ARTICLE pubs.acs.org/Langmuir
Compartmentalization and Separation of Aqueous Reagents in the Water Droplets of Water-in-Oil High Internal Phase Emulsions Timothy S. Dunstan and Paul D. I. Fletcher* Surfactant & Colloid Group, Department of Chemistry, University of Hull, Hull HU6 7RX, United Kingdom ABSTRACT: We have compartmentalized aqueous reagents and indicator species within the micrometer-sized water droplets of mixed high internal phase emulsions (HIPEs). Mass transport of the reagents across the micrometer-thickness oil films separating the water droplets followed by reaction with the indicator species produces a visible color change which provides a simple method to measure the trapping times of the reagents. Trapping times have been measured for an uncharged reagent (hydrogen peroxide) and charged reagents (HCl and NaClO) in different HIPEs. The trapping times are discussed in terms of a model in which the transferring species partitions from the water to the oil film followed by a ratedetermining step of diffusion across the oil film. Rather surprisingly, it is found that trapping times are of similar orders of magnitude for both uncharged and charged aqueous species transferring across liquid oil films.
’ INTRODUCTION Water-in-oil (w/o) emulsions consist of a thermodynamically unstable dispersion of micrometer-sized water drops in a continuous oil phase. In general, emulsion phase separation can occur through a combination of the processes of sedimentation or creaming, flocculation, Ostwald ripening, and droplet coalescence. Kinetic stability against the processes can be provided by adsorption of suitable stabilizers, which may be either low molar mass surfactants, polymers, or particles, at the oil-water interfaces of the emulsion drops. Emulsions have a very wide range of application in sectors including household products, pharmaceuticals, agrochemicals, oilfield chemicals, printing, paints, and others. In addition to these relatively “low-tech” applications, there is interest in the use of emulsion water drops for the compartmentalization and confinement of chemical reactions and biological processes for new applications.1-4 These exploit the small confinement volumes (of the order of 109 times smaller than typical microtiter-plate systems) and large numbers of individual droplets which enable massively parallel processing of up to 1010 reactions per milliliter of emulsion. Co-compartmentalization of genes and the molecules they encode within emulsion drops has been successfully used to achieve the directed evolution of a range of proteins and RNAs.3 Novel applications are being further extended through the use of monodisperse emulsion drops produced using microfluidic devices which allow greater control of the droplet processing.3 With this background in mind, we have investigated the extent to which different aqueous reagents can be compartmentalized and held separated within the water drops of an emulsion. For such formulations to be environmentally acceptable and costeffective, it is important to minimize the amount of continuous oil phase in the emulsions. For this reason, we focus on emulsions containing a volume fraction of the continuous oil phase which is less than about 30 vol %, that is, with a volume fraction of water drops greater than about 70 vol% which exceeds the packing limit r 2011 American Chemical Society
for undeformed spheres. These high internal phase emulsions (HIPEs) consist of deformed, polyhedral water drops separated by thin films of oil with an overall structure of a liquid-in-liquid foam. The formation, structure, rheology, and other physical properties of HIPEs are discussed in refs 5-16. In the present study, we prepare a first HIPE containing aqueous reagent A in its water drops and mix it with a second HIPE containing aqueous reagent B. We have selected various pairs of aqueous reagents (A and B) such that their mass transport into the same droplet produces a chemical reaction which gives a visible color change in the mixed HIPE. The aim is to develop a simple methodology which enables us to determine the “trapping time” for which A and B are retained in separate water drop compartments. The limited literature information on reagent mass transfer between water drops in oil-continuous emulsions17-26 mainly relates to water as the transferring species, generally requires fairly complex methodology to determine the mass transfer rates, and often relates to systems in which the geometry is incompletely defined which limits the extent of data analysis and interpretation. The main aims of the present work are to (i) develop a simple methodology to measure the rates of mass transport between water drops; (ii) compare experimental data with suitable model calculations (which requires the geometry of the oil film barriers between the water drops are sufficiently well-defined); and (iii) extend the data set of transferring species to include a wider range of both uncharged and charged aqueous solutes.
