Chemical Engineering Science 73 (2012) 79–85

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Science

0009-25

doi:10.1

n Corr

E-m1 Pr

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

Article history:

Received 15 November 2011

Received in revised form

16 January 2012

Accepted 18 January 2012Available online 28 January 2012

Keywords:

Clathrate hydrate

Gas hydrate

Crystallization

Solutions

Surfactant

Energy

09/$ - see front matter & 2012 Elsevier Ltd. A

016/j.ces.2012.01.038

esponding author. Tel.: þ81 45 566 1522; fax

ail address: yhmori@mech.keio.ac.jp (Y.H. Mo

esent address: Daihatsu Motor Co. Ltd., Ikeda

a b s t r a c t

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 hydrate-

forming reactor, which was initially composed of a 300-cm3 aqueous phase and a �640-cm3 methane-

gas 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-to-

hydrate 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.

1. Introduction

This study is concerned with the effects of surfactant additives onthe formation of a clathrate hydrate (abbreviated hydrate, hereafter)in a system containing an aqueous phase and a methane gas phasein mutual contact. Many studies, mostly experimental studies, havebeen reported so far about the effects of various surfactant additiveson the formation of hydrates, crystalline solid compounds formedfrom water and various guest substances including light hydrocar-bons, carbon dioxide and some fluorocarbons. The fact that thehydrate formation is substantially promoted by the addition of sometypes of surfactants is of potential importance for the industrialapplications of such hydrates, for example, the storage and transportof natural gas or hydrogen, the separation of carbon dioxide fromflue gas at coal-fired power plants, the recovery of clean water fromthe waste water generated at paper-making mills, and the cool

ll rights reserved.

: þ81 45 566 1495.

ri).

-shi, Osaka 563-8651, Japan.

storage for residential air conditioning. The point of particularinterest in the hydrate formation in surfactant-containing systemsis that, even in the absence of any mechanical means for mixing theaqueous and the guest-gas phases inside a hydrate-forming reactor,high-rate hydrate formation continues, generating porous hydratelayers climbing on the wall of the reactor from the level of thehorizontal 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-formingprocess while the guest gas is continuously replenished (just as inordinary isobaric, semi-batch hydrate-forming operations), the high-rate hydrate formation lasts until the aqueous phase reduces to�5�20% of its initial volume (Okutani et al., 2007, 2008). Thesefindings indicate the possibility of developing an economical, high-performance industrial hydrate-forming technology utilizing appro-priate surfactant additives.

One of the fundamental and, at the same time, practicallyimportant question still left for a better understanding about thehydrate formation in surfactant-containing systems is the effect

