- Email: [email protected]
liquid interfaces
Journal of Electroanalytical Chemistry 805 (2017) 91–97
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
Spontaneous emulsification at surfactantless liquid/liquid interfaces Barry R. Silver, Karel Holub, Vladimir Mareček
MARK
⁎
Department of Biophysical Chemistry, J. Heyrovsky Institute of Physical Chemistry, Dolejškova 2155/3, 182 23 Prague 8, Czech Republic
A R T I C L E I N F O
A B S T R A C T
Keywords: Spontaneous emulsification ITIES Water-in-oil Open circuit potential
Liquid/liquid interfaces can undergo spontaneous emulsification when a common-ion is distributed between both phases. By careful selection of electrolyte and solvent, stable nano-sized water-in-oil emulsions containing electrolyte can be produced. The emulsification process can be monitored indirectly via transient open circuit potential measurements using an electrochemical measuring cell, initially poised at equilibrium, within the organic phase. Theoretical analysis of experimental data indicates that the interfacial emulsification process is probably controlled by diffusion.
1. Introduction Spontaneous emulsification can occur when two immiscible liquids, out of equilibrium, are placed in physical contact [1,2]. Spontaneous emulsification can proceed without the aid of external energy inputs and without surfactant. This is because chemical potential energy gradients, in liquid-liquid systems which are out of equilibrium, are large enough to drive interfacial emulsification [2]. Spontaneous emulsification is currently an active area of research, although as noted by Aoki [3], a large proportion of previous work dealing with surfactantless spontaneous emulsification at immiscible liquid-liquid interfaces does not emphasise an electrochemical aspect to any great extent. Conventional ITIES (interface between two immiscible electrolyte solutions) may also be able to undergo spontaneous emulsification without external polarization [3]. Much of the previous work in this regard has however, dealt almost exclusively, with spontaneous oscillations in interfacial tension and general interfacial instability in the presence of added surfactant [4–9]. Under externally polarised conditions though, interfacial emulsification and interfacial instabilities in the presence of an ionic surfactant, arise from the formation of an ‘instability window’ [8]. This ‘window’ is caused by potential-dependent adsorption and partitioning of surfactant at the ITIES [10]. Within a conventional ITIES-type experimental arrangement, both phases contain electrolyte. Electrolytes used within the aqueous phase are hydrophilic salts such as lithium chloride. The organic phase (usually nitrobenzene or 1, 2-dichloroethane) contains hydrophobic electrolyte, such as tetrabutylammonium tetraphenylborate for example (TBATPB). A recent small angle neutron scattering (SANS) study by Sadakane et al. [11] demonstrated that in the presence of tetraphenylborate the water/3-methylpyridine interface can produce
⁎
periodic mesoscopic structures [12]. Water soluble tetraphenylborate salts (such as NaTPB) are sometimes termed as “antagonistic” [13,14]. In other words, the salt is comprised of both strongly hydrophilic (Na+) and strongly hydrophobic (TPB−) ions. In the presence of a composition heterogeneity, such as that found at an ITIES, antagonistic salts can lead to the formation of mesophase structures with charge density waves and decreases in interfacial tension [14]. Similar structures have also been recently reported at the water/2,6-dimethylpyridine interface with the addition of a tetra alkyl-ammonium bromide [15]. In this present article we demonstrate that ITIES, possessing a common-ion distributed between both phases, are able to undergo spontaneous emulsification and form stable water-in-oil nanoemulsions. This effect is exemplified using two model systems. Methodology detailed herein allows one to monitor the emulsification process indirectly using an electrochemical monitoring cell located within the organic phase. A theoretical framework is provided to analyse results of one of the experimental systems. 2. Materials and methods Tetrabutylammonium tetraphenylborate (TBATPB, > 99%), tetramethylammonium chloride (TMACl, > 98%), tetraphenylarsonium chloride (TPhAsCl, polarography grade), sodium tetraphenylborate (NaTPB, 99.5%) and agar-agar were all supplied by Fluka Analytical, Czech Republic. Tetrabutylammonium chloride (TBACl, > 97%) was supplied by Sigma Aldrich, Czech Republic. Tetramethylammonium tetraphenylborate (TMATPB) was prepared by metathesis. Tetraphenylarsonium bis-1,2-dicarbollylcobaltate (TPhAsDCC) was prepared at the Institute of Inorganic Chemistry ASCR, v.v.i. Chemicals and was used as received. Highly purified, deionised water (DI,
Corresponding author. E-mail address: [email protected] (V. Mareček).
http://dx.doi.org/10.1016/j.jelechem.2017.10.027 Received 1 August 2017; Received in revised form 9 October 2017; Accepted 12 October 2017 Available online 14 October 2017 1572-6657/ © 2017 Elsevier B.V. All rights reserved.
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
conductance: < 0.1 μS cm− 1, GORO system, Czech Republic) was used to prepare all aqueous solutions. 1, 2-dichloroethane (1,2-DCE, 99%) and nitrobenzene (NB, 99.5%) were supplied by Fluka Analytical (Czech Republic) and were used to prepare organic solutions.
Electrolyte in the indicator electrode was gelled to the tip with 2% (w/ v) agar for mechanical stability. Water-in-oil droplets impact the indicator electrode at the very tip of the capillary. The reference interface, in contrast, is situated well inside the Luggin capillary, to protect it from the droplet impact. (Fig. 1). Two experimental systems are used for experiments, namely: The TPB−/NB system DFI: x mM NaTPB(w)//10 mM TBATPB(NB). MC: Ag/AgCl/10 mM TBACl (in 2% agar)(w)//10 mM TBATPB(NB)// 10 mM TBACl(w)/AgCl/AgCl. The TPhAs+/1,2-DCE system DFI: x mM TPhAsCl(w)//10 mM TPhAsDCC (1,2-DCE) + 5 mM TMATPB (1,2-DCE). MC: Ag/AgCl/5 mM TMACl (in 2% agar)(w)//10 mM TPhAsDCC(1,2DCE) + 5 mM TMATPB(1,2-DCE)//5 mM TMACl(w)/AgCl/Ag. Two Ag/AgCl wires were used as electrodes. The OCP of the MC is continuously monitored/recorded. A concentrated aliquot (normally 100 mM) of aqueous phase electrolyte was injected at 60s (in all cases) after the initiation of OCP monitoring. The OCP was monitored using a CHI 660c potentiostat (CHI Instruments).
2.1. Glass plate experiments Equal aliquots (as much as 75 μl) of an aqueous phase and an organic phase (of varied compositions) were placed in close proximity to one another at the centre of a clean glass plate. The glass plate was of rectangular geometry and was approximately 1.5 mm thick. Glass cover slips of approximately 100 μm thickness were placed in each corner. A second glass plate (of the same thickness and geometry) was placed on top of the first which allows the two phases to contact. This ‘sandwich’ arrangement formed a thin-layer cell and allowed oil/water interfaces to be readily viewed using optical microscopy. Evaporation of the two phases was negligible in this arrangement over the course of a typical experiment. 2.2. Open circuit potential (OCP) monitoring experiments A cartoon depiction of the glass apparatus used in OCP monitoring experiments is illustrated in Fig. 1. The glass apparatus comprises two electrochemical cells. The first cell is formed by an aqueous phase and an organic phase. These two phase form a large area interface. This water/oil interface is the droplet forming interface (DFI). The cell is initially filled with organic phase (5 ml), upon which, deionized water (4 ml) is slowly layered. At the beginning of the OCP measurement, a concentrated aliquot of aqueous electrolyte is carefully injected into the distilled water yielding the required concentration of electrolyte in the aqueous phase. The injection initiates the spontaneous emulsification process. Water-in-oil droplets are homogenised within the organic phase by slow stirring, 2–3 revolutions per second, with a magnetic stir bar. The second electrochemical cell (MC), consisting of two Luggin capillaries (inner diameter 1.5 mm) placed in the organic phase, is used to monitor the emulsification process. The upper Luggin capillary with the reference interface inside the capillary serves as a reference electrode. The lower Luggin capillary serves as an indicator electrode.
