Coagulation of Na-Montmorillonite by Inorganic ... - ACS Publications

Article pubs.acs.org/Langmuir

Coagulation of Na-Montmorillonite by Inorganic Cations at Neutral pH. A Combined Transmission X‑ray Microscopy, Small Angle and Wide Angle X‑ray Scattering Study Laurent J. Michot,*,† Isabelle Bihannic,† Fabien Thomas,† Bruno S. Lartiges,†,∥ Yves Waldvogel,† Céline Caillet,† Juergen Thieme,‡,¶ Sérgio S. Funari,§ and Pierre Levitz⊥ †

Laboratoire Interdisciplinaire des Environnements Continentaux, UMR 7360 CNRS - Université de Lorraine, 15, Avenue du Charmois, BP40 54501 Vandœuvre Cedex, France ‡ Institut für Röntgenphysik, Universität Göttingen, Geiststrasse 11, 37073 Göttingen, Germany § HASYLAB, NotkeStrasse 85, D-22603 Hamburg, Germany ⊥ Physico-chimie des Electrolytes, Colloïdes et Sciences Analytiques UMR 7195 - UPMC - CNRS - ESPCI, UPMC bâtiment F74 place Jussieu, 75252 Paris Cedex 5, France ABSTRACT: The coagulation of sodium montmorillonite by inorganic salts (NaNO3, Ca(NO3)2 and La(NO3)3) was studied by combining classical turbidity measurements with wide-angle-X-ray scattering (WAXS), small-angle-X-ray scattering (SAXS), and transmission X-ray microscopy (TXM). Using size-selected samples, such a combination, associated with an original quantitative treatment of TXM images, provides a true multiscale investigation of the formed structures in a spatial range extending from a few ångstroms to a few micrometers. We then show that, at neutral pH and starting with fully Na-exchanged samples, coagulation proceeds via the formation of stacks of particles with a slight mismatch between layers. These stacks arrange themselves into larger porous anisotropic particles, the porosity of which depends on the valence of the cation used for coagulation experiments. Face−face coagulation is clearly dominant under those conditions, and no evidence for significant face−edge coagulation was found. These structures appear to arrange as larger clusters, the organization of which should control the mechanical properties of the flocs.

■

INTRODUCTION The coagulation of clay minerals plays a major role in numerous environmental and industrial processes. For instance, millions of tons of bentonite muds used in civil engineering have to be coagulated each year for the waste to be safely released in the environment (e.g. ref 1). In estuaries, clay materials transported in the river can coagulate and sediment upon their interaction with salted water, leading to composition-dependent spatial deposition patterns in estuaries.2,3 These processes also control to some extent the fate of particles discharged into coastal waters,4 and recent studies have, for instance, investigated the influence of clay coagulation on the dispersion of manufactured nanoparticles.5 In soils, the formation of clay-based organomineral aggregates plays a major role for optimal plant growth.6,7 On a more fundamental point of view, the role of coagulation processes in the phase diagram of sodium-exchanged swelling clay minerals remains a subject of debate. Indeed, some of the features of the sol−gel transition at moderate ionic strength (typically above 10−3 M) can be interpreted as resulting from microflocculation or retarded aggregation processes.8−11 For the various reasons abovedescribed, studies on the coagulation behavior of swelling clay minerals were initiated many decades ago.12−17 The general © 2013 American Chemical Society

knowledge about the mechanisms of clay coagulation by inorganic salts was summarized by Lagaly and co-workers.18,19 Three types of arrangements of clay platelets in the coagulated state are basically considered: (i) the edge-face (EF) configuration leading to the so-called house of cards, (ii) the edge−edge (EE) configuration leading to the formation of banded-like structures, and (iii) the face−face (FF) configuration. At relatively low pH, i.e. when the edges are supposedly having a positive charge, the EF configuration would be favored, whereas higher concentration and higher pH would favor the EE configuration. Finally, high charge density on the edges would tend to favor the formation of FF-like structures. As mentioned in a recent paper, however,20 surprisingly little direct experimental evidence for the existence of such configurations is provided in papers dealing with clay coagulation. Furthermore, computer simulations carried out in the case of laponite particles tend to show that the EF configuration, supposed to be dominant in coagulated Received: January 18, 2013 Revised: February 15, 2013 Published: February 19, 2013 3500

dx.doi.org/10.1021/la400245n | Langmuir 2013, 29, 3500−3510

Langmuir

Article

flocs sedimented at the bottom of the test tube was checked out by visual observation. SAXS and WAXS experiments were carried out on beamline A2 at Hasylab, Hamburg, Germany, using a fixed wavelength of 0.15 nm and sample-to-detector distances of 3 m and 0.3 m for SAXS and WAXS experiments, respectively. Bidimensional scattering patterns were collected on a CCD camera, and the curves intensity vs q (q = (4π sin θ)/λ, where 2θ is the scattering angle and λ the wavelength) were obtained by radial integration of the data. The samples were conditioned in cylindrical glass capillaries 1 mm in diameter. X-ray microscopy experiments were carried out on the TXM microscope set up on undulator U41 at the Bessy II storage ring in Berlin (Germany). Samples of flocculated clays (approximately 10 μm thick) were placed between two polymer membranes and observed by TXM. The total field of view was around 12 μm with a resolution of 20 nm. Typical exposure times were around 5 s. After background subtraction, the images were thresholded manually in order to derive average values of length and width for the observed structures. These two parameters were obtained on the thresholded images by assuming an elliptical shape for the particles. The curvature of the particles was then estimated by dividing the longest axis of the ellipse by the Feret diameter of the particles. Values of this ratio higher than 1.1 were considered as indicating significantly curved particles. Though only estimative, such a method provides a quick index. As shown in a recent study,25 it is also possible to derive SAXS curves from X-ray transmission images. Using the Fourier slice theorem,26 the power spectrum of the logarithmic transform of a two-dimensional (2D) TXM projection image and the SAS pattern are related as:

suspensions does not represent the global free energy minimum.20,21 The aim of the present paper is then to try to provide novel structural information on coagulated clay systems at various observation scales by carrying out coagulation experiments on size-selected montmorillonite suspensions using mono, di and trivalent salts in a range of initial solid concentrations located in the isotropic region of the phase diagrams of these particles.9 Wide angle and small angle X-ray scattering (WAXS and SAXS) were used to provide information at the scale of the arrangements of a few clay layers (from a few Å to a few hundredths of Å). In parallel transmission X-ray microscopy (TXM) was used to visualize clay aggregates at the scale of a few micrometers with a lateral resolution of around 20 nm. This latter technique22,23 is particularly well adapted to the present case, as it allows examining the aggregates directly in water without any particular sample preparation.

■

MATERIALS AND METHODS Wyoming montmorillonite SWy2 was purchased from the Source Clays Minerals repository at Purdue University, Indiana. Its structural formula can be written as (Si7.94 Al0.06 )(Al2.88Fe0.5Mg0.62)O20(OH)4 Na0.68. A 50 g/L clay suspension was first exchanged three times in 1 M NaCl. This suspension was washed by dialysis against Milliqwater using Visking membranes with a cutoff value of 14000 Da until a conductivity of less than 5 μS was obtained. The suspensions were then placed in Imhoff cones for 24 h in order to discard the major mineralogical impurities (iron oxide and quartz, mainly). Size fractionation procedures were then applied by centrifuging the stock suspension under different gravitational fields (7000g, 17000g, and 35000g). Using such a procedure, three size fractions referred to as sizes 1−3 were obtained.9 Mineralogical purity was checked by X-ray diffraction and infrared spectrometry, whereas sizes were determined by transmission electron microscopy. Table 1 summarizes the properties of the

Pθ (kx , k y)αI(kx cos θ , k y , kz sin θ )

where the (x,y) plane is parallel to the detector plane; the z axis is the optical axis; θ is the sample tilt angle about the y axis perpendicular to the z axis; Pθ is the power spectrum of the logarithmic transform of the projection at angle θ; (kx and ky) are the components of the 2D wave vector; and I(qx,qy,qz) is the small-angle scattering spectrum of the object under consideration. For a statistically isotropic medium, the small-angle scattering spectrum depends on the modulus q = |q| of the scattering vector only and is independent of θ. Computation of eq 1 is performed using a fast Fourier transform. The extension of the SAS pattern is then defined by application of Shannon’s theorem.26 qmax = (π/pixel resolution) and qmin = [2π/ minimum(Lx,Ly)] where Lx and Ly are the x and the y global size of the TXM projection image.

Table 1. Morphological and Charge Parameters of Wyoming Montmorillonite Particles average diameter (nm) polydispersity (%) average thickness (nm) CEC (mequiv/100 g)

size 1

size 2

size 3

410 130 1 90.5

240 93 0.8 93

100 45 0.75 91.5

(1)

particles used in the present study. The average thickness values corresponding to isolated objects in aqueous suspensions were determined from a thorough analysis of the swelling laws presented elsewhere.24 The cationic exchange capacity (CEC) measured by exchange with cobaltihexamine are 90.5, 93, and 91.5 meq/100 g for sizes 1, 2, and 3, respectively, which proves that, in Wyoming montmorillonites, the average layer charge does not depend on size.9 Coagulation tests were carried out in classical test tubes using NaNO3, Ca(NO3)2, and La(NO3)3 without any pH adjustment . The pH in all the experiments was then close to 7.5. In all cases, the electrolyte solution was first prepared, and the clay suspension was then added. Stirring was simply carried out by turning the test tube two or three times. The turbidity of the suspension in the top section of the test tube was measured using a Hach turbidimeter after 15 h, whereas the presence of

■

RESULTS AND DISCUSSION Turbidity Measurements. Figure 1 presents the evolution of the turbidity after settling in the case of size 1 particles at a total solid concentration of 0.8 g/L for increasing concentrations of Na-, Ca-, and La-nitrate. The curves thus obtained are typical of such experiments with an increase in turbidity, due to the formation of small flocs followed at higher concentration by a strong decrease of the turbidity of the supernatant due to the settling of larger flocs at the bottom of the test tube. Assuming that the onset of coagulation can be located at the start of turbidity increase, the “coagulation concentrations” can be estimated around 10−2 M/ L for NaNO3, 2.10−4 M/L for Ca(NO3)2 and 6.10−5 M/L for 3501


Langmuir

Article

La(NO3)3. The definition that we use here is not that of the critical coagulation concentration (c.c.c.) but corresponds to the point where objects larger than the initial ones start forming in the suspension. The values thus obtained are close to the literature data for similar pH.17 It can be immediately seen that the concentrations thus derived do not follow the classical Schulze−Hardy rule, according to which the coagulation concentration should vary as the sixth power of ion valence, a conclusion that was already reached in the first papers dealing with clay coagulation.12 The effect of particle size on coagulation is illustrated in Figure 2 that displays, for solid fractions of 0.1, 0.8, and 5 g/L the evolution of turbidity for increasing concentrations in NaNO3 (Figure 2 A, D, G) Ca(NO3)2 (Figure 2 B, E, H), and La(NO3)3 (Figure 2 C, F, I). In general, it appears that, especially for the two higher solid concentrations used, a decrease in the size of the particles leads to a slight increase in

Figure 1. Evolution of the turbidity with increasing salt concentration for size 1 Wyoming montmorillonite at a solid concentration of 0.8 g/ L.