N. Ando et al. / Chemical Engineering Science 73 (2012) 79–8580

of surfactant micelles formed in the aqueous phase on thehydrate-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 issueis still not yet completely settled, which is briefly discussedbelow. Zhong and Rogers (2000) were probably the first to raisethis issue. For interpreting their experimental results about theeffect of ethane hydrate formation in the presence of an anionicsurfactant, sodium dodecyl sulfate (SDS), they suggested that SDSmicelles once formed would work as the carriers of ethanemolecules to the bulk of the aqueous phase and thereby promot-ing the hydrate formation. However, this hypothesis has beendisputed and denied by several research groups on the basis ofdifferent 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 againstthis hypothesis was probably that micelles could not be formedby many surfactants including SDS at temperatures used inordinary hydrate-forming operations, i.e., typically less than283 K (Di Profio et al., 2005; Watanabe et al., 2005a, 2005b;Zhang et al., 2007b, 2007c). That is, the lowest micelle-formingtemperature, known as the Krafft point, for each of such surfac-tants is generally higher than 283 K (Watanabe et al., 2005a).Thus, we can safely claim that the promotion of hydrate forma-tion observed in experimental systems containing SDS or itshomologues has no relation to any surfactant micelles. However,the following question still remains unanswered: if micelles wereactually formed in a hydrate-forming system containing a lowKrafft-point surfactant, how will the hydrate formation beaffected? The only previous study concerned with this pointwas reported by Di Profio et al. (2007). Using an electrical-conductometric technique, they confirmed that three anionicsurfactants, dodecylbenzene sulfonic acid (DBSA), sodium oleate(SO) and cetyltripropylammonium bromide (CTPABr), formedmicelles in the aqueous phase in contact with methane at apressure of p¼4 MPa and temperature of T¼275 K as far as theirinitial concentrations in the aqueous phase, c, were in excess ofthe relevant critical micelle concentrations (CMCs). Unexpectedly,Di Profio et al. (2007) found that, as for DBSA and SO, the hydrateformation in a magnetically stirred reactor was retarded whenc4CMC as compared to that in the range of coCMC. Thereduction of hydrate-formation rate in the range of c4CMC wasmoderate for DBSA, but significant for SO. The effect of CTPABr onthe hydrate formation was so low that we could hardly discernany substantial difference between the rates of hydrate formationat c4CMC and coCMC. Based on such observations, Di Profioet al. (2007) concluded that surfactant micelles do not promote,but inhibit the hydrate formation from methane. Considering thatthese observations were limited to a specific stirred system andthat the mechanism of inhibition has not yet been clarified, weplanned this study to observe methane-hydrate formation in anunstirred system containing a micelle-forming surfactant. Exceptfor the selection of the surfactants to use, we performed thisstudy based on the procedure of our previous study of methane-hydrate formation in a surfactant-containing system (Okutaniet al., 2008). Our observations obtained in such an unstirredsystem were significantly different, regarding the effect of surfac-tant micelles, from those reported by Di Profio et al. (2007).

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-tensionmetry-

based estimate, 120730 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.

2. Description of experiments

2.1. Materials

We selected three commercially available anionic surfactantsfor use in this study. They were lithium dodecyl sulfate (LDS),

dodecylbenzene sulfonic acid (DBSA), and sodium oleate (SO).DBSA and SO were two of the three surfactants used in theprevious study by Di Profio et al. (2007). We selected them for thepurpose of comparing the results of the experiments usingthe same surfactants and different setup/procedures. The selec-tion of LDS was due to its structural similarity to SDS, thesurfactant the most extensively used in previous hydrate studies,and its sufficiently low Krafft point (Smejkal et al., 2003). Weexpected that the effect of the counterions (Liþ or Naþ) of thedodecyl sulfate surfactants could be realized by comparing thehydrate formation observed in the presence of LDS to that wepreviously observed in the presence of SDS using the sameexperimental apparatus and procedure (Okutani et al., 2008).

The three surfactants were used as received from the suppli-ers. They were LDS (C12H25LiO4S) having a certified purity of 0.99in 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 byTokyo Chemical Industry Co., Ltd., Tokyo). Each of these chemicalswas weighed on an electronic balance (A&D model ER-180A) witha 0.1 mg readability and dissolved in a known volume of deio-nized and distilled water to prepare each solution sample for thehydrate-forming experiments. The methane used in the experi-ments was a research-grade gas with a purity of 0.999 in massfraction supplied by Toyoko Kagaku Co., Tokyo.