’ EXPERIMENTAL SECTION Materials. Water was purified by passing through an Elgastat Prima reverse osmosis unit followed by a Millipore Milli-Q reagent water Received: January 6, 2011 Revised: February 11, 2011 Published: March 10, 2011 3409
dx.doi.org/10.1021/la200058d | Langmuir 2011, 27, 3409–3415
Langmuir system. Its surface tension was 71.9 mN m-1 at 25 °C, in good agreement with literature. The oils n-dodecane (Sigma Aldrich, > 99%), squalane (Sigma Aldrich, 99%), 0.65 cSt polydimethylsiloxane (PDMS, Dow Corning 200 Fluid, Sigma Aldrich), and rapeseed oil (Tesco) were columned either twice or four times over neutral aluminum oxide 90 (Merck) to remove polar impurities. The surfactant Emulsogen OG consists of oleyl hydrophobic chains bonded to polyglcerol and has an average molecular structure of two oleyl chains bonded to two glycerol units and so is denoted here with the abbreviation O2G2. The average structure was determined from the manufacturer’s (Clariant) information on the average number of glycerol units per molecule and the measured saponification number to derive the number of oleyl chains. Anfomul 2887, a polyisobutenyl succinic anhydride amine derivative, is a surfactant which is commonly used to stabilize emulsion explosive HIPEs16 and was supplied by Croda. Sodium bis-2-ethylhexylsulphosiccinate (AOT, Aldrich, 98%) was used as received. The transferring species sodium hypochlorite (BDH, 12 w/v available chlorine), hydrochloric acid (Fisher Chemicals, 36% solution), and hydrogen peroxide (FMC corporation, 50 wt % solution) and indicator species potassium permanganate (Sigma Aldrich, >99%), Congo Red (Acros Chemicals, 85%), and methyl orange (Sigma Aldrich, 85%) were used as received. Prior to use, all hydrogen peroxide solutions were titrated using cerium sulfate with ferroin indicator to determine their accurate concentrations. Additional reagents including urea (Fisher Chemicals, >99%), acetic acid (Fisher, >99%), potassium iodide (Sigma, >99%), sodium thiosulphate (Sigma, 99.99%), starch indicator (Sigma), ferroin (Sigma, 0.1 wt % solution), cerium sulfate (Reidel de Haen, >98%), sulfuric acid (Fisher, 98%), and chloroform (Fisher, >99%) were all used as received. Methods. A total of 10 mL of each high internal phase emulsion was prepared by adding the required volumes of aqueous phase (containing the required dissolved reagents), stabilizer, then oil phase to a 25 75 mm (diameter height) glass tube. After incubating for about 30 min, the samples were emulsified by either vigorous handshaking for 30 s or by using a simple overhead paddle stirrer with a single plastic-coated metal stirrer blade of 20 30 mm operating at approximately 100 rpm for 1 min. These rather low energy input methods of emulsion preparation were used in order to produce relatively large emulsion drops and correspondingly thick oil films separating the water drops. Repeated preparation of the emulsions produced mean drop sizes which were reproducible within approximately 10%. Mean drop diameters were measured by optical microscopy. A small volume of the emulsion was extracted from the middle of the emulsion by pipet and diluted into a large volume of the oil used as the continuous phase. Diluted samples were held in a 25 75 mm cavity microscope slides with a single cavity of 16 mm diameter and 0.2 mm depth which was covered with a coverslip. Micrographs of this diluted emulsion were obtained using a Leica DME transmission microscope equipped with a Leica DFC 290 camera. The entire sample field was scanned before acquisition of a micrograph to ensure the final image was representative of the total emulsion drop size distribution which was determined by measuring the diameters of all the drops appearing in a single micrograph (typically 50-100 drops) using Leica LAS image analysis software. All parent emulsion drop size distributions were monomodal with polydispersities (equal to the standard deviation divided by the mean) of approximately 50%. The final quoted mean drop radii refer to the number average. For the different systems, mean drop radii were measured for the “parent” HIPEs containing either the transferring or indicator species and the mixed HIPE containing both species immediately after preparation. The evolution of the drop size in the mixed HIPE over time was also measured. To study trapping times, equal volumes of the two “parent” HIPEs containing the transferring and indicator species were mixed by pouring all of one emulsion into another then gently stirring with a spatula four or five times. The vessels were sealed and the mixed emulsions were
