betaFeedback

Reading v1View latest version

WorksPhysics and Astronomy

Geometry-Controlled Dynamic Tensiometry Resolves Intrinsic Surfactant Adsorption Kinetics

Abstract

At short times, interfacial tension depends on experimental geometry because surfactant transport to the interface depends on the mass-transfer conditions.A predictive description therefore requires more than a dynamic tension curve or a fitted adsorption constant: interfacial thermodynamics, diffusion, and adsorption kinetics must be identified separately. Here, we combine equilibrium and diffusion measurements with a microfluidic EDGE tensiometer that provides a nearly stationary interface and controlled micrometer-scale transport. Equilibrium properties and diffusion are determined independently, leaving adsorption kinetics as the key unknown. Dynamic tension is then calculated using a nonequilibrium thermodynamic description, without assuming instantaneous equilibrium between the adsorbed layer and the subsurface solution. For the nonionic surfactant C10_{10}E8_{8}, equilibrium thermodynamics and transport are independently constrained, and a single intrinsic adsorption rate constant describes several concentrations. We extend the framework to SDS by including electrostatic interactions and subsurface-concentration dynamics, capturing transient depletion and replenishment. Once thermodynamic, transport, and kinetic parameters are identified, the model predicts dynamic interfacial tension beyond the geometry and conditions used to determine them. The microfluidic EDGE tensiometer thus provides both a reliable short-time tensiometry method and a quantitative framework for identifying the physical mechanisms governing surfactant mass transfer at interfaces.

Introduction

Dynamic interfacial tension controls droplet and bubble formation in emulsification, foaming, spraying, coating, and many pharmaceutical, cosmetic, and food-processing operations [1, 2, 3]. In these processes, interfaces are often created within milliseconds, long before surfactants reach equilibrium. Their behavior therefore depends on how rapidly surfactants are transported through the liquid and transferred to the newly created interface.

Accessing short interface ages is not sufficient to determine these kinetics. The measured tension relaxation combines bulk transport and molecular adsorption. An intrinsic adsorption rate constant can be identified only when bulk transport is independently constrained or is effectively faster than interfacial transfer. Otherwise, different combinations of diffusion and adsorption parameters may produce similar tension–time curves. The challenge is therefore to probe newly formed interfaces in a geometry with a well-defined rate-limiting mechanism [4].

Classical pendant- and rising-drop tensiometers provide accurate liquid–liquid measurements but involve transport over macroscopic distances. At dilute concentrations, their long-time response is commonly diffusion-controlled [5, 6, 7]. These measurements can determine the diffusion coefficient, but become insensitive to molecular adsorption when interfacial transfer is faster than bulk replenishment. Maximum-bubble-pressure tensiometry reaches millisecond surface ages [8, 9, 10], but the interface continuously expands and the measured pressure also depends on bubble hydrodynamics, surface dilution, and the definition of surface age. Shorter times therefore do not necessarily provide direct access to adsorption kinetics.

Microfluidics offers a promising alternative because small dimensions and convective renewal can accelerate bulk transport. Alvarez et al. developed a mic tensiometer in which flow is imposed around a spherical air–water interface held at the end of a glass capillary [11, 12, 13]. Reducing the interface radius and the convective boundary-layer thickness was intended to make adsorption rate-limiting. For C12E8\mathrm{C}_{12}\mathrm{E}_{8} and C14E8\mathrm{C}_{14}\mathrm{E}_{8}, however, the measured dynamics remained controlled by convection–diffusion, and only lower bounds on the adsorption constants could be obtained.

Brosseau, Vrignon, and Baret used the same timescale-separation principle with convected microdroplets [14, 15]. Water droplets were transported through a delay line containing repeated narrow channels and planar expansions. Their age was set by the travel time, and the interfacial tension was inferred from their deformation at each expansion. For the PEG–PFPE surfactant investigated, transport was estimated to be faster than molecular transfer, allowing adsorption and desorption constants to be extracted. The tension, however, was reconstructed from the deformation of a confined moving droplet through a calibrated hydrodynamic relation, rather than directly measured.

Differential-pressure measurements during microfluidic droplet formation provide another route to short-time interfacial tension [16]. In this geometry, the droplet radius, interfacial area, surface-dilution rate, and surrounding flow evolve simultaneously. The dominant mechanism must therefore be identified a posteriori from operating conditions and model comparison, and the inclusion of an adsorption constant does not ensure that it is experimentally identifiable.

A geometry is thus still needed that combines direct oil/water tension measurement, short interface ages, controlled subsurface renewal, and direct assessment of the mechanism governing the response. Here, we introduce a microfluidic EDGE (Edge-based Droplet GEneration) tensiometer designed to meet these requirements [17]. An oil meniscus is confined inside a shallow pore and exposed to a continuously renewed aqueous surfactant solution. During most of the adsorption period, the meniscus remains pinned and its geometry varies only weakly. Adsorption lowers the interfacial tension and the Laplace pressure retaining the meniscus. When the applied pressure exceeds this capillary threshold, the interface advances and a droplet is released. The waiting time defines the interface age, while the Young–Laplace relation gives the corresponding tension.

We first compare the characteristic transport and adsorption times of the principal tensiometric geometries. This analysis reveals the advantages of EDGE: a short and well-defined transport distance, continuous subsurface renewal, and a nearly stationary interface with weak variations in area and curvature. It establishes the conditions under which the EDGE response is sensitive to molecular adsorption rather than dominated by bulk transport.

The kinetic parameters are then determined sequentially. Equilibrium tensiometry provides the adsorption isotherm and interfacial equation of state, while dilute pendant-drop measurements in a diffusion-controlled regime determine the diffusion coefficient DD. This independently measured value is fixed in the EDGE analysis. Adsorption-limited, diffusion-limited, and mixed models are then confronted with the EDGE interfacial-tension measurements to identify the adsorption and desorption parameters without compensating for an unknown bulk-transport rate.

We apply this strategy to the nonionic surfactant C10_{10}E8_8 and sodium dodecyl sulfate (SDS) at the water–hexadecane interface. The kinetic parameters are then determined sequentially.

Microfluidic EDGE tensiometer

Microchip and set-up

The partitioned-Edge-based Droplet GEneration (EDGE) microchips have been produced by Micronit Microtechnologies B.V (Enschede, The Netherlands), and they have been previously used by Deng et al. [18] and Santos et al. [19] as microfluidic tensiometers. The partitioned-EDGE microchip consists of two deep channels, one straight channel for the dispersed phase and one meandering channel for the continuous phase (see Figure 1). The deep channels with 175 μ\mum height (HH) and 400 μ\mum (WdW_d) of width are connected in the parallel-running section by a shallow plateau of 200 μ\mum in length (LL) and 500 μ\mum in width (WW). The shallow plateau is further partitioned into eight identical parallel pores; these pores are 40 μ\mum in length (ll) and have a rectangular cross-section with 40 μ\mum width (ww). The plateau and pores have a height of h=0.93 μh = 0.93\ \mum. The partition width (ss) is 20 μ\mum.

Schematic illustration of the partitioned-EDGE device and pore geometries

Figure 1. Schematic illustration of the partitioned-EDGE device (A) and of the plateau and the pores (B). Side schematic views of curvatures in the two different directions in a pore (C). Adapted from Santos et al. [19].

The chip was set in a chip holder (Micronit B.V., The Netherlands) and connected via PEEK tubing (0.75 mm, BGB, Switzerland) to a digital pressure controller operated through Smart Interface software (Elveflow, France). The pressure of the continuous phase PcP_c was kept constant at 100 mbar, while that of the dispersed phase PdP_d was modified manually, after closing its outlet (Figure 1).

By keeping PcP_c constant, any effects related to variations in shear flow are avoided. Based on the estimation of flow resistance in the system, the effective pressure Pd∗P_d^* across the plateau and the pores is estimated as: Pd∗=Pd−Pc/2P_d^* = P_d - P_c/2. Interfacial tension γ\gamma is calculated by the following equation [20]:

γ=Pd∗1R1+1R2=Pd∗2(whw+hcos⁡(θ))(1)\gamma= \frac{P_d^*}{\frac{1}{R_1}+\frac{1}{R_2}} = \frac{P_d^*}{2}\left(\frac{wh}{w+h}\cos(\theta)\right) \tag*{(1)}

where R1R_1 and R2R_2 are the principal radii of curvature for the top and front view, respectively (Figure 1). A value of ∼30∘\sim30^\circ was measured for θ\theta and used here.

Surfactant transport and adsorption at the EDGE meniscus

We now examine how the EDGE geometry controls surfactant supply to the interface by combining continuous convective renewal above the pore with diffusion towards the recessed and nearly stationary meniscus. Surfactant transfer occurs through three successive steps. First, the flowing aqueous phase continuously renews the solution above the pore and limits the growth of the local depletion layer. Second, surfactant molecules diffuse through this depletion layer and across the recessed region separating the renewed liquid from the oil–water meniscus. Third, once they reach the meniscus, the surfactant molecules adsorb at the interface.

The first two steps determine the rate at which the surfactant is supplied to the interface. Their combined transport resistance is represented by an effective transport length defined as the following (see Supporting Information Fig.S1 and S2 for details):

LD=w∫−w/2w/2dy[πD(y+w/2)/U]1/2+ℓm(2y/w)2(2)L_D = \frac{w}{\int_{-w/2}^{w/2}\frac{dy}{\left[\pi D(y+w/2)/U\right]^{1/2}+\ell_m(2y/w)^2}} \tag*{(2)}

where UU is the effective renewal velocity, ww is the pore width, DD is the surfactant diffusion coefficient, and ℓm\ell_m is the maximum recession length of the meniscus (Fig. S1). For fixed flow conditions and meniscus geometry, LDL_D remains approximately constant during the measurement and can be calculated independently of the dynamic interfacial-tension curves. Increasing the renewal velocity reduces the contribution of the depletion layer to LDL_D and therefore increases the surfactant supply rate.