3. Results 3.1. Glass plate experiments A vigorous and spontaneous interfacial emulsification process is observed at the ITIES shortly after placing the two phases in physical contact (both TPB−/NB and TPhAs+/1,2-DCE systems). The sub-micron sized water-in-oil droplets (see DLS results, Supplementary section) formed exclusively on the oil side of the ITIES, makes proper resolution impossible using optical microscopy. Although one can, with the aid of microscopy, observe masses of sub-micron, spherical and translucent entities originating from the interface and pervade the bulk of the organic phase. Some larger micron-sized droplets are observed in bulk after a period of a few minutes. Fig. 2 illustrates the extent of the spontaneous emulsification process within the TPB−/NB system. No obvious signs of precipitate are observable. Droplets remain intact for long periods and for relatively long distances away from the liquid/liquid interface. This indicates droplet stability. Droplet stability allowed extensive droplet accumulation within the volume of the organic phase (Fig. 2 and Supplementary section). Within the TPB−/NB system, droplet accumulation results in organic phase turbidity (with slow stirring of the organic phase) over a period of a few hours (Fig. 2). Although a similarly vigorous and spontaneous interfacial emulsification process was observed for the TPhAs+/1,2-DCE system, organic phase turbidity was not as extensive as that of the TPB/NB system over the same period of time. Standard transfer potentials are similar for the transfer of TPB− from water to nitrobenzene and for its transfer from water to 1,2-dichloroethane [16]. Hence, by using the same electrolytes in each phase, but changing only the solvent (from NB to 1,2-DCE), similar Galvani potentials across the ITIES are expected. Both a TPB−/NB system and a TPB−/1,2-DCE system (using the same electrolytes), produce vigorous spontaneous emulsification at the interface. Although, within the TPB−/1,2-DCE system, water droplets formed within the oil phase are observed to clearly ‘pop’ and coalesce on the surface of the glass after a few minutes and at a short distance away from the interface (Figs. 1S and 2S). Furthermore, no accumulation in bulk, and thus no observable turbidity is evident after 12 h of stirring (as in Fig. 2). Rather, glass walls of the vial in the case of TPB−/1,2-DCE system, took on a ‘frosted’
Fig. 1. Cartoon depiction of the glass apparatus used for OCP monitoring experiments. The dashed line within the upper Luggin capillary (indicated by an arrow) shows the position of the liquid/liquid reference interface. Shading in both capillaries indicates an aqueous phase. The dark ellipsoidal shape at the bottom of the organic phase represents the small Teflon stir bar.
92
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Fig. 2. A) The extent of the spontaneous emulsification a short time after contacting the two phases within the glass plate arrangement. Copious amounts of water-in-oil droplets are formed in the NB phase which are hard to resolve using optical microscopy. No droplets were observed within the aqueous phase. B) Equal aliquots of both aqueous and organic phase (3 ml) are layered on top of one another in a small glass vial. A small Teflon-coated magnetic stir bar allows gentle stirring/homogenization of the organic phase. The extreme turbidity of the organic phase demonstrates the extent of droplet accumulation and further indicates droplet stability after a period of 12 h. The aqueous phase and the organic phase shown in A and B comprised 10 mM NaTPB in water and 10 mM TBATPB in nitrobenzene respectively. Scale bar in (A) is 10 μm.
appearance. ‘Frosting’ is caused by mass coalescence of water droplets on the interior surface of the glass within the volume of the organic phase (Figs. 1S and 2S). This indicates that the TPB−/NB system produces water-in-oil emulsions which are stable in the bulk of the organic phase and that the TPB−/1,2-DCE system does not. This observation is further strengthened using OCP measurements (Fig. 3). The OCP rise in the case of TPB−/1,2-DCE is only 50 mV after 10,000 s. This 50 mV OCP rise is in comparison to that of the 600 mV rise in approximately 500 s in the case of NB. The extremely low increase in OCP for the TPB−/1,2-DCE system further indicates that droplets formed at this DFI do not have the required stability in the bulk of the organic phase to cause significant changes in potential at the MC.
Spontaneously formed water-in-oil droplets contain aqueous phase electrolyte. A bright white precipitate was observed at the tip of the indicator electrode during experiments within the TPB−/NB system. Aqueous droplets containing NaTPB deposit at the indicator electrode and form an insoluble precipitate. The solubility of TBATPB in water is extremely low (0.259 μM, [17]). Precipitate is not observed at the tip of the indicator electrode within the TPhAs+/1,2-DCE system. This is because water droplets containing water soluble TPhAsCl depositing at the indicator electrode do not form a precipitate with water soluble TMACl. Another important observation which further emphasises that droplets produced at the DFI contain aqueous phase electrolyte, is that droplets are stable within the organic phase for long periods (Fig. 2). Extensive turbidity can be maintained within the organic phase (TPB−/ NB system) for well over a week (as per Fig. 2). This stability can be a product of the common-ion present within the droplet and within the bulk organic phase outside the droplet. The common-ion may cause a surface potential to be created on the droplet, which in turn could lead to an electrostatic repulsion between individual droplets [18]. Droplets should contain equal amounts of cations and anions and are therefore electroneutral. The potential of the indicator electrode (at t = 0, and just before aqueous electrolyte injection) is maintained by an equal concentration of TBA+ on either side of the interface in the case of the TPB−/NB system, and by an equal concentration of TMA+ in the case of the TPhAs+/1,2-DCE system. As water droplets containing electrolyte, start to impact the indicator electrode, the potential becomes perturbed from that of equilibrium. Droplet impact is therefore the first step in a process which alters the OCP and ultimately leads to the distinct sigmoidal shape.
3.2. OCP monitoring experiments All OCP voltage-time curves, in those experimental systems which produce stable water-in-oil droplets in the bulk of the organic phase, were sigmoidal in shape (Fig. 3 and Fig. 4). The sigmoid has three distinct regions. The first region begins at a condition of equilibrium before aqueous electrolyte injection at 0 V. After injection, a gentle rise in OCP from the equilibrium potential is observed due to droplet collision with the indicator electrode. The gentle rise gives way to a very steep rise, which tends toward or actually reaches a constant OCP. At the highest concentration of aqueous electrolyte, the final value of the OCP quickly approaches that of standard ion transfer potential of the common-ion initially distributed across the DFI.
4. Theoretical For reasons of brevity, our analysis herein considers only the TPB−/ NB system. As we monitor the DFI indirectly by means of the MC located within the organic phase, any processes occurring at the indicator electrode would be a direct function of, and in response to, processes occurring at the DFI. In this way one is able to make inference about the DFI through interpretation of OCP changes at the indicator electrode. The potential of the monitoring cell can be written as
E=
RT ⎛ c oTBA + ⎞ ln ⎜ w ⎟ nF ⎝ c TBA +⎠
(1)
The initial potential at the indicator electrode is maintained by an equal concentration of TBA+ on the aqueous side of the indicator o electrode cw TBA + and in the bulk of the organic phase cTBA +. When droplets start impacting the indicator electrode, a precipitate is formed in the aqueous phase between TBA+(w) and TPB−(w). The
Fig. 3. OCP experiments (explained within the text) reveal that by changing the organic phase in the TPB−/NB system to 1,2-DCE results in a droplet which is unstable in the organic phase. A comparison between NB and 1,2-DCE using the same aqueous phase and organic phase electrolytes comprising a DFI of 20 mM NaTPB(w) and 10 mm TBATPB(o) respectively.
93
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Fig. 4. OCP experiments with two model experimental systems namely (A) the TPB−/NB system and (B) the TPhAs+/1,2-DCE system. Experimental cells/arrangement are given within the text. A concentrated aliquot of aqueous electrolyte was injected at 60s (in all cases) into the upper phase of the DFI. The concentration of the aqueous phase after injection was, from left to right, in (A) 32,16,8,4,2,1 and 0.5 mM NaTPB, and from left to right in (B), 16,8,6,4 and 2 mM TPhAsCl. Potentials quoted are referred to the TBA+(A) and TMA+(B) potential scales respectively. Black dots on (A) indicate the position of the characteristic time (tC) for each concentration (details within theoretical section).
product of tC and the concentration of TPB− in the aqueous phase of the DFI, is nearly constant with a normalised standard deviation of 0.078. We have assumed proportionality between the concentration of TPB− in the aqueous phase of the DFI, and the overall concentration of TPB− found within the organic phase as water droplets, therefore the resultant flux of TPB− to the surface of the indicator electrode must also be proportional. Moreover, the flux of TPB− into the surface of the indicator electrode determines the flux of TBA+ on the aqueous side of the indicator electrode (see rhs of Eq. (3)). This further implies that the integral in Eq. (3) is proportional to time. In this way the product of concentration and tC can be taken as a constant. This requirement implies that g(t) in Eq. (3) is proportional to t . Namely, if one lets g (t ) = P t then Eq. (3) becomes
creation of precipitate causes a decrease in TBA+ concentration on the aqueous side of the indicator electrode. It therefore follows from Eq. (1).