ARTICLE
incubated at 25 °C until the color change indicating transfer had gone to completion. The mixed emulsions were not stirred during the incubation; HIPEs are gel systems in which possible effects due to sedimentation or creaming of the drops are absent. A HIPE in which the two aqueous phases had been premixed was prepared for each system to serve as the reference to show the indicator color change. The sample and reference emulsions were compared visually to determine the time for the color change to go to completion. Trapping times were measured at room temperature of 20 ( 3 °C. Values of the partition coefficient of hydrogen peroxide between water and dodecane under different conditions were determined using one of two methods. In the first, low sensitivity method, equal volumes of aqueous hydrogen peroxide and dodecane were equilibrated with gentle stirring for 48 h. An amount of 1 mL of the equilibrated dodecane phase was mixed with 5 mL of 50:50 glacial acetic acid/chloroform and 10 mL of aqueous KI solution (excess). The iodine liberated was titrated with aqueous sodium thiosulfate solution with starch as indicator.27 For the initial peroxide concentrations used here, partition coefficients greater than 5 10-4 could be determined by this method. In the second, high sensitivity method, 1 mL of aqueous hydrogen peroxide solution was equilibrated with 99 mL of dodecane with gentle stirring for 9 days. This produces a large oil volume with a low equilibrium peroxide concentration. A total of 90 mL of the equilibrated dodecane phase was then equilibrated with 1 mL of water for 14 days which extracts virtually all the available peroxide from the oil and gives a 90-fold increase in peroxide concentration, thereby enhancing the measurement sensitivity. Samples (0.25 mL) of this equilibrated aqueous phase were titrated with aqueous cerium sulfate using ferroin as indicator.27 Using this method, values of partition coefficient lower than 10-7 could be measured (dependent on the initial peroxide concentration). Samples containing no peroxide were also measured as controls.
’ RESULTS AND DISCUSSION Theoretical Considerations. A typical experiment to determine the time over which aqueous reagents are maintained in separate water drop compartments of a HIPE was performed as follows. A first HIPE (HIPE1) is prepared containing a concentration cA1 of an aqueous reagent A and volume fraction of oil φo1 with mean water drop radius r1. The concentration cA1 is expressed as moles of A per unit volume of water, not the overall emulsion volume. A second HIPE (HIPE2) containing cB2 of a different aqueous reagent B and volume fraction of oil φo2 with mean water drops radius r2. Equal volumes of HIPE1 and HIPE2 are mixed to produce a mixed HIPE with overall oil volume fraction φo, mean drop radius r = (r1 þ r2)/2 and comparable numbers (depending on the relative values of the radii r1 and r2) of water drops containing either A (at concentration cA1) or B (at concentration cB2). The mean drop radius r for the mixed HIPE was used since r1 and r2 were generally similar and this approximation greatly simplifies the theoretical analysis given below. The oil volume fractions of HIPE1 and HIPE2 were generally kept equal such that φo1 = φo2 = φo. Various pairs of aqueous reagents A and B were examined with A referring to the main transferring species at high concentration and B referring to an appropriate colorimetric indicator species which was present at low concentration. The mixed emulsions were observed visually to determine the time taken for the indicator species B was observed to undergo a color change. This measured time (t*) corresponds to the time taken for sufficient A species to transfer from the A-containing droplets into the B-containing droplets such that the color change occurs. 3410
dx.doi.org/10.1021/la200058d |Langmuir 2011, 27, 3409–3415
Langmuir
Figure 1. Schematic of a mixed water-in-oil HIPE consisting of water droplets containing the transferring species (white) and indicator species (colored red before reaction and white after reaction) separated by films of the oil continuous phase (green). The inset shows a blow-up of an oil film (mean thickness h) separating white and red droplets. Sufficient mass transfer across the oil films to produce the red-to-white indicator color change occurs in the measured trapping time t* and produces an overall color change of the HIPE.