The third step is the molecular adsorption of the surfactant at the oil–water interface. The progressive accumulation of adsorbed molecules lowers the interfacial tension and consequently decreases the Laplace pressure retaining the meniscus inside the pore. When the Laplace pressure becomes equal and exceeds the imposed effective pressure due to surfactant adsorption, the meniscus advances and a droplet is released. The droplet-formation time therefore measures the time required for transport and adsorption to reduce the interfacial tension to the selected threshold value. The regime analysis introduced below is used to determine whether this measured time is primarily controlled by bulk supply or by molecular adsorption.

Because droplet formation is highly periodic, the initial state of the meniscus is reproducible from one cycle to the next for a given concentration and applied pressure. Further experimental details are provided in section 6, including the operating conditions of the EDGE device, the data treatment and the experimental limits of the technique.

Geometry and characteristic kinetic times

To determine whether the measured interfacial dynamics are controlled by molecular adsorption or by bulk transport, we coupled Langmuir adsorption kinetics with a geometry-dependent mass-transfer model. We compared three different geometries: pendant drop, convected drop, and EDGE by constructing the regime maps shown in Figure 2. The surface coverage (ϕ\phi) evolves according to a Langmuir kinetics as the following expression:

Regime maps for pendant drops, convected drops, and the EDGE geometry

Figure 2. Regime maps for pendant drops, convected drops, and the EDGE geometry. The maps show log⁡10(Π)\log_{10}(\Pi) as a function of interface age tt and bulk surfactant concentration CbC_b, with Π=Γ∞ka(1−ϕ)/km\Pi= \Gamma_{\infty} k_a(1-\phi)/k_m. The black contour corresponds to Π=1\Pi= 1 and separates the adsorption-sensitive regime, Π<1\Pi< 1 (blue–cyan), from the transport-sensitive regime, Π>1\Pi> 1 (orange). Surface coverage is calculated from (3), coupled to the interfacial mass balance given by (4). The upper row corresponds to pendant drops with R=50R = 50, 150150, and 20002000 μm\mu\mathrm{m}, from left to right, using km=D/(πt)+D/Rk_m = \sqrt{D/(\pi t)} + D/R. The middle row corresponds to convected drops with R=20R = 20, 200200, and 20002000 μm\mu\mathrm{m} and Sh=20\mathrm{Sh} = 20, using km=ShD/Rk_m = \mathrm{Sh}D/R. The lower row corresponds to EDGE geometries with LD=3L_D = 3, 77, and 3030 μm\mu\mathrm{m}, using km=(Af/Ai)D/LDk_m = (A_f/A_i)D/L_D, with Af/Ai=1/1.38A_f/A_i = 1/1.38 (Calculations can be found in Supporting Information S1 and S2). All maps were calculated over 10−3≤t≤102 s10^{-3} \le t \le10^{2}\,\mathrm{s} and 10−2≤Cb≤104 μM10^{-2} \le C_b \le10^{4}\,\mu\mathrm{M}, with D=1.0×10−9 m2 s−1D = 1.0 \times10^{-9}\,\mathrm{m^2\,s^{-1}}, Γ∞=2.0×10−6 mol m−2\Gamma_{\infty} = 2.0 \times10^{-6}\,\mathrm{mol\,m^{-2}}, K=2000 m3 mol−1K = 2000\,\mathrm{m^3\,mol^{-1}}, ka=100 m3 mol−1 s−1k_a = 100\,\mathrm{m^3\,mol^{-1}\,s^{-1}}, and kd=ka/K=0.05 s−1k_d = k_a/K = 0.05\,\mathrm{s^{-1}}. For pendant drops, the decrease of kmk_m with interface age can produce re-entrant trajectories between adsorption-sensitive and transport-sensitive regimes. By contrast, kmk_m remains constant at fixed operating conditions for convected drops and EDGE, so that a system initially in the adsorption-sensitive regime tends to remain there as surface coverage increases.

dϕdt=kacs(1−ϕ)−kdϕ(3)\frac{d\phi}{dt}=k_{\mathrm{a}}c_{\mathrm{s}}(1-\phi)-k_{\mathrm{d}}\phi \tag*{(3)}

where csc_{\mathrm{s}} is the surfactant concentration in the subsurface, and is not assumed to remain equal to the bulk concentration. It is obtained from the following interfacial mass balance:

km(cb−cs)=Γ∞dϕdt(4)k_{\mathrm{m}}(c_{\mathrm{b}}-c_{\mathrm{s}})=\Gamma_{\infty}\frac{d\phi}{dt} \tag*{(4)}

The mass-transfer coefficient kmk_{\mathrm{m}} accounts for the geometry and hydrodynamic conditions of each experiment. We used km=D/(πt)+D/Rk_{\mathrm{m}}=\sqrt{D/(\pi t)}+D/R for a pendant drop, (km=ShD/(2R)k_{\mathrm{m}}=\mathrm{Sh}D/(2R)) for a convected drop. The Sherwood number Sh\mathrm{Sh} was estimated from the Ranz–Marshall correlation, Sh=2+0.6Re1/2Sc1/3\mathrm{Sh}=2+0.6\mathrm{Re}^{1/2}\mathrm{Sc}^{1/3}. Here, Re=ρUrel(2R)/μ\mathrm{Re}=\rho U_{\mathrm{rel}}(2R)/\mu is the Reynolds number based on the relative velocity (UrelU_{\mathrm{rel}}) between the drop and the surrounding liquid, and (Sc=μ/(ρD)\mathrm{Sc}=\mu/(\rho D)) is the Schmidt number. For the EDGE geometry, we used km=(Af/Ai)D/LDk_{\mathrm{m}}=(A_{\mathrm{f}}/A_{\mathrm{i}})D/L_D, where AfA_{\mathrm{f}} is the area of the renewed liquid film, AiA_{\mathrm{i}} is the interfacial area, and LDL_D is the effective transport length.

To compare the relative contributions of interfacial adsorption and bulk mass transfer, we defined the dimensionless ratio: Π=Γ∞ka(1−ϕ)/km\Pi=\Gamma_{\infty}k_{\mathrm{a}}(1-\phi)/k_{\mathrm{m}}. Here, Γ∞\Gamma_{\infty} is the maximum surface excess, kak_{\mathrm{a}} is the adsorption rate constant, ϕ\phi is the fractional surface coverage, and kmk_{\mathrm{m}} is the mass-transfer coefficient. The ratio Π\Pi compares the characteristic rates of interfacial adsorption and bulk mass transfer: Π<1\Pi<1 indicates that bulk transport is sufficiently fast for the measured dynamics to remain sensitive to the intrinsic adsorption rate, whereas Π>1\Pi>1 indicates that surfactant supply limits the response. For each geometry, Π\Pi was calculated as a function of bulk concentration and interface age to construct the corresponding regime maps. The coupled equations were integrated numerically for each bulk concentration, interface age, and geometry.

Interpretation of the regime maps

Figure 2 compares the adsorption- and transport-sensitive regimes obtained for pendant drops, convected drops, and the EDGE geometry. The black contour corresponds to Π=1\Pi=1 and separates the adsorption-sensitive domain (Π<1\Pi<1), shown in blue-cyan, from the transport-sensitive domain (Π>1\Pi>1), shown in orange. At fixed geometry, the regime evolves with bulk concentration and interface age. Decreasing the drop radius or increasing the Sher-age as the surface coverage increases, while the influence of the experimental geometry and hydrodynamic conditions is captured by kmk_m. These maps therefore identify the experimental geometries and operating conditions under which the measured dynamics remain sensitive to the intrinsic adsorption rate constant rather than being limited by bulk transport.

For pendant drops, shown in the upper row, the mass-transfer coefficient is km=D/πt+D/Rk_m = \sqrt{D/\pi t} + D/R. The first contribution decreases continuously with interface age because the diffusion layer grows around the drop. Consequently, the transport resistance increases with time, and a system that is initially adsorption-sensitive may progressively become transport-sensitive. Increasing the drop radius from 5050 to 20002000 μm\mu\mathrm{m} also decreases the contribution D/RD/R and therefore expands the transport-controlled domain. Some trajectories (for CbC_b high enough) cross the regime boundary twice: the system initially transitions from adsorption-sensitive to transport-limited dynamics as kmk_m decreases, before returning to the adsorption-sensitive regime at long times as progressive interfacial saturation strongly reduces the (1−ϕ)(1-\phi) factor in Π\Pi. Classical pendant-drop tensiometry is thus particularly sensitive to bulk diffusion for large drops, at long times, and at dilute concentrations.

For convected drops, shown in the middle row, the mass-transfer coefficient is approximated by km=(ShD)/(2R)k_m = (\mathrm{Sh}D)/(2R). At fixed flow conditions, kmk_m is independent of interface age.

Decreasing the drop radius or increasing the Sherwood number enhances mass transfer and shifts the system toward the adsorption-sensitive domain. Conversely, increasing RR decreases kmk_m and favors transport control. Because kmk_m remains constant while ϕ\phi increases during adsorption, the quantity Γ∞ka(1−ϕ)\Gamma_{\infty}k_a(1-\phi) decreases with time. A convected-drop experiment that starts in the adsorption-sensitive regime therefore remains in this regime throughout the relaxation.