2.2. Surfactant concentrations

For each of the three surfactants (LDS, DBSA and SO), weintended to vary the surfactant concentration c in the aqueoussolution to be used in each hydrate-forming experiment from runto run over a range extending from the sub-CMC regime to thesuper-CMC regime. For this purpose, we should know, evenroughly, the CMC for each surfactant under the thermodynamiccondition to be adjusted in the relevant hydrate-forming experi-ments, i.e., the condition in which an aqueous phase and amethane-gas phase are in mutual contact at p¼3.9 or 4.0 MPaand T¼275 K. Di Profio et al. (2007) reported the CMC values forDBSA and SO determined by electrical conductivity measure-ments of the solutions in contact with methane gas at p¼4.0 MPaand T¼275 K. We recently estimated the CMC values for LDS,DBSA and SO based on our own surface tension measurementsusing pendant drops suspended in a methane-gas phase adjustedat p¼3.9 MPa (for LDS) or 4.0 MPa (for DBSA and SO) andT¼275 K (Ando et al., 2012). The CMC values due to the abovetwo sources are listed in Table 1. Unexpectedly, we find signifi-cant disagreements in the CMC values for DBSA and SO betweenthe two sources. Although it is empirically known that surfacetensiometry is apt to provide lower CMC values than electricalconductometry (Jana and Moulik, 1991; Das and Das, 2008), thedifferences between the two sources, particularly that for SO,seem to be too large to be simply interpreted by the above natureof surface tensiometry in comparison to electrical conductometry.This issue has not yet been clarified (Ando et al., 2012). We thusextended the c range of our hydrate-forming experiments for eachsurfactant such that it safely included, except for SO, the relevantCMC value whichever source we may rely on.2

Table 1Critical 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/kg¼103 ppm) g (mN/m) at cZCMC

LDS Ando et al. (2012) Tensiometry 3.170.3 36.870.6

DBSA Ando et al. (2012) Tensiometry 0.3970.01 31.970.5

Di Profio et al. (2007) Conductometry 0.574a –

SO Ando et al. (2012) Tensiometry 0.1270.03 20.270.3

Di Profio et al. (2007) Conductometry 0.435b –

a Converted from original CMC value given as 1.76 mM.b Converted from original CMC value given as 1.43 mM.

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 v̂g , the molar volume of methane (NTP). According to an NIST database

(Lemmon et al., 2002), rg¼0.71746 kg/m3 and v̂g¼22.361 m3/kmol.

N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85 81

2.3. Apparatus

The experimental apparatus used in this study was just thesame as that used in our previous study of methane-hydrateformation in the presence of sodium alkyl sulfate surfactants, i.e.,SDS and its homologues having longer alkyl chains, and detailedelsewhere (Okutani et al., 2008). Thus, we only briefly describethis apparatus. The apparatus consists of a cylindrical stainless-steel test chamber, 100 mm ID and 120 mm height, used as thehydrate-forming reactor and auxiliary machinery and instru-ments. Connected to the top lid of the test chamber is the tubingfor evacuating the chamber with the aid of a vacuum pump,supplying an aqueous surfactant solution before each hydrate-forming experiment, and supplying methane gas before andduring the experiment. Two sheathed type-T thermocouples wereinserted into the chamber through the top lid to measure thetemperatures in the gas and aqueous phases. The chamber wasimmersed in a thermostated water bath.

The gas supply line connecting a high-pressure methanecylinder to the test chamber is equipped with a pressure regulator(Tescom model 44-5266), a mass-flow meter (Brooks model5860E or Oval model F-121), and a digital pressure gauge (Valcommodel VPMC-A4-4). This allowed us to measure the instanta-neous rate of gas supply into the test chamber within anuncertainty of 75 cm3/min NTP [converted to the volume flowrate at NTP (273.15 K and 101.3 kPa) condition] and the pressureinside 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 apparatussuch that the hydrate formation during each experimental runwas completed under a nearly constant pressure, i.e., within anarrow range of fluctuation about a prescribed pressure level.This was accomplished by allowing the methane gas from thehigh-pressure cylinder to flow through the pressure regulator tothe test chamber, thereby replenishing the loss in the methanegas inside the chamber due to hydrate formation. The prescribedpressure level for the experiments using LDS was 3.9 MPa, thesame as in our previous study using SDS and its longer-chainhomologues, but it was raised to 4.0 MPa for the experimentsusing DBSA or SO to make the experimental conditions coincidewith those imposed in the study by Di Profio et al. (2007) for thesake of comparing the results of both studies. Following the wayof our previous studies (Watanabe et al., 2005a; Okutani et al.,2007, 2008), we focused our attention on the macroscopicbehavior of hydrate formation following the nucleation whichcould be recognized by the visual observation, through the sightwindow of the test chamber, of the onset of hydrate formationand/or by the apparent start of methane flow into the chamber