In principle, mass transport of A f B droplets will be accompanied by mass transport of B f A droplets which will also cause a color change. However, separate experiments in which HIPE1 was carefully layered on top of HIPE2 followed by observations of where the color change developed (i.e., in HIPE1 or in HIPE2) showed that A f B dominates under the experimental conditions used here, that is, high aqueous concentrations of A and very low concentrations of B. If it is assumed that the so-called “lag time” required to establish the steady-state concentration gradient across the oil film is negligibly small, then this behavior is expected for diffusional mass transport because the relative initial mass transport fluxes are proportional to the concentration gradients which, in turn, are proportional to the initial concentrations. As described in ref 28, the lag time L is approximately equal to h2/6D where h is the oil film thickness and D is the diffusion coefficient of the permeating species in the oil film. For h = 2 10-6 m and D = 1.5 10-9 m2 s-1, we estimate L to be approximately 0.4 ms, which is indeed negligibly small relative to the overall mass transfer times observed here. The situation of the mixed HIPE containing similar numbers of A and B water drops held at fixed but random relative locations by the weak gel nature of the HIPE is shown schematically in Figure 1. In general, A and B initially located in separate droplets can meet and react together by combination of water drop coalescence and dissolution in and diffusion across the surfactant-coated oil films separating the water drops. Dependent on the HIPE composition, the oil films separating the droplets may or may not contain excess surfactant in the form of aggregated species including either inverse microemulsion droplets or lyotropic liquid crystalline phases.5 The presence of such surfactant aggregates may solubilize transporting A molecules in the oil and hence cause facilitated transport of A across the oil films. A further complication can arise from the fact that, in addition to the mass transport of the species A, it is expected that mass transport of water between the emulsion drops will also occur due to osmotic pressure differences between the two droplet types. As will be shown later, the experimental systems can be manipulated to suppress contributions arising from water drop coalescence, facilitated mass transport by surfactant aggregates
ARTICLE
present in the oil films and water mass transport. We consider here the limiting, idealized situation in which t* is determined solely by a process in which the uncharged species A partitions into and diffuses across the oil films separating the drops. It is assumed here that the adsorbed surfactant films coating the oil films play no significant role in controlling the rate of permeation of A between droplets and there is no significant energy barrier to the permeant species crossing the oil-water interface; that is, the rate-determining step in the mass transfer is diffusion across the oil film. Under these idealized conditions, the permeation in the mixed HIPE structure is equivalent to a system comprising an aqueous donor compartment initially containing species A at concentration cA1 separated from an aqueous receiving compartment (containing zero A initially) by a liquid oil membrane of thickness h. The area of the oil film A (per unit total volume of the emulsion) is approximately half that of the area of the water drops. Neglecting the distortion of the drops from sphericity gives an approximate expression for A. A
3ð1 - φo Þ 2r
ð1Þ
An approximate expression for the average oil film thickness h in terms of the oil volume fraction and mean drop radius r (corresponding to undistorted, spherical drops of the same volume as the actual distorted drops present in the HIPE) is then: 2φo r ð2Þ 3ð1 - φo Þ If it is assumed that the passive mass transport of A is determined solely by diffusion (i.e., that the transfer of A across the oil-water interfaces is not rate-limiting), then application of Fick’s laws gives the initial (but steady-state) rate of mass transport of the species A across the oil film E as h
E¼
AVDcA1 Kow h
ð3Þ
where D is the diffusion coefficient of the species A in the oil film, V is the total volume of the mixed HIPE, and Kow is the equilibrium partition coefficient of the species A between oil and water (i.e., Kow = [A]oil/[A]water at equilibrium). The time for the indicator color to change (t*) is the time taken for the transfer of a number of moles of A equal to that required to fully react with all the reagent B present, that is, a reaction stoichiometric factor (S) times the number of moles of B. In a volume V of the mixed HIPE containing equal volumes of HIPE1 and HIPE2, the number of moles of B (nB) is V ð1 - φo2 ÞcB2 2
ð4Þ
V ð1 - φo2 ÞcB2 S 2cB2 r 2 φo S 2E 9ð1 - φo ÞDcA1 Kow
ð5Þ
nB ¼ and hence t ¼
The final eq 5 predicts the time taken for the indicator color change (t*) for mixed HIPE emulsions in which • equal volumes of HIPE1 (containing transferring reagent) and HIPE2 (containing indicator species) have been mixed; • mass transfer between water drops occurs only by passive diffusion across the oil film separating the drops; that is, no water drop coalescence occurs and there is no facilitated 3411
dx.doi.org/10.1021/la200058d |Langmuir 2011, 27, 3409–3415
Langmuir
ARTICLE
transport by (for example) reversed micelles of the surfactant in oil; • within the passive diffusion process, the rate-determining step is the diffusion process across the oil film. The processes of entry and exit of transferring molecules into and out of the oil film are both relatively fast. Transfer of Hydrogen Peroxide between Water Drops in Water-in-Dodecane HIPEs. We first discuss measurements of the “trapping time” t* for the transfer of hydrogen peroxide across dodecane oil films into aqueous drops containing potassium permanganate as indicator. Peroxide and permanganate react according to27 2MnO4 - þ 5H2 O2 þ 6Hþ f 2Mn2þ þ 5O2 þ 8H2 O In the mixed HIPE, this indicator reaction produces a loss of the brownish/pink color of the permanganate to the white appearance of the HIPE in the absence of added indicator. Referring to eq 5, the stoichiometry factor S is 2.5 for this reaction. Measurements of t* were performed by mixing equal volumes of HIPE1 containing 0.75 volume fraction of aqueous phase containing various concentrations of hydrogen peroxide and HIPE2 containing 0.75 volume fraction of aqueous phase containing 0.1 mM KMnO4 indicator. Both HIPE1 and HIPE2 and the final mixed HIPE contained 0.25 volume fraction of dodecane as oil continuous phase and were stabilized using 1 wt % of Anfomul 2887 surfactant. As shown later, this surfactant concentration was carefully selected to be the minimum required such that no emulsion drop growth was observed over 2 days, that is, longer than the time scale of the t* measurements. Hence, this choice of surfactant concentration is expected to minimize possible peroxide mass transport due to drop coalescence and facilitated transport across the oil films by excess surfactant present in the form of reversed micelles or other aggregates present in the oil. In order to compare measured t* values with those estimated using eq 5, we require values of the diffusion coefficient of hydrogen peroxide in dodecane (D) and the partition coefficient (Kow) for hydrogen peroxide between water and dodecane. D was estimated using a literature value of Dw for pentane in water at 25 °C of 1.06 10-9 m2 s-1 (ref 29) and assuming that D scales as (molar volume)-0.4, where the scaling exponent was taken to be intermediate between -1/3 (spherical molecules) and -1/2 (random coil chains). The value was further scaled by the ratio of the viscosity of water at 25 °C and of dodecane at 20 °C (= 0.8899/1.374)30 to yield a final estimated value of 1.3 10-9 m2 s-1 for hydrogen peroxide in dodecane at 20 °C. D was also estimated using a measured value of the diffusion coefficient of water in hexane at 25 °C (= 9.53 10-9 m2 s-1)31 as an alternative starting point. Applying a similar rescaling to take account of differences in solvent viscosities and the molar volumes of the diffusing species gives a second estimate of D for peroxide in dodecane at 20 °C equal to approximately 1.7 10-9 m2 s-1. For the calculations shown below, a final value for D of 1.5 10-9 m2 s-1 was used which is estimated to be reliable within (30% or so. The value of Kow for hydrogen peroxide between pure water and dodecane was measured for a range of aqueous peroxide concentrations using the high sensitivity, backextraction method described in the Experimental Section. The results, shown in Figure 2, show that Kow is approximately independent of the peroxide concentration (i.e., nonideality effects are not significant) and is 1.1 10-7. This value of Kow is 3-4 orders of magnitude lower than literature values for hydrogen
Figure 2. Variation of partition coefficient of hydrogen peroxide between water and dodecane with aqueous phase peroxide concentration.
Table 1. Variation of the Partition Coefficient of Hydrogen Peroxide between Water (with and without 2 M Urea) and Dodecane with the Concentration of Anfomul 2887 Surfactant [Anfomul 2887]/wt % 0
Kow (no urea)
Kow (2 M urea)
-7
