empirically known that surface tensiometry is apt to provide lower CMC values than electrical conductometry (Jana and Moulik, 1991; Das and Das, 2008), the differences between the two sources, particularly that for SO, seem to be too large to be simply interpreted by the above nature of surface tensiometry in comparison to electrical conductometry. This issue has not yet been clarified (Ando et al., 2012). We thus extended the c range of our hydrate-forming experiments for each surfactant such that it safely included, except for SO, the relevant CMC value whichever source we may rely on.2
2. Description of experiments 2.1. Materials We selected three commercially available anionic surfactants for use in this study. They were lithium dodecyl sulfate (LDS),
2 As for SO, the c range extended up to 100 ppm on the lower side which was not definitely lower than the relevant CMC value due to our surface-tensionmetrybased estimate, 120 7 30 ppm (Ando et al., 2012). This is because we had planned the hydrate-forming experiments with SO after the CMC value reported by Di Profio et al. (2007) and had finished the experiments before obtaining the relevant surface-tension data by ourselves.
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
81
Table 1 Critical micelle concentrations (CMCs) for the three surfactants each contained in an aqueous phase in equilibrium with a methane-gas phase at a temperature of 275 K. The pressure is 3.9 MPa for the LDS-containing system and 4.0 MPa for the other two systems. The surface tension g measured at and above the CMC for each system is indicated in the right-hand column. Sufactant
Data source
Measurement method
CMC (g/kg ¼ 103 ppm)
g (mN/m) at c ZCMC
LDS
Ando et al. (2012)
Tensiometry
3.1 70.3
36.8 70.6
DBSA
Ando et al. (2012) Di Profio et al. (2007)
Tensiometry Conductometry
0.39 70.01 0.574a
31.9 70.5 –
SO
Ando et al. (2012) Di Profio et al. (2007)
Tensiometry Conductometry
0.12 70.03 0.435b
20.27 0.3 –
a b
Converted from original CMC value given as 1.76 mM. Converted from original CMC value given as 1.43 mM.
2.3. Apparatus The experimental apparatus used in this study was just the same as that used in our previous study of methane-hydrate formation in the presence of sodium alkyl sulfate surfactants, i.e., SDS and its homologues having longer alkyl chains, and detailed elsewhere (Okutani et al., 2008). Thus, we only briefly describe this apparatus. The apparatus consists of a cylindrical stainlesssteel test chamber, 100 mm ID and 120 mm height, used as the hydrate-forming reactor and auxiliary machinery and instruments. Connected to the top lid of the test chamber is the tubing for evacuating the chamber with the aid of a vacuum pump, supplying an aqueous surfactant solution before each hydrateforming experiment, and supplying methane gas before and during the experiment. Two sheathed type-T thermocouples were inserted into the chamber through the top lid to measure the temperatures in the gas and aqueous phases. The chamber was immersed in a thermostated water bath. The gas supply line connecting a high-pressure methane cylinder to the test chamber is equipped with a pressure regulator (Tescom model 44-5266), a mass-flow meter (Brooks model 5860E or Oval model F-121), and a digital pressure gauge (Valcom model VPMC-A4-4). This allowed us to measure the instantaneous rate of gas supply into the test chamber within an uncertainty of 75 cm3/min NTP [converted to the volume flow rate at NTP (273.15 K and 101.3 kPa) condition] and the pressure inside the chamber within a 75-kPa uncertainty. 2.4. Experimental procedure As in our previous studies in this series (Watanabe et al., 2005a; Okutani et al., 2007, 2008), we operated the apparatus such that the hydrate formation during each experimental run was completed under a nearly constant pressure, i.e., within a narrow range of fluctuation about a prescribed pressure level. This was accomplished by allowing the methane gas from the high-pressure cylinder to flow through the pressure regulator to the test chamber, thereby replenishing the loss in the methane gas inside the chamber due to hydrate formation. The prescribed pressure level for the experiments using LDS was 3.9 MPa, the same as in our previous study using SDS and its longer-chain homologues, but it was raised to 4.0 MPa for the experiments using DBSA or SO to make the experimental conditions coincide with those imposed in the study by Di Profio et al. (2007) for the sake of comparing the results of both studies. Following the way of our previous studies (Watanabe et al., 2005a; Okutani et al., 2007, 2008), we focused our attention on the macroscopic behavior of hydrate formation following the nucleation which could be recognized by the visual observation, through the sight window of the test chamber, of the onset of hydrate formation and/or by the apparent start of methane flow into the chamber