For EDGE, shown in the lower row, the mass-transfer coefficient is km=(Af/Ai)D/LDk_m = (A_f/A_i)D/L_D where AfA_f is the area of the renewed liquid film, AiA_i is the interfacial area, and LDL_D is the effective transport length. Increasing LDL_D from 33 to 3030 μm\mu\mathrm{m} decreases the surfactant supply rate and progressively enlarges the transport-sensitive domain. Conversely, reducing LDL_D shifts complete relaxation toward adsorption control. As in the convected-drop geometry, kmk_m remains approximately constant during the measurement. Therefore, when EDGE starts in the adsorption-sensitive regime, it remains adsorption-sensitive as the interface becomes progressively covered.

Convected drops and EDGE consequently share an important advantage over classical pendant drops: their mass-transfer resistance does not increase continuously with interface age. Their practical implementations, however, impose different experimental limitations.

In the microtensiometer developed by Alvarez et al. [11, 12], the flow is imposed around an interface attached to the end of a capillary. Increasing the liquid velocity reduces the thickness of the convective boundary layer and accelerates surfactant transport. However, the maximum accessible velocity is limited by the requirement that the interface remains attached, stable, and sufficiently weakly deformed. Increasing the velocity beyond this range risks displacing or detaching the interface. This limitation restricts the extent to which transport can be accelerated, and the reported dynamics for the studied nonionic surfactants remained controlled by convection–diffusion.

In the approach developed by Baret and coworkers [21, 15], the droplet is transported with the flow through a delay line containing successive constrictions and expansions. Its age is determined from its travel time, while the interfacial tension is inferred from its deformation in the expansion regions. This method can access short interface ages, but the interfacial tension measurement is indirect because it relies on a calibrated hydrodynamic relation between droplet shape and interfacial tension. Moreover, the droplet is successively compressed and dilated as it passes through constrictions and expansions. Its interfacial area and shape therefore change during the measurement. The tension is determined only at selected positions along the device, rather than continuously, which complicates the reconstruction of a complete and weakly perturbed tension–time curve.

EDGE avoids the limitations of convected droplets discussed above by separating the renewal flow from the motion of the measured interface. The aqueous solution flows alongside the pores, whereas the oil meniscus remains pinned inside the shallow cavity during most of the adsorption period. Its area and curvature therefore vary only weakly. Increasing the aqueous velocity UU reduces the convective depletion-layer thickness according to δU∝U−1/2\delta_U \propto U^{-1/2}, and consequently decreases the effective transport length LDL_D. Within the limits imposed by pressure losses and meniscus stability, the renewal velocity can therefore be increased to reduce LDL_D and move the experiment toward adsorption control without forcing the measured interface to travel through the device.

The EDGE geometry also provides a more direct measurement of interfacial tension. The applied pressure selects a capillary threshold, while the waiting time before droplet release defines the corresponding interface age. The interfacial tension is therefore obtained from the Young–Laplace relation without reconstructing it from the transient deformation of a moving droplet. EDGE thus combines a fixed micrometer-scale transport distance, tunable convective renewal, a nearly stationary interface, and a direct capillary measurement of dynamic interfacial tension.

Sequential identification of interfacial thermodynamics, transport, and adsorption kinetics

The interfacial tension measurements contain information about three distinct physical contributions: the thermodynamics of the adsorbed layer, surfactant transport through the liquid, and molecular transfer at the interface. These contributions cannot be determined reliably by fitting all parameters simultaneously to a single dynamic curve. We therefore use a sequential strategy in which the equilibrium, transport, and kinetic parameters are determined from complementary measurements before being combined in a common nonequilibrium model.

A central difficulty, that is also encountered in most studies in the literature [4, 22], is that interfacial tension is measured directly, whereas the surface coverage ϕ\phi and the subsurface concentration csc_s are not. Converting a measured tension–time curve into an adsorption trajectory therefore requires an interfacial thermodynamic model. In particular, the equilibrium equation of state cannot in general be applied at each instant of a dynamic experiment, because this would implicitly assume that the adsorbed layer remains in local equilibrium with the subsurface solution. This assumption is inappropriate when molecular adsorption is the rate-limiting process. We therefore use the Diamant–Andelman framework [23, 24], which expresses the interfacial tension as a function of both csc_s and ϕ\phi.

Step 1: equilibrium interfacial thermodynamics

Equilibrium interfacial-tension measurements are first used to determine the adsorption isotherm and the equation of state. At equilibrium, cs=cbc_s=c_b and ϕ=ϕeq\phi=\phi_{\mathrm{eq}}.

For a nonionic surfactant, the equilibrium coverage and interfacial tension are described by the following equations:

ϕeq1−ϕeq=Kcbexp⁡(β∗ϕeq)(5)\frac{\phi_{\mathrm{eq}}}{1-\phi_{\mathrm{eq}}}=Kc_b\exp\left(\beta^*\phi_{\mathrm{eq}}\right) \tag*{(5)}
γeq−γ0=RTΓ∞[ln⁡(1−ϕeq)+β∗2ϕeq2](6)\gamma_{\mathrm{eq}}-\gamma_0=RT\Gamma_\infty\left[\ln\left(1-\phi_{\mathrm{eq}}\right)+\frac{\beta^*}{2}\phi_{\mathrm{eq}}^2\right] \tag*{(6)}

where KK is the equilibrium adsorption constant, Γ∞\Gamma_\infty is the maximum surface excess, and β∗\beta^* accounts for lateral interactions within the adsorbed layer. The Langmuir limit is recovered when β∗=0\beta^*=0. Fitting γeq(cb)\gamma_{\mathrm{eq}}(c_b) therefore determines KK, Γ∞\Gamma_\infty, and, when required, β∗\beta^*.

Outside equilibrium, the interfacial tension is calculated from the complete Diamant–Andelman [23, 24] expression given by the following:

γ(cs,ϕ)−γ0=RTΓ∞[ϕln⁡ϕ+(1−ϕ)ln⁡(1−ϕ)−ϕln⁡(NAa3cs)−α∗ϕ−β∗2ϕ2](7)\gamma(c_s,\phi)-\gamma_0=RT\Gamma_\infty\left[\phi\ln\phi+(1-\phi)\ln(1-\phi)-\phi\ln\left(N_Aa^3c_s\right)-\alpha^*\phi-\frac{\beta^*}{2}\phi^2\right] \tag*{(7)}

where aa is the molecular length scale and α∗\alpha^* represents the affinity of the surfactant for the interface. This relation provides the required conversion between the calculated coverage and the measured dynamic interfacial tension without imposing local equilibrium.

For an ionic surfactant, adsorption is additionally opposed by the electrostatic cost of charging the interface. The equilibrium isotherm is written as follows [25, 26, 27]:

ϕeq1−ϕeq=Kcbexp⁡[β⋆ϕeq−χelF∣ψ0∣RT](8)\frac{\phi_{\mathrm{eq}}}{1-\phi_{\mathrm{eq}}}=Kc_b\exp\left[\beta^{\star}\phi_{\mathrm{eq}}-\chi_{\mathrm{el}}\frac{F|\psi_0|}{RT}\right] \tag*{(8)}

where ψ0\psi_0 is the self-consistent surface potential and χel\chi_{\mathrm{el}} weights the electrostatic contribution to the adsorption free energy. The ionic interfacial tension is calculated using:

γ(cs,ϕ)=γchem(cs,ϕ)+γel(cs,ϕ)(9)\gamma(c_s,\phi)=\gamma_{\mathrm{chem}}(c_s,\phi)+\gamma_{\mathrm{el}}(c_s,\phi) \tag*{(9)}

where γchem\gamma_{\mathrm{chem}} is given by the nonionic Diamant–Andelman expression (Equation 7) and γel\gamma_{\mathrm{el}} accounts for the diffuse-layer free energy. The detailed expressions for ψ0\psi_0 and γel\gamma_{\mathrm{el}} are given in the Supplementary Materials. Equilibrium measurements are used to determine the thermodynamic and electrostatic parameters before any dynamic curve is analyzed.

Step 2: independent determination of bulk diffusion

The diffusion coefficient DD is determined from dilute classical rising-drop measurements. Under these conditions, the large transport distance and the growth of the diffusion layer make the tension relaxation predominantly diffusion-controlled. The curves are analyzed with the Ward–Tordai model [28, 29], and the calculus from Nele et al. [7], using the previously determined equilibrium isotherm (Equation 5) and equation of state (Equation 6). All interfacial thermodynamic parameters are therefore fixed, leaving DD as the only transport parameter determined from the dilute drop data.

Step 3: identification of adsorption kinetics with EDGE

The EDGE measurements are used to determine the molecular adsorption kinetics after the equilibrium thermodynamic parameters and the diffusion coefficient have been determined independently.

Nonionic surfactant: adsorption-controlled regime

In the adsorption-controlled limit, transport from the bulk is sufficiently fast to maintain the subsurface concentration close to the bulk concentration so cs≃cbc_s \simeq c_b. Therefore, no significant depletion develops next to the interface. For a nonionic surfactant, the surface coverage evolves according to

dϕdt=RadsN(cb,ϕ)=kacb(1−ϕ)−kdϕ(10)\frac{\mathrm{d}\phi}{\mathrm{d}t}=R_{\mathrm{ads}}^{\mathrm{N}}(c_b,\phi)=k_a c_b(1-\phi)-k_d\phi \tag*{(10)}

With kd=ka/Kk_d=k_a/K and the equilibrium constant KK already determined from the equilibrium measurements, kak_a is the only kinetic parameter adjusted from the EDGE curves, while kdk_d defined by the equilibrium constraint. The interface is assumed to be initially clean, ϕ(0)=0\phi(0)=0. A dedicated calculation presented in the Supporting Information (S3 and Fig. S2) confirms that the residual initial coverage is negligible under the experimental conditions considered. At each time, the calculated coverage is converted into interfacial tension using the full nonequilibrium Diamant–Andelman expression γ(cb,ϕ)\gamma(c_b,\phi) given by Equation 7 rather than the equilibrium equation of state. This distinction is essential because, in an adsorption-controlled regime, the adsorbed layer is not assumed to remain at equilibrium with the adjacent solution during relaxation.

