8954
Langmuir 2008, 24, 8954-8958
Transmission X-ray Microscopy (TXM) Reveals the Nanostructure of a Smectite Gel Marek S. Z˙bik,*,† Wayde N. Martens,† Ray L. Frost,† Yen-Fang Song,‡ Yi-Ming Chen,‡ and Jian-Hua Chen‡ Inorganic Materials Research Program, School of Physical and Chemical Sciences, Queensland UniVersity of Technology, 2 George Street, GPO Box 2434, Brisbane Qld 4001, Australia, and National Synchrotron Radiation Research Center, 101 Hsin-Ann Road, Hsinchu Science Park, Hsinchu 30076, Taiwan, R.O.C. ReceiVed March 29, 2008. ReVised Manuscript ReceiVed May 26, 2008 The unusual behavior of smectites, the ability to change volume when wetted (swelling) or dried (shrinking), makes soil rich in smectites very unstable and dangerous for the building industry because of the movement of building foundations and poor slope stability. These macroscopic properties are dominated by the structural arrangement of the smectites’ finest fraction. Here, we show in three dimensions how the swelling phenomenon in smectite, caused by a combination of hydratation and electrostatic forces, may expand the dry smectite volume not 10-fold, as previously thought, but to more than 1000-fold. A new technique, transmission X-ray microscopy, makes it possible to investigate the internal structure and 3-D tomographic reconstruction of clay aggregates. This reveals, for the first time, the smectite gel arrangement in the voluminous cellular tactoid structure within a natural aqueous environment.
Introduction Smectites are clay minerals commonly found as components of soils in temperate climates. Smectites are formed as a result of the weathering of volcanic glass that is abundant in ash beds and basic rocks such as basalts. Smectites are useful in dam bed impregnation, improving water retention properties, and as drilling mud to seal the cut, thus preventing fluid loss. They are also popular stabilizing additives in engine oils and cosmetics and in the pharmaceutical and chemical industries. Smectite’s unusual macroscopic properties are dominated by its structural arrangement and the morphology of its finest fraction. These clays are extremely disperse and have very high surface areas of several hundred square meters per gram. The first attempt to describe the microstructure of clays was made by Terzaghi,1 who proposed the honeycomb model as the structural basis of water-saturated clays. Subsequent investigations2–5 introduced several models of clay structures in an aqueous environment, which were based on conjectural notions of clay microstructures. The first experimental confirmation was obtained much later, with the advent of transmission electron microscopy (TEM) and scanning electron microscopy (SEM).6 SEM micrographs confirmed the existence of the “card house” structure with Rosenquist,7 Bowles,8 and Pusch9 confirming the presence of the honeycomb microstructures in wet clay sediments. Cryo* Corresponding author. E-mail:
[email protected]. † Queensland University of Technology. ‡ National Synchrotron Radiation Research Center. (1) Terzaghi, K. Erdbaummechanik auf Bodenphysikalischer Grundlage; Franz Deuticke Press: Leipzig, 1925. (2) Casagrande, A. J. Boston Soc. CiVil Eng. 1932, 19, 168–208. (3) Goldshmidt, V. M. Nord. Jordbrugsforsk. 1926, 4-7, 434–445. (4) Lambe, T. W. Proc. Am. Soc. CiVil Eng. 1953, 79, 315, 3–49. (5) Sloane, R. L.; Kall, T. R. J. Soil Mech. Found. DiV. Am. Soc. CiVil Eng. 1966, 85, 5, 87–128. (6) Van Olphen, H. An Introduction to Clay Colloid Chemistry; Interscience Publishers: New York, 1963. (7) Rosenquist, J. T. J. Soil Mech. Found. DiVi., Proc. ASCE Sm. 2 1959, 85, 31–53. (8) Bowles, F. A. Science 1968, 159, 1236–1237. (9) Pusch, R. Clay Microstructure; National Swedish Building Research, Document D8, 1970.
