Article pubs.acs.org/Langmuir

Interfacial Tension and Surface Elasticity of Carbon Black (CB) Covered Oil−Water Interface Kristin Conrad Powell and Anuj Chauhan* Department of Chemical Engineering, University of Florida, Gainesville, Florida 32611, United States ABSTRACT: Carboxyl-terminated carbon black (CB) particles have been proposed as readily available, biocompatible dispersants to stabilize oil-in-water emulsions after an oil spill. Since the reduction in interfacial tension and the increase in interfacial elasticity are the key parameters which relate interfacial mechanics to emulsion stability, this investigation explores the effect of CB adsorption and surface coverage on oil−water interfacial tension and elasticity. Flocculation of CB was explored as ionic strength was increased from 0 to 0.6 M, approximately the salinity of seawater. As salinity increases, CB aggregates into larger particles from 100 nm to 6 μm. The interfacial tension and dilational viscoelasticity were measured for two systems: a drop of a CB suspension in oil and an inverted oil drop in a CB suspension. For the arrangement of a CB suspension drop in oil, most of the CB settles and accumulates toward the bottom of the drop with only small surface adsorption and no appreciable effect is observed on the dynamic interfacial tension or the dilational viscoelasticity. On reversing the arrangement to an inverted oil drop in CB suspension and increasing the convection of the outer phase, the surface coverage increases considerably. The CB coverage becomes more uniform with higher convection with an average value of approximately 2.6 g/m2, which is representative of the coverage in Pickering emulsions stabilized by CB particles. The CB coverage decreases the surface tension from about 30 to 8.5 mN/m accompanied by an increase in the surface elasticity to 20.7 mN/m. The sharp contrast between the results from the CB suspension drop and the oil drop could be partially due to the effect of the wetting characteristics of the particles or due to the significant differences between the convection in the two cases.

1. INTRODUCTION The 2010 Deepwater Horizon oil spill in the Gulf of Mexico was the largest in United States history and resulted in the release of 4.9 million barrels of crude oil, 11 deaths, and large scale ecological damage.1 The remediation efforts following the oil spill pointed to several gaps in knowledge and a need for further research to design novel approaches to manage such oil spills.1 For effective oil spill mitigation, dispersants such as molecular surfactants can be used to emulsify oil droplets either at the source of a deep-sea release or at the surface. This emulsification extends the retention time of oil within the water column to increase biodegradation and decrease the probability of oil slicks and kill zones.2,3 While dispersants are a key tool in minimizing the damage from many oil spills and molecular surfactants are excellent emulsifiers, recent events demonstrate the need for further research. Large background dilution provided by seawater causes the surfactant to eventually leave the oil−water interfaces and partition into the water, destabilizing the drops. Additionally, due to high aqueous solubility, large volumes of molecular-based dispersant must be used. In Deepwater Horizon, approximately 1 100 000 US gallons (4200 m3) of Corexit surfactant was utilized as a dispersant, causing concern about potential toxicity.4 While the potential toxicity from currently used dispersants is still under exploration, it is clear that designing more effective biocompatible dispersants is critical for effective management of oil spills. © 2014 American Chemical Society

Particle stabilized emulsions, commonly called Pickering emulsions, have been known for over a century.5 Pickering emulsions are very stable because the energy, Wa, required to remove an adsorbed particle from an interface into either the oil or water phase is very high. This energy can be expressed as Wa = πrp2γo/w(1 − |cos Θ|)2

(1)

where rp is the particle radius, γo/w is the oil−water interfacial tension, and particle wetting is described by the contact angle, Θ.6,7 Particles of size 500 nm, interfacial tension 50 mN/m (for octane8), and contact angle 90° bind with Wa ∼ 106kT/particle, or essentially irreversibly. This large binding energy implies that even under extreme dilute conditions, the surface-lodged particles will not partition into the water or oil. A large body of literature on the behavior of spherical particles at oil−water interfaces is available, including more recent research on the contribution of capillarity on the stabilization of Pickering emulsions,9 the role of adsorption energy and repulsion of particles at the oil−water interface,10 the similarities and differences of particles and chemical surfactants,6 the use of particles in combination with ionic surfactants,11 the activation barrier for interfacial adsorption, and the role of colloidal stability.12 Emulsion stability has been explored for a range of Received: July 31, 2014 Revised: September 22, 2014 Published: September 25, 2014 12287