The fitted value of kak_a is retained as an intrinsic adsorption constant only if the same value describes several concentrations and if transport remains faster than adsorption throughout the measured relaxation. This condition is checked using the regime criterion introduced above, which must remain in the adsorption-sensitive domain, Π<1\Pi<1. If this condition is not satisfied, subsurface depletion must be explicitly accounted for using the mixed adsorption–transport model described in the Supporting Information S5.

Ionic surfactant: electrostatic adsorption model

For an ionic surfactant, we follow the same sequential procedure. The transport criterion of Π\Pi will be checked first, and the mixed adsorption–diffusion model is used only when the transport criterion shows that cs≃cbc_s\simeq c_b is not valid. We first use the electrostatic adsorption law derived from the Diamant–Andelman [23, 24] description. The interfacial free-energy contribution is the following:

uDA(cs,ϕ)=β⋆ϕ−χelF∣ψ0(cs,ϕ)∣RT(11)u_{\mathrm{DA}}(c_s,\phi)=\beta^{\star}\phi-\chi_{\mathrm{el}}\frac{F|\psi_0(c_s,\phi)|}{RT} \tag*{(11)}

where β⋆ϕ\beta^{\star}\phi accounts for lateral interactions and the second term represents the electrostatic cost associated with charging the interface. The corresponding adsorption rate can be written as:

RadsI(cs,ϕ)=kacs(1−ϕ)exp⁡[puDA(cs,ϕ)]−kdϕexp⁡[−(1−p)uDA(cs,ϕ)](12)R_{\mathrm{ads}}^{\mathrm{I}}(c_s,\phi)=k_a c_s(1-\phi)\exp\left[p u_{\mathrm{DA}}(c_s,\phi)\right]-k_d\phi\exp\left[-(1-p)u_{\mathrm{DA}}(c_s,\phi)\right] \tag*{(12)}

The coefficient pp specifies how the interfacial free-energy difference is distributed between the adsorption and desorption pathways. A symmetric partition, p=1/2p=1/2, is used in the following. In the adsorption-controlled limit, the subsurface concentration is set to cs=cbc_s=c_b and the surface coverage is directly obtained from:

dϕdt=RadsI(cb,ϕ)(13)\frac{\mathrm{d}\phi}{\mathrm{d}t}=R_{\mathrm{ads}}^{\mathrm{I}}(c_b,\phi) \tag*{(13)}

The calculated coverage is converted into interfacial tension using the complete nonequilibrium ionic equation of state (9), including both the chemical and electrostatic contributions. The equilibrium, electrostatic, and transport parameters remain fixed in both the adsorption-controlled and mixed analyses. For more details on the calculations, refer to Supporting Information S4.

Results and discussion

Microfluidic EDGE vs classical rising-drop tensiometry: experimental results

Figure 3 compares the dynamic interfacial tension of C10_{10}E8_8 solutions measured with the EDGE tensiometer and with classical rising-drop tensiometry. The two techniques probe the same water/hexadecane interface and the same concentration range, from 0.01 CMC to 10 CMC, but they access very different regimes.

Dynamic interfacial tension of C10E8 solutions at the water/hexadecane interface

Figure 3. Dynamic interfacial tension of C10_{10}E8_8 solutions at the water/hexadecane interface, measured using the EDGE tensiometer (circles) and the rising drop tensiometry (ADT) (triangles). The concentration of the C10_{10}E8_8 solutions ranges from 0.01 CMC to 10 CMC, with CMC = 1.4 mol m−3^{-3}. The error bars of the rising drop tensiometry measurements are included within the triangles. The orange and yellow lines represent the diffusion model from Ward–Tordai [28, 29] using Nele’s calculus [7] for D=1.0×10−9 m2 s−1D = 1.0 \times10^{-9}\ \mathrm{m^2\,s^{-1}}, for c=0.01c = 0.01 CMC and c=0.025c = 0.025 CMC respectively.

A first striking feature is the much faster relaxation measured in the EDGE device. For a given concentration, the decrease of γ\gamma occurs at significantly shorter times than in the rising-drop experiment. This difference is particularly clear at intermediate and high concentrations, where the EDGE curves reach their plateau within a few tens to hundreds of milliseconds, whereas the rising-drop data evolve over seconds to tens of seconds. The effect is not a change of equilibrium adsorption: at long times, both techniques tend towards comparable limiting surface tensions. Instead, it is a consequence of the different transport conditions. At the lowest concentrations, the diffusion coefficient was extracted using the diffusion-limited Ward–Tordai model [28, 29] applied to the rising-drop data. The best-fit gave D=1.0×10−9 m2 s−1D = 1.0 \times10^{-9}\ \mathrm{m^2\,s^{-1}}, a value used in the analysis that follows.

Figure 4 reports the same comparison for SDS solutions at the water/hexadecane interface. In this case, the contrast between the two techniques is even more pronounced. The EDGE measurements display a rapid decrease of the interfacial tension over short times, with a concentration-dependent relaxation occurring between a few milliseconds and about one second. Increasing the SDS concentration accelerates the decrease of γ\gamma and lowers the final measured value, as expected from the larger surfactant supply to the interface.

Dynamic interfacial tension of SDS solutions at the water/hexadecane interface

Figure 4. Dynamic interfacial tension of SDS solutions at the water/hexadecane interface, measured using the microfluidic EDGE tensiometer (circles) and the rising drop tensiometer (triangles). The concentration of the surfactant solutions in SDS ranges from 0.05% wt (= 0.2 CMC) to 0.3% wt (= 1.2 CMC). Note that CMCSDS_{\mathrm{SDS}} = 0.25% wt = 8.7 mol m−3^{-3}.

By contrast, the rising-drop measurements show almost horizontal curves over the accessible time window: the classical tensiometer fails to resolve the early adsorption dynamics and essentially only captures the equilibrium interfacial tension. The dynamic regime is therefore accessible only with the EDGE device, which probes much shorter times. This comparison shows that the dynamic surface tension is not an intrinsic function of time for a given surfactant solution: it depends on how surfactant molecules are transported to the interface — from a nearly quiescent, macroscopic reservoir in the rising-drop tensiometer, to a micrometer-scale reservoir continuously renewed by convection in the EDGE device.

The two techniques therefore probe distinct transport regimes: diffusion-controlled mass transfer over macroscopic distances for the rising drop, versus adsorption under rapid convective renewal and micrometer-scale depletion for EDGE. This is why the same C10_{10}E8_8 solution gives different apparent dynamic curves in the two instruments, especially at low concentration, even though the underlying equilibrium adsorption isotherm is identical.

In the following, we use the modeling framework developed above to analyze the experimental data and to extract both the thermodynamic parameters of the adsorption isotherm and the kinetic parameters governing the mass transport of both studied surfactants.

Model analysis in the C10_{10}E8_8 situation

Sequential analysis of nonionic C10_{10}E8_8 adsorption

The C10_{10}E8_8 data were analyzed sequentially to separate interfacial thermodynamics, bulk diffusion, and molecular adsorption. The equilibrium measurements were first used to determine the adsorption isotherm and the interfacial equation of state. The diffusion coefficient was obtained independently from classical rising-drop measurements. Finally, the short-time EDGE curves were used to determine the adsorption and desorption rate constants, and the validity of the adsorption-controlled interpretation was verified a posteriori.

Equilibrium interfacial thermodynamics

The equilibrium interfacial tension was determined from the long-time plateau measured at each C10_{10}E8_8 concentration. The equilibrium surface coverage was described by the Langmuir isotherm, and the corresponding interfacial tension was calculated from the equilibrium equation of state, given by (5) and (6) respectively. These equations are simplified by assuming an ideal adsorbed layer with β∗=0\beta^{*}=0, and give the following:

ϕeq=Kcb1+Kcb(14)\phi_{\mathrm{eq}} = \frac{Kc_b}{1+Kc_b} \tag*{(14)}
γeq−γ0=RTΓ∞ln⁡(1−ϕeq)\gamma_{\mathrm{eq}}-\gamma_0 = RT\Gamma_{\infty}\ln(1-\phi_{\mathrm{eq}})

The equilibrium data were well described by this model, as shown in Figure 5. The best-fit gave: Γ∞=2.20×10−6\Gamma_{\infty}=2.20\times10^{-6} mol m−2^{-2} and K=2074K=2074 m3^{3} mol−1^{-1}. These values correspond to a molecular length a=8.70a=8.70 Å and an adsorption affinity α/(kBT)=15.47\alpha/(k_{\mathrm{B}}T)=15.47. The RMS deviation between the equilibrium measurements and the model was 0.45 mN m−1^{-1}. These thermodynamic parameters were subsequently fixed in all dynamic analyses.

Equilibrium interfacial tension as a function of concentration

Figure 5. Equilibrium interfacial tension obtained from the long-time plateau of the interfacial tension at the water/hexadecane interface with the best-fit from the classical Langmuir model (line). The colored circles refer to the data obtained with the EDGE tensiometer at long time, while the black circles are data points measured with the classical rising drop tensiometer. Equilibrium was reached at t=50t=50 s.

Diffusion coefficient from classical drop tensiometry

The diffusion coefficient of C10_{10}E8_8 was determined independently from the dilute rising-drop measurements. In this geometry, the diffusion distance increases continuously with time and the measured relaxation is primarily controlled by transport from the bulk solution.