SEM investigations by O’Brien10 resulted in the collection of a large amount of microstructural data. Given the size of the clay constituents, SEM was found to be the tool of choice used by scientists studying the microstructure of smectitic clays.11,12 Sample preparation methods available for these investigations, such as partial freeze drying, critical-point drying, and cryofixation, have been found to introduce many artifacts, especially when applied to the study of smectite structure.13 These artifacts result as a consequence of the low thermal conductivity of water and ice, which allows only a slow rate of heat withdrawal from the specimen. Thus, the size of the gelled smectite sample must be small enough to freeze quickly, limiting the inevitable damage associated with the sampling process.
Experimental Section Recently, a revolutionally new technique has been implemented in the study of nanomaterial sciencesnanotomography.14 This method is based on transmission X-ray microscopy (TXM), which works with a synchrotron photon source at the BL01B1 endstation of the National Synchrotron Radiation Research Center (NSRRC).15 This new technique16 has recently been established to investigate clay suspension in an aqueous environment without sample preparation. The major advantage of TXM tomography is that without sample pretreatment it is possible, for the first time, to observe clay microstructure in an aqueous environment, artifact-free. This method has recently been tested and described in detail in the study of another clay mineral, kaolinite,17,18 of which the discrete structure of the clay-water aggregates was successfully reproduced as an image. (10) O’Brien, N. R. Clays Clay Miner. 1971, 19, 353–359. (11) Grabowska-Olszewska, B.; Osipov, V.; Sokolov, Vi. Atlas of the Microstructure of Clay Soils; PWN: Warszawa, 1984. (12) Smart, P.; Tovey, N. K. Electron Microscopy of Soils and Sediments: Techniques; Clarendon Press: Oxford, U.K., 1982. (13) Smart, R. St. C.; Zb˙ik, M.; Morris, G. E. 43rd Annual Conference of Metallurgists of CIM; Laskowski, J. S., Ed.; 2004, Vol. 21, pp 5-228. (14) Chenu, C.; Tessier, D. Scanning Microsc. 1995, 9, 989–1010. (15) Attwood, D. Nature 2006, 442, 642–643. (16) Yin, G. C.; Tang, M. T.; Song, Y. F.; Chen, F. R.; Liang, K. S.; Duewer, F. W.; Yun, W.; Ko, D. H.; Shieh, H.-P. D. Appl. Phys. Lett. 2006, 88, 2411151–241115-3. (17) Zbik, M. S.; Frost, R. L.; Song, Y.-F. J. Colloid Interface Sci. 2008, 319, 169–174. (18) Zbik, M. S.; Frost, R. L.; Song, Y.-F.; Chen, Y.-M.; Chen, J.-H. J. Colloid Interface Sci. 2008, 319, 457–461.
10.1021/la800986t CCC: $40.75 2008 American Chemical Society Published on Web 07/12/2008
Nanostructure of a Smectite Gel The same methodology of the TXM technique was adopted in the present experiment. The smectite used in this study was the well-known Namontmorillonite from Wyoming, obtained from the Clay Minerials Society. The original clay sample (SWy-2) has been well described,19 and the two samples were prepared from this original clay. First, the colloidal fraction was separated by centrifugation, and second, all cations in exchangeable positions were ion exchanged with Na+ and Ca2+. Measurements were performed in distilled water. Note that it is difficult to control the Debye length in “water” because there is always some low level, 0.01 mM or less, of background electrolyte (including ions from the self-dissociation of water) that is difficult to quantify or control. The most important difference in the bonding between Ca- and Na-montmorillonite sheets is that in dilute salt solutions the spacing between most of the former is restricted to 0.95 nm, whereas the spacing of the latter is unlimited.20 The dispersions of both studied clay samples for AFM/colloidal probe experiments were prepared from a 0.02 wt % water suspension. A few drops of suspension were placed on a clean silicon wafer substrate to produce a clay-topped surface, followed by drying in a laminar flow cabinet at room temperature overnight. Sample film thickness in both samples spread on a silicon wafer substrate was studied in SEM micrographs and was estimate to be 10-20 nm for flat-lying sheets completely covering the substrate. The imaging and force measurements were conducted using a Nanoscope III AFM (Digital Instruments) in contact and force modes, respectively, utilizing a J scan head. A standard fluid cell with a scan rate of 1 Hz was used for the measurements. The AFM cantilever was triangular, tipless silicon nitride with a spherical colloidal probe (2.5 µm in diameter) purchased from NOVASCAN. The spring constant was a nominal 0.12 N/m. The clay-coated flat substrate surface was displaced in a controlled manner toward and away from the colloidal probe in aqueous solutions. The interacting force between the probe and the clay-coated silicon wafer was obtained from the deflection of the cantilever. The deflection of the cantilever versus the displacement of the flat surface was converted into surface force versus separation by assuming that the zero point of separation was defined as the compliance region where the probe and the flat surface are in contact, and the zero force is determined at large surface separations. For the microstructural cryo-SEM studies of the suspensions, the cryo-transfer method of sample preparation was used to avoid structural rearrangement caused by surface tension during oven or room-temperature drying. The vitrified samples were kept in a frozen state and placed onto the liquid-nitrogen-cooled specimen stage of the Philips XL30 field emission gun scanning electron microscope with an Oxford CT 1500HF cryo-stage. The samples were fractured under vacuum and sublimated, after which they were coated with gold and palladium to a thickness of 3 nm. The samples were then examined in the FESEM using a cathode voltage of 5 kV.