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

surface area of ∼200 mm2/g.17 The modified CB aggregates are insoluble but readily dispersible in water; the CB suspension contains no additional surfactants.22 1-Dodecene (95%) and n-dodecane (99%) were purchased from Aldrich Chemical Co., Inc., and n-hexadecane (92+%) was purchased from Alfa Aesar. Soybean oil (U.S.P.) was purchased from Spectrum Chemical Corporation. 2.2. Methods. 2.2.1. Flocculation of CB Solutions. The addition of Na+ ions to the aqueous CB solution results in increased particle hydrophobicity due to the “salting out” of some of the carboxylate groups on the surface of CB particles. Thus, in the presence of NaCl, the CB particles could flocculate. The particle aggregate sizes were measured by dynamic light scattering and validated by measuring settling heights of the suspension. Particle Size Measurements. Particle size distributions were analyzed for suspensions of varying NaCl and CB concentrations via dynamic light scattering with a Malvern Instruments Zetasizer Nano ZS. At constant NaCl concentration of 0.6 M, CB concentration was varied from 0.000 75 to 0.0075 wt %. Similarly, at 0.0075 wt % CB, NaCl concentration was varied from 0.0 to 0.6 M. To obtain repeatable results, suspensions were well mixed directly before analysis. All measurements were taken in triplicate. CB Settling. The CB suspension supplied by the Cabot Corp. is stable but is destabilized on increasing the salinity, leading to aggregation of particles and subsequent settling. This rate of settling can be related to the size of the flocs. To explore the settling on increasing salinity, a 0.0075 wt % and 0.1 M NaCl aqueous suspension was vortexed and then divided into six vials; two vials were agitated for 20 min with a stir bar at 1200 rpm, two vials were sonicated for 20 min, and two vials were not subjected to any additional agitation. All six vials were then left undisturbed and photographed until the particles flocculated and settled, which took about 6 h. The CB settling rate was determined visually and then used to calculate the floc size. 2.2.2. Interfacial Tension and Elasticity. Drop of CB Solutions Suspended in Oil. Dynamic interfacial tension was measured with a pendant-drop approach. The surface tension measurements require imaging the drop and fitting the shape to the Young−Laplace equation. Because of the opacity of the CB solutions, it is significantly simpler to image a drop of CB solution suspended in oil than it is to image an inverted drop of oil in CB solution, and so this arrangement was used to measure the dynamic surface tension. Specifically, the CB suspension was diluted with HPLC water to obtain solutions varying from 0.0 to 15 wt % CB for measurements of interfacial tensions. NaCl was also added to the suspension at concentrations ranging from 0.0 to 0.6 M. The oils were contacted with HPLC grade water for three successive 24 h periods to remove water-soluble impurities and saturate the oil phase. The aqueous drop was suspended at the tip of a 1.6 mm i.d. diameter needle in a quartz cuvette 28 mm × 26 mm × 40 mm in size, and the shape was imaged, digitized, and fit to the Young− Laplace equation using the commercial Kruss DSA 100 tensiometer.23 When surface-active substances adsorb at an interface, they form an interfacial layer that is resistive to perturbations and imparts dilational properties to the interface.24 For crude oil−water emulsions, it has been shown that surface-active substances that increased dilational modulus, elasticity, and viscosity also protect emulsions from coalescence.25−27 The interfacial elasticity can be obtained by oscillating the interface area (A) and obtaining a value of the surface tension (γ) from the fitted surface shape. These values include both contributions from elasticity and the true surface tension of the interface. The complex dilational modulus of the interface, E, is defined by

particles, including silica, latex, polystyrene, glass, maghemite, silver, and crystalline iron oxide.6,13−15 Recently, Saha et al.16 demonstrated that commercially available carboxyl-terminated CB particles can stabilize oil− water emulsions at low pH or high salt concentrations. The promising results on emulsion stabilization along with the relatively low cost, ease of availability, and low potential for toxicity suggest that the CB particles could serve as very effective dispersants after an oil spill. Several other researchers have also explored emulsion formation with CB particles with potential use as dispersants.17,18 While there is now significant data on emulsion stabilization with CB particles, there is very little data on quantifying the interfacial properties of the interfaces with CB particles. In particular, the reduction in interfacial tension and the increase in interfacial elasticity are the key parameters which relate the interfacial mechanics to the emulsion stability.19,20 Thus, we are interested in measuring the surface tension and elasticity of oil−water interfaces covered with CB particles. A similar study has been reported in the literature, but the results showed negligible changes in both surface tension and elasticity, which is surprising in view of the emulsion stabilization by CB particles.21 The interfacial tension and surface elasticity of a particle-covered interface depend on the surface coverage so it is possible that the surface coverage in the interfacial studies21 was lower than that in the emulsion studies,16 and significant surface tension reduction and elasticity increase could be measured by increasing the surface coverage. Therefore, our goal is to explore the effect of CB adsorption on the surface tension reduction and surface elasticity. We hypothesize that due to its relatively large size, the diffusion time scales for the carbon black particles to reach the interface are very long, and thus convection is necessary to achieve high surface coverage. This hypothesis is consistent with the observation that emulsions are prepared under high convection16 while rheology measurements are conducted under static conditions,21 which could limit the CB surface coverage. To explore the validity of our hypothesis, we first measure interfacial tension and surface elasticity under static conditions for two different arrangements: a drop of CB suspension in oil and an inverted oil drop in CB suspension. The two arrangements were explored to determine whether differences in the curvature (concave vs convex) could impact the adsorption. The CB suspension drop studies were useful in comparison with previous literature while the oil drop studies are more representative of the application of the CB particles as stabilizers for oil-in-water emulsions. We also explore the flocculation of the CB particles because relative fluxes from the diffusion and convection depend on the particle size and the diffusivity. The results of our study should be useful in designing the optimal system for contacting the CB dispersant with a rising plume of oil from an underwater leak. Additionally, the results could be useful in several other applications of Pickering emulsions such as in food, agriculture, and pharmaceuticals.

2. EXPERIMENTAL SECTION

E=

2.1. Materials. Enzyme grade sodium chloride and HPLC grade, submicron filtered water were purchased from Fisher Scientific. The (4-carboxyphenyl)-surface-modified carbon black, CB, suspension in water (14.87% solids, filtered to 0.5 μm particle size) was provided by Cabot Corporation. The CB in the suspension are aggregates of primary particles of 10−20 nm with a fractal morphology, a nominal aggregate size of ∼120 nm, and a nitrogen adsorption BET specific

d(γ ) d(ln A)

(2)

When energy is dissipated during the interface relaxation process, there is a phase difference (θ) between the surface tension and area oscillations, and the complex dilational modulus can be expressed as E = E′ + iE″ = Ed + iωηd 12288

(3)

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

Figure 1. Schematic of the pendant drop setup for measuring interfacial tension and surface rheology. To drive an interfacial perturbation, a piezo actuator was mounted in a lab-made Teflon assembly (a) and integrated with the pendent drop apparatus (b). where the real and the imaginary components E′ and E″ represent the elastic/storage and viscous/loss modulus, respectively. The loss component can be expressed as the product of the oscillation frequency ω and the interfacial dilational viscosity ηd.24,19 The drop volumes can be oscillated using the Kruss DSA 100 syringe pump, but this method results in noise in the surface tension measurements due to vibration. To minimize the vibrations and allow for precisely controlled drop oscillations at high frequencies, a piezo actuator was integrated into the device. The piezo actuator (PI P840.3B) and the piezo driver/amplifier (PI E-660) were purchased from Physik Instrumente, and the sine wave signal was generated by a BK Precision 4010A. The piezo was integrated with the DSA system through a lab-built three-way Teflon connector with a syringe connected to the inlet and a needle connected to the outlet as detailed in Figure 1a. The overall setup used for the surface tension measurements is shown in Figure 1b. In our experiment, the sinusoidal variation in surface area, interfacial tension response, and drop volume can be represented as

