A STUDY OF THE SURFACE CHEMISTRY OF CHLORITE

A thesis submitted in conformity with the requirements for the degree of Master of Science Graduate Department of Geology University of Toronto

© Copyright by Gordon Ante Vrdoljak 1994

A Study of the Surface Chemistry of Chlorite

Master of Science Thesis

1994

by Gordon Ante Vrdoljak

Graduate Department of Geology, in the University of Toronto

Abstract

The {001} cleavage surface of clinochlore in air, water, and oil was examined at near-atomic resolution with an Atomic Force Microscope (AFM). The AFM resolves the surface structure of both the talc-like and brucite-like layers composing chlorite. The imaged surface agrees with the expected bulk structure with mean unit cell distances of 5.31±.31 and 9.74±.80 Å. In some cases a slight structural relaxation of the talc-like layer occurs with shortening and elongation of both a and b unit cell dimensions. This probably results from a "concertina-like" readjustment of the surface after cleavage to compensate for the lack of an overlying brucite-like layer.

Exposing the {001} cleavage surface of clinochlore to CsCl and Cs2CO3 aqueous solutions resulted in specific ion adsorption of cesium onto the surface. This was observed in situ, using AFM. Cesium ions adsorb in an island pattern with a c(1x1) R60º unit mesh at negatively charged sites arising from the substitution of Al3+ for Si4+. The adsorption is temporal, with removal of the adsorbate by repeated scanning with the AFM tip. This temporality implies that we have imaged the highly exchangeable adsorption sites of clinochlore. X-ray Photoelectron Spectroscopy (XPS) of the surface at various stages confirm the adsorption of cesium, with a peak shift indicative of a Cs-OH bond. Computer simulation of the binding site geometry for cesium adsorbingto the clinochlore surface supports the AFM findings.

Acknowledgments:

I would also like to acknowledge the support of my family, especially my sister, Eva, for her advice and my father for his guidance. Thank you to my friends; Richard, Greg, and Jeremy for being there when I needed help or just a break from the work. A special thanks to Luvi for her companionship.

Permission for inclusion of two papers, one published in the journal, American Mineralogist, 79, pages 107-112. and one currently in press in the journal, Colloids and Surfaces (A):Physicochemical and Engineering Aspects, is kindly given from the editors.

Table of Contents:

List of Plates:

Figure; page
[2-1] Diagram of AFM operation. ; 12
[2-2] AFM operational settings. ; 14
[2-3] Height and deflection image comparison. ; 16
[3-1] The structure of chlorite. ; 27
[3-2] AFM images of the talc-like layer of clinochlore. ;29
[3-3] Graphs of measured unit cell distances. ; 31
[3-4] AFM images of the brucite-like layer of clinochlore. ;33
[4-1] AFM images of clinochlore in water. ; 44
[4-2] AFM images of clinochlore in Cs2CO3 solution. ; 46
[4-3] AFM images of clinochlore in CsCl solution. ; 48
[4-4] a)XPS of clinochlore after water exposure.
b)XPS of clinochlore after exposure to a Cs2CO3 solution. ;50
[4-4] c)Low resolution XPS survey of cesium peak region.
d)XPS of clinochlore after exposure to a CsCl solution. ; 52
[4-5] Computer simulation of cesium binding geometry. ;54

I. Introduction and Overview:

Overview:

Many geochemical phenomena such as mineral dissolution, precipitation and growth, cation adsorption and desorption, and trace element cycling are highly dependent upon the physical and chemical behavior of mineral surfaces (Hochella and White, 1990). The structure and topology of the surface itself are particularly important as they directly affect reaction rates and mechanisms.

This thesis presents the results of an investigation into the surface structure of chlorite using Atomic Force Microscopy (AFM). Chlorite was chosen because the flat, easily cleaved {001} surface facilitates study with AFM and it has two chemically different surfaces produced by cleaving along this plane. For our work, a clinochlore-magnesium rich chlorite was used. In addition, the adsorption of cesium on clinochlore was also investigated. Cesium was utilized because of its high affinity for the negatively charged silicate surface. Cesium forms strong inner-sphere complexes to such surfaces, and is not likely to be removed by the AFM during image acquisition.

The present chapter provides an outline of surface and interface chemistry relevant to this study. For a more complete review, see Hochella and White (1990) or Myers (1991). The explicit experimental procedures used in obtaining AFM images are given in chapter two. Chapter three presents the result of an AFM study of clinochlore, and its surface modification in air, water, and oil (published in the journal, American Mineralogist, 79, pages 107-112). Chapter four describes the results of the investigation into clinochlore reactivity with cesium adsorbents (currently in press in the journal, Colloids and Surfaces (A):Physicochemical and Engineering Aspects). Each section begins with an introduction, laboratory methods, results, and ends with a discussion of the results. An overview of the full spectrum of results is presented in chapter five.