responding to the methane uptake into the hydrate. The proce-dure of each hydrate-forming experiment as well as the prepara-tion of the apparatus in advance of the experiment were almostthe 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 surfactantsolution and cooled to the target temperature of 275 K. Afterevacuating the chamber as well as the methane-supply tubingconnected to the chamber, methane was supplied to the chamberto pressurize its interior to 3.9 MPa (when using LDS) or 4.0 MPa(when using DBSA or SO). Because the phase-equilibrium tem-peratures corresponding to the pressures of 3.9 MPa and 4.0 MPaare predicted by CSMGem, a phase-equilibrium calculation pro-gram (Sloan and Koh, 2008), to be 277.2 K and 277.5 K, respec-tively, the temperature driving force for hydrate formation wasestimated to be �2.2 K for the experiments using LDS and �2.5 Kfor the experiments using DBSA and SO. The onset of hydrateformation was detected by both the visual observations throughthe sight windows of the test chamber and monitoring of V 0g , thevolumetric rate (at NTP condition) of the methane inflow into thetest chamber. Throughout the experiment, V 0g as well as thepressure p and temperature T inside the test chamber wascontinuously 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. Themagnitudes of the variations in p and T during the hydrate-forming period in each experimental run were 0.01�0.02 MPaand 0.4�2.7 K, respectively. The variations in p and T throughoutthe hydrate-forming periods for all the experimental runs werewithin 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 eachexperimental run were those of V 0gðtÞ, the instantaneous rate ofmethane supply [converted to the volume flow rate at NTP(273.15 K and 0.1013 MPa) condition]3 to the test chamber, orVg(t), the cumulative volume (NTP) of methane supplied to thechamber 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 0gðtÞ and Vg(t) do not necessarily approximate V 0g,uptakeðtÞ,the instantaneous volumetric rate (NTP) of methane uptake intothe hydrate, and Vg,uptake(t), the cumulative volume (NTP) of

N. Ando et al. / Chemical Engineering Science 73 (2012) 79–8582

methane taken into the hydrate, with sufficient accuracy becauseof the change in the gas-phase volume inside the chamber withthe progress of the liquid-to-hydrate conversion. As detailedelsewhere (Mori and Komae, 2008), V 0gðtÞ and Vg(t) during anisobaric hydrate-formation process may be converted toV 0g,uptakeðtÞ and Vg,uptake(t), respectively, by multiplying them bythe following factor:

z�1

ð1�ðrG=rHÞÞ� �

þnwðMw=MgÞ=ððrG=rLÞ�ðrG=rHÞÞ, ð1Þ

where rG, rL and rH are the mass densities of the gas, liquid andhydrate phases, respectively, inside the test chamber; nw is thehydration number of the hydrate formed in the chamber; and Mg

and Mw are the molar masses of the guest gas (i.e., methane incase of this study) and water, respectively. Thus, we deducedV 0g,uptakeðtÞ and Vg,uptake(t) from the V 0gðtÞ and Vg(t) data for eachexperimental run using Eq. (1). For the hydrate-forming condi-tions set in this study, V 0g,uptakeðtÞ and Vg,uptake(t) should be higherthan the corresponding V 0gðtÞ and Vg(t) by 5�6%.

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 3.9070.01 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 (c¼0 ppm) is not plotted here because it almost overlaps with

the abscissa.