ΔA = A − A 0 = A a sin(ωt )

(4)

Δγ = γ − γ0 = γa sin(ωt + θ )

(5)

ΔV = V − V0 = Va sin(ωt )

(6)

Dynamic surface tension and dilational viscoelasticity measurements were carried out independently. Oil Drop in Aqueous CB Solutions. The interfacial tension measurements were also conducted with an inverted oil drop in an aqueous CB. These experiments were complicated by the opacity of the aqueous solutions even for very low CB concentration. In these experiments, the oil drop was created at the tip of a 0.6 mm i.d. needle in a quartz cuvette 28 mm × 26 mm × 40 mm in size filled with the opaque 0.0075 wt % CB, 0.6 M NaCl solution. It was hypothesized that the deposition of the CB at the drop interface will depend on the degree of convection, and so the solution was stirred at speeds varying from 0 to 60 rpm for several hours, during which time the CB particles settled, allowing imaging of the oil drop. After the settling of the CB flocs, the interface was imaged and surface tension was determined by fitting the shape to the Young−Laplace equation. Even after the settling of the majority of the flocs, however, the commercial software was unable to accurately perform drop edge detection and fitting of the Young−Laplace equation. Thus, custom software was written in Matlab following classical axisymmetric drop shape analysis methodologies20 to determine the interfacial tension for each drop image. Next, the drop volume was oscillated using the lab-built system described above, and the variation of the interfacial tension during oscillation was used to determine the surface elasticity by following the same process as described above. In some cases, the interface showed a highly nonuniform coverage of CB, which prevented quantification of the surface tension.

where A is the surface area at a given time, A0 is the equilibrium surface area, Aa is the amplitude of surface area oscillation, t is time, γ is the interfacial tension at a given time, γ0 is the equilibrium interfacial tension, γa is the amplitude of interfacial tension oscillation, θ is the phase shift between area and interfacial tension oscillation, V is the drop volume at any given time, V0 is the equilibrium drop volume, and Va is the amplitude of drop volume oscillation. To ensure that the oil− water interface is in the linear viscoelastic regime and that inertial and viscous effects could be neglected, area perturbations were limited to 10% of a drop’s equilibrium surface area, and it was ensured that both the capillary number, Ca, and the Weber number, We, were very small (O(10−6)). Based on the definition in eq 2, the magnitude of the interfacial elasticity and the real and imaginary components can be calculated as γa E(ω) = A a /A 0 (7)

E′(ω) = Ed =

γa A a /A 0

E″(ω) = ωηd =

cos(θ)

γa A a /A 0

sin(θ)

3. RESULTS AND DISCUSSION 3.1. Particle Size. Figure 2 shows the particle size distributions for solutions of varying CB concentrations ranging from 0.0019 to 0.0075 wt % at a fixed NaCl concentration of 0.6 M and solutions of varying salinity from 0 to 0.6 M at a fixed CB concentration of 0.0075 wt %. For varying CB concentration at constant salinity, the distributions are similar in each case with sizes ranging from approximately 1 to 7 μm, which are significantly larger than the size of the CB fractal particle, suggesting that the CB particles aggregate due to electrostatic screening from the high ionic concentrations. The measured size of CB particles without salt, at approximately 100 nm, is in excellent agreement with the material product page’s report (mean size: 130 nm, 100% < 600 nm) and the TEM image of the fractal included in the inset of Figure 2.28 On increasing the salt concentration, the size distributions show a significant increase likely due to aggregation driven by the screening of the electrostatic interactions, which stabilize the

(8)

(9) 12289

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

500 nm, which is in good agreement with the DLS measurements. To check whether the method of mixing significantly affects floc size, this experiment was performed with suspensions that were created in one of three ways: vortexing, sonicating, or stirring. All vials were then left undisturbed and photographed until the particles flocculated and settled, which took about 6 h. For the three mixing techniques, there was no quantifiable difference in settling rate, suggesting that the CB aggregate sizes are independent of mixing technique. Thus, the data in Figure 2 can be considered as the size distribution that one would expect in a well-stirred CB suspension, while a nonstirred suspension could have larger sizes due to slow flocculation. Stirring a flocculated or even a completely settled CB system gives particle sizes consistent with the data in Figure 2. 3.2. Aqueous CB Drop Suspended in Oil. Dynamic Interfacial Tension. Figure 3a shows the dynamic interfacial tension of the oil−water interfaces for n-dodecane, nhexadecane, 1-dodecene, and soybean oil. These experiments were conducted to establish a control with no addition of surfactant and assess the effect of impurities on surface tension measurements. Measurements were completed in triplicate; error bars represent 90% confidence intervals. A pure oil−water interface is expected to not show a steady decline in the interfacial tension, and so this effect is attributed to impurities. After significant efforts to purify the oils, the magnitude of this decrease is consistent with prior studies.20 To minimize this impurity effect, 95% pure 1-dodecene was chosen as the oil phase for subsequent studies because it exhibited the slowest decline in interfacial tension a rate at ∼0.9 (mN/m)/h. The interfacial tension values for the oil water interface are 49.8, 40.6, 29.7, and 20.7 mN/m for n-dodecane, n-hexadecane, 1dodecene, and soybean oil, respectively, at the 30 min stage. Figure 3b plots the transient interfacial tension at the 1dodecene/water interface with (4-carboxyphenyl)-surfacemodified CB concentrations varying from 0.0 to 15 wt % and NaCl varying from 0.0 to 0.6 M. Measurements were completed in triplicate; error bars represent 90% confidence intervals. Upon the addition of NaCl, CB, or both NaCl and CB, there is no statistical difference in the dynamic interfacial

Figure 2. Dynamic light scattering results for aqueous suspensions of CB and NaCl. Error bars represent maximum and minimum values. The inset is a TEM image of the CB fractal.