Surfaces and interfaces:

The surface is the site where a reaction occurs between two or more media (liquid/solid, solid/solid, gas/solid, or a gas/liquid interface). To date, most work has been done on the reactivity of bulk crystals or molecules. Such studies are important, but they cannot account for the control surfaces have on many chemical processes in natural systems. For example, water acting slowly on a rock will preferentially dissolve and weather minerals. As the water will first encounter the outer-most layers o f a crystal, it is the composition and structure of the mineral surface that determines if, and how, a reaction takes place.

Various types of interfaces, such as the air-liquid, liquid-liquid, air-solid, or liquid solid, are important in surface chemical reactions. The solid-liquid, or mineral-water interface is of most concern in geology. The Earth's surface consists of a n inestimable quantity of solid surfaces of geological and biological origin in contact with water. Thus, the mineral-water interface is the mediator, controlling reactions occurring on land, in freshwater and the oceans. Consequently, it is important t o develop an understanding of the physico-chemical behaviour of mineral surfaces and the mineral-water interface.

The location of atoms, bond lengths/angles, and defects make up the topography of the surface and control chemical reactivity. The chemical environment of an atom at the surface of a material is different to that experienced by a similar atom within a three dimensional crystal. As a consequence of this difference, surfaces often have different 'relaxed' structures. These 'relaxed' structures have some similarity to the bulk structure, but the positions of atoms are altered slightly to relieve stress . As well as these changes are the hills and valleys of rough surfaces, grain boundaries, twinned crystals, screw dislocations, vacancies, steps, and kinks. Usually, the reactivity of the mineral surface increases with the number of higher energy sites such as dislocations, steps, or kinks. These higher energy sites, may promote adsorption of a species, initiating a chemical reaction. Also, as roughness at the molecular or atomic scale, increases on the surface of the mineral, reactivity increases. The effects of microtopography (hills, valleys, etc., created by imperfect cleavage or fracturing) are too complex for complete analysis and may increase or decrease reactivity. A detailed understanding of the complex surface structure is essential in or der to comprehend surface chemistry .

Surface structures are often poorly characterized and the crystal edges, steps, kink sites, and defects, controlling reactivity, can be hard to locate. In addition, reactions at interfaces are difficult to study exhaustively with conventional experime ntal techniques and surface sensitive techniques (of which there is a limited but growing number) need to be employed. Finally, it is difficult to study a reaction which occurs at the surface on an atomic scale. A new generation of laboratory equipment, such as the scanning tunneling microscope (STM) and the atomic force microscope (AFM), with their high resolution capabilityat the atomic scale, are providing new insight into this problem.

Studies of the mineral-water interface are difficult and explicit information about the molecular nature of the reaction is often lacking. In our investigations, we chose to study the mineral-water interface of chlorite. Chlorite is a common, althoug h minor, component in soils and has a reactive surface similar to a variety of other clay mineral structures. Thus, results from the study of chlorite can be applied to other minerals. In addition, chlorite has a high degree of isomorphic substitution, generating a charge imbalance at the surface which often binds ions to specific reactive sites. This binding controls the elemental speciation of soils.

We utilized an atomic force microscope to fully characterize the surface structure of clinochlore (a trioctahedral, magnesium rich chlorite) in air and water. To study the negatively charge talc-like layer of clinochlore, the reaction of clinochlore a nd cesium ions at the mineral-water interface was monitored. X-ray photoelectron spectroscopy was used to determine the chemical composition of the clinochlore surface at various stages.

II. Experimental:

Microscopy:

Scanning Probe Microscopy (SPM) was developed in the mid '80's out of the desire to overcome the resolution limits of traditional microscopes which use lenses to focus light or other forms of radiation. All forms of radiation become diffracted or scat tered as they encounter objects of similar dimension to their wavelength. Thus each form of radiation has a limit of resolution defined by the wavelength of the radiation. To overcome the limit created by diffraction, a specimen is visualized with a sca nning near-field microscope which does not use lenses. In this method of microscopy , a probe is used to investigate a minute portion of the surface, record it, and then move onto a neighboring area of the sample. By analyzing a sample stepwise in this fashion, the resolution limit set by diffraction is avoided and the size of the probe is the limiting factor.

Various probes have been used to measure properties of the surface, but the measurement of quantum-mechanical tunneling of electrons between the probe and sample has received the greatest interest. Detection of a minute tunneling current of electrons flowing between the probe and a sample were first used by Binnig et al (1982) in the first Scanning Tunneling Microscope (STM). By generating a charge bias between the closely spaced sample and tip, electrons can quantum mechanically tunnel into or out o f the tip (depending on the bias). Tunneling current is dependent upon distance between the sample and tip. A topograph of the sample surface can be recorded by measuring differences in tunneling current over the surface while scanning with the tip in a n x-y raster pattern. The images of the surfaces of conducting and semiconducting materials can be recorded at atomic resolution with STM. However, there were difficulties in analyzing insulator materials, most minerals belong to this class, because of the requirement of a controllable charge on the surface. Developments made in using a physical probe to contact the surface instead of measuring tunneling current brought forth the Atomic Force Microscope (Binnig et al. 1986).