∂E 1 F = where f = w w ∂c TBA f c TBA RT + +
(2)
As cTBA+ → 0, a sharp rise in potential is observed for all concentrations of NaTPB injected into the upper phase of the DFI. The time at which this appreciable potential increase occurs allows one to define a characteristic time (tC) for each injected concentration (Fig. 4). The inflection point of the OCP sigmoid provides a convenient location to define this measure. We make the additional reasonable assumption that the concentration of TPB− injected into the aqueous phase of the DFI and the concentration of TPB− present in water droplets in the organic phase as a result thereof, are proportional to one another for a given time. For this analysis we will use the following expression in conjunction with axis definition 1 (Supplementary section) w
c (0, t ) = c∗ −
1 πD
∫0
t
g (τ ) dτ t−τ
c (0, t ) = c∗ −
tC , n = (3)
∂c ∂ 2c =D 2 ∂t ∂x
∂c ∂x
(8)
where index n relates to the n concentration (injected into the DFI), and Pn is a constant which depends on the nth injected concentration. As the organic phase is homogenised by slow stirring, one may consider solution convection at the surface of the indicator electrode to be one of laminar boundary layer flow. As such, the actual flux of TPB− in water droplets to the indicator electrode (at y = 0) could be roughly described by the equation [20,21] (axis definition 2 in Supplementary section)
(4)
(5)
j=D
where D is the diffusion coefficient and that the amount of diffusing substance crossing unit area of the surface at x = 0 per unit time or the diffusion flux g(t) at x = 0 is given by Ficks first law
g (t ) = D
2 D c∗ 1 π Pn th
Further it is assumed that the concentration c is governed by
x > 0:
(7)
and tC therefore becomes
which is valid for the one dimensional diffusion of an ion. Concentration c(x, t) is the concentration at distance x from a plane at time t. Initially, (t = 0), the concentration is assumed to be constant and at
x > 0: c (x , 0) = c∗
π Pt 2 D
∂c v WD‐ = 0.33872Dc TPB 3 ∂y Dd
⎛ U ⎜ 1 vz ⎜ ⎜3 1 − ⎝
()
3
h 4 z
⎞ ⎟ ⎟ ⎟ ⎠
(9)
This expression describes a stationary diffusion flux caused by laminar boundary flow over a flat plane (Q, see axis definition 2) for an electroactive surface which begins at some distance h in the direction of flow from the beginning of the flat plane (z). Here v is the kinematic viscosity of the nitrobenzene phase, U is the linear flow rate out of the boundary layer and y is the perpendicular distance to the plane. D is the apparent diffusion coefficient of water droplets containing TPB− in the organic phase and cTPB−WD is the concentration of TPB− in the organic phase as water droplets. As the indicator electrode has circular geometry (as per typical Luggin capillary), Eq. (9) is not strictly applicable. However, this approach forms the basis for a more geometrically applicable approximation. Therefore, the flux per unit area of the circular Luggin capillary at y = 0 can be more properly expressed as
(6)
Eq. (3) can be derived via Laplace transformation with respect to time (t) subject to conditions given in Eqs. (4),(5) and (6) [19]. Below we apply Eq. (3) to the water side of the indicator interface of the measuring cell. In that case x is the distance from the interface at which the concentration of TBA+ is c(x, t) and the diffusion flux of TBA+ to the interface is g(t). Assuming c(0, t) to be the concentration of TBA+ on the aqueous side of the indicator electrode, we further define the characteristic time (tC) as the time at which TBA+ or c(0, tC) = 0. Simple analysis of the experimental OCP vs time data (Fig. 4) leads to the conclusion that tC depends on the concentration of TPB− in the aqueous phase of the DFI as a hyperbolic function. In other words, the 94
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al. −6
⎛ 1.827 × 10 ⎞ −7 ⎜ 9.657 × 10 ⎟ 4.573 × 10−7 ⎟ ⎜ P= ⎜ 2.451 × 10−7 ⎟ ⎜1.340 × 10−7 ⎟ ⎜ −8 ⎟ ⎝ 5.191 × 10 ⎠
(14) − 10
W
where we used DTBA+ = 5.3 × 10 It can be shown that
PNB, n = 0.33872D 3
v D
2
m ·s
1 U 50 WD c − (tC , n )− 2 v m TPB n
−1
.
(15)
where PNB has been derived from Eqs. (10) and (12). We estimate U to be 0.14 m/s and used 2.27 × 10− 4m2 as the area of the DFI, and 5 × 10− 6m3as the volume of the NB phase, D as 7.7 × 10− 10m2 · s− 1, and 1.537 × 10− 6m2 · s− 1 for the kinematic viscosity of the NB phase. 3 We thus obtain a vector of values for PNB (in m−2·s− 2 ). −6
Fig. 5. The effect of increasing of the DFI aqueous phase viscosity on tC. (a) DI water layered on top of the NB phase, an aliquot of concentrated NaTPB solution is injected slowly into the DI water at 60s. (b) DI water (with 50% w/v sucrose) and a concentrated aliquot of NaTPB are mixed beforehand. At 60 s the entire solution is layered on top of the NB containing electrolyte. Final concentrations are 32 mM NaTPB (in all cases) + 34% (w/v) sucrose (after injection). Organic phase contained 10 mM TBATPB in all cases. tC is marked by a thick black dot.
1 π
d 2 2
()
WD ∫ D ∂∂yc dA = 0.33872DcTPB
−3
v Dd
U M v
PNB
(10)
32
⎛31.997 ⎞ ⎛32 ⎞ 16.914 16 ⎟ ⎜ ⎜ ⎟ P 8.01 conc = ⎜ 8 ⎟ and P0 = ⎜ ⎟ 4.292 4 ⎜ 2.347 ⎟ ⎜2⎟ 32 ⎜ ⎟ ⎟ ⎜ ⎝1⎠ ⎝ 0.909 ⎠
2 A W − cn (DTPB )t π V
5. Discussion Our analysis can explain the experimental data reasonably well. Our main finding was that the flux of TPB− from the DFI (within droplets) into the organic phase is likely controlled by diffusion. By increasing the viscosity of the aqueous phase in the DFI, using sucrose, an interesting effect is observed (Fig. 5a and b). The increase in viscosity appears to slow down droplet production at the DFI. This is reflected by a delay in tC. Our theoretical framework predicts an approximate 2.9 times increase in tC due to the viscosity increase. Experimentally, the ratio observed is closer to an approximate 4.4 times. More work is required to determine the physical significance of this difference. A cartoon (Fig. 6) depicts interfacial processes which cause sigmoidal OCP behaviour in the case of the TPhAs+/1,2-DCE system. At equilibrium, before injection (Fig. 6A), the interfacial potential is maintained by an equal concentration of TMA+ in the organic phase and the aqueous phase at the tip of the indicator electrode. Moreover, before injection of electrolyte into the aqueous phase of the DFI, organic phase ions such as TPhAs+, DCC− and TPB− do not contribute significantly to the interfacial potential. After injection of electrolyte into the DFI (Fig. 6B), water-in-oil droplets are formed and become homogenised within the bulk of the organic phase by gentle stirring. Droplets containing TPhAsCl begin to impact the indicator electrode and deposit TPhAs+ and Cl− on the aqueous side of the indicator electrode. TPhAs+ is now present on both sides of the interface and its concentration increases with time. A
(11)
(12)
where the interfacial area of the DFI is A, V is the volume of the nitrobenzene phase and DTPB−W is the diffusion coefficient of TPB− in the aqueous phase of the DFI. From experiment we are able to construct two vectors representing both tC (seconds) and the corresponding initial concentrations (in mM) of TPB− injected into the aqueous phase of the DFI.
⎛142.2 ⎞ ⎛32 ⎞ 269 16 ⎟ ⎜ ⎜ ⎟ 568 8 = ; tC = ⎜ conc ⎟ ⎜ ⎟ 1060 4 ⎜ 1939 ⎟ ⎜2⎟ ⎜ ⎟ ⎟ ⎜ 5004 ⎝1⎠ ⎠ ⎝
(17)
Concluding, this analysis indicates that the assumption of proportionality between the concentration of TPB− injected into the DFI, and the concentration of TPB− appearing in the organic phase in water droplets as a result was reasonable.