The dynamic curves were fitted with the Ward–Tordai model [28, 29], assuming local equilibrium between the adsorbed layer and the subsurface concentration. The equilibrium parameters KK, Γ∞\Gamma_{\infty}, and β∗=0\beta^{*}=0 were fixed to the values obtained above, leaving DD as the only adjustable transport parameter. The best fit using Nele’s calculus [7] gave D=1.0×10−9D=1.0\times10^{-9} m2^{2} s−1^{-1}, a reasonable value given the size of the molecule. The agreement between the model and the dilute drop measurements (solid lines in Figure 3) confirms that the classical-drop relaxation can be used to determine the diffusion coefficient independently of the adsorption kinetics. This value of DD was therefore fixed in the subsequent EDGE analysis.

Adsorption kinetics from EDGE measurements

The EDGE measurements were then analyzed using the thermodynamic parameters and diffusion coefficient determined independently above. The effective EDGE transport length was calculated using (2) with the fixed value of DD, from the device geometry and under the hydrodynamic renewal conditions, giving LD=7.3L_{\mathrm{D}}=7.3 μm.

We first analyzed the EDGE curves of C10_{10}E8_8 using (10) with an initially clean interface, ϕ(0)=0\phi(0)=0. Because KK was fixed by the equilibrium analysis, kak_a was the only dynamically adjusted parameter. At each time, the calculated coverage was converted into interfacial tension using (7).

A single pair of kinetic constants described the EDGE curves over the investigated concentration range, gave by the best-fit: ka=17.2k_a=17.2 m3^{3} mol−1^{-1} s−1^{-1} and kd=8.3×10−3k_d=8.3\times10^{-3} s−1^{-1}, with the corresponding RMS deviation of 1.75 mN m−1^{-1}. Figure 6A shows the good agreement between the adsorption-limited model and the experimental dynamic interfacial tension data measured by the EDGE device. The adsorption-controlled assumption was verified a posteriori using the fitted coverage trajectory. A dedicated numerical routine was used to calculate, at every experimental time and for each bulk concentration, the regime parameter Π(t)\Pi(t) as follows:

Π(t)=Γ∞ka(1−ϕ(t))km(15)\Pi(t)=\frac{\Gamma_{\infty}k_a(1-\phi(t))}{k_m} \tag*{(15)}

with km=(Af/Ai)D/LDk_m=(A_f/A_i)D/L_D. As shown in Figure 6B, the calculated value of Π(t)\Pi(t) remained below unity throughout the analyzed relaxation and for all concentrations included in the fit. Bulk transport was therefore sufficiently fast for the EDGE response to remain sensitive to the intrinsic molecular adsorption rate. This a posteriori verification supports the identification of kak_a and kdk_d as intrinsic kinetic constants rather than apparent parameters compensating for an unknown transport limitation. Note that the analysis was also performed using a mixed model, but no significant improvement or change in the kinetic parameters was observed, as detailed in the Supplementary Materials. The thermodynamic and kinetic parameters obtained for C10_{10}E8_8 are summarized in Table 1. Importantly, even though numerous parameters are reported in this table, kak_a is the only parameter fitted to the EDGE dynamics; all thermodynamic and transport quantities are independently constrained. This minimizes parameter compensation and allows the extracted adsorption rate to be interpreted as an intrinsic kinetic constant rather than an effective fit parameter. This is a key strength of our method.

ParameterSymbolValue
Maximum surface excessΓ∞\Gamma_\infty2.20×10−6 mol m−22.20 \times10^{-6}\ \mathrm{mol\ m^{-2}}
Adsorption equilibrium constantKK2.07×103 m3 mol−12.07 \times10^{3}\ \mathrm{m^3\ mol^{-1}}
Lateral interaction parameterβ∗\beta^*0
Molecular lengthaa8.70 Å
Adsorption affinityα/(kBT)\alpha/(k_\mathrm{B}T)15.47
Equilibrium-fit coefficientR2R^20.998
Equilibrium-fit errorRMSEeq\mathrm{RMSE}_{\mathrm{eq}}0.45 mN m−10.45\ \mathrm{mN\ m^{-1}}
Diffusion coefficientDD1.0×10−9 m2 s−11.0 \times10^{-9}\ \mathrm{m^2\ s^{-1}}
EDGE transport lengthLDL_D7.3 μm
Mass-transfer coefficientkmk_m1.89×10−4 m s−11.89 \times10^{-4}\ \mathrm{m\ s^{-1}}
Adsorption rate constantkak_a17.1 m3 mol−1 s−117.1\ \mathrm{m^3\ mol^{-1}\ s^{-1}}
Desorption rate constantkdk_d8.3×10−3 s−18.3 \times10^{-3}\ \mathrm{s^{-1}}
Dynamic-fit errorRMSEdyn\mathrm{RMSE}_{\mathrm{dyn}}1.75 mN m−11.75\ \mathrm{mN\ m^{-1}}

Table 1. Thermodynamic, transport, and kinetic parameters obtained for C10_{10}E8_8 at the water/hexadecane interface.

Adsorption-controlled dynamics of C10E8 in the EDGE geometry

Figure 6. Adsorption-controlled dynamics of C10_{10}E8_8 in the EDGE geometry. (A) Dynamic interfacial tension measured for C10_{10}E8_8 concentrations ranging from 0.01 to 1 CMC (symbols). Solid lines are simultaneous fits of the adsorption-limited model, yielding ka=17.2 m3 mol−1 s−1k_a = 17.2\ \mathrm{m^3\ mol^{-1}\ s^{-1}} and kd=8.3×10−3 s−1k_d = 8.3 \times10^{-3}\ \mathrm{s^{-1}}, with an overall RMS error of 1.75 mN m−11.75\ \mathrm{mN\ m^{-1}}. (B) Regime map showing log⁡10Π\log_{10} \Pi as a function of interface age and normalized bulk concentration cb/CMCc_b/\mathrm{CMC}, where Π=Γ∞ka(1−ϕ)/km\Pi= \Gamma_\infty k_a(1-\phi)/k_m compares the characteristic adsorption demand with mass transport toward the interface. Symbols indicate the experimental conditions used in the fit. The dashed line corresponds to Π=0.1\Pi= 0.1; Π<1\Pi< 1 (and log⁡(Π)<0\log(\Pi) < 0) identifies conditions for which interfacial adsorption kinetics, rather than bulk transport, controls the measured tension dynamics.

Model analysis for SDS

Unlike C10_{10}E8_8, SDS is an ionic surfactant. Its adsorption therefore involves not only transport through the liquid and molecular transfer to the interface, but also the electrostatic cost associated with charging the adsorbed layer. The SDS analysis consequently requires an ionic thermodynamic and kinetic description. A second important difference is experimental. For C10_{10}E8_8, the rising-drop measurements provide a sufficiently extended diffusion-controlled relaxation to determine the diffusion coefficient independently before analyzing the EDGE data. This is not possible for SDS. Over the accessible time window, the rising-drop curves are already close to their long-time values and do not resolve the early adsorption dynamics. They are therefore used primarily to constrain equilibrium interfacial thermodynamics, while the SDS diffusion coefficient is fixed from literature data. We use D=5.0×10−10 m2 s−1D = 5.0 \times10^{-10}\ \mathrm{m^2\ s^{-1}}, as reported for SDS at a closely related water–alkane interface [30]. With the EDGE geometry and renewal conditions, the effective transport length is LD=5.77 μmL_D = 5.77\ \mathrm{\mu m} and km=6.67×10−5 m s−1k_m = 6.67 \times10^{-5}\ \mathrm{m\ s^{-1}}. Thus, as for C10_{10}E8_8, the transport parameters entering the EDGE analysis are fixed independently of the dynamic fits. The analysis can then focus on the additional thermodynamic and kinetic effects associated with the ionic nature of SDS.

Equilibrium thermodynamics of SDS

We first determine the ionic interfacial thermodynamics independently from the dynamic EDGE measurements. For an ionic surfactant, adsorption is opposed by the electrostatic work required to charge the interface. The equilibrium isotherm and the corresponding interfacial tension are given by (8) and (9) respectively. The details of the calculations are given in the Supporting Information S4.

The ionic equilibrium model contains several thermodynamic and electrostatic parameters that may be strongly correlated if they are all adjusted simultaneously. We therefore begin with the simplest description compatible with the data. Lateral interactions are neglected by fixing β⋆=0\beta^\star= 0, so that deviations from Langmuir adsorption arise here from the electrostatic contribution rather than from an additional adjustable Frumkin interaction.

We also fix the electrostatic adsorption factor to χel=0.14\chi_{\mathrm{el}} = 0.14. This value is physically consistent with the strong counterion association known for SDS. Indeed, if α\alpha denotes the effective ionization fraction, the interfacial charge density is reduced according to σeff=αFΓ\sigma_{\mathrm{eff}} = \alpha F\Gamma and, because the Gouy–Chapman surface potential depends nonlinearly on the surface charge, χel\chi_{\mathrm{el}} can be approximated as χel≃α∣ψGC(ασ)∣/∣ψGC(σ)∣\chi_{\mathrm{el}} \simeq\alpha\lvert\psi_{\mathrm{GC}}(\alpha\sigma)\rvert/\lvert\psi_{\mathrm{GC}}(\sigma)\rvert. Using α≃0.20\alpha\simeq0.20, consistent with reported counterion dissociation fractions for SDS aggregates [31], gives χel≃0.14\chi_{\mathrm{el}} \simeq0.14 over the present concentration range. We therefore use χel\chi_{\mathrm{el}} as an effective weighting of the electrostatic contribution to the adsorption free energy, rather than as a direct measurement of the ionization degree of the planar interface.