Results Smectite represents a 2:1-type layer silicate with an expandable structure comprising sheets carrying a certain number of excess negative layer charges that are linked by weak van der Waals forces. The layered structure consists of an octahedral alumina sheet sandwiched between two tetrahedral silica sheets. From our microscopy observations (AFM micrograph in Figure 1A and SEM micrograph in Figure 1B), smectite consists of relatively large, flexible sheets with a lateral dimension of ∼800-1000 nm and a thickness of ∼1-10 nm. The platelet assembly consists of sheets stacked on top of each other in a parallel fashion, called a tactoid (stacks of parallel clay platelets at -10 Å separation). Because the bonding between individual sheets is very weak, the sheets easily slide along the top of each (19) Van Olphen, H.; Fripiat J. J. Data Handbook for Clay Materials and Other Non-Metallic Minerals. Pergamon Press: Oxford, U.K., 1979; p 183. (20) Emerson, W. W. Inter-particle bonding. In Soils an Australian Viewpoint; Academic Press: London, 1983; pp 477-498.
Langmuir, Vol. 24, No. 16, 2008 8955
other and are quite flexible. Individual smectite sheets, which are about 1 nm thick, can be seen to slide on top of each other and resemble a house-of-cards shape with the sheets showing undulations. Smectite in a water suspension forms a gel where the tactoid sheets are very flexible individual particles and interact by a combination of edge attraction and basal plane repulsion. These properties can build an expanded and extremely voluminous cellular network composed of chainlike sheet assemblies similar to those presented in the cryo-SEM micrograph (Figure 1C). In such an extended cellular network, flexible smectite sheets encapsulate water within cellular voids with dimensions of up to 0.5-2 µm. This flocked cellular structure can span the entire volume of the clay slurry. In such a case, the suspension is gelled; there is no free settling in the system, and further compacting may proceed slowly by structural rearrangement of the entire network. In a 3-D structure, smectite sheets are not oriented in any particular direction; therefore, their most probable orientation in the volume of suspension is random. In 2-D, TXM micrographs (Figure 2) reveal the gel structure in water. In the micrograph shown in Figure 2A, elongated montmorillonite sheets form a cellular network that is 0.6-1.5 µm in diameter (average 940 nm). In the 2-D TXM micrograph of the montmorillonite gel (Figure 2B), the cellular structure shows much smaller cells dimensions of 300-600 nm. The average distances measured between smectite sheets was around 450 nm. This is less than half of the cellular distance measured in Na-montmorillonite. Direct measurements of forces acting between studied Naand Ca-montmorillonite-coated surfaces were performed using colloidal probe atomic force microscopy (AFM) in water. Results in Figure 3 show force, F, plotted against the distance, D, of separation between Na- and Ca-smectite-coated surfaces. Direct measurements of forces acting between the studied Naand Ca-montmorillonites were performed. Results plotted in Figure 3 showed that (a) there is a long-range repulsion between these two surfaces; (b) the decay of the force is quasi-exponential; and (c) there is no adhesion between these surfaces. The longrange repulsive forces for Na-montmorillonite have been detected from distances of >1000 nm in surface separation whereas for Ca-montmorillonite the repulsive forces were detected from a distance ∼400 nm. Strangely, these