In the AFM a sharp tip is brought into contact with a sample and scanned in a x-y raster pattern. The probe is normally mounted underneath a flexible cantilever of known spring constant. As the probe rides over the surface, variations in height of th e probe are easily measured as flexing of the cantilever. A laser beam, bounced off of the cantilever and into a split photodiode produces a quantification of the flexing of the cantilever as variations in the photodiode signal. This gives a 3D profile map with the third dimension being height or tip deflection and the other dimensions the x-y postion of the probe (figure [2-1]). Piezoelectric controls are used to control sample position and ideally, an atomically sharp probe (actual radius of the prob e apex is at best 20 nm (Sheiko et al. 1993)) is used to provide image resolution of 0.1 Å.

The theory of how the AFM works, is understandable physically with large low resolution scans, but at high or atomic resolution (<100 nm), the physical theory breaks down. The current estimation for the average radius of curvature of most AFM tips is < 300 Å (Albrecht et al., 1990) and the theory of how such a large probe can resolve atomic structure at less than 1Å scale is not complete. Microtips may be present on the end of the tip, as local sharp protrusions providing high resolution imaging. Cu rrent theories on the forces involved are based upon Born internuclear repulsion and Van der Waals forces between atoms of the tip and the surface being scanned. There are, however, more complex forces such as magnetic and quantum mechanical forces gener ated via the exclusion principle (Hans Christian von Baeyer 1992). Not until all of these effects on the force are accounted for will there be a theory which can explain the apparent atomic resolution capability of the AFM. The AFM does provide real ato mic resolution of surfaces as evidenced by the work of numerous laboratories which have been able to correlate AFM images with expected atomic structure obtained via other techniques (STM, X-ray or electron diffraction, or Surface Extended X-ray Absorptio n Fine Structure spectroscopy (SEXAFS)). Some of the materials studied at atomic or molecular resolution include; graphite (Binnig et al., 1986), albite (Drake and Hellmann, 1991), calcite (Rachlin et al., 1992 and Ohnesorge and Binnig, 1993), and a vari ety of phyllosilicate minerals (illite-smectite (Lindgreen et al., 1991), lizardite (Wicks et al., 1992)).

The procedure used to establish the operating conditions, calibration and preparation of the AFM for imaging, were based on the procedures given in the operating manual, "Digital Instruments NanoScope® III Scanning Probe Microscope Control System User' s Manual, Version 2.1 June 15, 1992." Silicon nitride integrated pyramidal tips were purchased directly from the manufacturer and stored in covered petrie dishes to protect from dust or damage. The cantilevers used for imaging clinochlore were 200 µm lo ng with narrow or wide legs each having spring constants of 0.06 N/m and 0.12 N/m respectively. Cantilevers with the lowest possible spring constants were used to ensure as little damage as possible was done to the soft clinochlore surface during scannin g. Alignment of the laser was performed using the magnifier method described in the user's manual and gains were typically set at +3 volts for the A+B or total signal to the split photo diode detector and -2 volts for the A-B or differential signal to th e photo diode. Before engaging the tip to the surface, the operating controls of the microscope were typically set as shown in figure [2-2]. Proper engagement was obtained when the sample stage was as level as possible. After testing for proper engagem ent with the scanner on a desktop, ie.: the tip tracks the surface smoothly, the tip was withdrawn and the scanner protected from building noise by placing it on a vibration isolation platform (a concrete slab suspended from a tripod by three bungee cord s).

The AFM has two modes of operation, the deflection and height modes (figure [2-3]). In deflection mode the height of the piezo is kept constant and differences in surface topography are measured as fluctuations in the tip or cantilever. Height inform ation obtained with this method is not usually accurate because the careful calibration that is necessary is not available at all height levels. In the constant height mode of operation, the tip deflection is kept constant by a feedback loop and differen ces in height are measured by differences in piezo height. Height information obtained with this method is more accurate than in the deflection mode because an error-correcting feedback loop provides the data. Both modes are accurate in the x-y plane to 0.1 Å through careful calibration with a muscovite standard.

Samples were imaged in air to characterize their air-stable surface structure, then in water, in parrafin oil, and finally in various aqueous solutions of known chemical composition. In this way, an analysis of any change in surface composition and st ructure responding to the new chemical environment could be made. Consistent reliable imaging was obtainable by using the height mode with feedback gains as low as possible so that no feedback signal effects were visible in the images. For atomic resolu tion images, a large 1 m image of the surface would be obtained and the instrument slowly zoomed in to smaller and smaller scales until at 20 nm x 20 nm, rough structural images should be visible. The scan angle, setpoint (or operating force), feedback g ains, and scan rate were optimized to obtain the clearest images possible without artifacts. The instrument settings for optimal imaging of clinochlore were: 256 data samples per scan line; 30.5 Hertz; Z-limit of 220 Volts, integral and proportional gai n of approximately 8; and a setpoint of -1 to -2 volts. No real-time filters are applied during data acquisition except for a plane fitting routine which removes image bow derived from the scanner moving out of the plane of the sample during scanning. S can angle, scan rate, and force were varied for each series of images captured to verify the validity of the image and reduce the chance of artifacts being analyzed.