The interfacial tension contains a separate diffuse-layer contribution, γel\gamma_{\mathrm{el}}. Its magnitude is scaled by the prefactor χγ\chi_\gamma, which enters the interfacial equation of state but not the adsorption free-energy barrier. Thus, χel\chi_{\mathrm{el}} and χγ\chi_\gamma play distinct roles in the model: the former controls the electrostatic penalty for adsorption, whereas the latter sets the magnitude of the diffuse-layer contribution to the measured interfacial tension.

With β⋆\beta^\star and χel\chi_{\mathrm{el}} fixed, the equilibrium fit determines Γ∞\Gamma_\infty, KK, and χγ\chi_\gamma. The best fit gives: Γ∞=6.0×10−6 mol m−2\Gamma_\infty= 6.0 \times10^{-6}\ \mathrm{mol\,m^{-2}}, K=5.18 m3 mol−1K = 5.18\ \mathrm{m^3\,mol^{-1}} and χγ=0.15\chi_\gamma= 0.15.

With the present sign convention, the diffuse-layer free-energy contribution γel\gamma_{\mathrm{el}} is positive and therefore partially compensates the decrease in interfacial tension produced by adsorption. As shown in Figure 7, this contribution reaches approximately 15 mN m−115\ \mathrm{mN\,m^{-1}} over the investigated concentration range. All equilibrium parameters are subsequently fixed before analyzing the dynamic EDGE measurements.

Equilibrium interfacial tension of SDS at the water–hexadecane interface

Figure 7. Equilibrium interfacial tension of SDS at the water–hexadecane interface as a function of normalized bulk concentration cb/CMCc_b/\mathrm{CMC}. Colored circles correspond to the equilibrium values associated with the EDGE concentrations, and the black circle at cb=0c_b = 0 corresponds to the clean interface, γ0=48.9 mN m−1\gamma_0 = 48.9\ \mathrm{mN\,m^{-1}}. The solid black line is the total equilibrium fit obtained with the ionic adsorption model and β⋆=0\beta^\star= 0. The dashed line represents the chemical contribution γchem\gamma_{\mathrm{chem}}. The shaded region represents γel=γeq−γchem\gamma_{\mathrm{el}} = \gamma_{\mathrm{eq}} - \gamma_{\mathrm{chem}}, i.e. the positive free-energy contribution associated with charging the interface and forming the diffuse ionic layer.

Adsorption-controlled description

We first test whether the SDS dynamics can be described in the same adsorption-controlled limit identified for C10_{10}E8_8. In this limit, transport is assumed to be sufficiently fast to maintain cs=cbc_s = c_b. No additional phenomenological kinetic barrier is introduced.

With β⋆=0\beta^\star= 0, the ionic adsorption rate is given by RadsI(cb,ϕ)R^I_{\mathrm{ads}}(c_b,\phi) from (12) with uDA=−χelF∣ψ0∣/RTu_{\mathrm{DA}} = -\chi_{\mathrm{el}}F\lvert\psi_0\rvert/RT and p=1/2p = 1/2. Because KK and all other equilibrium parameters have already been determined, kak_a is the only dynamically adjusted parameter. A simultaneous fit of the four SDS concentrations gives: ka=9.63 m3 mol−1 s−1k_a = 9.63\ \mathrm{m^3\,mol^{-1}\,s^{-1}}, kd=k_d = 1.86 s−11.86\ \mathrm{s}^{-1}, with RMSEads=1.32 mN m−1\mathrm{RMSE}_{\mathrm{ads}} = 1.32\ \mathrm{mN}\,\mathrm{m}^{-1} and Rads2=0.978R_{\mathrm{ads}}^{2}=0.978. The adsorption-only model therefore captures the overall timescale and concentration dependence of the SDS relaxation reasonably well. However, fig:S3 shows that it does not reproduce the complete shape of the experimental curves: systematic deviations remain during the relaxation. This suggests that the approximation cs=cbc_s=c_b may not be fully satisfied. Because kmk_m is independently known, this assumption can be tested without readjusting the transport rate. For the ionic adsorption law we define:

Πeff(t)=Γ∞ka(1−ϕ(t))kmexp⁡(pu)(16)\Pi_{\mathrm{eff}}(t)=\frac{\Gamma_{\infty}k_a(1-\phi(t))}{k_m}\exp(pu) \tag*{(16)}

Over all experimental conditions, as shown in Figure 8, Πeff,max≃0.50\Pi_{\mathrm{eff,max}}\simeq0.50. The criterion remains below unity, confirming that the measurements remain sensitive to molecular adsorption. However, Πeff\Pi_{\mathrm{eff}} is not much smaller than unity. Transport is therefore not sufficiently faster than adsorption to guarantee cs=cbc_s=c_b throughout the complete relaxation. The systematic shift observed in fig:S3, together with this independent regime criterion, motivate a mixed description in which the same interfacial adsorption kinetics are retained but the subsurface concentration is allowed to evolve in time.

Interfacial dynamics of SDS in the EDGE geometry

Figure 8. Interfacial dynamics of SDS in the EDGE geometry. Regime map showing log⁡10(Πeff)\log_{10}(\Pi_{\mathrm{eff}}) as a function of interface age and normalized bulk concentration, using the best-fit model from the adsorption model (assuming cb=csc_b = c_s) with ka=9.63 m3 s−1k_a = 9.63\ \mathrm{m^3\,s^{-1}}, and kd=1.86 s−1k_d = 1.86\ \mathrm{s^{-1}} with an overall RMS error of RMSEads=1.32 mN m−1\mathrm{RMSE}_{\mathrm{ads}} = 1.32\ \mathrm{mN\,m^{-1}}. Symbols indicate the experimental conditions included in the fit. Πeff<1\Pi_{\mathrm{eff}} < 1 identifies conditions for which the EDGE measurements remain sensitive to interfacial adsorption kinetics, whereas Πeff≪1\Pi_{\mathrm{eff}} \ll1 would correspond to a strictly adsorption-controlled regime. Here, Πeff\Pi_{\mathrm{eff}} reaches values of order 0.5, indicating that adsorption remains identifiable but transport effects are not negligible.

Mixed adsorption–transport dynamics

We therefore introduce a finite subsurface reservoir whose concentration can transiently depart from the bulk value. Its dynamics is written as follows:

dcsdt=DLDLs(cb−cs)−AiAfΓ∞LsRadsI(cs,ϕ)(17)\frac{d c_s}{d t}=\frac{D}{L_D L_s}(c_b-c_s)-\frac{A_i}{A_f}\frac{\Gamma_{\infty}}{L_s}R_{\mathrm{ads}}^{I}(c_s,\phi) \tag*{(17)}

together with

dϕdt=RadsI(cs,ϕ)(18)\frac{d\phi}{d t}=R_{\mathrm{ads}}^{I}(c_s,\phi) \tag*{(18)}

Here, LsL_s is an effective storage length. It should not be interpreted as the thickness of a molecular interfacial layer. Rather, it represents the effective depth of the near-interface liquid reservoir whose concentration can be transiently depleted by adsorption before being replenished from the bulk.

In the absence of adsorption, (17) gives a characteristic replenishment time:

τs=LDLsD(19)\tau_s=\frac{L_D L_s}{D} \tag*{(19)}

Importantly, no new interfacial kinetic mechanism is introduced in this second description. The same electrostatic adsorption law is retained for RadsI(cs,ϕ)R_{\mathrm{ads}}^{I}(c_s,\phi) using (12), with β⋆=0\beta^{\star}=0, χel=0.14\chi_{\mathrm{el}}=0.14, and all equilibrium and transport parameters fixed. Only kak_a and LsL_s are adjusted.

The simultaneous fit gives ka=10.62 m3 s−1k_a=10.62\ \mathrm{m}^{3}\,\mathrm{s}^{-1}, Ls=6.95 μmL_s=6.95\ \mu\mathrm{m}, and therefore kd=2.05 s−1k_d=2.05\ \mathrm{s}^{-1}. The agreement improves to RMSEmix=0.988 mN m−1\mathrm{RMSE}_{\mathrm{mix}}=0.988\ \mathrm{mN}\,\mathrm{m}^{-1} and Rmix2=0.99R_{\mathrm{mix}}^{2}=0.99. Figure 9 shows that introducing a finite subsurface response improves not only the global fit but also the shape of the relaxation. The calculated cs/cbc_s/c_b trajectories in Figure 9B provide a direct interpretation: rapid

adsorption initially depletes the near-interface reservoir, followed by diffusive replenishment on a characteristic timescale of approximately 70 ms.

The fitted adsorption constant changes only moderately, from 9.639.63 to 10.62 m3 mol−1 s−110.62\ \mathrm{m}^{3}\,\mathrm{mol}^{-1}\,\mathrm{s}^{-1}, when the transient transport contribution is introduced. This relative stability supports the conclusion that EDGE remains primarily sensitive to the interfacial adsorption kinetics, while the improved description of the curve shape requires accounting for a finite subsurface transport memory. Thus, for SDS, electrostatic effects are required to describe the ionic adsorption thermodynamics and kinetics, but electrostatics alone do not fully account for the measured relaxation. A transient subsurface depletion and replenishment provides the additional contribution needed to reproduce the dynamics without introducing an extra coverage-dependent phenomenological kinetic barrier. To summarize our results, the thermodynamic, transport, and kinetic parameters obtained for SDS are summarized in Table 2.