3. Results and discussion

3.1. Qualitative observations

In general, the behavior of macroscopic hydrate-phase growthobserved in the experiments using LDS or DBSA was almost thesame as what we observed in our previous experiments usingSDS, STS or SHS (Okutani et al., 2008) and also in qualitativeagreement 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 theirobservations with gaseous guests (propane, ethane, HFC-32, andmethane) and SDS. That is, in the LDS- or DBSA-containingsystem, hydrate crystals first formed discrete porous layers onthe surface of the aqueous phase, then generated porous layers onthe chamber wall above the aqueous phase. The latter layerslaterally grew, while thickening at the same time, and integratedinto a single riftless layer spreading over the entire wall of thechamber. Such an observation indicates that, as we alreadypointed out based on similar observations with SDS, STS andSHS (Okutani et al., 2008), the major mechanism of the hydrate-formation enhancement due to the addition of LDS or DBSA is thecapillary-driven supply of water into the porous hydrate layersgrowing on the chamber wall. In contrast, we observed only theformation of a thin hydrate film lying on the surface ofthe aqueous phase in the SO-containing system. Apparently, theaddition of SO to any concentration c up to almost 4.2 times thesurface-tensiometry-based CMC exhibited no appreciable changein the hydrate-forming behavior. This fact was unexpected inview of the intensive function of SO for decreasing the surfacetension of the aqueous solution (see Table 1).

3.2. Quantitative evaluation of hydrate formation

Fig. 1 shows the methane-uptake data obtained in the threeexperimental runs performed with aqueous LDS solutions ofdifferent initial concentrations. Each curve represents the timeevolution of Vg,uptake, the cumulative volume (NTP) of methanetaken into the hydrate formed in the test chamber, during eachrun after the instant t¼0 at which the hydrate formation was firstdetected. In general, the Vg,uptake versus t curve initially exhibits agentle increase, then a much steeper slope, and finally a plateau.This pattern of the Vg,uptake evolution is essentially the same asthose observed in our previous experiments (Okutani et al., 2008)

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 dur-ing each experimental run, we defined the effective hydrate-formation time as the duration from the inception of hydrateformation (t¼0) to a specific moment, t¼tt, at which the rate ofmethane uptake V 0g,uptakeð ¼ dVg,uptake=dtÞ had decreased to 20% ofits maximum value recorded in the same run. We also definedVgt,uptake, the total amount of methane uptake, as the Vg,uptake valueat t¼tt. The average rate of methane uptake during the effectivehydrate-formation time, _V g,uptake, is deduced by dividing Vgt,uptake

by tt. Besides, Vgt,uptake may be normalized by dividing it byVgt,uptake,max, the amount of methane that would be consumed ifthe water inside the test chamber was completely converted intothe structure-I (sI) hydrate. The ratio of Vgt,uptake to Vgt,uptake,max

indicates the fraction of the water that had been converted fromthe liquid state into the hydrate inside the test chamber duringthe effective hydrate-formation time, tt. For calculating Vgt,uptake,-

max, we estimated the hydration number for the sI methanehydrate at the pressure relevant to the experiments (3.90 MPafor 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 andVgt,uptake/Vgt,uptake,max, respectively, versus c, the initial surfactantconcentration in the aqueous phase. Here we note that neither_V g,uptake nor Vgt,uptake/Vgt,uptake,max appreciably change with thevariation in c from 2000 ppm to 5000 ppm crossing the CMC(E3100 ppm). The level of _V g,uptake observed here is almostcomparable 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. 4–6correspondingly to those for LDS shown in Figs. 1–3. The experi-ments were performed at seven different DBSA concentrations toclosely examine the possible variation in the hydrate-formation ratewith an increase in c from the sub-CMC to the super-CMC regimeswith reference to either of the two estimates of the CMC, i.e., about390 ppm due to the surface tensiometry (Ando et al., 2012)

Fig. 3. Water-to-hydrate conversion ratio at the end of the effective hydrate-

formation 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 tem-

perature T inside the test chamber were held at 4.0070.05 MPa and 275.0 (þ2.7/

�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).