responding to the methane uptake into the hydrate. The procedure of each hydrate-forming experiment as well as the preparation of the apparatus in advance of the experiment were almost the same as those in our previous study (Okutani et al., 2008), hence we only briefly describe the procedure. The chamber was first charged with a 300-cm3 surfactant solution and cooled to the target temperature of 275 K. After evacuating the chamber as well as the methane-supply tubing connected to the chamber, methane was supplied to the chamber to pressurize its interior to 3.9 MPa (when using LDS) or 4.0 MPa (when using DBSA or SO). Because the phase-equilibrium temperatures corresponding to the pressures of 3.9 MPa and 4.0 MPa are predicted by CSMGem, a phase-equilibrium calculation program (Sloan and Koh, 2008), to be 277.2 K and 277.5 K, respectively, the temperature driving force for hydrate formation was estimated to be 2.2 K for the experiments using LDS and 2.5 K for the experiments using DBSA and SO. The onset of hydrate formation was detected by both the visual observations through the sight windows of the test chamber and monitoring of V 0g , the volumetric rate (at NTP condition) of the methane inflow into the test chamber. Throughout the experiment, V 0g as well as the pressure p and temperature T inside the test chamber was continuously recorded by a data logger (Eto Denki Co., Cadac 21). Once the hydrate formation started inside the chamber, p generally tended to decrease while T tended to increase. The magnitudes of the variations in p and T during the hydrateforming period in each experimental run were 0.01 0.02 MPa and 0.4 2.7 K, respectively. The variations in p and T throughout the hydrate-forming periods for all the experimental runs were within 3.9070.05 MPa and 275.0 ( þ2.7/ 0.1) K, respectively. 2.5. Data processing The quantitative data that we directly obtained during each experimental run were those of V 0g ðtÞ, the instantaneous rate of methane supply [converted to the volume flow rate at NTP (273.15 K and 0.1013 MPa) condition]3 to the test chamber, or Vg(t), the cumulative volume (NTP) of methane supplied to the chamber after the inception of hydrate formation, where t denotes the time lapse after the inception of hydrate formation. As we pointed out earlier (Mori and Komae, 2008; Okutani et al., 2008), V 0g ðtÞ and Vg(t) do not necessarily approximate V 0g,uptake ðtÞ, the instantaneous volumetric rate (NTP) of methane uptake into the hydrate, and Vg,uptake(t), the cumulative volume (NTP) of 3 Both of the two types of thermal mass-flow meters used in this study directly indicate the measured quantities in terms of V 0g ðtÞ or Vg(t). These volumetric values can readily be converted to the mass- or mole-base values by multiplying them by rg, the mass density of methane (NTP), or by dividing them by v^ g , the molar volume of methane (NTP). According to an NIST database (Lemmon et al., 2002), rg ¼ 0.71746 kg/m3 and v^ g ¼ 22.361 m3/kmol.
82
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
methane taken into the hydrate, with sufficient accuracy because of the change in the gas-phase volume inside the chamber with the progress of the liquid-to-hydrate conversion. As detailed elsewhere (Mori and Komae, 2008), V 0g ðtÞ and Vg(t) during an isobaric hydrate-formation process may be converted to V 0g,uptake ðtÞ and Vg,uptake(t), respectively, by multiplying them by the following factor:
z
1 , ð1ðrG =rH ÞÞ þnw ðMw =M g Þ=ððrG =rL ÞðrG =rH ÞÞ
ð1Þ
where rG, rL and rH are the mass densities of the gas, liquid and hydrate phases, respectively, inside the test chamber; nw is the hydration number of the hydrate formed in the chamber; and Mg and Mw are the molar masses of the guest gas (i.e., methane in case of this study) and water, respectively. Thus, we deduced V 0g,uptake ðtÞ and Vg,uptake(t) from the V 0g ðtÞ and Vg(t) data for each experimental run using Eq. (1). For the hydrate-forming conditions set in this study, V 0g,uptake ðtÞ and Vg,uptake(t) should be higher than the corresponding V 0g ðtÞ and Vg(t) by 5 6%.