CB particles. At approximately the salinity of seawater, CB particle initial rapid aggregation creates diameters of approximately 1−7 μm. These measurements represent quasi-steady distributions as the floc sizes keep increasing slowly with time due to further aggregation, but stirring the system even after a considerable period of settling returns the sizes to the same distributions as shown in Figure 2. Additionally any degree of stirring including sonication does not reduce the sizes below those in Figure 2. These sizes should, however, be considered as only the approximate values because the DLS measurements assumed the shapes to be spheres, which is certainly not accurate for the CB fractal under no-salt conditions and may not be accurate for flocs as well. To check the DLS results, a rough estimate for particle size was obtained from the rate of settling in a CB suspension by setting Stokes’ law for drag force on a sphere equal to the force of gravity for the falling particles and assuming particles are traveling at their terminal velocity. For a 0.0075 wt % CB, 0.1 M NaCl suspension, the initial settling velocity was measured to be 2 mm/h, which yielded a particle radius of approximately

Figure 3. (a) Dynamic interfacial tension of the n-dodecane, n-hexadecane, 1-dodecene, and soybean oil in contact with water. Both oils and water were pure so the slow reduction in the interfacial tension is due to impurities. (b) Effect of concentration and salinity on dynamic interfacial tension of (4-carboxyphenyl)-surface-modified carbon black particles at the 1-dodecene/water interface. Water salinity adjusted with sodium chloride to the molarities listed. 12290

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

Figure 4. (a) Change in interfacial tension of the 1-dodecene/CB-solution interface and drop area and volume in response to sinusoidal oscillations along with model least-squares fits (solid lines). (b) Magnified view of the changes in the interfacial tension. (c) Discrete Fourier transformation of the interfacial tension data in (b). ω = 0.16 rad/s, 0.0075 wt % CB, 0.6 M NaCl.

tension measurements from a pure water/dodecene interface. Thus, in this experimental setup, CB particles do not appear to cause any additional decrease in surface tension, which is in agreement with literature29 and suggests negligible surface coverage. Negligible surface coverage is strongly supported by visual observation of the CB containing drops at various times (Figure 6a). These images clearly show that the drop is almost opaque at short times due to the CB particles, but slowly the CB particles settle toward the bottom and the drop becomes almost transparent, signifying negligible surface adsorption. In fact, the bottom of the oil drop shows deposition of the CB particles, which can be attributed to the slow aggregation and settling. The surface tension data along with the visual observations very strongly suggest that the CB particles settle over time but do not adsorb at the interface. Surface Elasticity. Increasing the elasticity of the oil−water interface is important to the stabilization of the oil−water emulsions. To further probe the interfacial deposition of the CB, it was decided to measure the surface elasticity by periodically oscillating the volume using the system described in Figure 1. The dynamic surface tension, drop area, and drop volume for ω = 0.16 rad/s oscillations in a 0.0075 wt % CB, 0.6 M NaCl suspension are plotted in Figure 4a, and more detailed surface tension data from a small time segment are shown in Figure 4b. Figure 4c shows the discrete Fourier transform of the surface tension data, which has only one peak at the same frequency as that of the piezo oscillator, showing that there are no higher order harmonics in the data. The discrete Fourier

transform also gives the phase difference and the amplitude of the surface tension oscillation, which can be used in eqs 5−9 to calculate elasticity. It has, however, been reported that this approach is not accurate because of noise in the response and the finite length of signal.30 Instead, it is preferable to fit the area and surface tension data to eqs 4 and 5 via a nonlinear least-squares curve-fitting procedure. Per Myrvold and Hansen,31 this procedure is able to universally obtain the best results even from noisy data, assuming that good initial parameter estimates are obtained. Figure 4a illustrates the model fit of eqs 4−6 to area, volume, and interfacial tension data of the response in Figure 4a, using initial parameters obtained in Figure 4c. Figure 5 depicts the viscoelastic moduli, as a function of angular frequency, for the interface between purified 1dodecene and water with 0.6 M NaCl and either 0 or 0.0075 wt % CB. Each point represents the average of several trials. Error bars represent 90% confidence intervals. The viscous modulus is significantly smaller than the elastic component, except at very low frequencies where both components are small but comparable in magnitude. The elastic modulus increases as a power law with frequency, but the viscous modulus remains small for all frequencies. In some cases, errors in determination of the phase shift lead to small but negative values for the viscous modulus. Negative values reflect the errors bounds on this approach, and similar observations have been reported previously in other studies.32,33 This limitation is 12291

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

particles in the 0.0075 wt % suspension with 0.6 M salt aggregate into larger particles about a few micrometers in size, which then further aggregate into larger flocs and settle. The settling particles eventually deposit near the bottom of the drop, and there is finite but negligible adsorption at the interface, resulting in negligible changes in interfacial tension and surface elasticity, consistent with published results.21 It is possible that there is a small effect of finite CB adsorption on the surface tension and elasticity, but that effect may not be clearly evident due to the impurities. As noted, these results are surprising considering that CB particles are effective at stabilizing oil-in-water emulsions. There are a few potential explanations to resolve the observations described above. First, it is well-known that emulsion stabilization by surfactant favors oil-in-water emulsification for high HLB surfactants and water-in-oil emulsification for low HLB surfactants. Similarly, the outer and inner phases of particle stabilized Pickering emulsions are controlled by the contact angleor wettabilityof the solid, with oil-in-water emulsions favored for contact angles less than 90° and water-in-oil emulsions favored for contact angles greater than 90°.34 Kruglyakov assigned an equivalent HLB value = [(1 + cos(θ))/(1 − cos(θ))]2 to particles on this basis.35 Because of the wettability of CB particles and therefore the preferred curvature of the oil−water interface, it is feasible that the rate of adsorption of the particles will be different for a CB suspension drop suspended in oil and oil drops dispersed in particle suspension, which is the arrangement used in the emulsification experiments. Second, due to the very low diffusivity of the particles, convection may be essential to

Figure 5. Interfacial dilational viscous and elastic moduli of (4carboxyphenyl)-surface-modified carbon black particles at the 1dodecene/HPLC water interface.