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. 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 hydrate-

formation time in each experimental run, Vgt,uptake/Vgt,uptake,max, versus the DBSA

concentration c.

N. Ando et al. / Chemical Engineering Science 73 (2012) 79–85 83

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. 3in our previous paper (Okutani et al., 2008), we note a significantdifference in the time evolution of Vgt,uptake in the DBSA-contain-ing system from those in the LDS-, SDS-, STS- and SHS-containingsystems. This difference may be realized in the following twopoints: (a) except for the result obtained at the lowest concentra-tion, 200 ppm, Vgt,uptake(t) in the DBSA-containing system exhib-ited an almost linear increase from start, not accompanied by aslowly-increasing start-up period, and (b) the hydrate formationalmost ceased within 4 h, a much shorter period compared to thehydrate-formation periods observed in the other systems. Such adifference 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,uptakeE200 cm3/min (NTP) as realized inFig. 5 is about 45% higher than the level of _V g,uptake obtained in thecorresponding SDS-containing system (see Fig. 4 in Okutani et al.,2008). As for Vgt,uptake/Vgt,uptake,max, the DBSA-containing systemslightly surpasses the LDS-containing system (Fig. 3 in this paper)and is slightly inferior to the SDS-containing system (Fig. 5 inOkutani et al., 2008). The other point of our concern, which wecan clearly recognize in Figs. 4–6, is that the hydrate formation isinsensitive 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. Morespecifically, we do not find in Figs. 4–6 any sign of suppressionof the hydrate formation in the super-CMC regime as compared tothe hydrate formation in the sub-CMC regime. This fact is some-what inconsistent with the experimental observations by Di

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 (þ0.5/�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).

N. Ando et al. / Chemical Engineering Science 73 (2012) 79–8584

Profio et al. (2007) for a stirred DBSA-containing system, in whichthe super-CMC regime was found to be inferior to the sub-CMCregime not only in delaying the hydrate nucleation (i.e., providinga longer induction time), but also in slightly decreasing the finalwater-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 esti-mates of the CMC, i.e., about 120 ppm due to the surfacetensiometry (Ando et al., 2012) and 435 ppm due to the electricalconductometry (Di Profio et al., 2007). All the Vg,uptake(t) recordsobtained in the experiments using SO are shown in Fig. 7. Thevariation in Vg,uptake(t) with c is erratic, and it is practicallyinsignificant in view of the order of magnitude of Vg,uptake(t)during tr7 h. Note that even the highest value of Vg,uptake att¼7 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 notsubstantially promote the hydrate formation irrespective as towhether c is below or above the CMC. Again, this finding isapparently inconsistent with the experimental observation by DiProfio et al. (2007) in a stirred SO-containing system, in which theaddition of SO to the concentration of c¼219 ppm (0.72 mM)4 inthe sub-CMC regime significantly promoted the hydrate formation.At present, we cannot provide any physical interpretation aboutthis inconsistency between the observations obtained in unstirredand stirred experimental systems.

4. Conclusions

We have experimentally examined the effects on the hydrateformation 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.

surfactants—lithium dodecyl sulfate (LDS), dodecylbenzene sul-fonic acid (DBSA) and sodium oleate (SO)—each of which couldform, if added to an aqueous phase at a sufficiently highconcentration, micelles even under the hydrate-forming thermo-dynamic condition (�275 K in temperature and 3.9�4.0 MPa inpressure). The major findings obtained in this study may besummarized as follows:

(1)

LDS and DBSA yield qualitatively the same change in thehydrate-formation behavior, which is in turn qualitatively thesame as that we observed with the other three anionicsurfactants—sodium dodecyl sulfate (SDS), sodium tetradecylsulfate (STS) and sodium hexadecyl sulfate (SHS)—in ourprevious study (Okutani et al., 2008). That is, porous hydratelayers grow on the chamber wall above the level of theaqueous phase, and their growth accounts for the majorportion of the total hydrate formation inside the chamber.