The resultant concentration of TPB−, in water droplets in the nitrobenzene phase after time t is thus WD −, n = c TPB
(16)
We note the close correspondence between the two vectors (14) and (16). Agreement between vector values of P and PNB were obtained by adjusting the diffusion coefficient D. The standard deviation of the difference P − PNB divided by the mean value of P was found to be 0.039. Also the ratio of PP0 is comparable to the ratio of concentrations.
where d is the diameter of the active area of the indicator electrode and 1 M is a numerical factor estimated to be 50m− 2 to allow for the difference in geometries between that of an idealised flat plate and that of the circular Luggin capillary. It was shown above, that the flux of TPB− to the indicator electrode is proportional to t , which applies to the rhs of Eq. (10). Therefore cTPB−WD also has to be proportional to t . Let us now assume that we have a concentration cn of TPB− in the aqueous phase of the DFI at t = 0. Then via diffusion, the maximum concentration of TPB−, after a time t, which can be transferred per unit area of the DFI into the organic phase can be given by
2 W − cn (DTPB )t π
⎛ 1.87 × 10 ⎞ −7 ⎜ 9.35 × 10 ⎟ 4.675 × 10−7 ⎟ ⎜ = ⎜ 2.337 × 10−7 ⎟ ⎜1.169 × 10−7 ⎟ ⎜ −8 ⎟ ⎝ 5.844 × 10 ⎠
(13)
Using relation (8) we can now define a corresponding vector Pn 3 (m−2·s− 2 ) 95
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Fig. 6. A cartoon depiction of interfacial processes which alter the boundary potential at the indicator electrode/organic phase interface within the TPhAs+/1,2-DCE system before and after injection of electrolyte into the upper phase of the DFI. Panel (A) indicates conditions at the indicator electrode interface before injection and panel (B) indicates conditions at the indicator electrode interface after injection. The mechanism illustrated in panel B gives rise to the sigmoidal shape of the OCP in Fig. 1B. The organic phase is 1,2-DCE. A more detailed explanation is given within the text.
was qualitatively explained and operates in an altogether different manner than that of the TPB−/NB system at the level of the indicator electrode. The general behaviour of the two systems at the DFI level are assumed, at this time, to be similar.
perturbation in the interfacial potential is created which results in the reaction TMA+(w) → TMA+(o). This reaction decreases the concentration of TMA+(w). To maintain electroneutrality, the transfer TMA+(w) → TMA+(o) is accompanied by TPhAs+(o) → TPhAs+(w) which further increases the concentration of TPhAs+ in the indicator electrode. The rapid increase in the concentration of TPhAs+(w) and a concomitant decrease in TMA+(w) leads to a negative potential shift toward the standard transfer potential of TPhAs+. Clearly, this process is different from that of the TPB−/NB system at the level of the indicator electrode and requires separate analysis. It is however not at this time the opinion of the authors that the general behaviour of the DFI within the TPhAs+/1,2-DCE system would be vastly different from that of the presently analysed DFI within the TPB−/NB system. Formation of droplets at the DFI may lead to concentration fluctuations. We have indeed observed such fluctuations/drifts in potential at the DFI in very recent unpublished experiments and hope to report our findings more fully in due course. Proposed concentration fluctuations appear similar to those observed during previous fluctuation analysis experiments [22].
Acknowledgements We gratefully acknowledge funding from the Czech Science Foundation (project number 17-09980S). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jelechem.2017.10.027. References [1] K. Tauer, S. Kozempel, G. Rother, The interface engine: experimental consequences, J. Colloid Interface Sci. 312 (2) (2007) 432–438. [2] C. Solans, D. Morales, M. Homs, Spontaneous emulsification, Curr. Opin. Colloid Interface Sci. 22 (2016) 88–93. [3] K. Aoki, M. Li, J. Chen, T. Nishiumi, Spontaneous emulsification at oil-water interface by tetraalkylammonium chloride, Electrochem. Commun. 11 (2) (2009) 239–241. [4] Y. Kitazumi, T. Kakiuchi, Electrochemical instability in liquid-liquid two-phase systems, Bull. Chem. Soc. Jpn. 84 (12) (2011) 1312–1320. [5] Y. Kitazumi, T. Kakiuchi, A model of the electrochemical instability at the liquid ∣ liquid interface based on the potential-dependent adsorption and Gouy's double layer theory, J. Electroanal. Chem. 648 (1) (2010) 8–14. [6] T. Kakiuchi, Electrochemical instability at liquid/liquid interfaces, in: H. Watarai, N. Teramae, T. Sawada (Eds.), Interfacial Nanochemistry: Molecular Science and Engineering at Liquid—Liquid Interfaces, Springer US, Boston, MA, 2005, pp. 155–170. [7] T. Kakiuchi, Avalanche transfer of charged particles across the electrochemical liquid | liquid interface, Electrochem. Commun. 2 (5) (2000) 317–321. [8] T. Kakiuchi, Electrochemical instability of the liquid ∣ liquid interface in the presence of ionic surfactant adsorption, J. Electroanal. Chem. 536 (1–2) (2002) 63–69. [9] M. Dupeyrat, E. Nakache, 205 - direct conversion of chemical energy into mechanical energy at an oil water Interface, Bioelectrochem. Bioenerg. 5 (1) (1978) 134–141. [10] T. Kakiuchi, M. Chiba, N. Sezaki, M. Nakagawa, Cyclic voltammetry of the transfer of anionic surfactant across the liquid–liquid interface manifests electrochemical instability, Electrochem. Commun. 4 (9) (2002) 701–704. [11] K. Sadakane, H. Seto, H. Endo, M. Shibayama, A periodic structure in a mixture of D2O/3-Methylpyridine/NaBPh4 induced by solvation effect, J. Phys. Soc. Jpn. 76 (11) (2007) 113602. [12] K. Sadakane, A. Onuki, K. Nishida, S. Koizumi, H. Seto, Multilamellar structures induced by hydrophilic and hydrophobic ions added to a binary mixture of ${\mathbf{D}}_{2}\mathbf{O}$ and 3-methylpyridine, Phys. Rev. Lett. 103 (16) (2009) 167803. [13] D. Michler, N. Shahidzadeh, M. Westbroek, R. van Roij, D. Bonn, Are antagonistic salts surfactants? Langmuir 31 (3) (2015) 906–911. [14] A. Onuki, S. Yabunaka, T. Araki, R. Okamoto, Structure formation due to antagonistic salts, Curr. Opin. Colloid Interface Sci. 22 (2016) 59–64. [15] M. Witala, S. Lages, K. Nygard, Mesoscale ordering in binary aqueous solvents induced by ion size asymmetry, Soft Matter 12 (21) (2016) 4778–4782. [16] K. Kontturi, L. Murtomäki, J.A. Manzanares, Ionic Transport Processes: In Electrochemistry and Membrane Science, OUP Oxford, 2008.
6. Conclusions We have found that ITIES, with a common-ion distributed between both phases, can undergo vigorous and spontaneous emulsification. The rate of emulsification is concentration dependent. Theoretical analysis of the TPB−/NB system indicates the rate at which TPB− enters the organic phase, as a water-in-oil emulsion, may be controlled by diffusion. By increasing the viscosity of the DFI aqueous phase, our theoretical framework predicts an increase in tC. This increase is experimentally confirmed, although the experimental shift in tC is larger than that expected by theory. Two well-defined experimental platforms have been developed for further study. The spontaneous interfacial emulsification process can produce stable nano- to micron-sized water-in-oil droplets solely within the organic phase. In some cases, by changing the organic solvent, yet by keeping electrolytes in each phase exactly the same, unstable waterin-oil droplets are formed. The reported emulsification process, when stable emulsions are produced, can be monitored indirectly using transient open circuit potential measurements of an electrochemical cell, initially at equilibrium, located within the organic phase. Droplet stability in bulk is caused by the presence of a common-ion both inside the droplet and outside the droplet in bulk organic solution. Impact of water-in-oil droplets with the indicator electrode can polarise the indicator electrode through a large potential range. Sigmoidal OCP behaviour observed in the TPhAs+/1,2-DCE system 96
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Chem. 36 (4) (1964) 706–723. [20] V.G. Levich, Physicochemical hydrodynamics: (by) Veniamin G. Levich. Transl. by Scripta Technica, Inc., Prentice-Hall 1962. [21] H. Schlichting, K. Gersten, E. Krause, H. Oertel, K. Mayes, Boundary-Layer Theory, Springer, 1960. [22] V. Mareček, A. Lhotský, S. Račinský, Fluctuation analysis and faradaic impedance at micro liquid/liquid interface—II, Electrochim. Acta 40 (18) (1995) 2909–2912.
[17] O. Popovych, O.P.O. Popovych, Solubility Data Series, Tetraphenylborates, Vol. 18 Pergamon Press, Oxford, UK, 1981. [18] B.V. Derjaguin, N.V. Churaev, V.M. Muller, The Derjaguin—Landau—Verwey—Overbeek (DLVO) Theory of Stability of Lyophobic Colloids, Surface Forces, Springer US, Boston, MA, 1987, pp. 293–310. [19] R.S. Nicholson, I. Shain, Theory of stationary electrode polarography. Single scan and cyclic methods applied to reversible, irreversible, and kinetic systems, Anal.