Figure 2. Evolution with particle size of the turbidity as a function of salt concentration for solid concentrations of 0.1 g/L: (A) NaNO3, (B) Ca(NO3)2, and (C) La(NO3)3; 0.8 g/L: (D) NaNO3, (E) Ca(NO3)2, and (F) La(NO3)3; and 5 g/L: (G) NaNO3, (H) Ca(NO3)2, and (I) La(NO3)3. 3502


Langmuir

Article

Figure 3. Evolution with solid concentration of the salt concentration corresponding to the onset of coagulation. (A) Size 1. (B) Size 2. (C) Size 3.

Figure 4. Evolution of the scattered intensity as a function of scattering vector for size 1 Wyoming montmorillonite (solid concentration 2 g/L). A− E: NaNO3 (A) 3.10−3 M/L; (B) 7.10−3 M/L; (C) 10−2 M/L; (D) 3.10−2 M/L; (E) 1 M/L. F−K: Ca(NO3)2. (F) 10−5 M/L; (G) 10−4 M/L; (H) 3.10−4 M/L; (I) 7 × 10−4 M/L; (J) 10−3 M/L; (K) 10−1 M/L. L−Q La(NO3)3. (L) 10−6 M/L; (M) 3.10−5 M/L; (N) 7.10−5 M/L; (O) 2.10−4 M/L; (P) 10−3 M/L; (Q) 10−2 M/L.

concentration. For all salts and all sizes, a significant increase of the salt concentration required to trigger coagulation is observed at low solid content. Such an effect appears to be classical of the behavior of swelling clay minerals.18 SAXS Experiments. Figure 4 illustrates some of the SAXS results obtained by presenting the evolution of the scattered intensity as a function of the scattering vector q for size 1 Wyoming montmorillonite at a solid concentration of 2 g/L. In all of these curves the peak observed for a q value of 4.3 nm−1 and marked by a star on the figures corresponds to a parasitic reflection arising from the Kapton polymer used as a window material. Very similar curves are obtained for lower sizes, and common features are observed for all the investigated systems whatever the particle size. For low electrolyte concentrations, the scattered intensity decays as q−2, indicating the bidimensional character of the scattering objects. In those conditions, the system can then be viewed as a collection of individual clay platelets in suspension. However, the existence of a few

the salt concentration corresponding to the onset of coagulation. This trend is the reverse to that observed for various particles such as hematite,27 latex, or silica.28 In that context, it must be emphasized that the relationship between size and coagulation remains relatively ill-defined. Some of the experiments carried out with silica particles29 indeed reveal an inverse dependency on size, the smaller particles being stabilized, which is interpreted as resulting from surface chemical effects. In the case of swelling clay minerals, the surface chemistry does not depend on size (although the proportion of edge to basal faces increases with decreasing size), which certainly makes these materials extremely suitable for investigating in depth the possible non-Derjaguin−Landau−Verwey− Overbeek (nonDLVO) effects linked to the size dependence of coagulation. This point will be further examined in the following sections. Figure 3 presents the evolution of the salt concentration corresponding to the onset of coagulation as a function of solid 3503


Langmuir

Article

Figure 5. Evolution with salt concentration of the full width at half-maximum of the correlation peak observed in the SAXS experiments. (Filled circles) NaNO3. (Empty squares) Ca(NO3)2. (Filled triangles) La(NO3)3. Blue: Solid concentration 0.1 g/L. Red: Solid concentration 0.8 g/L. Green: Solid concentration 2 g/L. Black: Solid concentration 5 g/L. (A) Size 1. (B) Size 2. (C) Size 3. Lines are guides to the eye.

Figure 6. Evolution with salt concentration of the low angle slope observed in the SAXS experiments. (Filled circles) NaNO3. (Empty squares) Ca(NO3)2. (Filled triangles) La(NO3)3. Blue: Solid concentration 0.1 g/L. Red: Solid concentration 0.8 g/L. Green: Solid concentration 2 g/L. Black: Solid concentration 5 g/L. (A) Size 1. (B) Size 2. (C) Size 3.

nanometers, for the three salts, four solid concentrations, and three clay sizes investigated. In the case of size 3 particles (Figure 5C), the correlation peaks for sodium and calcium salts are rather faint, and a width can only be safely derived for samples coagulated with lanthanum nitrate. A few trends can be deduced from Figure 5. (i) The average peak width increases with decreasing particle size. In the case of a face−face coagulation, this appears rather logical as larger particles have a higher probability to get stuck in their face−face configuration. (ii) The average peak width decreases with increasing cation valence. Such trends were also observed in recent studies dealing with tactoid formation in montmorillonite suspensions.32,33 Once again this behavior appears rather logical as the correlation forces increase with cation valence as shown by simulation in the framework of the primitive model.34−36 Still, the difference remains rather limited as, for a given state of coagulation, such as the highest concentration investigated for each salt, the difference in width is around 30% at most. (iii) For a given solid concentration, the peak width slightly diminishes with increasing salt concentration, before plateauing at a higher concentration. This shows that, for higher salt concentration, slightly more clay layers stack together to form the building blocks of the flocs. The appearance of a correlation peak at high q is systematically associated to a change of slope in the low q region of the SAXS curves. Figure 6 presents the evolution of this slope with salt concentration for the three salts, three sizes and four solid concentrations used in the present study. If one compares the salt concentration values corresponding to a