After the AFM images were recorded, a certain amount of digital filtering was performed to remove noise and clarify the periodic structure present in the images. Filtering of data is a contentious issue with many workers in the field feeling that it a mounts to editing data, thus creating the results you want for an experiment (Andrew Gratz, personal communication). Therefore, for all of our AFM images, we perform as little filtering as possible and only periodic information is attenuated. All period ic features present in raw or unfiltered images are present in the filtered images and no new features are created. No periodic data is removed so that only noise and image artifacts are edited out. The typical sequence of digital filtering applied is a uto flattening, planefit, lowpass filtering, and 2-dimensional fast-Fourier transform (2DFFT). Auto flattening removes image bow by calculating a least squares fitted second order polynomial for each scan line and subtracts it from the scan line. Planef it eliminates the effect of sample tilt by calculating a best fit, second order polynomial plane and subtracting it from the image. Usually it is applied once in both the x and y directions. Lowpass filtering removes high frequency noise by replacing ea ch data point in the image with a weighted average of the 3x3 cell of points surrounding and including the point. A 2DFFT is used to select the periodic information from an AFM image and suppress non-periodic information. Fourier filtering is a powerful technique which, if judiciously and carefully used, can greatly improve images. Incorrect fourier filtering an image can alter the observed surface structure by selective removal of spots at reciprocal lattice sites in the fourier transformed image. Fa st fourier filtering is an accepted technique in Transmission Electron Microscopy (TEM) and other areas of science and there is no reason that its judicious use in AFM should be questioned. After the image is processed, it is analyzed and any statements made about the surface structure such as distances, periodic structure, etc., are verified by comparison with the unfiltered image.

The acquisition of AFM images at atomic or molecular resolution is not easy and usually requires hours of operation and adjustment of controls and settings of the instrument for one particular sample. A variety of effects may be responsible for the di fficulty in obtaining AFM images. One of the largest problems with the AFM is its sensitivity. Because it is designed to sense minute fluctuations in a cantilever tip, it also senses building noise. The instrument must be carefully isolated from noise and vibration usually utilizing a vibration dampening system. Bad tips used in the AFM are also a problem as the manufacturing process and/or shipping may damage them slightly. Also, tip preparation is difficult and a slight error in handling can, and o ften does, damage the tip. Contamination of surfaces in air is prevalent for a variety of materials and often interferes with successful imaging of surface structure at molecular resolution. Furthermore, surfaces which are not 'flat ' pose problems wit h tip engagement and tracking over the surface when the piezo range is exceeded. Electrical noise from machinery or local subway transit system operations may also affect the instrument as we often obtain better resolution (a higher signal to noise ratio ) with the AFM at night. Finally, the worst difficulty we have encountered is the problem of static electricity during the winter. Static generated on materials induces a charge on the AFM tip through static field interaction and interferes with the ope ration of the instrument. The laser alignment settings drift as charging pulls the tip toward the sample surface. We have overcome this problem by the careful use of static control measures (anti-static mats, grounding straps, and ion sources) in the wi nter. Images were often collected at night or during low auto traffic flow to avoid noise created by cars, buses, and subways.

The AFM lacks wide acceptance in the scientific community as it is a relatively new instrument and its operation is not well understood. Images can be readily obtained, but are often subject to a variety of effects distorting images. Common effects a rise because of drift of the piezo, noise, calibration problems with the piezo, the shape of the AFM tip creating artifacts, tip contamination, vibration, thermal fluctuations, and deformation of the surface by the tip. However, these effects are no diff erent than problems encountered with a typical TEM, such as; astigmatism correction, charging problems, aperture alignment, operating current/voltage. Similar to the operator having a good knowledge of effects present in TEM images and their sources and correction, the AFM operator must have a good understanding of image artifacts and corrections to obtain good images. Filtering is often performed on images to get rid of excessive noise and to remove instrument effects (such as image bow due to sample t ilt). With all of these artifacts generated by improper use of the AFM and the digital filtering operations, many suspect the validity of AFM images. I have reported here the step, by step, procedure used for the acquisition of AFM images in this thesis with the Digital Instruments Nanoscope III AFM. In this fashion, I hope to relate that carefully obtained AFM images are no more difficult to obtain than a good TEM image by a trained operator. Additionally, AFM images are valid provided careful calibr ation of the instrument and judicious filtering is performed.

Although the operation of the AFM is difficult, time consuming, and almost impossible for some materials, its potential for the imaging of surface processes is promising and worth the effort. Atomic resolution imaging of a calcite atomic step ledge by Ohnesorge and Binnig (1993), shows the appearance of a 'defect' where atoms are missing. Recording surface atomic structure and defects demonstrates that the AFM images are real atomic representations of surfaces. Through careful technique and patient operation of the instrument, we have been successful at imaging a wide variety of materials. Advances in design and operation of the instrument and simplification of its setup will improve reliability so that the AFM will become a routine laboratory tool for surface analysis.
Figure [2-1] Diagram of the AFM optical sensing system. (From Digital Instruments Nanoscope User's Manual.)
Figure [2-2] Image of clinochlore {001} surface structure with the AFM. Both the constant height and deflection mode of obtaining data is shown. Molecular scale corrugations can be seen in both.
Figure [2-3] Settings on the AFM controls prior to engagement of the probe with the surface.