numbers correspond to the cellular dimensions observed in the TXM 3-D reconstruction as well as the 2-D TXM micrographs. This fact is important in how the microstructures of Na- and Ca-montmorillonite differ in dilute water gel, which may be due to the difference in Na and Ca hydration. Callaghan and Ottewill21 believed that the electrostatic forces described by the DLVO theory are responsible for the long-range interactions in Namontmorillonite gels. Low and Margheim22 concluded, from the data of Callaghan and Ottewill, that the forces affecting the swelling behavior of montmorillonite platelets are due to the structure of water and not electric double-layer forces. This opinion has been disputed by Bares.23 Data from our AFM measurements clearly show the existence of the long-range, exponentially decaying repulsion, which looks similar to the double-layer interaction. According to Chan, Pashley, and White,24 the maximum force felt is more likely related to the amount of charge on the surface, and the slope of the exponential (21) Callaghan, I. C.; Ottewill, R. H. Faraday Discuss. Chem. Soc. 1974, 57, 110–118. (22) Low, P. F.; Margheim, J. F. Soil Sci. Soc. Am. J. 1979, 43, 473–481. (23) Barnes, C. J. Soil Sci. Sac. Arner. J. 1980, 44, 658–659. (24) Chan, D. Y. C.; Pashley, R. M.; White, L. R. J. Colloid Interface Sci. 1980, 77, 283.
8956 Langmuir, Vol. 24, No. 16, 2008
Z˙bik et al.
Figure 1. Micrographs of the montmorillonite from the Na-Wyoming SWy-2 sample. (A) AFM micrographs showing large, thin, flexible sheets assembled from the tail of the cards. (B) SEM micrograph showing a dried-out thin layer of montmorillonite where the sheets covering each other and shrinking wrinkles are visible. (C) Cryo-SEM micrograph of the Na-montmorillonite Swy-2 sample showing a partially freeze-dried montmorillonite gel where most of the surface ice was sublimed, leaving a cellular, highly porous gel microstructure.
Figure 2. TXM micrographs of a 5 wt % montmorillonite colloidal gel cellular structure in water. (A) Na saturated at an exchangeable position, with larger cells and thinner walls. (B) Ca saturated at an exchangeable position, with smaller cells and thicker walls.
decay at large separation is related more to the Debye length. If the observed phenomenon corresponds to the Debye length, then it should have the same distance in water. Observed differences may be explained by differences in the hydratation of the exchangeable counterions, allowing these ions to then set up an electric double layer. It is also possible that these longdistance forces, similar to electrostatic repulsion, have a steric origin and reflect the flexibility of smectite flakes. In effect, the hindrance of other platelets against the interaction between two concerned platelets would be unavoidable. If this mechanism is
Figure 3. Force-separation curves for the interaction between Swy-2 on a silicon wafer on approach: (---) Na+ exchangeable cations form; (-) Ca2+ exchangeable cations form.
responsible for our force measurement results, then it reflects the squeezing of the cellular monolayer of the smectite gelled structure, which has dimensions that are consistent with microscopy observations. Significant differences between smectite with Na+- and Ca2+exchangeable cations may be due to the differences in the rigidity
Nanostructure of a Smectite Gel
Langmuir, Vol. 24, No. 16, 2008 8957
Figure 4. (A) Three-dimensional reconstruction of the montmorillonite gel Na sorption complex and (B) Ca sorption complex sample as seen within the aqueous solution.