doublets, especially for larger-size particles can be deduced from the thickness values displayed in Table 1. Above a certain concentration, the value of which depends on cation valence, particle size, and solid concentration, the low angle slope of the scattering curves starts increasing, whereas in the high angle region, this feature is accompanied by the appearance of a correlation peak corresponding to distances around 19 Å. This can be interpreted as corresponding to the formation in the suspension of stacks of layers with repetition distances equivalent to the intercalation of approximately three water layers. It shows that, in the conditions used in our work (natural pH), Na-montmorillonite coagulation clearly occurs mainly through the formation of face-to-face objects. This is consistent with studies of the charge characteristics of swelling clay minerals30,31 which show that, at pH 8, the edge faces of montmorillonite are likely negatively charged. In that context, then, it is particularly relevant to try to follow the evolution with salt concentration of the correlation peak width for the various concentrations and clay sizes used in the present study. Indeed, as the position of the peak on the q-axis is almost constant, changes in the full width at half-maximum can be interpreted as corresponding to changes in the average number of clay layers stacked together inside the flocs. In principle, it could even be possible to use Scherrer’s formula to derive an average size of the coherent scattering domains and hence an average number of clay layers per stack. However, in view of the relatively poor accuracy on peak width, we will here restrict ourselves to analyze the relative evolutions. Figure 5 presents the evolution of the peak width measured in inverse 3504


Langmuir

Article

needed for coagulation to start occurring. Considering that the experimental conditions used in the present work are those of Brownian coagulation (without any specific stirring) and that the main “sticking” mechanism is face-to-face, the reason for such a combined behavior may be geometrical. Indeed, higher particle numbers and/or larger particle size increases the probability with which two particles find themselves in each other’s vicinity. In the case of anisometric particles, such as those considered in the present study, this can be captured by using the equivalent spherical volume fraction of the particles,39−41 defined as the volume of the equivalent sphere of radius R generated by the free rotation of the disklike particle.

slope increase with the evolution of turbidity displayed in Figure 2, it clearly appears that the onset of coagulation deduced from turbidity measurements also corresponds to the salt concentration where the low angle slope of the SAXS curves start increasing. The values of the slope for the highest salt concentrations corresponding to the formation of flocs strongly depend on cation valence with average values of 2.6, 3.0, and 3.3 for sodium nitrate, calcium nitrate and lanthanum nitrate, respectively. The slope values therefore appear to increase with the thickness of the stacks of particles. According to the literature, such slope values can be interpreted as corresponding to mass fractals or surface fractals. However, alternate explanations can be linked to a very polydisperse distribution of heterogeneities.37,38 We would tend to favor this latter interpretation, as the system appears rather disorganized (see below). However, more work is required before a definite answer regarding the local structure of flocs can be provided. Furthermore, the influence of particle size on coagulation can be further analyzed from the results provided in Figure 6 that display the same trends as those described in Figure 2. As SAXS results reveal that before coagulation isolated platelets are present in the suspension, one can determine the total number of particles/L for all the conditions used in the present study on the basis of the morphology data presented in Table 1. As the charge of the clay platelets is known, it is also straightforward to determine the total amount of exchangeable charge present in the suspension for all cases. It is then possible to calculate for each salt, at the onset of coagulation, the ratio between the charge brought by the salt and the initial exchangeable charge. Figure 7 displays the evolution of this

ϕsphere =

Vsphere Vdisk

ϕ=

4 R3 ϕ 3 r 2e

(2)

The value of R depends on the thickness (e) and diameter (d = 2r) of the particles and R = ((d2 + e2)1/2/2). In the specific case of strongly anisotropic particles such as those used here, d ≫ e, the average radius R of the equivalent sphere is then close to r, the radius of the disklike particle. The expression of ϕsphere therefore reduces as: 4r ϕsphere ≈ ϕ (3) 3e Figure 8 displays the evolution of the charge ratio previously defined as a function of ϕsphere for the three salts used in the

Figure 8. Evolution with ϕsphere of the charge ratio corresponding to the onset of coagulation. (Filled circles) NaNO3. (Empty squares) Ca(NO3)2. (Filled triangles) La(NO3)3. Red: Solid concentration 0.8 g/L. Green: Solid concentration 2 g/L. Black: Solid concentration 5 g/ L.