III. Preamble: Structural Relaxation of the Chlorite Surface Imaged by the Atomic Force Microscope

This chapter describes the first AFM study of the mineral clinochlore. Clinochlore was used as the presence of distinct layers (talc-like and brucite-like) in its structure serves as a model to which other clay minerals can be compared. This material was studied in air, under water, and in paraffin oil; each of which provide a different chemical environment for the clinochlore surface. The structure of the {001} surface of clinochlore was shown through careful measurements to undergo slight expansio n with each media and slight relaxation compared to the bulk structure. This study provides a basis for the further study of the surface chemistry of layer silicates.

Material for this chapter has been published in the journal, American Mineralogist (volume 79, pages 107-112). This was the first pursuit in our study of the surface chemistry of layered silicates. Understanding the surface structure at the atomic le vel in different environments aided us later in our study of specific ion adsorption on the surface of the talc-like layer of chlorite.

Structural Relaxation of the Chlorite Surface Imaged by the Atomic Force Microscope

GORDON A. VRDOLJAK, GRANT S. HENDERSON, J. JEFFREY FAWCETT

Department of Geology
22 Russell St., University of Toronto
Toronto, Ontario, Canada, M5S 3B1

FREDERICK J. WICKS

Abstract

Introduction

The development of the atomic force microscope (AFM) has enabled the structure of insulator surfaces such as silicate minerals to be readily imaged at very high resolution. Images of the surfaces of calcite (Rachlin et al., 1992); zeolite (Weisenhorn e t al., 1990); albite (Drake and Hellmann, 1991); illite-smectite (Lindgreen et al., 1991); and lizardite (Wicks et al., 1992) have been obtained at near atomic or molecular resolution. Phyllosilicate minerals, composed of flat sheets of three corner shar ing SiO4 tetrahedra, are particularly suitable for study by AFM because of their perfect cleavage along the {001} plane. Imaging of the {001} surface of the 2:1 sheet silicate muscovite shows molecular resolution of the individual tetrahedra for the tetr ahedral sheet (Drake et al., 1989).

Lizardite is a 1:1 phyllosilicate consisting of a single tetrahedral and octahedral sheet as its basic repeat unit. Atomic resolution of the hydroxyl groups and the magnesium atoms has been obtained in the octahedral sheet and molecular resolution of the SiO4 tetrahedra in the tetrahedral sheet (Wicks et al., 1992). The 1:1 sheet silicates and the 2:1 sheet silicates with an interlayer hydroxide sheet exhibit three structurally different {001} surfaces (octahedral, tetrahedral and interlayer hydroxid e sheets), and it should be possible to image both the tetrahedral and octahedral or interlayer hydroxide surfaces within the same sample.

Chlorite group minerals are 2:1 phyllosilicates with an interlayer brucite--like sheet that incorporates medium sized cations such as Mg, Al and Fe in the octahedral sites. The 2:1 talc-like layer consists of two tetrahedral sheets of SiO4 tetrahedra with an octahedral sheet sandwiched between the two tetrahedral sheets. In chlorite this talc-like layer has a general composition of (R2+, R3+)3(Si4-xAlx)O10(OH)2, where R= Mg, Fe or Al. The excess negative charge is neutralized by the positively charg ed octahedral brucite-like interlayer sheet of composition (R2+, R3+)3(OH)6 (R= Mg/Fe). The chlorite structure is presented in Figure [3-1] (Bailey, 1988). In this study we have investigated a clinochlore IIb chlorite with composition (Mg4.4Fe0.6Al)[(Si 2.9Al1.1)]O10(OH)8 obtained from northern Ontario.

Experimental Details

Samples were cleaved along {001} using a scalpel and both halves were examined in air using a Digital Instruments Nanoscope III AFM with a vibration isolation platform. Contact forces were on the order of <20--150 nN and were adjusted to optimize ima ge resolution. A wide range of scan rates and scan directions were used. The AFM scanner was calibrated against mica and gold ruling standards and thermal drift was minimized by allowing the AFM to equilibrate with the surrounding air temperature. Imag es identified as having excessive drift were discarded.

It was found that Au coated Si3N4 cantilevers with 200 µm wide legs were best for imaging the tetrahedral sheet while the 100 µm leg versions were more effective for the brucite--like interlayer sheet. All structural features discussed in the text wer e visible in the raw (=unfiltered) images but standard digital filtering techniques were performed to improve the quality of the images and every effort was made not to introduce artifacts into the processed image. This was checked by careful monitoring of interatomic distances and atom positions before and after application of the filtering routines. Observed images were compared with talc- and brucite-like structures calculated from a II-b chlorite (corrected for our composition) of Bailey, (1975).