Figure 5. Schematic model of a Na-montmorillonite colloidal gel, where electrostatic repulsion pushing individual smectite tactoid sheets outward results in an expanded structure.
of the smectite flakes. This was shown by Amorim et al.,25 who used XRD analyses of the influence of sodium and calcium on the swelling of smectite flakes, which determine the flakes’ thickness and resulting rigidity. Calcium produced strong control over clay swelling, and d spacing was maintained at around 2 nm, even at very low concentrations in solution, whereas sodium caused a d spacing >5 nm and significant peak broadening, proving singular platelet delamination. This finding is consistent with data presented by Emerson.20 In our case, it may suggest that thicker and more rigid flakes made with Ca2+-exchangeable cations will build much smaller cellular structures than the thinner, more flexible smectite flakes made with Na+-exchangeable cations. However, the nature of the long-distance repulsive force that is responsible for the enormous swelling between flakes to build the cellular structure remains obscure. Because TXM based on the synchrotron photon source is relatively new and is in rapid development, in this work we report the first attempt to study smectite gel structure in an aqueous environment. Although the images obtained from this technique may not be very impressive compared with those from other well-established electron microscopy techniques, they enable us to observe clay aggregates within water, which has not previously been possible. A 3-D space reconstruction is shown in Figure (25) Amorim, C. L. G.; Lopes, R. T.; Barroso, R. C.; Queiroz, J. C.; Alves, D. B.; Perez, C. A.; Schelin, H. R. Nucl. Instrum. Methods Phys. Res. A 2007, 580, 768–770.
4. It was obtained from 2-D images of particles as shown in Figure 2, observed from angles +70 to -70°. (The gelled suspension is able to be observed from different angles.) Such a reconstruction reveals, for the first time, the cellular orientation of associated mineral sheets within a water-based electrolyte. The reconstruction also allowed for the observation of the difference in sheet thickness between single Na and Camontorillonite flakes. This was done from different angles. Individual colloidal-sized particles are not clearly visible in this image; however, distinctive spongy, cellular structure is visible when carefully observing this reconstruction (Figure 4) upon rotating this image, which cannot be demonstrated here in 2-D prints. Shown here is a cellular network of highly bendable elastic flakes with diameter in excess of the 1 µm continuous structural pattern of the montmorillonite gel with the Na sorption complex (Figure 4A) and is represented in the simplified graphical image in Figure 5. The thickness of the sheet-assembled cells walls is below the resolution of TXM but is visible when rotating this image. In Figure 4B, a much denser gel assembled from 20-50nm-thick flakes builds a cellular network with a cell diameter of around 0.5 µm. Such flakes are not banded to the degree observed in the Na-montmorillonite gel. Disputes about the nature of the long-range repulsion have remained. However, the presence of such forces is supported by our TXM and SEM micrographs. The observed swollen cellular structure (cellular tactoid) has void diameters consistent in value
Z˙bik et al.
8958 Langmuir, Vol. 24, No. 16, 2008
with the distances of the AFM-measured repulsion in Na and Ca montmorillonite samples. The cellular tactoid structure is summarized in Figure 5, where the combination of the repulsion of negatively charged basal surfaces and the attraction of positively charged edges toward negatively charged basal surfaces creates the swollen cellular tactoid form. This structure is unique to smectites and especially to montmorillonite gel with the Na sorption complex as a result of the thinner sheets’ flexibility. The deformed structure may correspond to the well-known Rosenquist7 card house structure, described for more rigid, plateletshaped minerals such as kaolinite.
Conclusions Our TXM investigations reveal swollen cellular structures (cellular tactoids) in an aqueous montmorillonite gel with void diameters on the micrometer scale. Most of the previous studies on the swelling of smectites concentrated on the hydratation of the mineral interlayer that expands a few nanometres, forming
a clay tactoid or in some cases resulting in total delamination. In our findings, we can conclude, from consistent TXM observations and AFM force measurements, that the initial hydratation may trigger complex electrostatic and steric (sheet elasticity) repulsion that pushes swollen smectite flakes apart significantly long distances that exceed 1 µm. This forms thickly gelled cellular tactoids in Na-montmorillonite with cells up to 1 to 2 µm in diameter or coagulated aggregates in Camontmorillonite with thicker cells of around 0.4 µm in diameter. The nature of long-distance repulsive forces, which look similar to electrostatic repulsion and are responsible for enormous swelling, remains uncertain; however, elastic interactions cannot be ruled out. Acknowledgment. This work was supported by the Australian Synchrotron Research Program (ASRP) and partially funded by an ARC linkage grant. LA800986T