3. Results and discussion 3.1. Qualitative observations In general, the behavior of macroscopic hydrate-phase growth observed in the experiments using LDS or DBSA was almost the same as what we observed in our previous experiments using SDS, STS or SHS (Okutani et al., 2008) and also in qualitative agreement with the descriptions given by Kutergin et al. (1992), Mel’nikov et al. (1998), Zhong and Rogers (2000), Watanabe et al. (2005a), Gayet et al. (2005) and Pang et al. (2007) of their observations with gaseous guests (propane, ethane, HFC-32, and methane) and SDS. That is, in the LDS- or DBSA-containing system, hydrate crystals first formed discrete porous layers on the surface of the aqueous phase, then generated porous layers on the chamber wall above the aqueous phase. The latter layers laterally grew, while thickening at the same time, and integrated into a single riftless layer spreading over the entire wall of the chamber. Such an observation indicates that, as we already pointed out based on similar observations with SDS, STS and SHS (Okutani et al., 2008), the major mechanism of the hydrateformation enhancement due to the addition of LDS or DBSA is the capillary-driven supply of water into the porous hydrate layers growing on the chamber wall. In contrast, we observed only the formation of a thin hydrate film lying on the surface of the aqueous phase in the SO-containing system. Apparently, the addition of SO to any concentration c up to almost 4.2 times the surface-tensiometry-based CMC exhibited no appreciable change in the hydrate-forming behavior. This fact was unexpected in view of the intensive function of SO for decreasing the surface tension of the aqueous solution (see Table 1). 3.2. Quantitative evaluation of hydrate formation Fig. 1 shows the methane-uptake data obtained in the three experimental runs performed with aqueous LDS solutions of different initial concentrations. Each curve represents the time evolution of Vg,uptake, the cumulative volume (NTP) of methane taken into the hydrate formed in the test chamber, during each run after the instant t ¼0 at which the hydrate formation was first detected. In general, the Vg,uptake versus t curve initially exhibits a gentle increase, then a much steeper slope, and finally a plateau. This pattern of the Vg,uptake evolution is essentially the same as those observed in our previous experiments (Okutani et al., 2008)
Fig. 1. Time evolution of Vg,uptake, the cumulative volume (NTP) of methane uptake into the hydrate, in each of the three experimental runs performed with different LDS concentrations. The pressure p and the liquid-phase temperature T inside the test chamber were held at 3.907 0.01 MPa and 275.0 (þ2.0/ 0.1) K, respectively, throughout the experiments. The CMC was estimated based on surface tension measurements (Ando et al., 2012) to be 3100 ppm. The Vg,uptake(t) curve for no surfactant addition (c¼ 0 ppm) is not plotted here because it almost overlaps with the abscissa.
in which three sodium alkyl sulfate surfactants (SDS, STS and SHS) were separately used. Following the method used in a previous study (Okutani et al., 2008) to quantitatively evaluate the hydrate-formation rate during each experimental run, we defined the effective hydrateformation time as the duration from the inception of hydrate formation (t ¼0) to a specific moment, t ¼ tt, at which the rate of methane uptake V 0g,uptake ð ¼ dV g,uptake =dtÞ had decreased to 20% of its maximum value recorded in the same run. We also defined Vgt,uptake, the total amount of methane uptake, as the Vg,uptake value at t ¼ tt. The average rate of methane uptake during the effective hydrate-formation time, V_ g,uptake , is deduced by dividing Vgt,uptake by tt. Besides, Vgt,uptake may be normalized by dividing it by Vgt,uptake,max, the amount of methane that would be consumed if the water inside the test chamber was completely converted into the structure-I (sI) hydrate. The ratio of Vgt,uptake to Vgt,uptake,max indicates the fraction of the water that had been converted from the liquid state into the hydrate inside the test chamber during the effective hydrate-formation time, tt. For calculating Vgt,uptake,max, we estimated the hydration number for the sI methane hydrate at the pressure relevant to the experiments (3.90 MPa for the case of experiments using LDS) on the basis of CSMGem (Sloan and Koh, 2008). Figs. 2 and 3 plot the data of V_ g,uptake and Vgt,uptake/Vgt,uptake,max, respectively, versus c, the initial surfactant concentration in the aqueous phase. Here we note that neither V_ g,uptake nor Vgt,uptake/Vgt,uptake,max appreciably change with the variation in c from 2000 ppm to 5000 ppm crossing the CMC (E3100 ppm). The level of V_ g,uptake observed here is almost comparable to that observed with SDS in the c range from 500 to 3000 ppm, while the level of Vgt,uptake/Vgt,uptake,max observed here is slightly less than that for SDS in this c range (Okutani et al., 2008). The experimental results for DBSA are shown in Figs. 4–6 correspondingly to those for LDS shown in Figs. 1–3. The experiments were performed at seven different DBSA concentrations to closely examine the possible variation in the hydrate-formation rate with an increase in c from the sub-CMC to the super-CMC regimes with reference to either of the two estimates of the CMC, i.e., about 390 ppm due to the surface tensiometry (Ando et al., 2012)
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
Fig. 2. Rate of methane uptake averaged over the effective hydrate-formation time, V_ g,uptake , in each experimental run versus the LDS concentration c.