a common disadvantage of oscillatory drop shape analysis when the interfacial tension response is small. While the complex viscoelastic modulus data appear slightly greater for samples with carbon particles compared to the control, at 90% confidence the difference is not statistically significant. Thus, the surface elasticity data also suggest negligible adsorption of CB at the interface. By combining the visual observations (Figure 6a) and size measurements (Figure 2) with surface tension (Figure 3b) and surface elasticity (Figure 5), it can be concluded that the CB

Figure 6. Images of CB settling to oil−water interface. All images for 1-dodecene and HPLC water with 0.0075 wt % CB and 0.6 M NaCl. (a) CB suspension drop on 14 gauge needle surrounded by 1-dodecene. No stirring. (b) 1-Dodecene drop on 20 gauge needle surrounded by CB suspension. No stirring. Convection increased by particle settling. (c) 1-Dodecene drop on 20 gauge needle surrounded by CB suspension. Minimal stirring. (d) 1-Dodecene drop on 20 gauge needle surrounded by CB suspension. Stirring at 60 rpm. Scale bars are 1 mm. 12292

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

adsorption to create a dense layer near the bottom with a clean interface on the top. For comparison with the oil drop suspended in a stagnant CB suspension (Figure 6b), the images of the aqueous CB suspension drop in oil are included (Figure 6a) to show a very clear contrast between the two cases. Because of the highly nonuniform distribution in Figure 6b, it was not possible to calculate the surface tension or the elasticity, and so we utilize the visual images to at least qualitatively understand the behavior of this system, particularly its clear differences from the water drop in oil. The surface coverage in this case can be visibly seen to be significantly larger than the case of CB aqueous drops suspended in oil (Figure 6a). As an initial hypothesis, the difference in fluid volumes of the drop vs the reservoir could be a reason for the significant differences between the two cases. However, this effect can be discounted by noting that the total mass of CB in the aqueous solution reservoir at 0.0075 wt % is less than the total mass of CB in the aqueous drops at the highest explored concentration of 15 wt %. Thus, the differences between the CB adsorption to the interface cannot be attributed completely to the total mass of CB in the system. As noted, this could be due to the effect of particle wettability. It is, however, also useful to note that the size of the CB flocs increases with time, and due to the large differences in the settling distance, the CB flocs are much larger for the case of oil drop in water (Figure 6b). The large size of the adsorbed flocs is also visually evident in Figure 7a, which is a magnified view of the images in Figure 6b. To quantify whether the inertia of the settling CB flocs could be important in facilitating adsorption, we measured the floc size and estimated the velocity by particle tracking. This analysis yielded an approximate floc size of 0.2 mm and a velocity of 0.35 ± 0.10 mm/s. Each tracked particle’s velocity was well-defined with model correlation coefficients of R2 > 0.995. The measured size and velocity give an approximate particle Reynolds number of 0.07, which implies that the particles have small but nonnegligible inertia. The experiments described above clearly show that a higher surface coverage is achieved in the case of the inverted oil drop in a CB suspension, but the coverage is highly nonuniform which prevents quantification of the surface tension or the viscoelasticity. We hypothesized that stirring the CB suspension could lead to a more uniform coverage due to the convection generated in the system. Below we describe the results from these experiments. Stirred CB Suspension. The CB suspension was stirred at 60 rpm using a 1.1 cm (Figure 6c) or 1.6 cm (Figure 6d) long stirrer and continuously videoed. Several frame captures from the video are included in the bottom two rows of Figure 6. The images clearly show that stirring the CB suspension leads to a more uniform coverage of the surface. The images in Figure 6c,d suggest well-covered interfaces as the entire drop appears opaque due to the black color of the CB particles. There are a few holes in Figure 6c even after 4 h of stirring, which preclude estimations of the surface tension, but the surface in Figure 4d is uniformly covered. The drop shapes from the highest convection set (Figure 6d) were fitted to the Young−Laplace equation to determine the surface tension. The CB-covered interface at 4 h demonstrated interfacial tensions of 10 ± 2 mN/m compared to approximately 30 mN/m for clean ndodecene−salt water interfaces. Also, at the 6 h stage, the drop volume was oscillated to determine the surface elasticity. Figure 8a shows the area, volume, and interfacial tension oscillations for a drop of n-dodecene that was initially surrounded by a

drive sufficient adsorption by thinning the convective boundary layer near the surface or by inertia driven adsorption. Thus, it was decided to measure the surface tension and rheology for an inverted oil drop suspended in a dispersion of carbon black. This arrangement is similar to that of emulsification and also affords the possibility of increasing the convection by stirring the aqueous dispersion, which could increase the adsorption. 3.3. Oil Drop in CB Aqueous Solution. The dynamics of the process cannot be directly observed in this case because the CB suspension is opaque, but it was observed that the CB particles flocculate and settle, allowing visual observations at long times. Stagnant CB Suspension. Figure 6b shows images of the oil drop suspended in a stagnant CB suspension of 0.0075 wt % CB and 0.6 M NaCl at various times. The images show that there is a large but nonuniform distribution of CB particles on the bottom of the inverted drop, reminiscent of the immobilized surfactant cap at the back of a rising bubble in a surfactant solution.36 As a suggested mechanism to explain the nonuniform distribution, we propose that the settling CB flocs adsorb on the interface but continue to settle even after adsorption, leading to a dense coverage near the bottom and a clean interface toward the top. To prove that the surface is covered with very large weakly held flocs, we created significant mixing in the system by carefully siphoning off the outer solution while pure water was added to dilute the CB concentration. The convection generated by siphoning caused shear stresses at the interface which caused rapid movement of the adsorbed flocs on the interface. Some flocs remained absorbed but became much smaller in size because the shear stresses overcame the cohesive forces between the floc fragments. This caused shearing off of fragments away from the interface, leaving a smaller floc still adsorbed. Thus, while Figure 7a shows a thick multilayer covering of CB on the

Figure 7. Images of an inverted dodecene drop on a 20 gauge inverted needle after 20 h in a 0.0075 wt % carbon black suspension: (a) 0.6 M NaCl; (b) 0.6 M NaCl, outer phase has been siphoned off and replaced with HPLC water; (c) 0.6 M NaCl, outer phase has been siphoned off and replaced with HPLC water, carbon black flocs have settled approximately 1 h. Scale bars are 1 mm.