QuantitySymbolValue
Critical micelle concentrationCMC8.67 mol m−38.67\ \mathrm{mol}\,\mathrm{m}^{-3}
Clean interfacial tensionγ0\gamma_048.9 mN m−148.9\ \mathrm{mN}\,\mathrm{m}^{-1}
Maximum surface excessΓ∞\Gamma_{\infty}6.0×10−6 mol m−26.0\times10^{-6}\ \mathrm{mol}\,\mathrm{m}^{-2}
Equilibrium adsorption constantKK5.18 m3 mol−15.18\ \mathrm{m}^{3}\,\mathrm{mol}^{-1}
Lateral interaction parameterβ⋆\beta^{\star}0
Electrostatic adsorption factorχel\chi_{\mathrm{el}}0.14
Electrostatic tension factorχγ\chi_{\gamma}0.15
Molecular lengthaa5.26 A˚5.26\ \text{\AA}
Adsorption affinityα/(kBT)\alpha/(k_{\mathrm{B}}T)10.99
Equilibrium-fit coefficientReq2R_{\mathrm{eq}}^{2}0.998
Equilibrium-fit errorRMSEeq\mathrm{RMSE}_{\mathrm{eq}}0.68 mN m−10.68\ \mathrm{mN}\,\mathrm{m}^{-1}
Diffusion coefficientDD5.0×10−10 m2 s−15.0\times10^{-10}\ \mathrm{m}^{2}\,\mathrm{s}^{-1}
EDGE transport lengthLDL_D5.77 μm5.77\ \mu\mathrm{m}
Mass-transfer coefficientkmk_m6.28×10−5 m s−16.28\times10^{-5}\ \mathrm{m}\,\mathrm{s}^{-1}
Adsorption rate, cs=cbc_s=c_bkak_a9.63 m3 mol−19.63\ \mathrm{m}^{3}\,\mathrm{mol}^{-1}
Desorption rate, cs=cbc_s=c_bkdk_d1.86 s−11.86\ \mathrm{s}^{-1}
Adsorption-only errorRMSEads\mathrm{RMSE}_{\mathrm{ads}}1.32 mN m−11.32\ \mathrm{mN}\,\mathrm{m}^{-1}
Maximum effective regime parameterΠeff,max\Pi_{\mathrm{eff,max}}0.50
Mixed-model adsorption ratekak_a10.62 m3 mol−110.62\ \mathrm{m}^{3}\,\mathrm{mol}^{-1}
Subsurface storage lengthLsL_s6.95 μm6.95\ \mu\mathrm{m}
Mixed-model desorption ratekdk_d2.05 s−12.05\ \mathrm{s}^{-1}
Subsurface relaxation timeτs=LDLs/D\tau_s=L_D L_s/D6.92×10−2 s6.92\times10^{-2}\ \mathrm{s}
Mixed-model errorRMSEmix\mathrm{RMSE}_{\mathrm{mix}}0.936 mN m−10.936\ \mathrm{mN}\,\mathrm{m}^{-1}
Mixed-model coefficientRmix2R_{\mathrm{mix}}^{2}0.988

Table 2. Thermodynamic, transport, and kinetic parameters obtained for SDS at the water/hexadecane interface.

Conclusion

We have developed a strategy to identify surfactant adsorption kinetics from dynamic interfacial-tension measurements by explicitly accounting for the transport conditions imposed by the experimental geometry. The central result is that access to short interface ages is not, by itself, sufficient to determine intrinsic adsorption kinetics. The measured relaxation reflects interfacial thermodynamics, molecular adsorption, and surfactant transport, and these contributions must be independently constrained before an adsorption rate constant can be identified. The EDGE geometry provides this separation by combining a nearly stationary interface with micrometer-scale transport distances and controlled hydrodynamic renewal. Equilibrium measurements first determine the adsorption isotherm and interfacial equation of state, while independent transport measurements or estimates constrain the diffusion coefficient and the EDGE mass-transfer resistance. A regime criterion then establishes whether a given dynamic measurement is actually sensitive to molecular adsorption. For the nonionic surfactant C10_{10}E8_8, the EDGE measurements remain in the adsorption-sensitive regime over the investigated conditions. A single adsorption rate constant, ka=17.1 m3 s−1k_a = 17.1\ \mathrm{m^3\,s^{-1}}, describes all concentrations once equilibrium thermodynamics and transport are independently fixed. The corresponding kinetic constants can therefore be interpreted as intrinsic interfacial parameters rather than effective quantities compensating for an unknown mass-transfer resistance. The ionic surfactant SDS illustrates a more complex regime. The electrostatic contribution to the adsorption free energy is required to describe its equilibrium and dynamic behavior, but an adsorption-only model does not reproduce the complete relaxation. Allowing the near-interface concentration to evolve accounts for a transient depletion followed by diffusive replenishment. The resulting storage length, Ls=6.95 μmL_s = 6.95\ \mathrm{\mu m}, corresponds to a characteristic replenishment time of approximately 70 ms and yields ka=10.62 m3 s−1k_a = 10.62\ \mathrm{m^3\,s^{-1}}. Importantly, introducing this transport memory changes the inferred adsorption rate only moderately and removes the need for an additional phenomenological coverage-dependent kinetic barrier. These results establish a general distinction between measuring a young interface and measuring intrinsic adsorption kinetics. The latter requires not only sufficient temporal resolution, but also quantitative control of the transport history experienced by the interface. By combining controlled geometry, independently constrained transport, and nonequilibrium interfacial thermodynamics, EDGE tensiometry provides a route to distinguish molecular adsorption from diffusion and transient subsurface depletion, and to determine when adsorption rate constants can be meaningfully transferred between experimental geometries.

Mixed-adsorption dynamics of SDS in the EDGE geometry

Figure 9. Mixed-adsorption dynamics of SDS in the EDGE geometry. (A) Dynamic interfacial tension measured for SDS concentrations ranging from 0.05% wt (0.2 CMC) to 0.2% wt (0.8 CMC) (symbols). Solid lines are simultaneous fits of the mixed model (cb≠csc_b \ne c_s), yielding ka=10.62 m3 mol−1 s−1k_a = 10.62\ \mathrm{m^3\,mol^{-1}\,s^{-1}}, kd=2.05 s−1k_d = 2.05\ \mathrm{s^{-1}} and Ls=6.95 μmL_s = 6.95\ \mathrm{\mu m}, with an overall RMSmix=0.936 mN m−1\mathrm{RMS}_{\mathrm{mix}} = 0.936\ \mathrm{mN\,m^{-1}}. (B) Analysis of transient subsurface storage of SDS dynamics in the EDGE geometry. Evolution of the normalized subsurface concentration cs/cbc_s/c_b. cs(t)c_s(t) is calculated using the mixed-model described above. The transient decrease of cs/cbc_s/c_b quantifies local depletion caused by adsorption, whereas its subsequent recovery reflects diffusive replenishment from the bulk.

Materials and methods

Materials

Hexadecane (Reagent Plus 99%, Sigma-Aldrich) was used as the dispersed phase for all experiments. Hexadecane was first brought into contact with pure water to remove any impurities into the aqueous phase. Two surfactants were used: nonionic 3,6,9,12,15,18,21,24-octaoxatetratriacontan-1-ol (or octaethylene glycol monodecyl ether) C10_{10}E8_8, (BioXtra ≥\geq 98.0%, Sigma-Aldrich); and ionic sodium dodecylsulfate SDS (ACS reagent ≥\geq 99.0%, Sigma-Aldrich). Aqueous solutions were prepared with ultrapure water (Milli-Q, Merck Millipore) and used as the continuous phase. For the microfluidic experiments, all aqueous solutions were filtered using 0.22 μm\mathrm{\mu m} (Merck, Germany) or 0.2 μm\mathrm{\mu m} PES filters (VWR) before use. Chip cleaning procedures utilize ethanol (96 % v/v, VWR International B.V., the Netherlands) and a piranha solution, prepared in a 3:1 volume ratio of sulfuric acid (96% purity, Sigma-Aldrich, USA) to 35 % wt hydrogen peroxide (Sigma-Aldrich, USA).

Rising drop tensiometry measurements

For long-term measurements, a classical rising drop tensiometer [6] was used. In order to define an initial state of the interface that is as free as possible from surfactant, it is preferable to place the surfactant in solution in the external phase and not in the droplet. Few droplets are always discarded before starting the measurement. As the oil is less dense than the surfactant solutions, we use a rising drop setup: a droplet of hexadecane rises into the water phase, allowing proper interaction with the surfactant. The accessible time range is limited to short periods by the formation of the drop and to long periods by mechanical vibrations or by an equilibrium tension that is too low, which can cause the drop to fall or rise. To conduct these experiments, we used a commercial setup from Tracker, Teclis, France, called an Automated Drop Tensiometer (ADT). This setup enables us to measure dynamic interfacial tension over a timescale ranging from 1 second to one hour. For all the solutions tested, equilibrium is reached within 50 s.

Limitations of the method

The microfluidic EDGE tensiometer is suitable for values of γ\gamma below ≈45\approx45 mN m−1^{-1}. Note that with the EDGE tensiometer, it is almost impossible to measure the interfacial tension of pure fluids at very long time. The rising drop tensiometer is best suited for measuring equilibrium interfacial tension γeq\gamma_{eq} while the microfluidic EDGE tensiometer excels at capturing short-time adsorption kinetics. Both techniques handle a wide range of surfactants and concentrations, and together they allow diffusion and adsorption to be resolved independently.