97
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
Spontaneous emulsification at surfactantless liquid/liquid interfaces Barry R. Silver, Karel Holub, Vladimir Mareček
MARK
⁎
Department of Biophysical Chemistry, J. Heyrovsky Institute of Physical Chemistry, Dolejškova 2155/3, 182 23 Prague 8, Czech Republic
A R T I C L E I N F O
A B S T R A C T
Keywords: Spontaneous emulsification ITIES Water-in-oil Open circuit potential
Liquid/liquid interfaces can undergo spontaneous emulsification when a common-ion is distributed between both phases. By careful selection of electrolyte and solvent, stable nano-sized water-in-oil emulsions containing electrolyte can be produced. The emulsification process can be monitored indirectly via transient open circuit potential measurements using an electrochemical measuring cell, initially poised at equilibrium, within the organic phase. Theoretical analysis of experimental data indicates that the interfacial emulsification process is probably controlled by diffusion.
1. Introduction Spontaneous emulsification can occur when two immiscible liquids, out of equilibrium, are placed in physical contact [1,2]. Spontaneous emulsification can proceed without the aid of external energy inputs and without surfactant. This is because chemical potential energy gradients, in liquid-liquid systems which are out of equilibrium, are large enough to drive interfacial emulsification [2]. Spontaneous emulsification is currently an active area of research, although as noted by Aoki [3], a large proportion of previous work dealing with surfactantless spontaneous emulsification at immiscible liquid-liquid interfaces does not emphasise an electrochemical aspect to any great extent. Conventional ITIES (interface between two immiscible electrolyte solutions) may also be able to undergo spontaneous emulsification without external polarization [3]. Much of the previous work in this regard has however, dealt almost exclusively, with spontaneous oscillations in interfacial tension and general interfacial instability in the presence of added surfactant [4–9]. Under externally polarised conditions though, interfacial emulsification and interfacial instabilities in the presence of an ionic surfactant, arise from the formation of an ‘instability window’ [8]. This ‘window’ is caused by potential-dependent adsorption and partitioning of surfactant at the ITIES [10]. Within a conventional ITIES-type experimental arrangement, both phases contain electrolyte. Electrolytes used within the aqueous phase are hydrophilic salts such as lithium chloride. The organic phase (usually nitrobenzene or 1, 2-dichloroethane) contains hydrophobic electrolyte, such as tetrabutylammonium tetraphenylborate for example (TBATPB). A recent small angle neutron scattering (SANS) study by Sadakane et al. [11] demonstrated that in the presence of tetraphenylborate the water/3-methylpyridine interface can produce
⁎
periodic mesoscopic structures [12]. Water soluble tetraphenylborate salts (such as NaTPB) are sometimes termed as “antagonistic” [13,14]. In other words, the salt is comprised of both strongly hydrophilic (Na+) and strongly hydrophobic (TPB−) ions. In the presence of a composition heterogeneity, such as that found at an ITIES, antagonistic salts can lead to the formation of mesophase structures with charge density waves and decreases in interfacial tension [14]. Similar structures have also been recently reported at the water/2,6-dimethylpyridine interface with the addition of a tetra alkyl-ammonium bromide [15]. In this present article we demonstrate that ITIES, possessing a common-ion distributed between both phases, are able to undergo spontaneous emulsification and form stable water-in-oil nanoemulsions. This effect is exemplified using two model systems. Methodology detailed herein allows one to monitor the emulsification process indirectly using an electrochemical monitoring cell located within the organic phase. A theoretical framework is provided to analyse results of one of the experimental systems. 2. Materials and methods Tetrabutylammonium tetraphenylborate (TBATPB, > 99%), tetramethylammonium chloride (TMACl, > 98%), tetraphenylarsonium chloride (TPhAsCl, polarography grade), sodium tetraphenylborate (NaTPB, 99.5%) and agar-agar were all supplied by Fluka Analytical, Czech Republic. Tetrabutylammonium chloride (TBACl, > 97%) was supplied by Sigma Aldrich, Czech Republic. Tetramethylammonium tetraphenylborate (TMATPB) was prepared by metathesis. Tetraphenylarsonium bis-1,2-dicarbollylcobaltate (TPhAsDCC) was prepared at the Institute of Inorganic Chemistry ASCR, v.v.i. Chemicals and was used as received. Highly purified, deionised water (DI,
Corresponding author. E-mail address: [email protected] (V. Mareček).
http://dx.doi.org/10.1016/j.jelechem.2017.10.027 Received 1 August 2017; Received in revised form 9 October 2017; Accepted 12 October 2017 Available online 14 October 2017 1572-6657/ © 2017 Elsevier B.V. All rights reserved.
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
conductance: < 0.1 μS cm− 1, GORO system, Czech Republic) was used to prepare all aqueous solutions. 1, 2-dichloroethane (1,2-DCE, 99%) and nitrobenzene (NB, 99.5%) were supplied by Fluka Analytical (Czech Republic) and were used to prepare organic solutions.
Electrolyte in the indicator electrode was gelled to the tip with 2% (w/ v) agar for mechanical stability. Water-in-oil droplets impact the indicator electrode at the very tip of the capillary. The reference interface, in contrast, is situated well inside the Luggin capillary, to protect it from the droplet impact. (Fig. 1). Two experimental systems are used for experiments, namely: The TPB−/NB system DFI: x mM NaTPB(w)//10 mM TBATPB(NB). MC: Ag/AgCl/10 mM TBACl (in 2% agar)(w)//10 mM TBATPB(NB)// 10 mM TBACl(w)/AgCl/AgCl. The TPhAs+/1,2-DCE system DFI: x mM TPhAsCl(w)//10 mM TPhAsDCC (1,2-DCE) + 5 mM TMATPB (1,2-DCE). MC: Ag/AgCl/5 mM TMACl (in 2% agar)(w)//10 mM TPhAsDCC(1,2DCE) + 5 mM TMATPB(1,2-DCE)//5 mM TMACl(w)/AgCl/Ag. Two Ag/AgCl wires were used as electrodes. The OCP of the MC is continuously monitored/recorded. A concentrated aliquot (normally 100 mM) of aqueous phase electrolyte was injected at 60s (in all cases) after the initiation of OCP monitoring. The OCP was monitored using a CHI 660c potentiostat (CHI Instruments).
2.1. Glass plate experiments Equal aliquots (as much as 75 μl) of an aqueous phase and an organic phase (of varied compositions) were placed in close proximity to one another at the centre of a clean glass plate. The glass plate was of rectangular geometry and was approximately 1.5 mm thick. Glass cover slips of approximately 100 μm thickness were placed in each corner. A second glass plate (of the same thickness and geometry) was placed on top of the first which allows the two phases to contact. This ‘sandwich’ arrangement formed a thin-layer cell and allowed oil/water interfaces to be readily viewed using optical microscopy. Evaporation of the two phases was negligible in this arrangement over the course of a typical experiment. 2.2. Open circuit potential (OCP) monitoring experiments A cartoon depiction of the glass apparatus used in OCP monitoring experiments is illustrated in Fig. 1. The glass apparatus comprises two electrochemical cells. The first cell is formed by an aqueous phase and an organic phase. These two phase form a large area interface. This water/oil interface is the droplet forming interface (DFI). The cell is initially filled with organic phase (5 ml), upon which, deionized water (4 ml) is slowly layered. At the beginning of the OCP measurement, a concentrated aliquot of aqueous electrolyte is carefully injected into the distilled water yielding the required concentration of electrolyte in the aqueous phase. The injection initiates the spontaneous emulsification process. Water-in-oil droplets are homogenised within the organic phase by slow stirring, 2–3 revolutions per second, with a magnetic stir bar. The second electrochemical cell (MC), consisting of two Luggin capillaries (inner diameter 1.5 mm) placed in the organic phase, is used to monitor the emulsification process. The upper Luggin capillary with the reference interface inside the capillary serves as a reference electrode. The lower Luggin capillary serves as an indicator electrode.