bottom of the oil drop, turbulence from siphoning off the outer phase removed all CB except the flocs that were truly adsorbed to the interface (Figure 7b). Essentially the forces generated by the fluid convection were not sufficient to overcome the very high energy of particle adsorption at the interface, which is at least of the order of 106kT, but were adequate to overcome the cohesive energy between the various fragments of the flocs. After stopping the convection, the smaller flocs remained adsorbed even though the concentration of CB in the fluid was almost zero, proving the irreversibility of adsorption under dilution. We were then able to observe the remaining CB settle with time to the bottom of the oil drop (Figure 7c), proving our hypothesis that the flocs settle along the interface after 12293

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

Figure 8. (a) Model fits using nonlinear, least-squares, curve-fitting procedure for the area, volume, and interfacial tension data for oscillations of a drop of 1-dodecene surrounded by HPLC water/CB/NaCl suspension after 6 h of settling. ω = 1.13 rad/s, 0.0075 wt % CB, 0.6 M NaCl. (b) Discrete Fourier transformation of the interfacial tension response.

water interface. Further, this measured surface coverage correlates to an average CB layer thickness of 2 μm. From published cryo-SEM images of CB stabilized octane-in-water emulsions at high salt concentrations,16 CB layers in emulsions appear to range from 0.25 to 2 μm in thickness. We assume that this range is due to varying degrees of CB flocculation, the nonspherical shape of actual CB flocs, and the orientation of adsorbed flocs. The CB layer thicknesses for stirred CB suspensions (Figure 6d) thus appear to be representative of the CB thicknesses in octane-in-water emulsions, suggesting that the surface tension and elasticity measurements from the oil pendant drop are relevant in the formation of the CB stabilized emulsions. The degree of convection is clearly a vital variable in particle adsorption to oil−water interfaces. For the measured CB size of about 3 μm, the Stokes−Einstein diffusivity is approximately 7 × 10−14 m2/s. The diffusion time scales as the ratio of the square of the characteristic distance and the diffusivity. For diffusion-controlled adsorption, the characteristic distance is the size of the depletion zone, i.e., the ratio of the surface and the bulk concentration. Thus, in the absence of convection, the time required for diffusion-controlled adsorption of particles from the bulk solution can be estimated as (π/4D)(Γ/c)2, where D is the diffusivity and Γ and c are the surface and the bulk concentrations, respectively.37,38 For the adsorption of 3 μm CB particles from the 0.0075 wt % solution, this time is of order 105 days. Therefore, without convection, it is not feasible to obtain a high surface coverage of the particles. In the absence of stirring, the settling of the particles in the reservoir generates convection which brings the particles near the interface leading to adsorption. The settling leads to preferential adsorption near the top of the oil drop, leaving the lower part uncovered, which in turn drives settling of the adsorbed particles on the interface, to create a nonuniform coverage. Forced stirring increases inertia and makes the convection more uniform, thereby leading to a relatively uniform coverage. The particles that are convected in the vicinity of the surface can adsorb stochastically. Additionally, the particle inertia could cause a lateral drift across streamlines, allowing it to reach the interface even as the fluid streamlines curve near the surface of the oil drop. To quantify whether the inertia of the CB flocs could be important in this case, we estimate the Reynolds number by using the sizes reported earlier and using the product of the

0.0075 wt % CB, 0.6 M NaCl solution. Area and volume oscillations were clearly defined, oscillatory frequency and phase shift are consistent with known values, and a Fourier transform of the interfacial tension data shows that while there is noise, the interfacial tension response is still predominately one frequency (Figure 8b). Average interfacial tension and viscoelastic modulus were 8.5 and 20.7 mN/m, respectively. Interestingly, it was oberved that in any oscillation cycle the fitting of the images to the Young−Laplace equation worsened over time, and the fitted IFT did not show the cyclic pattern with a single frequency. While we cannot conclude on the reasons for this behavior, it appears that the oscillarions lead to clumping of the CB particles. Thus, the duration of the oscillatory times was restricted, and Fourier transforms were computed to ensure that the response had only a single harmonic. While this method yielded reliable values of surface tension and viscoelasticity, there were several challenges that limited our ability to obtain detailed kinetic data. Because of the opacity of CB, the dynamics of CB adsorption and the dynamic interfacial tension cannot be observed. With long time studies, the effects of gravity and vibrations were amplified, and significant work was done to simply maintain a suspended oil drop long enough to produce a visible drop with sufficient coverage. Finally, due to CB opacity, image quality was significantly decreased, increasing error bars on measured interfacial tensions. The stirring of the CB suspension, however, clearly leads to a more uniform surface coverage. To estimate the surface coverage by the CB flocs, we increased the volume of the oil drop after the 4 h stage (Figure 6d) until the drop broke off from the needle tip. The rising oil drop was collected and diluted with 10 mL of HPLC water. The concentration of the CB particles in the water was calculated by comparing the turbidity with that of standard solutions. The concentration was then used to calculate the mass of the particles adsorbed on the drop surface. Using the known drop surface area, the surface concentration of CB on the oil drop in Figure 6d was estimated at 2.6 g/m2. For comparison, the calculated surface coverage for a single layer of CB on the interface, assuming spherical particles and a specific gravity of 1, is 2 g/m2. From this and visual observation, it is safe to conclude that stirring of the CB suspension created thin, near-uniform coverage of the oil− 12294

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir



half-length of the stirrer and the angular velocity as the velocity scale. This analysis yields Re values of 0.12 and 0.17 for 60 rpm stirring with the smaller and the larger stirrers, respectively, which again implies that the particles have small but nonnegligible inertia. It thus seems plausible that the inertia of the particles play a direct role in adsorption, but further investigations are needed to explore this possibility.

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]fl.edu; Tel (352) 392 2592; fax (352) 392 9513 (A.C.). Notes

The authors declare no competing financial interest.