Data treatment

Videos and images were recorded with a high-speed camera (FASTCAM SA-Z, Photron Limited, Japan) connected to the inverted microscope (Axiovert 200 MAT, Carl Zeiss B.V., The Netherlands). In each experiment, three videos were recorded (at a frame rate of 100 000 fps) for the measurement of droplet formation frequency (f0f_0) at each pore. f0f_0 was obtained by counting the number of droplets formed at the pore within a set period of time. The pore frequency f0f_0 value was first averaged over three independent recordings and then over the eight pores on the plateau. The surfactant mass transport time (i.e., the droplet formation time) is given by 1/f01/f_0. Error bars on time are calculated from the standard deviation of droplet counts of three recorded videos. Error bars for the measurement of interfacial tension are calculated using the uncertainty propagation method.

Authors contribution

C.B.: Writing - original draft, Visualisation, Validation, Methodology, Data acquisition, Data curation, Formal analysis, Conceptualization. B.D.: Writing - review & editing, Visualisation, Validation, Methodology, Conceptualization. C.D.: Writing - review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. K.S.: Writing - review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. A.C.: Writing - original draft, Writing - review & editing, Supervision, Methodology, Funding acquisition, Formal analysis, Conceptualization.

Acknowledgements

This investigation is supported by IFP Energies nouvelles doctorate funding. The authors thank the Laboratory of Process Engineering of Wageningen University and Research (WUR, The Netherlands) for access to the laboratory and equipments. C.B. warmly thanks Professor Benoit Scheid from the Université Libre de Bruxelles (ULB, Belgium) for very interesting discussions regarding the hydrodynamics in this specific geometry. C.B. acknowledges the support and guidance of Marie Marsiglia. C.B. acknowledges the help of Eric Trant on the optimization of the solvers.

References

References

  1. [1]A. A. Maan, K. Schroën, and R. M. Boom, “Spontaneous droplet formation techniques for monodisperse emulsions preparation: Perspectives for food applications,” Journal of Food Engineering, vol. 107, no. 3–4, pp. 334–346, 2011.
  2. [2]S. Schultz, G. Wagner, K. Urban, and J. Ulrich, “High-pressure homogenization as a process for emulsion formation,” Chemical Engineering & Technology, vol. 27, no. 4, pp. 361–368, 2004.
  3. [3]K. Malysa, M. Kraskowska, and M. Krzan, “Influence of surface active substances on bubble motion and collision with various interfaces,” Advances in Colloid and Interface Science, vol. 114–115, pp. 205–225, 2005.
  4. [4]V. Kovalchuk, E. Aksenenko, E. Schneck, and R. Miller, “Surfactant adsorption layers: Experiments and modeling,” Langmuir, vol. 39, no. 10, pp. 3537–3545, 2023.
  5. [5]S.-Y. Lin, K. McKeigue, and C. Maldarelli, “Diffusion-controlled surfactant adsorption studied by pendant drop digitization,” AIChE Journal, vol. 36, no. 12, pp. 1785–1795, 1990.
  6. [6]J. D. Berry, M. J. Neeson, R. R. Dagastine, D. Y. Chan, and R. F. Tabor, “Measurement of surface and interfacial tension using pendant drop tensiometry,” Journal of colloid and interface science, vol. 454, pp. 226–237, 2015.
  7. [7]M. Nele, “An analytically solvable numerical approximation to ward-tordai equation applied to langmuir adsorption,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 648, p. 129186, 2022.DOI
  8. [8]P. R. Garrett and D. R. Ward, “A reexamination of the measurement of dynamic surface tensions using the maximum bubble pressure method,” Journal of Colloid and Interface Science, vol. 132, no. 2, pp. 475–490, 1989.
  9. [9]N. C. Christov, K. D. Danov, P. A. Kralchevsky, K. P. Ananthapadmanabhan, and A. Lips, “Maximum bubble pressure method: Universal surface age and transport mechanisms in surfactant solutions,” Langmuir, vol. 22, no. 18, pp. 7528–7542, 2006.
  10. [10]N. Mishchuk, S. Dukhin, V. Fainerman, V. Kovalchuk, and R. Miller, “Hydrodynamic processes in dynamic bubble pressure experiments: Part 5. the adsorption at the surface of a growing bubble,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 192, no. 1–3, pp. 157–175, 2001.
  11. [11]N. J. Alvarez, L. M. Walker, and S. L. Anna, “A microsiometer to probe the effect of radius of curvature on surfactant transport to a spherical interface,” Langmuir : the ACS journal of surfaces and colloids, vol. 26, no. 16, pp. 13310–13319, 2010.
  12. [12]N. J. Alvarez, D. R. Vogus, L. M. Walker, and S. L. Anna, “Using bulk convection in a microtensiometer to approach kinetic-limited surfactant dynamics at fluid–fluid interfaces,” Journal of colloid and interface science, vol. 372, no. 1, pp. 183–191, 2012.
  13. [13]M. D. Reichert, N. J. Alvarez, C. F. Brooks, A. M. Grillet, L. A. Mondy, S. L. Anna, and L. M. Walker, “The importance of experimental design on measurement of dynamic interfacial tension and interfacial rheology in diffusion-limited surfactant systems,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 467, pp. 135–142, 2015.
  14. [14]Q. Brosseau, J. Vrignon, and J.-C. Baret, “Microfluidic dynamic interfacial tensiometry (µdit),” Soft Matter, vol. 10, no. 17, pp. 3066–3076, 2014.DOI
  15. [15]B. Riechers, F. Maes, E. Akoury, B. Semin, P. Gruner, and J.-C. Baret, “Surfactant adsorption kinetics in microfluidics,” Proceedings of the National Academy of Sciences of the United States of America, vol. 113, no. 41, pp. 11465–11470, 2016.
  16. [16]X. Liang, J. Zhang, M. Li, K. Wang, and G. Luo, “Dynamic interfacial tension and adsorption kinetics of nonionic surfactants during microfluidic droplet formation process,” Chemical Engineering Journal, vol. 445, p. 136658, 2022.
  17. [17]B. Deng, K. Schroën, M. Steegmans, and J. de Ruiter, “Capillary pressure-based measurement of dynamic interfacial tension in a spontaneous microfluidic sensor,” Lab on a Chip, vol. 22, no. 20, pp. 3860–3868, 2022.
  18. [18]B. Deng, K. Schroën, and J. de Ruiter, “Dynamics of bubble formation in spontaneous microfluidic devices: Controlling dynamic adsorption via liquid phase properties,” Journal of Colloid and Interface Science, vol. 622, pp. 218–227, 2022.
  19. [19]T. Porto Santos, B. Deng, M. Corstens, C. Berton-Carabin, and K. Schroën, “Interfacial protein adsorption behavior can be connected across a wide range of timescales using the microfluidic EDGE (Edge-based droplet GEneration) tensiometer,” Journal of Colloid and Interface Science, vol. 674, pp. 951–958, 2024.
  20. [20]J. Cao, T. P. Santos, P. C. del Prado Castillo, B. Deng, M. Corstens, and K. Schroën, “Heat treatment influences adsorption of potato protein at the oil-water interface and emulsion droplet stability at short timescales,” Journal of Colloid and Interface Science, p. 141023, 2026.
  21. [21]Q. Brosseau, J. Vrignon, and J.-C. Baret, “Microfluidic dynamic interfacial tensiometry (µdit),” Soft Matter, vol. 10, no. 17, pp. 3066–3076, 2014.
  22. [22]N. M. Kovalchuk and M. J. H. Simmons, “Review of the role of surfactant dynamics in drop microfluidics,” Advances in Colloid and Interface Science, vol. 312, p. 102844, 2023.
  23. [23]H. Diamant and D. Andelman, “Kinetics of surfactant adsorption at fluid-fluid interfaces,” The Journal of Physical Chemistry, vol. 100, no. 32, pp. 13732–13742, 1996.
  24. [24]H. Diamant, G. Ariel, and D. Andelman, “Kinetics of surfactant adsorption: the free energy approach,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 183–185, pp. 259–276, 2001.
  25. [25]S. Levine, J. Mingins, and G. Bell, “The discrete-ion effect and surface potentials of ionized monolayers,” The Journal of Physical Chemistry, vol. 67, no. 10, pp. 2095–2105, 1963.
  26. [26]A. Bonfillon, F. Sicoli, and D. Langevin, “Dynamic surface tension of ionic surfactant solutions,” Journal of colloid and interface science, vol. 168, no. 2, pp. 497–504, 1994.
  27. [27]A. J. Prosser and E. I. Franses, “Adsorption and surface tension of ionic surfactants at the air–water interface: review and evaluation of equilibrium models,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 178, no. 1-3, pp. 1–40, 2001.
  28. [28]A. F. H. Ward and L. Tordai, “Existence of time-dependence for interfacial tension of solutions,” Nature, vol. 154, no. 3900, pp. 146–147, 1944.
  29. [29]A. F. H. Ward and L. Tordai, “Time-dependence of boundary tensions of solutions i. the role of diffusion in time-effects,” The Journal of Chemical Physics, vol. 14, no. 7, pp. 453–461, 1946.
  30. [30]A. Javadi, N. Mucic, D. Vollhardt, V. Fainerman, and R. Miller, “Effects of dodecanol on the adsorption kinetics of sds at the water–hexane interface,” Journal of colloid and interface science, vol. 351, no. 2, pp. 537–541, 2010.
  31. [31]M. Bennraou, B. L. Bales, and R. Zana, “Effect of the nature of the counterion on the properties of anionic surfactants. 1. cmc, ionization degree at the cmc and aggregation number of micelles of sodium, cesium, tetramethylammonium, tetraethylammonium, tetrapropylammonium, and tetrabutylammonium dodecyl sulfates,” The Journal of Physical Chemistry B, vol. 107, no. 48, pp. 13432–13440, 2003.

Paper details

vlnc:W_az7drh8g
Published
Added
License
CC BY-SA 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents