Ion extraction mechanism studied in a lyotropic lamellar phase

amphiphilic molecules (vesicles 9, micelles and micro-emulsions 10). In the present work, we chose to use lyotropic lamellar phases, periodical stacki...

0 downloads 85 Views 1MB Size
Ion extraction mechanism studied in a lyotropic lamellar phase Amélie Banc, P. Bauduin, Bernard Desbat, Isabelle Ly, Olivier Diat

To cite this version: Amélie Banc, P. Bauduin, Bernard Desbat, Isabelle Ly, Olivier Diat. Ion extraction mechanism studied in a lyotropic lamellar phase. Journal of Physical Chemistry B, American Chemical Society, 2011, 115 (6), pp.1376-1384. �10.1021/jp108585f�. �hal-00578376�

HAL Id: hal-00578376 Submitted on 21 Mar 2011

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Ion extraction mechanism studied in a lyotropic lamellar phase Amélie Banc†, Pierre Bauduin†, Bernard Desbat§, Isabelle Ly‡, Olivier Diat†*

Institut de Chimie Séparative de Marcoule, UMR 5257 (CEA/CNRS/UM2/ENSCM), BP 17171, F-30206 Bagnols sur Cèze, France § Université Bordeaux-1, Laboratoire de Chimie et Biologie des Membranes et Nanoobjets, UMR 5248-CNRS, ENITAB, 2 rue Robert Escarpit, F-33607 Pessac, France ‡ Université Bordeaux-1 CNRS, Centre de Recherche Paul-Pascal, UPR 8641, 115 avenue A. Schweitzer, F-33600 Pessac, France

ABSTRACT In this paper we used a surfactant-stabilized lyotropic lamellar model system to study the interfacial behaviour of an ion-extracting agent: N1,N3 dimethyl- N1, N3-dibutyl-2-tetradecyl malonamide (DMDBTDMA). An analysis of small angle X-ray scattering (SAXS) and polarized Attenuated Total Reflectance - Fourier Transform Infra Red (ATR-FTIR) data enabled to describe the distribution of the malonamide extractant within the bilayers and its complexation state at the equilibrium. The lamellar phase was diluted with salt water containing varying amount of complexing salt, and each structural state measured was described using a thermodynamic model based on three elementary equilibria: (i) partition of the extractant polar heads between the core and the polar shell of the bilayers, (ii) the complexation of ions by extractants at the bilayers surfaces, and (iii) the partition of bonded extractants between the core and interfaces of bilayers. This model enabled to compare the energy cost of each step.

INTRODUCTION Reactions at liquid-liquid interfaces are of great fundamental interest because they play a key role in many important chemical and biological systems1, such as phasetransfer catalysis2, liquid-liquid extraction, micellar catalysis3, biomembrane activity, enzymatic reactions on fat lipases4. Systems in which such reactions take place are difficult to study because their global behaviour is the result of reactants diffusion in bulk and interfacial mechanisms. To study specifically liquid interfaces, different experimental systems were developed: systems with a unique and controlled interface between two immiscible liquids (interface of pendant drops5, interface developed for reflectance measurements6, interface created into a microchannel7, nanoscopic interface supported by nanopipette8, and systems with interfaces stabilized by amphiphilic molecules (vesicles9, micelles and microemulsions10). In the present work, we chose to use lyotropic lamellar phases, periodical stacking of oily layers (made by the surfactant tails) diluted by salt water. These systems with multiple interfaces, in thermodynamic equilibrium, are structurally and mechanically responsive to weak effects, such as the Hofmeister ion effect11, 12, 13 or the partitioning of host molecules within bilayers14, 15, 16, 17. In a previous paper18 we described and characterized the insertion of a lipophilic ion-complexing agent, N1,N3 dimethyl- N1, N3dibutyl-2-tetradecyl malonamide (DMDBTDMA), in a lamellar phase stabilized by a non-ionic surfactant: pentaethyleneglycol dodecyl ether (C12E5). DMDBTDMA is an extractant molecule used in nuclear industry to separate minor actinides from high level radioactive liquid wastes by liquid-liquid extraction processes19 (DIAMEX process) or in more conventional hydrometallurgy for the rare earths elements recycling20, 21. We showed that the structural parameters of an ideal lamellar system enabled to describe the partition of the extractant molecules between

the core and the interface of bilayers. This simple approach was based on the evolution of interfacial area with the addition of extractants, which was measurable thanks to the multitude of interfaces in lamellar phases, as for microemulsion systems22. In the present paper, we studied the effect of ion complexation on the system. First, we characterized the system using Small Angle X-Ray Scattering (SAXS) and transmission electron microcopy (TEM) coupled with freeze fracture to follow structural parameters, and polarized Attenuated Total Reflectance - Fourier Transform Infra Red (ATR-FTIR) spectroscopy to measure orientation of dipoles and to quantify the ion complexation phenomenon. In a second part, a thermodynamic approach of liquid-liquid extraction was developed on the basis of the experimental data by considering the ion extraction process as the sum of three equilibria (Scheme 1): (1) Positioning of free extractants polar head areas at interfaces (2) Complexation of cations with extractants at interfaces (3) Burying of bonded extractants within the bilayers Hence, the free energy contributions of each step was estimated and compared. This allowed to elucidate the precise mechanism of extraction by DMDBTDMA in this model system. EXPERIMENTAL SECTION C12E5 was obtained from Nikko Ltd. (high purity grade > 99%) and used as received. DMDBTDMA was obtained from Panchim and was purified on an alumina column to eliminate traces of surface active impurities. Aqueous solutions were prepared using deuterated water (99.9%, Eurisotop), anhydrous LiNO3 salt (purity>99%) from Strem, and Nd(NO3)3,6H2O salt (purity>99.9%) from Aldrich.







Scheme 1. Schematic view of cation extraction by extractant molecules. The global extraction reaction (KT) can be divided into three elementary mechanisms which are characterized by thermodynamic equilibrium constants: Adsorption of extractants from the organic phase to the interface (K1), complexation of cations with extractants which are at the interface (K2), and desorption of ionextractant complex from the interface to the organic phase (K3). To avoid introduction of H2O into samples, hydration water molecules of the neodymium salt were exchanged by D2O performing five D2O solubilisation (stirring at 50°C) – drying (under vacuum) cycles under inert atmosphere (Ar). The efficiency of the exchange was checked by FTIR looking at the disappearance of the OH stretching mode around 3400 cm-1. Samples were formulated adding the deuterated solutions to a premix of C12E5/DMDBTDMA (80/20 molar). Sample were homogenised with a succession of centrifugations and were left several days at 25°C to reach equilibrium. Dodecane (purity> 90%, Fluka) used for tensiometry experiments was previously purified on an alumina column. Microscopic observations were performed on a Zeiss Axio Imager A1m microscope equipped with crossed polarisers. To determine conditions of homeotropic orientation of lamellar samples (bilayers oriented perpendicular to the optical axis), different thicknesses of spacers between glass slide and cover slip were tested. It was observed that samples sheared between slide and cover slip spaced by 10µm thick mica sheets, display homeotropic orientation after few minutes (Figure 1). For freeze-fracture measurements, a drop of sample was placed on a gold holder and frozen by plunging into liquid propane. Fracture and replication were carried out in a freeze-fracture apparatus (Leica EM BAF060) at the temperature of −150 °C under vacuum. Pt/C was deposited at an angle of 45° and C at 90°. Replicas were recovered at room temperature by specimen dissolution in ethanol and water. Then, replicas were collected onto ATHENE copper 400 mesh grids to be examined on an Hitachi H-600 75kV’s transmission electron microscope equipped with an AMT’s CCD camera (DVC 1 megapixels).

Small angle X-ray scattering (SAXS) measurements using Mo-radiation (λ=0.71 Å) were performed on a bench built by XENOCS. The scattered beam was recorded using a large online scanner detector (diameter: 345 mm, from MAR Research) located at 750 mm from the sample stage. A large Q range (2.10-2-2.5Å-1) was covered thanks to an off centre detection. The collimation was applied using a 12:∞ multilayer Xenocs mirror (for Mo radiation) coupled to two sets of scatterless FORVIS slits23 providing a 0.8 x 0.8 mm x-ray beam at the sample position. Pre-analysis of data was performed using FIT2D software, taking into account the electronic background of the detector (the flatfield is homogeneous) and the empty cell subtraction. The scattered intensities were expressed versus the magnitude of scattering vector Q = [(4π)/λ].sin (θ/2), where λ is the wavelength of incident radiation and θ the scattering angle. Experimental resolution was ∆Q/Q=0.05. Silver behenate in a sealed capillary was used as the scattering vector calibration standard. The lamellar periodicity of samples (d) was deduced from the first Bragg peak position Q0:


2Π Q0


Surface tension experiments at the oil-water interface were performed on pendant drops using a drop shape analyser (Kruss DSA 100). The aqueous solution was either LiNO3 1M or Nd(NO3)3 0.33M, and organic solutions were DMDBTDMA solubilized into dodecane (from 0.005 to 1M). Before tensiometry measurements, each organic solution was emulsified with the aqueous solution of interest (50/50 v/v) by mechanical stirring for 2 minutes and then isolated after centrifugation. Densities of solutions were measured at 20°C using a vibrating tube densimeter, DSA 5000 from Anton Paar. After injection of the organic drop into the aqueous solution, surface tension was measured versus time. The values at equilibrium (<1% evolution) were selected for this study. The Gibbs equation was used to calculate the superficial excess of extractant (Γ):


1  ∂γ  (2)   RT  ∂ ln C  T , P

where γ is the surface tension and C is the concentration of the adsorbing specie (in the concentration region below cmc). Hence, the polar head area (A) was determined using:


1 NaΓ


where Na is the Avogadro constant. Polarized ATR-FTIR experiments were performed on a Perkin Elmer Spectrum-One Fourier Transform Infrared system equipped with an attenuated total mono reflection (ATR) device, using a diamond crystal. Lamellar phase samples, spread on the ATR crystal, were confined under a glass coverslip, 10µm-spaced from the crystal (Figure 1).


z Glass coverslip

y 10µm



n1 φ


where Ex², Ey² and Ez² are the time-averaged square electric field amplitudes of the evanescent wave in the sample at the crystal/sample interface. The electric field amplitudes on the crystal-sample interface were calculated using the thick film model of Harrick25, valid when the thickness of the sample is very large compared to the penetration depth:

E x2 =

2 4 cos 2 Φ.(sin 2 Φ − n 21 ) 2 2 2 (1 − n 21 ) 1 + n 21 sin ² Φ − n 21


E y2 = E z2 = Figure 1. Polarized FTIR-ATR. (a) Scheme of the experimental setup. The sample was confined between the ATR crystal and a glass cover slip to ensure an homeotropic orientation of the lamellar phase. Polarized light microscopy pictures of samples observed 1 minute (b) and 5 minutes (c) after shearing between flat surfaces separated by 10µm. This confinement ensured the orientation of the lamellar phase (bilayers parallel to the crystal) as demonstrated by polarized light microscopy and a sample thickness higher than the penetration depth of the evanescent electromagnetic waves, dp, defined by24:

dp =


2πn1 (sin 2 Φ − n 21 ) 2


where Φ is the incidence angle, λ is the wavelength, n1 is the refractive index of the ATR crystal, n2 the refractive index of the sample, and n21=n2/n1. In our case, dp varied between 0.2 and 1µm (n1=2.4, n2=1.45 and Φ=45°). For each sample, 64 scans spectra, with a resolution of 4 cm-1, were obtained for two polarizations of the incident beam: parallel (p) and perpendicular (s) to the incidence plane. A linear baseline correction was performed on each spectrum. Two kinds of analysis were performed with FTIR data: first, to obtain information about orientation of dipoles into samples, and second to quantify bonded extractants. Orientation information of dipoles is contained in the dichroic ratio RATR which is the ratio of the absorbance of the corresponding band measured with a parallel polarization of the incident light Ap to the absorbance measured with a perpendicular polarization As: RATR=

Ap As


The dichroic ratio is related to the dipole order parameter S, according to:


( E x2 − R ATR E y2 + E z2 ) ( E x2 − R ATR E y2 − 2 E z2 )



4 cos ²Φ 2 (1 − n 21 )






4 cos ² Φ sin ² Φ 2 2 (1 − n ) 1 + n 21 sin ² Φ − n 21 2 21



In our case, Ex=1.157, Ey=1.774 and Ez=2.227. The order parameter S contains the information on the mean orientation of the transition dipole and on the distribution of the dipole orientations at the mean value:


3 cos ²θ − 1 2


where θ is the angle between the dipole and the normal of the ATR plate, and

cos ²θ is the mean value of cos²θ for

a distribution of transition moment in the film. To perform quantitative analysis of bonded extractants, the isotropic spectrum of samples was calculated from the polarized spectra. The isotropic sprectrum can be defined as: Siso=Sx+Sy+Sz (11) where Si is the spectrum in the direction i. The sample being uniaxial, Sx=Sy. Polarized spectra can be defined as follow: Ss=Ex.Sx and Sp=Ey.Sy+Ez.Sz (12) Hence using our parameters, Siso=0.8345.Ss+0.449.Sp


The second analysis was related to the ion-bonding of extractants and was studied using the carbonyl band positioned around 1650 cm-1. This band was fitted using two Gaussian curves centred at 1640 and 1615 cm-1 assigned to free carbonyls (not complexed) and bonded carbonyls (complexed with Nd3+) respectively. Fitting procedure was realised with the Omnic software which uses a Fletcher-Powell-McCormick algorithm as convergence routine. To obtain the area of each contribution, positions and the full widths at half height were fixed and maintained constant for all samples. The evolution of the free carbonyl absorbance (area) in isotropic spectra enabled to quantify the complexation into samples. According to the BeerLambert law: A=ε.l.C (14)


where A is the absorbance, l the length of the pathway, C the concentration, and ε the extinction coefficient. Considering that for a vibration mode ε is constant, and that l is constant in our samples (the sample thickness is higher than the penetration depth which is constant for given wavelength and setup), the following relationship can be established: (A0-A)/A0=(C0-C)/C0


Hence, the fraction of free carbonyl consumed (bonded carbonyls created) was determined.


surfactant/water system26, enabled to roughly determine the lamellar region in the ternary system (Fig. 2). In the present work, we choose to work with a formulation located at the center of this domain: a premix C12E5/DMDBTDMA 80/20 molar hydrated with 40 wt% of aqueous solution. This composition lies in the concentrated region of the lamellar phase (aqueous layers are about 2 nm thick) mainly stabilized by the interplay of the attractive van der Waals force and the steric repulsion of the hydrated head groups (repulsive undulation forces are negligible)27. The effect of complexation was studied varying the nature of the aqueous solution from LiNO3 1M to Nd(N03)3 0.33M to keep 1 molar of ionic strength (range of ionic strength applied in extraction systems) in the different samples. The lamellar phases obtained with such formulations were characterized by freeze-fracture (see Figure 3). A characteristic lamellar structure with periodical steps, was obtained for both LiNO3 or Nd(NO3)3 hydrated extractantcontaining lamellar phases, like for the binary system. No structural change e.g. topological defects28 was detected after addition of extractant or salt to the lamellar phase.


20 C12E5 C12E5/DMDBTDMA 19 % l ar mo

δ (Å) HC


35 20

Aqueous solution






% w/w (c)

Figure 2. Molecular structures of C12E5 (a) and DMDBTDMA (b). (c) Ternary phase diagram of C12E5/DMDBTDMA/aqueous solution indicating in grey the monophasic lamellar phase domain at room temperature. The salty aqueous solution contains LiNO3 and Nd(NO3)3 salts in a 1 molar ionic strength. Crosses indicate experimental points and the star indicates the composition of samples studied in this paper (C12E5/DMDBTDMA = 80/20 molar, 40% hydration). RESULTS Preliminary experiments were carried out to identify the boundaries of the monophasic lamellar phase in the ternary phase diagram C12E5/DMDBTDMA/aqueous solution (1M). Macroscopic and microscopic (optical microscopy with crossed polarizers) observations performed on some formulations, derived from the lamellar domain of the (a)


15 0






% Nd(NO )

3 3 Figure 4. Evolution of bilayer thickness (δHC) with the percentage of Nd(NO3)3 in the aqueous solution. Empty dots display the reference system (C12E5) whereas full dots display the mixed system C12E5/DMDBTDMA (80/20 molar).

The periodicity (d) of samples was characterized using SAXS. Assuming an ideal lamellar structure (flat rigid bilayers without defect), the hydrocarbon thickness of bilayers (δHC) was determined using the following dilution law: δHC=d.ΦHC




Figure 3. TEM pictures of freeze-fracture replica: (a) C12E5 lamellar phase hydrated with 30% LiNO3 1M, (b) C12E5/DMDBTDMA lamellar phase hydrated with 40% LiNO3 (1M), and (c) C12E5/DMDBTDMA lamellar phase hydrated with 40% Nd(NO3)3(0.33M).

δ HC =

(1 − x )VC12 + xVDMDBTDMA V =2 A (1 − x) AC12 E5 + xκADMDBTDMA


where Vi are molecular volumes determined from density values, Ai are polar head areas, x is the molar fraction of DMDBTMA among the organic components and κ the fraction of DMDBTDMA participating to the interfacial area. The two extreme cases, κ=0 and κ=1, are plotted on Figure 5. The experimental points are intermediate (0<κ<1) and indicate roughly an equi-partitioning of extractant molecules between the core of the bilayers and the interfaces. The decrease of δHC with the percentage of complexing ions can be thus interpreted as an evolution of extractants partition within the bilayers; qualitatively, the interaction between extractant molecules and ions that can be complexed, pull the lipophilic extractants towards the interface.

28 26


δ (Å) HC

24 22 20

18Evolution of bilayer thickness Figure 5. +Nd3+as function of the κ=1 percentage of DMDBTDMA in the C12E5/DMDBTDMA 18 16 . The dotted line display the case for which mixture extractants are fully embedded within bilayers, whereas the 14 represents the case for which extractants fully dashed line 0 5 10 15 20 25 30 35 participate to the interfacial area as the surfactant % DMDBTDMA molecules. Full dots are experimental points obtained for a system hydrated with LiNO3 1M.The arrow indicates the decrease of the bilayers thickness by exchanging lithium with neodymium cations at constant DMDBTDMA percentage.

To estimate the contribution of extractants to the interfacial area we measured their polar head area at model oil-water interfaces: dodecane-LiNO3 1M and dodecane-Nd(NO3)3 0.33M, to consider free and bonded extractants respectively. Dodecane was chosen to mimic the C12 chain of surfactants which constitute the oily domains. Figure 6 displays the evolution of interfacial tensions versus DMDBTDMA concentration for both cases. Critical aggregation concentrations around 0.1M can be noticed, in accordance with literature30, 31. Using the Gibbs equation (2), we obtained Afree DMDBTDMA=80±5 Ų/molecule and Abonded DMDBTDMA=100±5 Ų/molecule. These values do not represent only a geometrical polar area but takes into account all the lateral interactions within the bilayers. Ion complexation charges the extractant molecules and induces additional electrostatic lateral repulsions that consequently increase the apparent extractant polar head area. 24


22 20

γ (mN/m)

where φHC is the volume fraction of DMDBTDMA and alkyl chains of C12E5 in the sample. The following density values were used: dC12=0.80329, dC12E5=0.963 (from supplier) and dDMDBTDMA=0.906 (measured). The interface was defined as being located between the aliphatic chains of surfactants, and the polar layer comprising the polar heads of surfactants and hydration water. Figure 4 displays the evolution of bilayer thickness versus the neodymium percentage in the hydration solution for C12E5 and C12E5/DMDBTDMA lamellar phases. By increasing the neodymium content, the bilayer thickness remains constant for the reference C12E5 samples, whereas it continuously decreases from 18.8 to 18.3 Å for samples containing DMDBTDMA. In our previous paper18, we demonstrated in analysing the bilayer thicknesses that the extractant molecule can be considered either as a co-surfactant or as a solvent molecule distributed within the aliphatic domain of the bilayers, both configurations being in equilibrium for a given surfactant/extractant ratio. Assuming an ideal mixing of extractant and surfactant, and constant polar areas per molecule at the hydrophilic/hydrophobic interface, the bilayer thickness was expressed as a function of the extractant repartition between the interfaces and the bulk of the bilayers:

18 16 14



molecules. This latter was decomposed into two main contributions centred at 1640 and 1615 cm-1 attributed to free (ν C= O f ) and bonded (ν C= Ob ) carbonyls respectively. Hence, we calculated the dichroic ratios corresponding to ν CH 2 , ν C= O f , ν C=Ob . These ratios are constant whatever S

the neodymium percentage and the mean values are displayed in Table 1. As the transition moment for the ν CH 2 lies perpendicular to the chain axis, the order S

Figure 7 displays the characteristic polarized ATR-FTIR spectra of oriented samples. We can notice the large OD elongation vibrations around 2500 cm-1, the OH elongation vibrations attributed to the polar head of surfactants around 3400 cm-1, the CH2 vibration modes around 2900 cm-1, and the amide I (mostly 0.2 the C=O elongation vibration) around 1650 cm-1. The symmetric p-polarisation vibration mode of CH2, ν s-polarization -1 positioned at 2855 cm (ν CH 2 ), was used to probe the OD S

Intensity (a.u.)

0.15 into bilayers, whereas the C=O order of aliphatic chains elongation vibration mode was used to probe extractant νCH2


2( E x2 − R ATR E y2 + E z2 ) (3 cos 2 α − 1)( E x2 − R ATR E y2 − 2 E z2 )


where α is the angle between the chain axis director and the transition dipole (α=90°). The low value obtained (S=0.1) is explained by the fluid character of the Lα bilayers and is consistent with literature results for C12E5/water lamellar phases32. The zero value of S for free carbonyls indicates an isotropic orientation of these bonds, as previously measured 0.05 polarized microRaman spectroscopy18, contrary to using bonded carbonyls which are characterized by a value close 0.04 to that of aliphatic chains. Intensity (u.a.)

Figure 6. Determination of polar-head areas of free and bonded DMDBTDMA by tensiometry measurements. Free DMDBTDMA is studied using an aqueous solution of LiNO3 1M, whereas bonded DMDBTDMA is studied using a 0,3M Nd(NO3)3 solution.

parameter of aliphatic chains was calculated using the following relationship:


0.03 0.02 0.01 0 1700 1680 1660 1640 1620 1600 1580 1560 -1

Wavenumber (cm )



ν OH 0 4000










Wavenumber (cm )

Figure 7. Polarized ATR-FTIR spectra of samples. The carbonyl band was decomposed into two mains contributions centred at 1640 cm-1 and 1615cm-1(in inset, yellow line : experimental spectrum, full dark line: fitted spectrum). 0.35 0% 10% 4

Intensity (u.a.)

4 10

20% 30% 40%


3 10

50% 60%


2 10

70% 80%


1 10



Fraction of bonded C=O


5 10

0.3 0.25 0.2 0.15 0.1

0 1700 1680 1660 1640 1620 1600 1580 1560 1540



0.05 0






Figure 8. (a) Evolution of the carbonyl stretching vibration band with the percentage of neodynium salts in the aqueous solution. (b) Fraction of bonded carbonyl versus the percentage of neodymium obtained using equation (15). This result must be analysed considering that the experimental order parameter measured from RATR, denoted Sexp, is the product of several order parameters related to a set of nested, uniaxial symmetric distributions33, 34: Sexp=SmembraneSmoleculeSdipole (19) where Smemb describes the distribution function of the membrane patches with respect to the internal reflection element, Smolecule describes the orientation of the molecules within the membrane plane, and Sdipole describes the orientation of the dipole relative to the molecule. Providing the good orientation of the samples sheared between 10µmspaced flat surfaces, we can estimate that Smembrane is close to 1. Hence, the variation of order parameter with complexation can be attributed either to a change of molecule orientation inside bilayers (Smolecule) or a modification of carbonyls orientation with respect to the molecule axis (Sdipole). It is difficult to consider a higher order parameter for the extractant compared to the one of the surfactant (S=0.1). Then, with SC=Obonded=0.096 and considering Smemb~1 and Smolecule≤0.1, Sdipole-bonded must be close to 1. This means that in a bonded malonamide, both carbonyls bonds are parallel. For free malonamides, SC=Ofree=0 is explained either by the trans/gauche conformation of carbonyl groups or by the isotropic orientation of molecules within bilayers. Using a Quantum Mechanics method, Boehme et al35 showed that free malonamides, in gas phase, roughly adopt trans conformation (β, the angle between the two carbonyl dipoles is 162°). The monodentate mode, rarely observed, displays a gauche conformation (β is about 90°). Both conformations would lead to Sdipole<<1. However, the bidentate coordination state with lanthanides induces an alignement of carbonyl bonds i.e. cis conformation with 30°<β<60° according to the stoichiometry considered. This latter conformation can be considered as the most probable for ion bonded extractants in our system. The evolution of the carbonyl band shape with the addition of neodymium (figure 8a) enabled to follow quantitatively the complexation. Figure 8b display the fraction of bonded carbonyls versus the percentage of neodymium salts in the aqueous solution. In 0.33M neodymium samples (100% of Li+ replaced by Nd3+ at constant ionic strength), the DMDBTDMA/Nd ratio is 2 and only 30% of carbonyls are bonded. Assuming a unique denticity of the coordination complex, the fraction of bonded carbonyls is equal to the fraction of bonded extractants. Free C=O

Bonded C=O

Dichroïc ratio R




Order parameter S




Table 1. Dichroïc ratios of ν C= O f


ν C=O


and ν CH 2



Order parameters of carbonyls and aliphatics chains. (*Dichroïc ratio of CH2 and order parameter of aliphatic chains).

ANALYSIS - DISCUSSION The liquid-liquid extraction of neodymium by DMDBTDMA is usually described by the following solvation reaction36 , characterized by an equilibrium constant KT: x DMDBTDMA + Nd3+ + 3NO3-  (DMDBTDMA)xNd(NO3)3 where upperlined species are solubilized in the organic phase and the other species are solubilized in water. Usually, above a given bulk concentration in organic phases the extractant forms reverse aggregates which are characterized by their aggregation number (from 4 to 1037, 38 , depending on concentration, pH, type of salt…). Within these aggregates the coordination number, x, is still under debate since the molecular conformation is not well defined, but it is established that 2


DMDBTDMA + Nd3+ + 3NO3-  (DMDBTDMA)Nd(NO3)3 K2 (3)

Burying of bonded extractants from interfaces

(DMDBTDMA)Nd(NO3)3 (DMDBTDMA)Nd(NO3)3 K3 where dash upperlined species are localized at interfaces, and Ki are the equilibrium constants. In the next part, we will show that these equilibria can be quantified knowing the concentration of each species. 19

κ1=0.43 κ3=0.55



K 2a 1 + K 2a


where θ is the relative occupancy of sites, each site being the polar head of an extractant at interface, K2 the adsorption constant, and a the complexing ion activity . The relative occupancy of sites, defined as the ratio between bonded extractants over the total number of extractants at the interface, was determined using FTIR data (y), κ1 and κ3:

δ (Å) HC





values must be used with care. However, this result indicates that bonded extractants, i.e. extractant forming a complex with Nd3+ are more amphiphilic than free extractants. The complexation must charge the lipophilic extractant and should make the head more polar. This is in accordance with tensiometry measurements which display a more important and faster reduction of surface tension with DMDBTDMA in presence of neodymium salt (bonded molecules) than with DMDBTDMA in presence of lithium salt (free extractant). The complexation of neodymium by DMDBTDMA at the interfaces (K2) can be described in a first approach as a cation adsorption isotherm at fluid interfaces. Thus, the Langmuir model can be used:



0.05 0.1 0.15 0.2 0.25 0.3 0.35

y Figure 9. Evolution of bilayer thicknesses (δHC) versus the fraction of bonded extractant (y). Experimental points are adjusted with equation (20), using κ1= 0.43±0.01 et κ3= 0.55±0.02. The variation in the bilayer thickness can be analysed considering the partition of extractants between the core (as the solvent molecule) and the interface of bilayers (as a cosurfactant), using the following relationship derived from (17):

κ 3 .y κ 3 . y + κ 1 .(1 − y )


This definition assumes that the saturation limit of the ion isotherm is Nd/DMDBTDMA=1/1 molar ratio. As previously, the molar concentration was used instead of the ion activity. As a consequence, an apparent adsorption constant K’2 was obtained:

K 2' = K 2 γ


where γ is the molar activity coefficient of the complexing ion (a=γC where C is the concentration). 1

0.8VC12 + 0.2VDMDBTDMA V =2 A 0.8 AC12 E5 + 0.2(κ1 (1 − y ).AfreeDMDBTDMA + κ 3 yAbondedDMDBTDMA )

(20) where y is the molar fraction of bonded extractants deduced from the FTIR analysis, κ1 the fraction of free extractants at interfaces and κ3 the fraction of bonded extractants at interfaces. Each polar head areas was considered constant and was evaluated by tensiometry for DMDBTDMA, and taken from the literature for C12E5 (AC12E5=50Ų/molecule)39, 18. Figure 9 displays the evolution of the bilayers thickness as a function of the fraction of bonded extractant, which was adjusted by the equation (20) giving κ1=0.43±0.01 and κ3=0.55±0.02 (correlation coefficient: 0.8). Approximating extractant activities to molar concentrations, the apparent related equilibrium constants can be calculated using:

κ K '1 = 1 1 − κ1


K '3 =

1 − κ3






δ HC =



0 0








[Nd(NO ) ] (mol/L) 3 3

Figure 10. Fraction of bonded extractant at the interface (θ) as a function of the Nd(NO3)3 concentration in the aqueous solution. Experimental points are well fitted with a Langmuir isotherm using K2=1.85±0.3.

Hence, K’1=0.75±0.3 and K’3=0.82±0.7. Because of the dispersion of experimental points in the studied scale,


Figure 10 shows that experimental data were well fitted with such model, using K’2=1.85±0.3. The apparent equilibrium constants characterizing the system (K’1, K’2, K’3) deduced from experimental data enabled to calculate the apparent free energy (∆G’) corresponding to the different equilibrium:

∆G ' = − RT ln K '


The different contributions to the free energy are displayed in a table and illustrated by an energy diagram in figure 11. We can observe that the free energy of the total reaction, the ion extraction from the aqueous domains towards the bilayers, is negative due to the complexation (2) reaction which is spontaneous. However this effect is reduced by the energy cost to position the free extractant at the interfaces (1) and to desorb bonded extractant from the interfaces (3), both being energetically unfavourable. The thermodynamics of liquid-liquid extraction by diamides was previously studied in classic biphasic systems. Taking into account the formation of supramolecular aggregates of extractant in the organic phase37, 38 and their associated potential surfaces on which ions can adsorb, Zemb et al40, 41 have obtained thermodynamical extraction parameters assuming a sum of Langmuir isotherms for each state of aggregation. This view enabled to associate thermodynamical values to the different aggregation states of extractants. Free energies of about few kBT were deduced, slightly higher than our values. However, in our model system, we cast off aggregation and only elementary mechanisms at interfaces were studied. Consequently, adsorption constants are not fully comparable. Nevertheless the complexation free energy value can be evaluated looking at the complexation values obtained for hydrosoluble malonamides. Couston et al.42 found an 1:1 complexation equilibrium constant K1:1 equals to 1.3 for tetraethylmalonamides with europium in water. The corresponding free energy value (∆G=-0.26kBT) has the same order of magnitude as ∆G’2.


Free extractant in bulk

Free extractant at interface

∆G1 ∆G2

Bonded extractant in bulk


∆G3 Bonded extractant at interface Extraction pathway












∆G (kBT)






Figure 11. (a) Energy diagram of the extraction reaction divided into three elementary equilibrium: positioning of free extractants at interfaces (1), complexation of extractants with cations at interfaces (2), burying of bounded extractants from interfaces (3). (b) Free energy values associated to each elementary steps and to the global extraction reaction. In our system we use surfactant molecules which must be in competition with extractants for the interfacial adsorption, like in model micellar systems10. Thus, the results are certainly function of the surfactant used, but for a given system, parameters such as the nature of extracted cations, the nature of extractant, the ionic force can be studied comparatively. Nevertheless, from these data, the amphiphilic property of extractant could be considered as the limiting step for extraction. In future studies, the extractant aggregation could be eventually introduced in the lyotropic system releasing sterical constraints by swelling bilayers with oil. In addition, the dehydration of cations, which is included in the mechanism of cation adsorption at interfaces, could be modulated decreasing the water thickness between bilayers. Indeed, the water activity in such confined environments was shown to be modulated by the distance between bilayers43. The cation desolvation energy could be thus studied using the lamellar phase. CONCLUSIONS In this work we proposed a lyotropic lamellar phase to study interfacial mechanisms of liquid-liquid extraction. Due to the numerous oriented interfaces and using a combination of SAXS and FTIR techniques, we determined the concentrations of free and bonded extractants molecules either buried into bilayers or localized and oriented at interfaces,. Hence, apparent free energies associated to each interfacial elementary mechanism were estimated. Results indicate that extraction by DMDBTDMA is partially controlled by the variation of amphiphilicity of the molecule during the ion complexation. Between bonded and free extractant, the former appears to be a more favourable state at interfaces. This observation could be considered as a bridle to an ideal extraction kinetics. With this study we demonstrated that a lyotropic lamellar phase can be a meaningful system to model reactive liquidliquid interfaces. In the future, other studies could be engaged varying parameters such as the nature of reactants or layers thicknesses. ACKNOWLEDGMENTS We thank Thomas Zemb, Laurence Berthon and MarieChristine Charbonnel for stimulating discussions, David Maurin for his assistance during polarized FTIR-ATR experiments, Olivier Mondain Monval for his expertise on cryofracture observations, Jérôme Maynadié for his help during deuteration of neodymium salts and Gaëlle GassinMartin for fruitful discussions concerning tensiometry measurements. We also thank Daniel Meyer for fruitful discussion on ion coordination with extractant molecules.


This work was supported by the European ACSEPT network. REFERENCES (1) Liquid interfaces in chemical, biological, and pharmaceutical applications; Volkov, A. G., Ed.; Marcel Dekker, New York, 2001. (2) Fiamegos, Y.; Stalikas, C. Anal. Chim. Acta 2005, 550, 1–12. (3) Zhang, J.; Meng, X.-G.; Zeng, X.-C.; Yu, X.-Q. Coord. Chem. Rev. 2009, 253, 2166–2177. (4) Reis, P.; Holmberg, K.; Watzke, H.; Leser, M.; Miller, R. Adv. Colloid Interface Sci. 2009, 147-148, 237 – 250. (5) Jeribi, M.; Almir-Assad, B.; Langevin, D.; Henaut, I.; Argillier, J. J. Colloid Interface Sci. 2002, 256, 268–272. (6) Zarbakhsh, A.; Webster, J. R. P.; Wojciechowski, K. Langmuir 2009, 25, 11569–11575. (7) Watanabe, M. Anal. Chem. 2009, 81, 8213–8218. (8) Li, Q.; Xie, S.; Liang, Z.; Meng, X.; Liu, S.; Girault, H. H.; Shao, Y. Angew. Chem. Int. Ed. 2009, 48, 8010–8013. (9) Mircheva, K.; Minkov, I.; Ivanova, T.; Panaiotov, I.; Proust, J.; Verger, R. Colloids Surf., B 2008, 67, 107 – 114. (10) Tondre, C.; Hebrant, M.; Watarai, H. J. Colloid Interface Sci. 2001, 243, 1 – 10. (11) Petrache, H. I.; Zemb, T.; Belloni, L.; Parsegian, V. A. Proc Natl Acad Sci U S A 2006, 103, 7982–7987. (12) Aroti, A.; Leontidis, E.; Dubois, M.; Zemb, T. Biophys. J. 2007, 93, 1580–1590. (13) Leontidis, E.; Aroti, A.; Belloni, L.; Dubois, M.; Zemb, T. Biophys. J. 2007, 93, 1591–1607. (14) Ladbrook.BD,; Williams, R.; Chapman, D. Biochim. Biophys. Acta 1968, 150, 333–&. (15) Tamai, N.; Uemura, M.; Goto, M.; Matsuki, H.; Kaneshina, S. Colloids Surf., B 2008, 65, 213–219. (16) Mills, T. T.; Tristram-Nagle, S.; Heberle, F. A.; Morales, N. F.; Zhao, J.; Wu, J.; Toombes, G. E. S.; Nagle, J. F.; Feigenson, G. W. Biophys. J. 2008, 95, 682–690. (17) Dhez, O.; Konig, S.; Roux, D.; Nallet, F.; Diat, O. Eur. Phys. J. E 2000, 3, 377–388. (18) Banc, A.; Bauduin, P.; Diat, O. Chem. Phys. Lett. 2010. (19) Advanced separation techniques for nuclear fuel reprocessing and radioactive waste treatment; Nash, K. L., Lumetta, G. J., Eds.; 2010. (20) Estler, F.; Eickerling, G.; Herdtweck, E.; Anwander, R. Organometallics 2003, 22(6), 1212–1222.

(21) Tian, Q.; Hugues, M. Hydrometallurgy 1994, 36(3), 315– 330. (22) Tchakalova, V.; Testard, F.; Wong, K.; Parker, A.; Benczédi, D.; Zemb, T. Colloids Surf., A 2008, 331, 40 – 47. (23) Li, Y.; Beck, R.; Huang, T.; Choi, M. C.; Divinagracia, M. J. appl. crystallogr. 2008, 41, 1134–1139. (24) Harrick, N. J.; du Pré, F. K. Appl. Opt. 1966, 5, 1739–1743. (25) Harrick, N. J. Opt. Soc. Am. 1965, 55, 851–857. (26) Mitchell, D. J.; Tiddy, G. J. T.; Waring, L.; Bostock, T.; McDonald, M. P. J. Chem. Soc., Faraday Trans. 1 1983, 79, 975 – 1000. (27) Varade, D.; Kunieda, H.; Strey, R.; Stubenrauch, C. J. Colloid Interface Sci. 2006, 300, 338–347. (28) Moreau, P.; Navailles, L.; Giermanska-Kahn, J.; MondainMonval, O.; Nallet, F.; Roux, D. Europhys. Lett. 2006, 73, 49–54. (29) Tanford, C. J. Phys. Chem. 1972, 76, 3020–3024. (30) Dozol, H.; Berthon, C. C. R. Chim. 2006, 9, 556–563. (31) Martinet, L. Ph.D. thesis, Université Paris XI, 2005. (32) Tonegawa, A.; Ohno, K.; Matsuura, H.; Yamada, K.; Okuda, T. J. Phys. Chem. B 2002, 106, 13211–13223. (33) Rothschild, K.; Clark, N. Biophys. J. 1979, 25, 473–487. (34) Ivanov, D.; Dubreuil, N.; Raussens, V.; Ruysschaert, J.-M.; Goormaghtigh, E. Biophys. J. 2004, 87, 1307–1315. (35) Boehme, C.; Coupez, B.; Wipff, G. J. Phys. Chem. A 2002, 106, 6487–6498. (36) Weigl, M.; Geist, A.; Gompper, K.; Kim, J. Solvent Extr. Ion Exch. 2001, 19, 215–229. (37) Erlinger, C.; Belloni, L.; Zemb, T.; Madic, C. Langmuir 1999, 15, 2290–2300. (38) Testard, F.; Bauduin, P.; Martinet, L.; Abecassis, B.; Berthon, L.; Madic, C.; Zemb, T. Radiochim. Acta 2008, 96, 265– 272. (39) Rosen, M.; Cohen, A.; Dahanayake, M.; Hua, X. J. Phys. Chem. 1982, 86, 541–545. (40) Testard, F.; Berthon, L.; Zemb, T. C. R. Chim. 2007, 10, 1034 – 1041. (41) Meridiano, Y. Ph.D. thesis, Université Paris XI – Orsay, 2009. (42) Couston, L.; Charbonnel, M.-C.; Flandin, J.-L.; Moulin, C.; Rancier, F. Radiochim. Acta 2004, 92, 1–8. (43) Lyle, I.; Tiddy, G. Chem. Phys. Lett. 1986, 124, 432–436.