3. Results 3.1. Glass plate experiments A vigorous and spontaneous interfacial emulsification process is observed at the ITIES shortly after placing the two phases in physical contact (both TPB−/NB and TPhAs+/1,2-DCE systems). The sub-micron sized water-in-oil droplets (see DLS results, Supplementary section) formed exclusively on the oil side of the ITIES, makes proper resolution impossible using optical microscopy. Although one can, with the aid of microscopy, observe masses of sub-micron, spherical and translucent entities originating from the interface and pervade the bulk of the organic phase. Some larger micron-sized droplets are observed in bulk after a period of a few minutes. Fig. 2 illustrates the extent of the spontaneous emulsification process within the TPB−/NB system. No obvious signs of precipitate are observable. Droplets remain intact for long periods and for relatively long distances away from the liquid/liquid interface. This indicates droplet stability. Droplet stability allowed extensive droplet accumulation within the volume of the organic phase (Fig. 2 and Supplementary section). Within the TPB−/NB system, droplet accumulation results in organic phase turbidity (with slow stirring of the organic phase) over a period of a few hours (Fig. 2). Although a similarly vigorous and spontaneous interfacial emulsification process was observed for the TPhAs+/1,2-DCE system, organic phase turbidity was not as extensive as that of the TPB/NB system over the same period of time. Standard transfer potentials are similar for the transfer of TPB− from water to nitrobenzene and for its transfer from water to 1,2-dichloroethane [16]. Hence, by using the same electrolytes in each phase, but changing only the solvent (from NB to 1,2-DCE), similar Galvani potentials across the ITIES are expected. Both a TPB−/NB system and a TPB−/1,2-DCE system (using the same electrolytes), produce vigorous spontaneous emulsification at the interface. Although, within the TPB−/1,2-DCE system, water droplets formed within the oil phase are observed to clearly ‘pop’ and coalesce on the surface of the glass after a few minutes and at a short distance away from the interface (Figs. 1S and 2S). Furthermore, no accumulation in bulk, and thus no observable turbidity is evident after 12 h of stirring (as in Fig. 2). Rather, glass walls of the vial in the case of TPB−/1,2-DCE system, took on a ‘frosted’
Fig. 1. Cartoon depiction of the glass apparatus used for OCP monitoring experiments. The dashed line within the upper Luggin capillary (indicated by an arrow) shows the position of the liquid/liquid reference interface. Shading in both capillaries indicates an aqueous phase. The dark ellipsoidal shape at the bottom of the organic phase represents the small Teflon stir bar.
92
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Fig. 2. A) The extent of the spontaneous emulsification a short time after contacting the two phases within the glass plate arrangement. Copious amounts of water-in-oil droplets are formed in the NB phase which are hard to resolve using optical microscopy. No droplets were observed within the aqueous phase. B) Equal aliquots of both aqueous and organic phase (3 ml) are layered on top of one another in a small glass vial. A small Teflon-coated magnetic stir bar allows gentle stirring/homogenization of the organic phase. The extreme turbidity of the organic phase demonstrates the extent of droplet accumulation and further indicates droplet stability after a period of 12 h. The aqueous phase and the organic phase shown in A and B comprised 10 mM NaTPB in water and 10 mM TBATPB in nitrobenzene respectively. Scale bar in (A) is 10 μm.
appearance. ‘Frosting’ is caused by mass coalescence of water droplets on the interior surface of the glass within the volume of the organic phase (Figs. 1S and 2S). This indicates that the TPB−/NB system produces water-in-oil emulsions which are stable in the bulk of the organic phase and that the TPB−/1,2-DCE system does not. This observation is further strengthened using OCP measurements (Fig. 3). The OCP rise in the case of TPB−/1,2-DCE is only 50 mV after 10,000 s. This 50 mV OCP rise is in comparison to that of the 600 mV rise in approximately 500 s in the case of NB. The extremely low increase in OCP for the TPB−/1,2-DCE system further indicates that droplets formed at this DFI do not have the required stability in the bulk of the organic phase to cause significant changes in potential at the MC.
Spontaneously formed water-in-oil droplets contain aqueous phase electrolyte. A bright white precipitate was observed at the tip of the indicator electrode during experiments within the TPB−/NB system. Aqueous droplets containing NaTPB deposit at the indicator electrode and form an insoluble precipitate. The solubility of TBATPB in water is extremely low (0.259 μM, [17]). Precipitate is not observed at the tip of the indicator electrode within the TPhAs+/1,2-DCE system. This is because water droplets containing water soluble TPhAsCl depositing at the indicator electrode do not form a precipitate with water soluble TMACl. Another important observation which further emphasises that droplets produced at the DFI contain aqueous phase electrolyte, is that droplets are stable within the organic phase for long periods (Fig. 2). Extensive turbidity can be maintained within the organic phase (TPB−/ NB system) for well over a week (as per Fig. 2). This stability can be a product of the common-ion present within the droplet and within the bulk organic phase outside the droplet. The common-ion may cause a surface potential to be created on the droplet, which in turn could lead to an electrostatic repulsion between individual droplets [18]. Droplets should contain equal amounts of cations and anions and are therefore electroneutral. The potential of the indicator electrode (at t = 0, and just before aqueous electrolyte injection) is maintained by an equal concentration of TBA+ on either side of the interface in the case of the TPB−/NB system, and by an equal concentration of TMA+ in the case of the TPhAs+/1,2-DCE system. As water droplets containing electrolyte, start to impact the indicator electrode, the potential becomes perturbed from that of equilibrium. Droplet impact is therefore the first step in a process which alters the OCP and ultimately leads to the distinct sigmoidal shape.
3.2. OCP monitoring experiments All OCP voltage-time curves, in those experimental systems which produce stable water-in-oil droplets in the bulk of the organic phase, were sigmoidal in shape (Fig. 3 and Fig. 4). The sigmoid has three distinct regions. The first region begins at a condition of equilibrium before aqueous electrolyte injection at 0 V. After injection, a gentle rise in OCP from the equilibrium potential is observed due to droplet collision with the indicator electrode. The gentle rise gives way to a very steep rise, which tends toward or actually reaches a constant OCP. At the highest concentration of aqueous electrolyte, the final value of the OCP quickly approaches that of standard ion transfer potential of the common-ion initially distributed across the DFI.
4. Theoretical For reasons of brevity, our analysis herein considers only the TPB−/ NB system. As we monitor the DFI indirectly by means of the MC located within the organic phase, any processes occurring at the indicator electrode would be a direct function of, and in response to, processes occurring at the DFI. In this way one is able to make inference about the DFI through interpretation of OCP changes at the indicator electrode. The potential of the monitoring cell can be written as
E=
RT ⎛ c oTBA + ⎞ ln ⎜ w ⎟ nF ⎝ c TBA +⎠
(1)
The initial potential at the indicator electrode is maintained by an equal concentration of TBA+ on the aqueous side of the indicator o electrode cw TBA + and in the bulk of the organic phase cTBA +. When droplets start impacting the indicator electrode, a precipitate is formed in the aqueous phase between TBA+(w) and TPB−(w). The
Fig. 3. OCP experiments (explained within the text) reveal that by changing the organic phase in the TPB−/NB system to 1,2-DCE results in a droplet which is unstable in the organic phase. A comparison between NB and 1,2-DCE using the same aqueous phase and organic phase electrolytes comprising a DFI of 20 mM NaTPB(w) and 10 mm TBATPB(o) respectively.
93
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Fig. 4. OCP experiments with two model experimental systems namely (A) the TPB−/NB system and (B) the TPhAs+/1,2-DCE system. Experimental cells/arrangement are given within the text. A concentrated aliquot of aqueous electrolyte was injected at 60s (in all cases) into the upper phase of the DFI. The concentration of the aqueous phase after injection was, from left to right, in (A) 32,16,8,4,2,1 and 0.5 mM NaTPB, and from left to right in (B), 16,8,6,4 and 2 mM TPhAsCl. Potentials quoted are referred to the TBA+(A) and TMA+(B) potential scales respectively. Black dots on (A) indicate the position of the characteristic time (tC) for each concentration (details within theoretical section).
product of tC and the concentration of TPB− in the aqueous phase of the DFI, is nearly constant with a normalised standard deviation of 0.078. We have assumed proportionality between the concentration of TPB− in the aqueous phase of the DFI, and the overall concentration of TPB− found within the organic phase as water droplets, therefore the resultant flux of TPB− to the surface of the indicator electrode must also be proportional. Moreover, the flux of TPB− into the surface of the indicator electrode determines the flux of TBA+ on the aqueous side of the indicator electrode (see rhs of Eq. (3)). This further implies that the integral in Eq. (3) is proportional to time. In this way the product of concentration and tC can be taken as a constant. This requirement implies that g(t) in Eq. (3) is proportional to t . Namely, if one lets g (t ) = P t then Eq. (3) becomes
creation of precipitate causes a decrease in TBA+ concentration on the aqueous side of the indicator electrode. It therefore follows from Eq. (1).