Figure 7. Evolution with the number of particles of the charge ratio corresponding to the onset of coagulation (see text for details) as derived from the change in the low angle slope in SAXS experiments. (Filled circles) NaNO3. (Empty squares) Ca(NO3)2. (Filled triangles) La(NO3 ) 3 . Red: Solid concentration 0.8 g/L. Green: Solid concentration 2 g/L. Black: Solid concentration 5 g/L.

present study. In all cases, a power law dependence is obtained, which shows that the notion of spherical volume fraction captures most of the size and concentration dependence of the system. It must be pointed out that such a feature should also appear in coagulation kinetics that have not been investigated in detail in the present study. The combination of SAXS and WAXS experiments clearly provides fruitful information about the coagulation mechanisms and both the local and semilocal organization of the flocs. Coagulation involves the formation of stacks of layers that likely organize themselves at higher scale as heterogeneous porous structures. Transmission X-ray microscopy experiments were

ratio as a function of the number of particles in the suspension. This graph reveals various trends. (i) The valence effect already described appears clearly. In some cases, especially for lanthanum, coagulation starts before complete neutralization of the charge of the platelets. (ii) For a given particle size, an increase in the number of particles leads to a decrease of the amount of salt required for coagulation. (iii) For a given solid concentration, the dependency of the charge ratio on particle number is not linear. It then appears that more particles or larger particles lead to a decrease of the amount of salt/charge 3505


Langmuir

Article

Figure 9. TXM images of coagulated size 1 Wyoming montmorillonite. Solid concentration 10 g/L. (A) NaNO3 1M/L. (B) Ca(NO3)2 0.1 M/L. (C) La(NO3)3 0.01 M/L.

Figure 10. TXM images of coagulated size 1 Wyoming montmorillonite. Solid concentration 0.5 g/L. (A) NaNO3 1 M/L. (B) Ca(NO3)2 0.1 M/L. (C) La(NO3)3 0.01 M/L.

then performed to obtain more detailed information about the large-scale organization of the flocs. Transmission X-ray Microscopy Experiments. Figure 9 presents some X-ray transmission images obtained on size 1 montmorillonite particles coagulated in NaNO3 1 M/L (Figure 9A), Ca(NO3)2 0.1 M/L (Figure 9B) and La(NO3)3 0.01 M/L (Figure 9C) for a solid concentration of 10 g/L. In all cases, strongly anisotropic structures are observed, thus confirming a dominant face−face coagulation. Even at large scale, there is no evidence for significant edge−face contacts. The anisotropic structures observed appear to have roughly similar thickness independently of the nature of the cation used in the coagulation experiments. Very similar structures are obtained for lower solid concentrations as evidenced in Figure 10 that displays the flocs obtained for the same salt concentrations but for a solid concentration of 0.5 g/L. In the same way, the salt concentration does not seem to have a significant impact on the observed features as shown in Figure 11 that displays the flocs obtained for a solid concentration of 0.5 g/L and a concentration in Ca(NO3)2 of 0.001 M/L. In order to obtain more quantitative information about the large-scale structure of the flocs, image analysis procedures were applied to the various images obtained on coagulated size 1 montmorillonite particles. The parameters determined were the average length and width of the particles as well as the percentage of significantly curved particles. This latter parameter was approached from the ratio of the longest axis of the particles to their Feret diameter. Values higher than 1.1 were considered as corresponding to significantly curved particles. Table 2 presents for different conditions the average width, length and percentage of curved particles derived from the image analysis procedure. In agreement with visual observa-

Figure 11. TXM image obtained on size 1 montmorillonite at a solid concentration of 0.5 g/L coagulated in Ca(NO3)2 0.001 M/L.

Table 2. Average Width, Length and Proportion of Curved Particles Deduced from the Image Analysis for Coagulated Size 1 Montmorillonite Particles width (nm) NaNO3 1 M 0.5 g/L Ca(NO3)2 0.1 M 0.5 g/L La(NO3)3 0.01 M 0.5 g/L NaNO3 1 M 10 g/L Ca(NO3)2 0.1 M 10 g/L La(NO3)3 0.01 M 10 g/L

3506

64 62 64 68 67 64

± ± ± ± ± ±

20 20 20 20 20 20

length (nm)

% curved particles

± ± ± ± ± ±

20 21 29 21 24 27

578 531 647 566 570 616

180 170 200 180 180 195


Langmuir

Article

Figure 12. Small angle scattering curves obtained from SAXS and from the treatment of TXM images displayed in Figure 9.

Figure 13. Small angle scattering curves obtained from SAXS and from the treatment of TXM images displayed in Figure 10.

Figure 14. TXM images of coagulated size 2 Wyoming montmorillonite. Solid concentration 0.5 g/L. (A) NaNO3 1 M/L. (B) Ca(NO3)2 0.1 M/L. (C) La(NO3)3 0.01 M/L.

simulation work.20 In addition the structures thus formed must be porous, which could well explain the slopes observed at low angle in the SAXS experiments. According to such an interpretation, the compacity or pore size heterogeneity would be different depending on the valence of the cation used in the coagulation experiments. Cryo-TEM experiments could provide the missing link between results derived from SAXS and those derived from TXM and help in confirming the assumptions that we make. To further analyze the large-scale structures, it appears particularly relevant to calculate the small angle scattering curves that can be obtained from the TXM images according to the application of eq 1 and to plot them together with the SAXS data. Figure 12 and 13 present the data thus obtained. The corresponding images are those of Figures 9 and 10, respectively. In both cases, the SAS curves derived from TXM images clearly extend the SAXS results toward low q values, i.e toward

tions, the width of the particles is nearly constant around 60 nm whatever the conditions. The average length of the particles is also nearly constant, though lanthanum appears to lead to slightly longer particles. It must be pointed out that the average length of the objects observed by TXM is significantly higher than that of the individual platelets (410 nm). Considering an interlayer distance of around 2 nm as observed by SAXS, the average width of the structures observed by TXM would correspond to the stacking of around 30 clay layers, provided that a regular stacking is achieved. The width of the correlation peak observed in SAXS experiments is not consistent with such a value, which would yield extremely sharp correlation peaks. Considering in addition that as the length of the objects is longer than that of individual clay layers, the structures observed by TXM must be viewed as corresponding to the larger scale arrangement of relatively thin particle stacks. In such arrangements, there is likely a partial mismatch between the layers. Such a conformation was recently proposed from 3507