4. CONCLUSIONS



ACKNOWLEDGMENTS



REFERENCES

This work was supported by the Gulf of Mexico Research Initiative and by Graduate School Fellowship from the University of Florida. We gratefully acknowledge the supply of carbon black from Cabot Corporation. We thank our collaborators Arijit Bose and Mindy Levine at the University of Rhode Island and Aubhav Tripathi at Brown University for useful discussions. We also thank Arijt Bose for providing the TEM image of the CB fractal.

We have examined interfacial tension and rheological properties of oil−water interfaces with adsorbed surface-modified carbon black particles. The surface rheology was explored by measuring the dynamic surface tension while oscillating the drop surface by integrating a piezo oscillator into the pendant drop system. For the commonly used setup of a drop of CB suspension surrounded by oil, neither surface tension nor rheology differs from that of a pure oil−water system, which appears to be surprising considering prior studies on emulsion stabilization by CB particles. We proposed that interfacial tension and surface elasticity of a particle-covered interface depend on the surface coverage, and therefore, the CB drop in oil studies is not representative of the conditions for CB-stabilized emulsions. We demonstrate that CB form flocs under high-salinity environments, which cannot diffuse to the interface on short time scales but reach the interface due to increased convection in both the intense mixing during emulsification and the above experiments for an inverted oil drop surrounded by CB suspension. On reversing the arrangement in this way, convection is increased by the settling of CB flocs, and the surface coverage increases considerably but is highly nonuniform with accumulation of the particles toward the bottom of the drop. This is likely due to settling of the adsorbed particles along the interface. Fragments of adsorbed flocs can be sheared off by introducing higher convection, leaving behind smaller adsorbed flocs that eventually also settle along the interface. Once adsorbed, CB remains at the interface even under extreme dilution due to the high energies required for desorption, which are typical for the Pickering emulsions. The surface coverage becomes more uniform on stirring the aqueous dispersion creating thin, near-uniform coverage of the oil−water interface that is representative of the larger end of CB thicknesses in octane-in-water emulsions. At a surface coverage of approximately 2.6 g/m2, the surface tension decreases from about 30 to 10 mN/m and the surface elasticity increases to approximately 20.7 mN/m. These results are consistent with the emulsion stabilization achieved by CB particles under high-salinity conditions. The sharp contrast between the results from the two experimental setupsa CB suspension drop surrounded by oil and an inverted oil drop surrounded by CB suspensioncould be partially due to the effect of the wetting characteristics of the particles or due to the significant differences between the convection in the two cases. The differences in oil drop surface coverage with increasing convection certainly supports that the degree of convection is a vital variable in particle adsorption to oil−water interfaces, particularly for CB flocs of a few micrometers in size. This improved understanding of the mechanisms of CB particle adsorption could be useful in emulsion stabilization for various applications including remediation of oil spills.

(1) Thibodeaux, L. J.; Valsaraj, K. T.; John, V. T.; Papadopoulos, K. D.; Pratt, L. R.; Pesika, N. S. Marine Oil Fate: Knowledge Gaps, Basic Research, and Development Needs; A Perspective Based on the Deepwater Horizon Spill. Environ. Eng. Sci. 2011, 28 (2), 87−93. (2) Kujawinski, E. B. The Impact of Microbial Metabolism on Marine Dissolved Organic Matter. Annu. Rev. Marine Sci. 2011, No. No.3, 567−599. (3) Reinsfeld, A.; Rosenberg, E.; Gutnick, D. Microbial Degradation of Crude Oil: Factors Affecting the Dispersion in Sea Water by Ixed and Pure Cultures. Appl. Microbiol. 1972, 24 (3), 363−368. (4) George-Ares, A.; Clark, J. R. Aquatic Toxicity of Two Corexit Dispersants. Chemophere 2000, 8 (897−906), 40. (5) Pickering, S. U. Emulsions. J. Chem. Soc., Trans. 1907, 91, 2001− 2021. (6) Binks, B. P. Particles as Surfactants - Similarities and Differences. Curr. Opin. Colloid Interface Sci. 2002, 7 (1−2), 21−41. (7) Lin, Y.; Boker, A.; Skaff, H.; Cookson, D.; Dinsmore, A. D.; Emrick, T.; Russell, T. P. Nanoparticle Assembly at Fluid Interfaces: Structure and Dynamics. Langmuir 2005, 21 (1), 191−194. (8) Amaya, J.; Rana, D.; Hornof, V. Dynamic Interfacial Tension Behavior of Water/Oil Systems Containing In situ-Formed Surfactants. J. Solution Chem. 2002, 31 (2), 139−148. (9) Denkov, N. D.; Ivanov, I. B.; Kralchevsky, P. A.; Wasan, D. T. A Possible Mechanism of Stabilization of Emulsions by Solid Particles. J. Colloid Interface Sci. 1992, 150 (2), 589−593. (10) Du, K.; Glogowski, E.; Emrick, T.; Russell, T. P.; Dinsmore, A. D. Adsorption Energy of Nano- and Microparticles at Liquid-Liquid Interfaces. Langmuir 2010, 26 (15), 12518−12522. (11) Binks, B. P.; Rodriques, J. A. Synergistic Interaction in Emulsions Stabilized by a Mixture of Silica Nanoparticles and Cationic Surfactant. Langmuir 2007, 23, 3626−3636. (12) Salari, J. W. O.; Leermakers, F. A. M.; Klumperman, B. Pickering Emulsions: Wetting and Colloidal Stability of Hairy Particles-A SelfConsistant Field Theory. Langmuir 2011, 27 (11), 6574−6583. (13) Saleh, N.; Sarbu, T.; Sirk, K.; Lowry, G. V.; Matyjaszewski, K.; Tilton, R. D. Oil-in-Water Emulsions Stabilized by Highly Charged Polyelectrolyte-Grafted Silica Nanoparticles. Langmuir 2005, 21 (22), 9873−9878. (14) Binks, B. P.; Lumsdon, S. O. Pickering Emulsions Stabilized by Monodisperse Latex Particles: Effects of Particle Size. Langmuir 2001, 17 (15), 4540−4547. (15) Tambe, D. E.; Sharma, M. M. Factors Controlling the Stability of Colloid-Stabilized Emulsions. J. Colloid Interface Sci. 1993, 157, 244−253. (16) Saha, A.; Nikova, A.; Venkataraman, P.; John, V.; Bose, A. Oil Emulsification Using Surface-Tunable Carbon Black Particles. ACS Appl. Mater. Interfaces 2013, 5, 3094−3100.

12295

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Langmuir

Article

(17) Katepalli, H.; John, V. T.; Bose, A. The Response of Carbon Black Stabilized Oil-in-Water Emulsions to the Addition of Surfactant Solutions. Langmuir 2013, 29 (23), 6790−6797. (18) Rodd, A. L.; Creighton, A.; Vaslet, C. A.; Rangel-Mendez, J. R.; Hurt, R. H.; Kane, A. B. Effects of Surface-Engineered NanoparticleBased Dispersants for Marine Oil Spills on the Model Organism Artemia franciscana. Environ. Sci. Technol. 2014, 48 (11), 6419−6427. (19) Sjöblom, J.; Oye, G.; Glomm, W. R.; Hannisdal, A.; Knag, M.; Brandal, O.; Ese, M.-H.; Hemmingsen, P. V.; Harve, T. E.; Oschmann, H.-J.; Kallevik, H. Modern Characterization Techniques for Crude Oils, Their Emulsions, and Functionalized Surfaces. In Emulsions and Emulsion Stability, 2nd ed.; Sjöblom, J., Ed.; Taylor & Francis Group: Bergen, Norway, 2006; pp 415−476. (20) del Rio, O. I.; Neumann, A. W. Axisymmetric Drop Shape Analysis: Computational Methods for the Measurement of Interfacial Properties from the Shape and Dimensions of Pendant and Sessile Drops. J. Colloid Interface Sci. 1997, 196, 136−147. (21) Santini, E.; Ravera, F.; Ferrari, M.; Alfe, M.; Ciajolo, A.; Liiggieri, L. Interfacial Properties of Carbon Particulate-Laden Liquid Interfaces and Stability of Related Foams and Emulsions. Colloids Surf., A 2010, 365, 189−198. (22) CAB-O-JET 300 Black Colorant; Material Safety Data Sheet; Cabot Corporation, Billerica, MA, Sept 26, 2012. (23) Beverung, C. J.; Radke, C. J.; Blanch, H. W. Protein Adsorption at the Oil/Water Interface: Characterization and Adsorption Kinetics by Dynamic Interfacial Tension Measurements. Biophys. Chem. 1999, 81, 59−80. (24) Wang, Y.; Zhang, L.; Sun, T.; Zhao, S.; Yu, J. A Study of Interfacial Dilational Properties of Two Different Structure Demulsifiers at Oil-Water Interfaces. J. Colloid Interface Sci. 2004, 270, 163− 170. (25) Callghan, I. C.; Neustadter, E. L. Foaming of Crude Oils: A Studly of Non-Aqueous Foam Stability. Chem. Ind. (London) 1981, 53−557. (26) McLean, J. D.; Kilpatrick, K. Effects of Asphaltene Solvency on Stability of Water-in-Crude-Oil Emulsions. J. Colloid Interface Sci. 1997, 189, 242−253. (27) McLean, J. D.; Kilpatrick, P. K. Effects of Asphaltene Aggregation in Model Heptane-Toluene Mixtures on Stability of Water-in-Oil Emulsions. J. Colloid Interface Sci. 1997, 196, 23−34. (28) Cab-O-Jet 300 Product Page, Cabot Corporation, 2009. (29) Satini, E.; Ravera, F.; Ferrari, M.; Alfe, M.; Ciajolo, A.; Liggieri, L. Interfacial Properties of Carbon Particulate-Laden Liquid Interfaces and Stability of Relate Foams and Emulsions. Colloids Surf., A 2010, 365, 189−198. (30) MATLAB version 6.5.1 [computer software], The MathWorks Inc., Natick, MA, 2003. (31) Myrvold, R.; Hansen, F. K. Surface Elasticity and Viscosity from Oscillating Bubbles Measured by Automatic Axisymmetric Drop Shape Analysis. J. Colloid Interface Sci. 1998, 207, 97−105. (32) Alexandrov, N.; Marinova, K. G.; Danov, K. D.; Ivanov, I. B. Surface Dilational Rheology Measurements for Oil/Water Systems with Viscous Oils. J. Colloid Interface Sci. 2009, 339 (2), 545−550. (33) Russev, S. C.; Alexandrov, N.; Marinova, K. G.; Danov, K. D.; Denkov, N. D.; Lyutov, L.; Vulchev, V.; Bilke-Krause, C. Instrument and Methods for Surface Dilational Rheology Measurements. Rev. Sci. Instrum. 2008, 79, 104102-1−104102-10. (34) Chevalier, Y.; Bolzinger, M.-A. Emulsions Stabilized with Solid Nanoparticles: Pickering Emulsions. Colloids Surf., A 2013, 439, 23− 34. (35) Kruglyakov, P. M.; Nushtayeva, A. V. Phase Inversion in Emulsions Stabilized by Solid Particles. Adv. Colloid Interface Sci. 2004, 108-109, 151−158. (36) Palaparthi, R.; Papageorgiou, D. T.; Maldarelli, C. Theory and Experiments on the Stagnant Cap Regime in the Motion of Spherical Surfactant-Laden Bubbles. J. Fluid Mech. 2006, 559, 1−44. (37) Ward, A. F. H.; Tordai, L. Time-Dependence of Boundary Tensions of Solutions I. The Role of Diffusion Time-Effects. J. Chem. Phys. 1946, 14 (453), 453−461.

(38) Liu, J.; Messow, U. Diffusion-Controlled Adsorption Kinetics at the Air/Solution Interface. Colloid Polym. Sci. 2000, 278 (2), 124−129.

12296

dx.doi.org/10.1021/la503049m | Langmuir 2014, 30, 12287−12296

Interfacial tension and surface elasticity of carbon black (CB) covered oil-water interface.

Carboxyl-terminated carbon black (CB) particles have been proposed as readily available, biocompatible dispersants to stabilize oil-in-water emulsions...
4MB Sizes 0 Downloads 5 Views