∂E 1 F = where f = w w ∂c TBA f c TBA RT + +
(2)
As cTBA+ → 0, a sharp rise in potential is observed for all concentrations of NaTPB injected into the upper phase of the DFI. The time at which this appreciable potential increase occurs allows one to define a characteristic time (tC) for each injected concentration (Fig. 4). The inflection point of the OCP sigmoid provides a convenient location to define this measure. We make the additional reasonable assumption that the concentration of TPB− injected into the aqueous phase of the DFI and the concentration of TPB− present in water droplets in the organic phase as a result thereof, are proportional to one another for a given time. For this analysis we will use the following expression in conjunction with axis definition 1 (Supplementary section) w
c (0, t ) = c∗ −
1 πD
∫0
t
g (τ ) dτ t−τ
c (0, t ) = c∗ −
tC , n = (3)
∂c ∂ 2c =D 2 ∂t ∂x
∂c ∂x
(8)
where index n relates to the n concentration (injected into the DFI), and Pn is a constant which depends on the nth injected concentration. As the organic phase is homogenised by slow stirring, one may consider solution convection at the surface of the indicator electrode to be one of laminar boundary layer flow. As such, the actual flux of TPB− in water droplets to the indicator electrode (at y = 0) could be roughly described by the equation [20,21] (axis definition 2 in Supplementary section)
(4)
(5)
j=D
where D is the diffusion coefficient and that the amount of diffusing substance crossing unit area of the surface at x = 0 per unit time or the diffusion flux g(t) at x = 0 is given by Ficks first law
g (t ) = D
2 D c∗ 1 π Pn th
Further it is assumed that the concentration c is governed by
x > 0:
(7)
and tC therefore becomes
which is valid for the one dimensional diffusion of an ion. Concentration c(x, t) is the concentration at distance x from a plane at time t. Initially, (t = 0), the concentration is assumed to be constant and at
x > 0: c (x , 0) = c∗
π Pt 2 D
∂c v WD‐ = 0.33872Dc TPB 3 ∂y Dd
⎛ U ⎜ 1 vz ⎜ ⎜3 1 − ⎝
()
3
h 4 z
⎞ ⎟ ⎟ ⎟ ⎠
(9)
This expression describes a stationary diffusion flux caused by laminar boundary flow over a flat plane (Q, see axis definition 2) for an electroactive surface which begins at some distance h in the direction of flow from the beginning of the flat plane (z). Here v is the kinematic viscosity of the nitrobenzene phase, U is the linear flow rate out of the boundary layer and y is the perpendicular distance to the plane. D is the apparent diffusion coefficient of water droplets containing TPB− in the organic phase and cTPB−WD is the concentration of TPB− in the organic phase as water droplets. As the indicator electrode has circular geometry (as per typical Luggin capillary), Eq. (9) is not strictly applicable. However, this approach forms the basis for a more geometrically applicable approximation. Therefore, the flux per unit area of the circular Luggin capillary at y = 0 can be more properly expressed as
(6)
Eq. (3) can be derived via Laplace transformation with respect to time (t) subject to conditions given in Eqs. (4),(5) and (6) [19]. Below we apply Eq. (3) to the water side of the indicator interface of the measuring cell. In that case x is the distance from the interface at which the concentration of TBA+ is c(x, t) and the diffusion flux of TBA+ to the interface is g(t). Assuming c(0, t) to be the concentration of TBA+ on the aqueous side of the indicator electrode, we further define the characteristic time (tC) as the time at which TBA+ or c(0, tC) = 0. Simple analysis of the experimental OCP vs time data (Fig. 4) leads to the conclusion that tC depends on the concentration of TPB− in the aqueous phase of the DFI as a hyperbolic function. In other words, the 94
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al. −6
⎛ 1.827 × 10 ⎞ −7 ⎜ 9.657 × 10 ⎟ 4.573 × 10−7 ⎟ ⎜ P= ⎜ 2.451 × 10−7 ⎟ ⎜1.340 × 10−7 ⎟ ⎜ −8 ⎟ ⎝ 5.191 × 10 ⎠
(14) − 10
W
where we used DTBA+ = 5.3 × 10 It can be shown that
PNB, n = 0.33872D 3
v D
2
m ·s
1 U 50 WD c − (tC , n )− 2 v m TPB n
−1
.
(15)
where PNB has been derived from Eqs. (10) and (12). We estimate U to be 0.14 m/s and used 2.27 × 10− 4m2 as the area of the DFI, and 5 × 10− 6m3as the volume of the NB phase, D as 7.7 × 10− 10m2 · s− 1, and 1.537 × 10− 6m2 · s− 1 for the kinematic viscosity of the NB phase. 3 We thus obtain a vector of values for PNB (in m−2·s− 2 ). −6
Fig. 5. The effect of increasing of the DFI aqueous phase viscosity on tC. (a) DI water layered on top of the NB phase, an aliquot of concentrated NaTPB solution is injected slowly into the DI water at 60s. (b) DI water (with 50% w/v sucrose) and a concentrated aliquot of NaTPB are mixed beforehand. At 60 s the entire solution is layered on top of the NB containing electrolyte. Final concentrations are 32 mM NaTPB (in all cases) + 34% (w/v) sucrose (after injection). Organic phase contained 10 mM TBATPB in all cases. tC is marked by a thick black dot.
1 π
d 2 2
()
WD ∫ D ∂∂yc dA = 0.33872DcTPB
−3
v Dd
U M v
PNB
(10)
32
⎛31.997 ⎞ ⎛32 ⎞ 16.914 16 ⎟ ⎜ ⎜ ⎟ P 8.01 conc = ⎜ 8 ⎟ and P0 = ⎜ ⎟ 4.292 4 ⎜ 2.347 ⎟ ⎜2⎟ 32 ⎜ ⎟ ⎟ ⎜ ⎝1⎠ ⎝ 0.909 ⎠
2 A W − cn (DTPB )t π V
5. Discussion Our analysis can explain the experimental data reasonably well. Our main finding was that the flux of TPB− from the DFI (within droplets) into the organic phase is likely controlled by diffusion. By increasing the viscosity of the aqueous phase in the DFI, using sucrose, an interesting effect is observed (Fig. 5a and b). The increase in viscosity appears to slow down droplet production at the DFI. This is reflected by a delay in tC. Our theoretical framework predicts an approximate 2.9 times increase in tC due to the viscosity increase. Experimentally, the ratio observed is closer to an approximate 4.4 times. More work is required to determine the physical significance of this difference. A cartoon (Fig. 6) depicts interfacial processes which cause sigmoidal OCP behaviour in the case of the TPhAs+/1,2-DCE system. At equilibrium, before injection (Fig. 6A), the interfacial potential is maintained by an equal concentration of TMA+ in the organic phase and the aqueous phase at the tip of the indicator electrode. Moreover, before injection of electrolyte into the aqueous phase of the DFI, organic phase ions such as TPhAs+, DCC− and TPB− do not contribute significantly to the interfacial potential. After injection of electrolyte into the DFI (Fig. 6B), water-in-oil droplets are formed and become homogenised within the bulk of the organic phase by gentle stirring. Droplets containing TPhAsCl begin to impact the indicator electrode and deposit TPhAs+ and Cl− on the aqueous side of the indicator electrode. TPhAs+ is now present on both sides of the interface and its concentration increases with time. A
(11)
(12)
where the interfacial area of the DFI is A, V is the volume of the nitrobenzene phase and DTPB−W is the diffusion coefficient of TPB− in the aqueous phase of the DFI. From experiment we are able to construct two vectors representing both tC (seconds) and the corresponding initial concentrations (in mM) of TPB− injected into the aqueous phase of the DFI.
⎛142.2 ⎞ ⎛32 ⎞ 269 16 ⎟ ⎜ ⎜ ⎟ 568 8 = ; tC = ⎜ conc ⎟ ⎜ ⎟ 1060 4 ⎜ 1939 ⎟ ⎜2⎟ ⎜ ⎟ ⎟ ⎜ 5004 ⎝1⎠ ⎠ ⎝
(17)
Concluding, this analysis indicates that the assumption of proportionality between the concentration of TPB− injected into the DFI, and the concentration of TPB− appearing in the organic phase in water droplets as a result was reasonable.