Langmuir

Article

the region of where the two techniques overlap. As far as Figure 15A is concerned, the low q region must be analyzed cautiously, considering the image quality and the presence of rings that are artifacts due to the beam coherence (Figure 14A). The differences obtained between calcium (Figure 15B) and lanthanum (Figure 15C) are much more significant. As shown in the insets, the SAS curves derived from TXM display a q−4 dependence at high q. Even if this region of the curve is sensitive to the filtering used in the image processing, this behavior can be interpreted as corresponding to the presence of sharp interfaces. According to simulations of the SAS patterns of isolated anisometric objects, the deviation from the q−4 region occurs at a q value for which π/q corresponds to the thickness of the objects. The q value obtained in the present experiments is close to 0.1 nm−1, i.e. distances around 35 nm, in agreement with the minimal thickness derived from image processing procedures (Table 3). The difference at lower q between calcium and lanthanum is clearly related to the significant length differences between the two systems. In the low q region, q scales as q−2 or q−1.8. Such exponents are very close to the one found for 3D cluster−cluster aggregation processes.42 This may indicate a hierarchical scale organization of coagulated clay systems. In that regard, it would have been relevant to examine the large-scale organization obtained when dealing with size 3 montmorillonite. However, as revealed by the comparison between images obtained for size 1 and size 2 particles, it becomes increasingly difficult to analyze systems where particle size is close to the experimental resolution of the system. Further improvement in focusing Fresnel lenses used in transmission X-ray microscopes should allow analyzing the behavior of smaller particles such as size 3 Wyoming montmorillonite in the near future.

larger distances with some continuity between both experiments. It shows that, for the coagulation conditions examined in these experiments, i.e. “concentrated” salt solutions, the structure appears constant for 10−2 < q < 1. For lower q values, the curves tend to level off. This may indicate a trend toward a finite size. Unfortunately, the curves being obtained in arbitrary units, this low q region cannot be treated quantitatively. Still, simulations of the scattering by elongated objects with dimensions compatible with those shown in Table 2, suggest that the downward modification of the scattering at low q start occurring for q values that are close to 2π/d where d is the length of the objects. In the present case, the downward deviation from the power law often occurs around q = 10−2 nm, i.e. for distances around 600 nm, that agree with the values measured by image analysis. Figure 14 displays typical TXM images obtained for size 2 particles, whereas Table 3 presents the results obtained by Table 3. Average Width, Length, and Proportion of Curved Particles Deduced from the Image Analysis for Coagulated Size 2 Montmorillonite Particles width (nm) NaNO3 1 M 0.5 g/L Ca(NO3)2 0.1 M 0.5 g/L La(NO3)3 0.01 M 0.5 g/L La(NO3)3 0.01 M 5 g/L

52 55 82 65

± ± ± ±

15 15 20 20

length (nm)

% curved particles

± ± ± ±

7 9 20 16

250 277 597 527

100 100 200 180

image analysis. Similar trends as those observed for size 1 particles can be deduced from the analysis carried out on size 2 montmorillonite, i.e. a nearly constant thickness around 60 nm and a length of the objects higher than the length of the individual clay layers. However, the differences between the samples coagulated with lanthanum nitrate and those coagulated with either sodium or calcium nitrate are more marked than in the case of size 1 particles. Indeed, the length of the lanthanum-coagulated objects is approximately twice that obtained for sodium or calcium-coagulated samples. In addition, it appears that the proportion of curved objects is nearly directly correlated with the total length. This was already the case for size 1 particles but was less obvious due to the lower variability in length. Figure 15 presents together with the SAXS data the small angle scattering curves obtained from the treatment of the images displayed in Figure 14. As in the case of size 1 particles, a clear continuity between SAXS and TXM is obtained. However, significant differences can be observed at low q and in

■

CONCLUSIONS AND PERSPECTIVES

The combination of WAXS, SAXS, and TXM appears as a very powerful approach for a better understanding of the coagulation mechanisms of sodium montmorillonite, especially when applying quantitative treatments on the TXM images. Such a statement can be extended to other aggregated or coagulated systems, as this combination allows a true multiscale examination of the various structures formed during the process. In the case of swelling clay minerals, using sizeselected samples provides additional information on the physicochemical mechanisms that control coagulation. In the conditions used in the present study, i.e. at neutral pH and starting with fully Na-exchanged samples, coagulation proceeds via the formation of stacks of particles with a slight mismatch

Figure 15. Small angle scattering curves obtained from SAXS and from the treatment of TXM images displayed in Figure 14. 3508


Langmuir

Article

illonite Clay. Influence of Particle Anisotropy. Langmuir 2004, 20, 10829−10837. (10) Michot, L. J.; Bihannic, I.; Maddi, S.; Funari, S. S.; Baravian, C.; Levitz, P.; Davidson, P. Liquid-Crystalline Aqueous Clay Suspensions. Proc. Natl Acad. Sci. U.S.A. 2006, 103, 16101−16104. (11) Michot, L. J.; Bihannic, I.; Maddi, S.; Baravian, C.; Levitz, P.; Davidson, P. Sol/Gel and Isotropic/Nematic Transitions in Aqueous Suspensions of Natural Nontronite Clay. Influence of Particle Anisotropy. 1. Features of the I/N Transition. Langmuir 2008, 24, 3127−3139. (12) Jenny, H.; Reitemeier, R. F. Ionic Exchange in Relation to the Stability of Colloidal Systems. J. Phys. Chem. 1935, 39, 593−604. (13) Kahn, A. J. The Flocculation of Sodium Montmorillonite by Electrolytes. J. Colloid Sci. 1958, 13, 51−60. (14) Hsi, H. R.; Clifton, D. F. Flocculation of Selected Clays by Various Electrolytes. Clays Clay Miner. 1962, 9, 269−275. (15) Fitzsimmons, R. D.; Posner, A. M.; Quirk, J. P. Electron Microscopic and Kinetic Study of the Flocculation of Calcium Montmorillonite. Isr. J. Chem. 1970, 8, 301−314. (16) Helmy, A. K.; Ferreiro, E. A. Flocculation of NH 4 Montmorillonite by Electrolytes. Electroanal. Chem. Interface Electrochem. 1974, 57, 103−112. (17) Swartzen-Allen, S. L.; Matijevic, E. Colloid and Surface Properties of Clay Suspensions. III. Stability of Montmorillonite and Kaolinite. J. Colloid Interface Sci. 1976, 56, 159−167. (18) Lagaly, G.; Ziesmer, S. Colloid Chemistry of Clay Minerals: The Coagulation of Montmorillonite Dispersions. Adv. Colloid Interface Sci. 2003, 100−102, 105−128. (19) Lagaly, G. In Handbook of Clay Science; Bergaya, F., Theng, B.K.G., Lagaly, G., Eds.; Elsevier: Amsterdam; London, 2006; Chapter 5, pp 141−245. (20) Jönsson, B.; Labbez, C.; Cabane, B. Interaction of Nanometric Clay Platelets. Langmuir 2008, 24, 11406−11413. (21) Meyer, S.; Levitz, P.; Delville, A. Influence of the Relative Orientation of Two Charged Anisotropic Colloidal Particles on Their Electrostatic Coupling: A (N,V,T) Monte Carlo Study. J. Phys. Chem. B 2001, 105, 10684−10690. (22) Schmahl, G.; Rudolph, D.; Niemann, B.; Guttman, P.; Thieme, J.; Schneider, G. X-ray Microscopy. Naturwissenchaften 1996, 83, 61− 70. (23) Thieme, J.; Sedlmair, J.; Gleber, S. C.; Prietzel, J.; Coates, J.; Eusterhues, K.; Abbt-Braun, G.; Salome, M. X-ray Spectromicroscopy in Soil and Environmental Sciences. J. Synchrotron Radiat. 2010, 17, 149−157. (24) Paineau, E.; Bihannic, I.; Baravian, C.; Philippe, A.-M.; Davidson, P.; Levitz, P.; Funari, S. S.; Rochas, C.; Michot, L. J. Aqueous Suspensions of Natural Swelling Clay Minerals. 1. Structure and Electrostatic Interactions. Langmuir 2011, 27, 5662−5573. (25) Brisard, S.; Chae, R. S.; Bihannic, I.; Michot, L. J.; Guttmann, P.; Thieme, J.; Schneider, G.; Monteiro, P. J. M.; Levitz, P. Hierarchical Materials under the X-ray Microscope. Filling in the Gap between Nano and Micro Scales. Am. Mineral. 2012, 97, 480−483. (26) Kak, A. C.; Slaney, M. Principles of Computerized Tomographic Imaging; Society for Industrial and Applied Mathematics: Philadelphia, 2001. (27) He, Y. T.; Wan, J.; Tokunaga, T. Kinetic Stability of Hematite Nanoparticles: The Effect of Particle Sizes. J. Nanopart. Res. 2008, 10, 321−332. (28) Swanton, S. W. Modelling Colloid Transport in Groundwater; the Prediction of Colloid Stability and Retention Behaviour. Adv. Colloid Interface Sci. 1995, 54, 129−208. (29) Kobayashi, M.; Juillerat, F.; Galletto, P.; Bowen, P.; Borkovec, M. Aggregation and Charging of Colloidal Silica Particles: Effect of Particle Size. Langmuir 2005, 21, 5761−5769. (30) Tombacz, E.; Szekeres, M. Colloidal Behavior of Aqueous Montmorillonite Suspensions: the Specific Role of pH in the Presence of Indifferent Electrolytes. Appl. Clay Sci. 2004, 27, 75−94. (31) Duc, M.; Gaboriaud, F.; Thomas, F. Sensitivity of the Acid-Base Properties of Clays to the Methods of Preparation and Measurement -

between layers. These stacks arrange themselves into larger, porous, anisotropic particles, the porosity of which depends on the valence of the cation used for coagulation experiments. Face−face coagulation is clearly dominant under those conditions, and no evidence for any face−edge coagulation was found. These structures appear to arrange as larger clusters, the organization of which should control the mechanical properties of the flocs. To go further in our understanding of coagulation features of swelling clay minerals, it would certainly be relevant to carry out a similar study by modifying the pH conditions in order to better assess the influence of this latter parameter that should change the edge-charge. Furthermore, all the experiments described in the present study were carried out in equilibrium conditions, and it would be worth analyzing the kinetics of floc formation in such systems.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: (33) 3 83 59 62 59. Fax: (33) 3 83 59 62 55. E-mail: [email protected]. Present Addresses ∥

Géosciences Environnement Toulouse, UMR 5563 (CNRS/ UPS/IRD/CNES), 14, avenue Édouard Belin, 31400 Toulouse, France. ¶ Brookhaven National Laboratory, NSLS-II, Building 817, Upton, NY 11973, United States. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Support of beamtime at Hasylab by the European Community Research Infrastructure Action under the FP6 “Structuring the European Research Area” Programme (through the Integrated Infrastructure Initiative “Integrating Activity on Synchrotron and Free Electron Laser Science”) Contract RII3-CT-2004506008 is gratefully acknowledged. We thank Drs. Alfred Delville and Jerôme F.L. Duval for enlightening discussions.

■

REFERENCES

(1) Harvey, C. C.; Lagaly, G. In Handbook of Clay Science; Bergaya, F.; Theng, B.K.G., Lagaly, G., Eds.; Elsevier: Amsterdam; London, 2006; Chapter 10.1, pp 501−540. (2) Edzwald, J. K.; O’Melia, C. R. Clay Distributions in Recent Estuarine Sediments. Clays Clay Min. 1975, 23, 39−44. (3) Edzwald, J. K.; Upchurch, J. B.; O’Melia, C. R. Coagulation in Estuaries. Environ. Sci. Technol. 1974, 8, 58−63. (4) Hunt, J. R. Particle Dynamics in Seawater: Implications for Predicting the Fate of Discharged Particles. Environ. Sci. Technol. 1982, 16, 303−309. (5) Zhou, D.; Abdel-Fattah, A. I.; Keller, A. A. Clay Particles Destabilize Engineered Nanoparticles in Aqueous Environments. Environ. Sci. Technol. 2012, 46, 7520−7526. (6) Labille, J.; Thomas, F.; Bihannic, I.; Santaella, C. Destabilization of Montmorillonite Suspensions by Ca2+ and Succinoglycan. Clay Min. 2003, 38, 173−185. (7) Labille, J.; Thomas, F.; Milas, M.; Vanhaverbeke, C. Flocculation of Colloidal Clay by Bacterial Polysaccharides: Effect of Macromolecule Charge and Structure. J. Colloid Interface Sci. 2005, 284, 149−156. (8) Mongondry, P.; Tassin, J. F.; Nicolai, T. Revised State Diagram of Laponite Dispersions. J. Colloid Interface Sci. 2005, 283, 397−405. (9) Michot, L. J.; Bihannic, I.; Porsch, K.; Maddi, S.; Baravian, C.; Mougel, J.; Levitz, P. Phase diagrams of Wyoming Na-Montmor3509


Langmuir

Article

2. Evidence from Continuous Potentiometric Titrations. J. Colloid Interface Sci. 2005, 289, 148−156. (32) Segad, M.; Anski, S.; Olsson, U.; Ruokolainen, J.; Åkeson, T.; Jönsson, B. Microstructural and Swelling Properties of Ca and Na Montmorillonite: (In Situ) Observations with Cryo-TEM and SAXS. J. Phys Chem C 2012, 116, 7596−7601. (33) Segad, M.; Jönsson, B.; Cabane, B. Tactoid Formation in Montmorillonite. J. Phys. Chem. C 2012, 116, 25425−25433. (34) Pellenq, R. J. M.; Caillol, J. M.; Delville, A. Electrostatic Attraction between Two Charged Surfaces: A (N,V,T) Monte Carlo Simulation. J. Phys. Chem. B 1997, 101, 8584−8594. (35) Delville, A.; Gasmi, N.; Pellenq, R. J.-M.; Caillol, J. M.; van Damme, H. Correlations between the Stability of Charged Interfaces and Ionic Exchange Capacity. A Monte Carlo Study. Langmuir 1998, 14, 5077−5082. (36) Segad, M.; Jönsson, B.; Åkeson, T.; Cabane, B. Ca/Na Montmorillonite: Structure, Forces, and Swelling Properties. Langmuir 2010, 26, 5782−5790. (37) Bihannic, I.; Delville, A.; Demé, B.; Plazanet, M.; Villiéras, F.; Michot, L. J. Clay Swelling: New Insights from Neutron-Based Techniques. In Neutron Applications in Earth, Energy and Environmental Sciences; Neutron Scattering Applications and Techniques; Springer: New York, 2008; Chapter 18, pp 521−546. (38) Botet, R.; Cabane, B. Simple Inversion Formula for the SmallAngle X-ray Scattering Intensity from Polydisperse Systems of Spheres. J. Appl. Crystallogr. 2012, 45, 406−416. (39) Baravian, C.; Vantelon, D.; Thomas, F. Rheological Determination of Interaction Potential Energy for Aqueous Clay Suspensions. Langmuir 2003, 19, 8109−8114. (40) Michot, L. J.; Baravian, C.; Bihannic, I.; Maddi, S.; Moyne, C.; Duval, J. F. L.; Levitz, P.; Davidson, P. Sol/Gel and Isotropic/Nematic Transitions in Aqueous Suspensions of Natural Nontronite Clay. Influence of Particle Anisotropy. 2. Gel Structure and Mechanical Properties. Langmuir 2009, 25, 127−139. (41) Paineau, E.; Michot, L. J.; Bihannic, I.; Baravian, C. Aqueous Suspensions of Natural Swelling Clay Minerals. 2. Rheological Characterization. Langmuir 2011, 27, 7806−7819. (42) Meakin, P. Reaction-Limited Cluster-Cluster Aggregation in Dimensionalities 2−10. Phys. Rev. A 1988, 38, 4799−4813.

3510


Coagulation of Na-Montmorillonite by Inorganic ... - ACS Publications

Recommend Documents