(2)

No appreciable change is detected in neither the qualitativehydrate-formation behavior nor quantitative hydrate-forma-tion-related quantities (such as the average rate of hydrateformation, the final water-to-hydrate conversion ratio) as theresult of an increase in the LDS or DBSA concentration in theaqueous phase crossing the relevant critical micelle concen-tration (CMC). That is, the micelle formation in the aqueousphase neither promotes nor retards the hydrate formation.

(3)

LDS is almost comparable to SDS as regard to its effect on therate of hydrate formation, and is slightly inferior to SDS asregards to its effect on the final water-to-hydrate conversionratio.

(4)

As regard to its effect on the rate of hydrate formation, DBSAis superior to any other surfactants tested in our previous andpresent studies using the same experimental apparatus/pro-cedure, which are SDS, STS, SHS, LDS, DBSA and SO. Thetypical rate of hydrate formation promoted by DBSA exceedsthat by SDS by about 45%.

(5)

The addition of SO to any concentration, either below orabove the CMC, hardly promotes the hydrate formation.

Acknowledgment

This study has been supported in part by a Grant-in-Aid forScientific Research from the Japan Society for the Promotion ofScience (Grant 20246040).

References

Ando, N., Kuwabara, Y., Kodama, T., Mori, Y.H., 2012. Surface tensions of aqueoussolutions of lithium dodecyl sulfate, sodium oleate, and dodecylbenzenesulfonic 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 behaviorof 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 onclathrate 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 forgas 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 to55 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 hexadecyltrimethylam-monium 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 themechanism 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 85

Lemmon, E., McLinden, M., Huber, M., 2002. NIST Reference Fluid Thermodynamicand Transport Properties Database (REFPROP), Ver. 7.0. US Department ofCommerce.

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 aclathrate 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 formationin an unstirred gas/liquid system: amendments to previous study using HFC-32 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 andsodium alkyl sulfates. Chem. Eng. Sci. 63, 183–194.

Pang, W.X., Chen, G.J., Dandekar, A., Sun, C.Y., Zhang, C.L., 2007. Experimentalstudy 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. CRCPress, Boca Raton, FL, USA.

Smejkal, G.B., Shainoff, J.R., Kottke-Marchant, K.M., 2003. Rapid high-resolutionelectrophoresis of multimeric von Willebrand Factor using a thermopiloted gel

apparatus. Electrophoresis 24, 582–587.

Sun, Z.-G., Ma, R.-S., Wang, R.-Z., Guo, K.-H., Fan, S.-S., 2003a. Experimentalstudying 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 inhydrates 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 liquidhydrocarbons on gas hydrate formation rate and storage capacity. Int. J. EnergyRes. 27, 747–756.

Watanabe, K., Imai, S., Mori, Y.H., 2005a. Surfactant effects on hydrate formation inan unstirred gas/liquid system: an experimental study using HFC-32 andsodium dodecyl sulfate. Chem. Eng. Sci. 60, 4846–4857.

Watanabe, K., Niwa, S., Mori, Y.H., 2005b. Surface tensions of aqueous solutions ofsodium alkyl sulfates in contact with methane under hydrate-forming condi-tions. J. Chem. Eng. Data 50, 1672–1676.

Zhang, J.S., Lee, S.Y., Lee, J.W., 2007a. Kinetics of methane hydrate formation fromSDS solution. Ind. Eng. Chem. Res. 46, 6353–6359.

Zhang, J.S., Lee, S., Lee, J.W., 2007b. Does SDS micellize under methane hydrate-forming 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 nearpropane and carbon dioxide hydrate-forming conditions. J. Chem. Eng. Data2007 (52), 2480–2483.

Zhong, Y., Rogers, R.E., 2000. Surfactant effects on gas hydrate formation. Chem.Eng. Sci. 55, 4175–4187.