The resultant concentration of TPB−, in water droplets in the nitrobenzene phase after time t is thus WD −, n = c TPB
(16)
We note the close correspondence between the two vectors (14) and (16). Agreement between vector values of P and PNB were obtained by adjusting the diffusion coefficient D. The standard deviation of the difference P − PNB divided by the mean value of P was found to be 0.039. Also the ratio of PP0 is comparable to the ratio of concentrations.
where d is the diameter of the active area of the indicator electrode and 1 M is a numerical factor estimated to be 50m− 2 to allow for the difference in geometries between that of an idealised flat plate and that of the circular Luggin capillary. It was shown above, that the flux of TPB− to the indicator electrode is proportional to t , which applies to the rhs of Eq. (10). Therefore cTPB−WD also has to be proportional to t . Let us now assume that we have a concentration cn of TPB− in the aqueous phase of the DFI at t = 0. Then via diffusion, the maximum concentration of TPB−, after a time t, which can be transferred per unit area of the DFI into the organic phase can be given by
2 W − cn (DTPB )t π
⎛ 1.87 × 10 ⎞ −7 ⎜ 9.35 × 10 ⎟ 4.675 × 10−7 ⎟ ⎜ = ⎜ 2.337 × 10−7 ⎟ ⎜1.169 × 10−7 ⎟ ⎜ −8 ⎟ ⎝ 5.844 × 10 ⎠
(13)
Using relation (8) we can now define a corresponding vector Pn 3 (m−2·s− 2 ) 95
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Fig. 6. A cartoon depiction of interfacial processes which alter the boundary potential at the indicator electrode/organic phase interface within the TPhAs+/1,2-DCE system before and after injection of electrolyte into the upper phase of the DFI. Panel (A) indicates conditions at the indicator electrode interface before injection and panel (B) indicates conditions at the indicator electrode interface after injection. The mechanism illustrated in panel B gives rise to the sigmoidal shape of the OCP in Fig. 1B. The organic phase is 1,2-DCE. A more detailed explanation is given within the text.
was qualitatively explained and operates in an altogether different manner than that of the TPB−/NB system at the level of the indicator electrode. The general behaviour of the two systems at the DFI level are assumed, at this time, to be similar.
perturbation in the interfacial potential is created which results in the reaction TMA+(w) → TMA+(o). This reaction decreases the concentration of TMA+(w). To maintain electroneutrality, the transfer TMA+(w) → TMA+(o) is accompanied by TPhAs+(o) → TPhAs+(w) which further increases the concentration of TPhAs+ in the indicator electrode. The rapid increase in the concentration of TPhAs+(w) and a concomitant decrease in TMA+(w) leads to a negative potential shift toward the standard transfer potential of TPhAs+. Clearly, this process is different from that of the TPB−/NB system at the level of the indicator electrode and requires separate analysis. It is however not at this time the opinion of the authors that the general behaviour of the DFI within the TPhAs+/1,2-DCE system would be vastly different from that of the presently analysed DFI within the TPB−/NB system. Formation of droplets at the DFI may lead to concentration fluctuations. We have indeed observed such fluctuations/drifts in potential at the DFI in very recent unpublished experiments and hope to report our findings more fully in due course. Proposed concentration fluctuations appear similar to those observed during previous fluctuation analysis experiments [22].
Acknowledgements We gratefully acknowledge funding from the Czech Science Foundation (project number 17-09980S). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jelechem.2017.10.027. References [1] K. Tauer, S. Kozempel, G. Rother, The interface engine: experimental consequences, J. Colloid Interface Sci. 312 (2) (2007) 432–438. [2] C. Solans, D. Morales, M. Homs, Spontaneous emulsification, Curr. Opin. Colloid Interface Sci. 22 (2016) 88–93. [3] K. Aoki, M. Li, J. Chen, T. Nishiumi, Spontaneous emulsification at oil-water interface by tetraalkylammonium chloride, Electrochem. Commun. 11 (2) (2009) 239–241. [4] Y. Kitazumi, T. Kakiuchi, Electrochemical instability in liquid-liquid two-phase systems, Bull. Chem. Soc. Jpn. 84 (12) (2011) 1312–1320. [5] Y. Kitazumi, T. Kakiuchi, A model of the electrochemical instability at the liquid ∣ liquid interface based on the potential-dependent adsorption and Gouy's double layer theory, J. Electroanal. Chem. 648 (1) (2010) 8–14. [6] T. Kakiuchi, Electrochemical instability at liquid/liquid interfaces, in: H. Watarai, N. Teramae, T. Sawada (Eds.), Interfacial Nanochemistry: Molecular Science and Engineering at Liquid—Liquid Interfaces, Springer US, Boston, MA, 2005, pp. 155–170. [7] T. Kakiuchi, Avalanche transfer of charged particles across the electrochemical liquid | liquid interface, Electrochem. Commun. 2 (5) (2000) 317–321. [8] T. Kakiuchi, Electrochemical instability of the liquid ∣ liquid interface in the presence of ionic surfactant adsorption, J. Electroanal. Chem. 536 (1–2) (2002) 63–69. [9] M. Dupeyrat, E. Nakache, 205 - direct conversion of chemical energy into mechanical energy at an oil water Interface, Bioelectrochem. Bioenerg. 5 (1) (1978) 134–141. [10] T. Kakiuchi, M. Chiba, N. Sezaki, M. Nakagawa, Cyclic voltammetry of the transfer of anionic surfactant across the liquid–liquid interface manifests electrochemical instability, Electrochem. Commun. 4 (9) (2002) 701–704. [11] K. Sadakane, H. Seto, H. Endo, M. Shibayama, A periodic structure in a mixture of D2O/3-Methylpyridine/NaBPh4 induced by solvation effect, J. Phys. Soc. Jpn. 76 (11) (2007) 113602. [12] K. Sadakane, A. Onuki, K. Nishida, S. Koizumi, H. Seto, Multilamellar structures induced by hydrophilic and hydrophobic ions added to a binary mixture of ${\mathbf{D}}_{2}\mathbf{O}$ and 3-methylpyridine, Phys. Rev. Lett. 103 (16) (2009) 167803. [13] D. Michler, N. Shahidzadeh, M. Westbroek, R. van Roij, D. Bonn, Are antagonistic salts surfactants? Langmuir 31 (3) (2015) 906–911. [14] A. Onuki, S. Yabunaka, T. Araki, R. Okamoto, Structure formation due to antagonistic salts, Curr. Opin. Colloid Interface Sci. 22 (2016) 59–64. [15] M. Witala, S. Lages, K. Nygard, Mesoscale ordering in binary aqueous solvents induced by ion size asymmetry, Soft Matter 12 (21) (2016) 4778–4782. [16] K. Kontturi, L. Murtomäki, J.A. Manzanares, Ionic Transport Processes: In Electrochemistry and Membrane Science, OUP Oxford, 2008.
6. Conclusions We have found that ITIES, with a common-ion distributed between both phases, can undergo vigorous and spontaneous emulsification. The rate of emulsification is concentration dependent. Theoretical analysis of the TPB−/NB system indicates the rate at which TPB− enters the organic phase, as a water-in-oil emulsion, may be controlled by diffusion. By increasing the viscosity of the DFI aqueous phase, our theoretical framework predicts an increase in tC. This increase is experimentally confirmed, although the experimental shift in tC is larger than that expected by theory. Two well-defined experimental platforms have been developed for further study. The spontaneous interfacial emulsification process can produce stable nano- to micron-sized water-in-oil droplets solely within the organic phase. In some cases, by changing the organic solvent, yet by keeping electrolytes in each phase exactly the same, unstable waterin-oil droplets are formed. The reported emulsification process, when stable emulsions are produced, can be monitored indirectly using transient open circuit potential measurements of an electrochemical cell, initially at equilibrium, located within the organic phase. Droplet stability in bulk is caused by the presence of a common-ion both inside the droplet and outside the droplet in bulk organic solution. Impact of water-in-oil droplets with the indicator electrode can polarise the indicator electrode through a large potential range. Sigmoidal OCP behaviour observed in the TPhAs+/1,2-DCE system 96
Journal of Electroanalytical Chemistry 805 (2017) 91–97
B.R. Silver et al.
Chem. 36 (4) (1964) 706–723. [20] V.G. Levich, Physicochemical hydrodynamics: (by) Veniamin G. Levich. Transl. by Scripta Technica, Inc., Prentice-Hall 1962. [21] H. Schlichting, K. Gersten, E. Krause, H. Oertel, K. Mayes, Boundary-Layer Theory, Springer, 1960. [22] V. Mareček, A. Lhotský, S. Račinský, Fluctuation analysis and faradaic impedance at micro liquid/liquid interface—II, Electrochim. Acta 40 (18) (1995) 2909–2912.
[17] O. Popovych, O.P.O. Popovych, Solubility Data Series, Tetraphenylborates, Vol. 18 Pergamon Press, Oxford, UK, 1981. [18] B.V. Derjaguin, N.V. Churaev, V.M. Muller, The Derjaguin—Landau—Verwey—Overbeek (DLVO) Theory of Stability of Lyophobic Colloids, Surface Forces, Springer US, Boston, MA, 1987, pp. 293–310. [19] R.S. Nicholson, I. Shain, Theory of stationary electrode polarography. Single scan and cyclic methods applied to reversible, irreversible, and kinetic systems, Anal.
97