Fig. 3. Water-to-hydrate conversion ratio at the end of the effective hydrateformation time in each experimental run, Vgt,uptake/Vgt,uptake,max, versus the LDS concentration c.
Fig. 4. Time evolution of Vg,uptake, the cumulative volume (NTP) of methane uptake into the hydrate, in each of the three experimental runs performed with different DBSA concentrations. The pressure p and the liquid-phase temperature T inside the test chamber were held at 4.007 0.05 MPa and 275.0 (þ 2.7/ 0.1) K, respectively, throughout the experiments. The CMC was estimated to be 390 ppm based on the surface tension measurements (Ando et al., 2012) and 574 ppm based on the electrical conductivity measurements (Di Profio et al., 2007).
83
Fig. 5. Rate of methane uptake averaged over the effective hydrate-formation time, V_ g,uptake , in each experimental run versus the DBSA concentration c.
Fig. 6. Water-to-hydrate conversion ratio at the end of the effective hydrateformation time in each experimental run, Vgt,uptake/Vgt,uptake,max, versus the DBSA concentration c.
and 574 ppm due to the electrical conductometry (Di Profio et al., 2007). Comparing Fig. 4 with Fig. 1 in this paper as well as Fig. 3 in our previous paper (Okutani et al., 2008), we note a significant difference in the time evolution of Vgt,uptake in the DBSA-containing system from those in the LDS-, SDS-, STS- and SHS-containing systems. This difference may be realized in the following two points: (a) except for the result obtained at the lowest concentration, 200 ppm, Vgt,uptake(t) in the DBSA-containing system exhibited an almost linear increase from start, not accompanied by a slowly-increasing start-up period, and (b) the hydrate formation almost ceased within 4 h, a much shorter period compared to the hydrate-formation periods observed in the other systems. Such a difference in Vgt,uptake(t) leads to a substantially higher level of V_ g,uptake in the DBSA-containing system than in the other systems. In fact, the level of V_ g,uptake E200 cm3/min (NTP) as realized in Fig. 5 is about 45% higher than the level of V_ g,uptake obtained in the corresponding SDS-containing system (see Fig. 4 in Okutani et al., 2008). As for Vgt,uptake/Vgt,uptake,max, the DBSA-containing system slightly surpasses the LDS-containing system (Fig. 3 in this paper) and is slightly inferior to the SDS-containing system (Fig. 5 in Okutani et al., 2008). The other point of our concern, which we can clearly recognize in Figs. 4–6, is that the hydrate formation is insensitive to c and there is no substantial change in either V_ g,uptake or Vgt,uptake/Vgt,uptake,max over the entire c range (200 1146 ppm) covered by the present experiments. More specifically, we do not find in Figs. 4–6 any sign of suppression of the hydrate formation in the super-CMC regime as compared to the hydrate formation in the sub-CMC regime. This fact is somewhat inconsistent with the experimental observations by Di
84
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
surfactants—lithium dodecyl sulfate (LDS), dodecylbenzene sulfonic acid (DBSA) and sodium oleate (SO)—each of which could form, if added to an aqueous phase at a sufficiently high concentration, micelles even under the hydrate-forming thermodynamic condition ( 275 K in temperature and 3.9 4.0 MPa in pressure). The major findings obtained in this study may be summarized as follows:
Fig. 7. Time evolution of Vg,uptake, the cumulative volume (NTP) of methane uptake into the hydrate, in each of the three experimental runs performed with different SO concentrations. The pressure p and the liquid-phase temperature T inside the test chamber were held at 4.0070.02 MPa and 275.0 (þ 0.5/ 0.1) K, respectively, throughout the experiments. The CMC was estimated to be 120 ppm based on the surface tension measurements (Ando et al., 2012) and 435 ppm based on the electrical conductivity measurements (Di Profio et al., 2007).
Profio et al. (2007) for a stirred DBSA-containing system, in which the super-CMC regime was found to be inferior to the sub-CMC regime not only in delaying the hydrate nucleation (i.e., providing a longer induction time), but also in slightly decreasing the final water-to-hydrate conversion ratio. As for SO, we performed hydrate-forming experiments in the c range from 100 to 500 ppm that almost covered the two estimates of the CMC, i.e., about 120 ppm due to the surface tensiometry (Ando et al., 2012) and 435 ppm due to the electrical conductometry (Di Profio et al., 2007). All the Vg,uptake(t) records obtained in the experiments using SO are shown in Fig. 7. The variation in Vg,uptake(t) with c is erratic, and it is practically insignificant in view of the order of magnitude of Vg,uptake(t) during tr7 h. Note that even the highest value of Vg,uptake at t ¼7 h shown in Fig. 7 is less than 3% of the corresponding Vg,uptake values for the LDS- and DBSA-containing systems (cf. Figs. 1 and 4). It may be claimed from the above results that SO did not substantially promote the hydrate formation irrespective as to whether c is below or above the CMC. Again, this finding is apparently inconsistent with the experimental observation by Di Profio et al. (2007) in a stirred SO-containing system, in which the addition of SO to the concentration of c¼219 ppm (0.72 mM)4 in the sub-CMC regime significantly promoted the hydrate formation. At present, we cannot provide any physical interpretation about this inconsistency between the observations obtained in unstirred and stirred experimental systems.
4. Conclusions We have experimentally examined the effects on the hydrate formation from methane in an unstirred chamber of three anionic 4 There is a typographical error in Table 1 in the paper by Di Profio et al. (2007). The two SO concentrations, 2.86 mM and 0.72 mM, indicated in the table should be exchanged. This correction is based on a private communication (an e-mail on November 13, 2008) from Di Profio to Y.H. Mori.
(1) LDS and DBSA yield qualitatively the same change in the hydrate-formation behavior, which is in turn qualitatively the same as that we observed with the other three anionic surfactants—sodium dodecyl sulfate (SDS), sodium tetradecyl sulfate (STS) and sodium hexadecyl sulfate (SHS)—in our previous study (Okutani et al., 2008). That is, porous hydrate layers grow on the chamber wall above the level of the aqueous phase, and their growth accounts for the major portion of the total hydrate formation inside the chamber. (2) No appreciable change is detected in neither the qualitative hydrate-formation behavior nor quantitative hydrate-formation-related quantities (such as the average rate of hydrate formation, the final water-to-hydrate conversion ratio) as the result of an increase in the LDS or DBSA concentration in the aqueous phase crossing the relevant critical micelle concentration (CMC). That is, the micelle formation in the aqueous phase neither promotes nor retards the hydrate formation. (3) LDS is almost comparable to SDS as regard to its effect on the rate of hydrate formation, and is slightly inferior to SDS as regards to its effect on the final water-to-hydrate conversion ratio. (4) As regard to its effect on the rate of hydrate formation, DBSA is superior to any other surfactants tested in our previous and present studies using the same experimental apparatus/procedure, which are SDS, STS, SHS, LDS, DBSA and SO. The typical rate of hydrate formation promoted by DBSA exceeds that by SDS by about 45%. (5) The addition of SO to any concentration, either below or above the CMC, hardly promotes the hydrate formation.
Acknowledgment This study has been supported in part by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (Grant 20246040).
References Ando, N., Kuwabara, Y., Kodama, T., Mori, Y.H., 2012. Surface tensions of aqueous solutions of lithium dodecyl sulfate, sodium oleate, and dodecylbenzene sulfonic acid in contact with methane under hydrate-forming conditions. Fluid Phase Equilibria 314, 146–151. Das, C., Das, B., 2008. Effect of tetraalkylammonium salts on the micellar behavior of lithium dodecyl sulfate: a conductometric and tensiometric study. J. Mol. Liq. 137, 152–158. Di Profio, P., Arca, S., Germani, R., Savelli, G., 2005. Surfactant promoting effects on clathrate hydrate formation: are micelles really involved? Chem. Eng. Sci. 60, 4141–4145. Di Profio, P., Arca, S., Germani, R., Savelli, G., 2007. Novel nanostructured media for gas storage and transport: clathrate hydrates of methane and hydrogen. Trans. ASME—J. Fuel Cell Sci. Technol. 4, 49–55. Gayet, P., Dicharry, C., Marion, G., Graciaa, A., Lachaise, J., Nestrov, A., 2005. Experimental determination of methane hydrate dissociation curve up to 55 MPa by using a small amount of surfactant as hydrate promoter. Chem. Eng. Sci. 60, 5751–5758. Jana, P.K., Moulik, S.P., 1991. Interaction of bile salts with hexadecyltrimethylammonium bromide and sodium dodecyl sulfate. J. Phys. Chem. 95, 9525–9532. Kutergin, O.B., Mel’nikov, V.P., Nesterov, A.N., 1992. Surfactant effect on the mechanism and kinetics of gas hydrate formation. Dokl. Akad. Nauk. 323, 549–553. (in Russian).
N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85
Lemmon, E., McLinden, M., Huber, M., 2002. NIST Reference Fluid Thermodynamic and Transport Properties Database (REFPROP), Ver. 7.0. US Department of Commerce. Mel’nikov, V.P., Nesterov, A.N., Feklistov, V.V., 1998. Formation of gas hydrates in the presence of additives consisting of surface-active substances. Khim. Interesakh. Ustoich. Razv. 6, 97–102. (in Russian). Mori, Y.H., Komae, N., 2008. A note on the evaluation of the guest-gas uptake into a clathrate hydrate being formed in a semibatch- or batch-type reactor. Energy Convers. Manage. 49, 1056–1062. Okutani, K., Kuwabara, Y., Mori, Y.H., 2007. Surfactant effects on hydrate formation in an unstirred gas/liquid system: amendments to previous study using HFC32 and sodium dodecyl sulfate. Chem. Eng. Sci. 62, 3858–3860. Okutani, K., Kuwabara, Y., Mori, Y.H., 2008. Surfactant effects on hydrate formation in an unstirred gas/liquid system: an experimental study using methane and sodium alkyl sulfates. Chem. Eng. Sci. 63, 183–194. Pang, W.X., Chen, G.J., Dandekar, A., Sun, C.Y., Zhang, C.L., 2007. Experimental study on the scale-up of gas storage in the form of hydrate in a quiescent reactor. Chem. Eng. Sci. 62, 2198–2208. Sloan Jr., E.D., Koh, C.A., 2008. Clathrate Hydrates of Natural Gases, third ed. CRC Press, Boca Raton, FL, USA. Smejkal, G.B., Shainoff, J.R., Kottke-Marchant, K.M., 2003. Rapid high-resolution electrophoresis of multimeric von Willebrand Factor using a thermopiloted gel apparatus. Electrophoresis 24, 582–587.
85
Sun, Z.-G., Ma, R.-S., Wang, R.-Z., Guo, K.-H., Fan, S.-S., 2003a. Experimental studying of additives effects on gas storage in hydrates. Energy Fuels 17, 1180–1185. Sun, Z., Wang, R., Ma, R., Guo, K., Fan, S., 2003b. Natural gas storage in hydrates with the presence of promoters. Energy Convers. Manage. 44, 2733–2742. Sun, Z., Wang, R., Ma, R., Guo, K., Fan, S., 2003c. Effect of surfactants and liquid hydrocarbons on gas hydrate formation rate and storage capacity. Int. J. Energy Res. 27, 747–756. Watanabe, K., Imai, S., Mori, Y.H., 2005a. Surfactant effects on hydrate formation in an unstirred gas/liquid system: an experimental study using HFC-32 and sodium dodecyl sulfate. Chem. Eng. Sci. 60, 4846–4857. Watanabe, K., Niwa, S., Mori, Y.H., 2005b. Surface tensions of aqueous solutions of sodium alkyl sulfates in contact with methane under hydrate-forming conditions. J. Chem. Eng. Data 50, 1672–1676. Zhang, J.S., Lee, S.Y., Lee, J.W., 2007a. Kinetics of methane hydrate formation from SDS solution. Ind. Eng. Chem. Res. 46, 6353–6359. Zhang, J.S., Lee, S., Lee, J.W., 2007b. Does SDS micellize under methane hydrateforming conditions below the normal Kraft point? J. Colloid Interface Sci. 315, 313–318. Zhang, J.S., Lee, S.Y., Lee, J.W., 2007c. Solubility of sodium dodecyl sulfate near propane and carbon dioxide hydrate-forming conditions. J. Chem. Eng. Data 2007 (52), 2480–2483. Zhong, Y., Rogers, R.E., 2000. Surfactant effects on gas hydrate formation. Chem. Eng. Sci. 55, 4175–4187.