To facilitate comparison between the talc- and brucite-like layers, images in figures [3-2] and [3-4] have been filtered in the same manner. Filtering routines used are flatten, low pass filter and 2-D fast fourier filtering (2DFFT). In some figures the intensity distribution across the image appears to vary. This variation is related to the color contrast/offset, and the 2DFFT parameters employed during the filtering. Different filtering and color contrast/offset parameters change the intensity dis tributions but do not alter the instrumentally induced ditrigonal symmetry nor the postulated Mg atom positions (see later).

The flatten routine subtracts the average value of height along each scan line from each point in the line and is applied to remove image bow. Lowpass filtering replaces each pixel with a weighted average of the 3x3 cell of pixels surrounding and inclu ding the original pixel and removes spot noise; The 2DFFT calculates periodicities observed in the image (it is somewhat comparable to a LEED image of the surface). These periodicities may then be selected (or rejected) for incorporation into the inverse transform which produces an image significantly reduced in noise. However, one must be careful that information is not lost and in general the 2DFFT is used to highlight features that are already observable in the unfiltered images.

Results and Discussion

Raw and filtered images of the talc- and brucite-like structures are shown in Fig. [3-2] and Fig. [3-4], respectively. The {001} surface of the 2:1 talc-like layer is seen to consist of an array of hexagonal rings of SiO4 tetrahedra (Fig. [3-2c]) cons istent with the known crystal structure. There is very good agreement between the observed and calculated structure of the silicate sheet of the talc-like layer and the mean observed unit cell distances (a=5.31±.31Å, b=9.74±.80Å) agree within 1 and 5%, r espectively, of the calculated values. This close correspondence between the structure observed in the images and the cell structure calculated from known atom positions determined by X-ray structure analysis is strong evidence that we are obtaining true atomic or near-atomic resolution (cf., Ohnesorge and Binnig, 1993).

Rotation of the tetrahedra within the {001} plane of the siloxane sheet (see Figure [3-1b]) reduces the hexagonal symmetry of the SiO4 rings to ditrigonal (Brown and Bailey, 1963). This ditrigonal symmetry is observed in Fig. [3-2d] as distortions of the hexagonal SiO4 rings with each ring appearing to ``point'' in a defined direction (see Figure [3-2d]). The magnitude of the in-plane rotation could not be determined as the images do not resolve the bridging oxygen between adjacent SiO4 tetrahedra.

However, the tetrahedra around each hexagonal ring also exhibit an ``apparent'' rotation of up to 18º (mean =6.96±4.38) out of the a-b plane. This is seen in the images as alternating highs and lows in image intensity for each tetrahedra within indivi dual 6-membered SiO4 rings. This ``apparent'' rotation further emphasizes the real ditrigonal symmetry but is purely an instrumental artifact. It is observed in other sheet structures such as, graphite, mica and 1:1 layer silicates (Gould et al., 1989), and theoretical calculations of tip--sample interactions indicate that it probably results from an asymmetric (2 atom) cantilever tip (Gould et al., 1989).

In an attempt to quantify surface relaxation effects, and to determine if there is a systematic spatial relationship to any relaxation, ~700 measurements of a and b distances were performed on a single image of the talc--like layer. It was found that the b unit cell distances tend to vary (variance=0.632) more than a (variance=0.1). In addition, both a and b dimensions essentially cluster around specific values (see Fig. [3-3]). Greater than 99% of the a and b measurements fall around the following values; a = 5.47, 5.86. 6.26 Å, and b= 8.98, 9.37, 9.77, 10.16, Å. These values indicate that the cell dimensions undergo a systematic change of ~±0.4 Å with both a and b varying equally. This is expected if the ditrigonal symmetry of the surface is m aintained; if the surface symmetry were not ditrigonal then the variation in a and b would be non--equivalent.

The variation in cell dimensions could result from thermal drift of the instrument, tip induced deformation of the surface, an artifact of the algebraic correction for non--linearity of the piezo and/or from calibration of the instrument to a standard with similar dimensions (in this case muscovite), or from relaxation of the surface. Thermal drift of the instrument and tip induced surface deformation seem unlikely given that only slight differences are observed for a and b measurements even when scan speed and direction, contact force, cantilever type, sample, and imaging medium are varied. Images of the same talc-like layer in air, oil and water exhibit essentially the same behaviour in a and b.

Figure [3-3C] is a plot of measured unit cell distances taken from a number of images of the talc--like layer in air, oil and water. It is clear that while the variation in both a and b is somewhat greater for these measurements relative to those from the single image, the a and b dimensions do exhibit similar behaviour to that of the single image in air. Further, linear regression of the data indicates that both cell dimensions appear to increase slightly on going from air to water. This slight inc rease of a and b in water presumably occurs as the talc-like layer is re-hydrated in solution, water molecules entering the siloxane ring and expanding the structure.

We believe that the observed variation in a and b is probably due to relaxation of the surface. However, we cannot rule out the possibility of an instrumental artifact produced by the piezo non-linearity correction or instrument calibration. But, a ny instrumental artifact should also generate similar behaviour in images of the brucite-like layer, but no such variation in the a cell dimension is observed (see below). If the variation in cell parameters is real then the surface relaxation is presum ably a consequence of surface decompression after cleaving and readjustment of the negatively charged talc-like layer to compensate for the lack of overlying charge balancing brucite-like interlayer sheet. The mechanism by which this occurs may be a "co ncertina-like'' compression of the surface producing both shortening and elongation of cell dimensions.

The brucite-like interlayer sheet (Fig. [3-4]) has trigonal symmetry as expected. The hydroxyl sheet is readily seen (Figs. 4c and 4d) and the measured mean a unit cell distance (shown in Fig. [3-4d]) is within 2% (5.40±.18Å) of the calculated a value of 5.31 Å. No corresponding rotation out of the a-b plane was observed at low contact forces but rotation was observed when higher contact forces were used. The measured a distances, although variable, are random and do not exhibit a systematic variati on similar to that observed for the talc-like layer. Comparison of the imaged structure with a calculated brucite-like interlayer sheet shows a very close correspondence between the OH atoms of the calculated structure (large spheres) and the OH position s in the image (Fig. [3-4c]).

There is also good correspondence between the calculated Mg positions (small open spheres) and regions of the image where we believe the Mg cations of the Mg(OH)6 groups can be seen in the centre of a triangle of 3 hydroxyls (cf., Fig. [3-1c]). This is most clearly seen in Fig. [3-4d]. Figure [3-4e] shows a portion of Fig. [3-4d] through which we have drawn a cross section. The OH positions are represented by topographic highs while the Mg atoms appear as lower secondary highs. The correlation bet ween these secondary highs and the Mg positions is confirmed when the image is overlain by a calculated brucite-like sheet and the observed and calculated Mg positions are compared (see Fig. [3-4c]). The very close correspondence between the calculated b rucite--like structure and the regions which we believe to be Mg atoms, makes it very unlikely that the features in the image are due to anything other than Mg. It would be truely fortuitous if these features were instrumental artifacts that always happe ned to occur at positions in the lattice structure where one would expect the Mg atoms to occur.

Much of the bonding between layers within phyllosilicate minerals is due to weak Van der Waals forces and hydrogen bonding. Scanning the surface of these materials with high contact forces strips away weaker bonded surface layers. Thus it is possible to progressively image the upper tetrahedral, octahedral, and bottom tetrahedral sheets of the 2:1 talc-like layer. Images obtained in this manner generally agreed with the expected bulk structure but tended to be damaged by the cantilever. However, the possibility exists that the AFM may be used to ``depth profile'' the structure of minerals and to directly determine the polytype.

It is clear that it is possible to obtain near atomic resolution images of both the talc-- and brucite-like layers in 2:1 phyllosilicates using the AFM. Careful analysis of the images is also able to show that although the surface structure closely re sembles that of the known mineral, there is significant relaxation at the surface. The AFM may be used to image other structural features on the near atomic scale such as domain structures, and to progressively image descending layers in the structure.

Acknowledgments

This work was funded by a Natural Science and Engineering Research Council of Canada operating and equipment grants to G.S.H. Reviews by two anonymous referees and S.W. Bailey greatly improved the manuscript.

Fig. [3-1]: a) Schematic of the unit structure of chlorite showing the 2:1 talc-like layer and brucite-like interlayer sheet. b) {001} projection of the siloxane sheet of the talc-like layer. c) {001} projection of the upper surface of the brucite-like interlayer sheet.

Fig. [3-2]. Images of the siloxane sheet of the talc-like layer: a) Typical raw image; b) Image after application of flattening and lowpass filtering routines; c) Image shown in (b) after application of 2--D FFT to further increase the resolution. The ov erlay is the talc-like layer (cf., Fig. [3-1b]) calculated from the chlorite structure of Bailey (1975) (see text). Note that the theoretical and experimental structures essentially match. d) Enlarged view of (c) clearly showing the tip-induced rotati ons out of the a-b plane (this is seen as alternating bright tetrahedra within the rings). Unit cell dimensions are indicated. The arrow indicates the direction a 6-membered SiO4 ring is ``pointing'' due to tetrahedral rotations described by Brown and B ailey (1963).

Fig. [3-3]: a) Histogram of measured a cell distances (Angstroms). b) Histogram of measured b cell distances (Angstroms). All measurements are from a single image of the talc-like layer; >90% of the measurements fall within 4 values. c) Measured cell distances (Angstroms) taken from some 250 different images of the talc-like layer in air (square), oil (triangle) and water (circle). Filled symbols are a cell distances and open symbols are b cell distances. Linear regression analysis indicates expansio n of the b dimension (r = .4587) and to a lesser extent a (r = .4544) upon the addition of water to the surface.

Fig. [3-4] Images of the brucite--like interlayer sheet: a) Typical raw image; b) Image after application of flatten and low pass filter routines; c) Image in (b) after application of 2DFFT. The overlay is a brucite-like layer calculated from a known c hlorite structure (see text); d) Enlarged image of (c) with Mg cation visible (indicated by arrow). The a distance is indicated; e) A cross section through the Mg site showing a topographic high at the OH position and a smaller high at the Mg location.

IV. Preamble: Specific Ion Adsorption at the Mineral-Water Interface: Cesium Adsorption on Chlorite

This chapter is currently in press in the journal, Colloids and Surfaces (A):Physicochemical and Engineering Aspects, although the reference scheme has been altered to make this thesis internally consistent. It presents our work in elucidating the ads orption mechanism of cesium at the chlorite-water interface. The mineral-water interface of chlorite was investigated using atomic force microscopy and X-ray photoelectron spectroscopy (XPS). Both techniques are surface specific with one providing atomi c structure information at the interface and the other providing quantitative chemical analysis so that both methods complement each other. Computer modeling to simulate our system was also performed and has been added to this chapter.

Specific Ion Adsorption at the Mineral-Water Interface: Cesium Adsorption on Chlorite

G.V. Vrdoljak, G.S. Henderson*

Abstract:

Key words for indexing: atomic force microscope; cesium; chlorite; adsorption; X-ray photoelectron spectroscopy

Introduction:

The cycling of elements in many natural systems is controlled by geochemical processes occurring at the mineral-water interface (Hochella and White 1990). Of particular importance are the mechanisms by which cations and anions attach or detach from th e surfaces of minerals. Cesium, a long lived and highly soluble nuclide produced as a product of nuclear fission, has been of much geochemical concern due to its potential for harm in the environment. Cesium's large ionic radius allows it to polarize it self and bond at negatively charged sites on mineral surfaces, forming strong inner sphere complexes. The large size and strong bonding properties of cesium make it an ideal ion for visualization at mineral surfaces with the atomic force microscope (AFM) developed by Binnig et al in 1986.

Studies of cesium adsorption on clay minerals have been performed by analyzing the change in cesium concentration in aqueous solution in contact with clay over time or by altering solution conditions via the addition of surface-site competitive cations or isotopes. Mechanistic interpretations are postulated for the observed reactions by analysis of kinetic data. The current theory of Comans et al. (1991) indicates cesium is released from rapidly accessible surface sites and slowly migrates toward a c ollapsed interlayer site resistant to exchange. Visualizing adsorption of cesium with the AFM provides fresh insight into the nature of cesium bonding at surface active sites without the need for inferences from geochemical kinetic data.

Chlorite, a common but minor clay mineral in soils (Bailey 1975), was studied because it has proven to be a suitable substrate on which to adsorb cations. The chlorite structure contains a 2:1 talc-like layer that possesses excess negative charge and can therefore adsorb cations to its surface.

Experimental Methods:

A previously studied (Vrdoljak et al 1994) clinochlore IIb chlorite of composition (Mg4.4Fe0.6Al)[(Si2.9Al1.1)]O10(OH)8 obtained from northern Ontario was studied with a Digital Instruments Nanoscope III AFM. Chlorite samples were cleaved along the {001 } plane and immersed in distilled water for seven days before imaging by the microscope. The {001} cleavage surface was then imaged in the AFM fluid cell in distilled water using a 200 m wide Si3N4 cantilever (k=0.12 N/m), and in a 0.07030.0006 M Cs2CO3 solution at pH 10.9. A further sample was immersed in a (5.9560.006)10-3 M CsCl solution of pH 7. A 200 m thin legged cantilever (k=0.06 N/m) was used to image this sample. Solutions in contact with the sample were allowed to stand for one to two hours before imaging. Images were collected with the operating force of the instrument minimized ( 20nN).

All images were collected at a variety of scan speeds and angles to reduce the possibility of image artifacts being created. No real-time filters were applied during image acquisition and any images with excessive noise, distortions, or obvious artifa cts arising from instrumental fluctuations or transients were discarded. Of these remaining images (about 10%), the ones showing the surface structure most clearly were chosen and if necessary, digital filtering performed (as noted in figures) to reduce noise. Care was taken in processing of images so that important information would not be deleted and the periodic nature of the images highlighted.

To characterize the surface chemical composition, X-ray photoelectron spectra were obtained of the chlorite samples: before exposure to water; after exposure to water for 48 hours; after exposure to a 0.07030.0006 M Cs2CO3 pH 10.9 solution for 48 hour s; after exposure to a saturated Cs2CO3 solution for 24 hours; and after exposure to a (5.9560.006)10-3 M CsCl (aq) solution at pH 7 for 48 hours. X-ray photoelectron spectroscopy (XPS) analysis was performed using a Leybold Max 200 X-ray photoelectron s pectrometer. Spectra were satellite subtracted to remove extraneous peaks generated by the X-ray source with an algorithm based on the program of van Attekum and Trooster (1977). The algorithm of Berresheim et al (1991) was used for unit detector transm ission function normalization (so that different energy electrons will have the same detectability at the detector). XPS spectra were shifted to match the carbon 1s binding energy of 284.6 eV and a Shirley background fit was applied for high resolution s pectra. Peak intensities were obtained from the normalized spectra using sensitivity factors empirically derived for these spectra and supplied with the instrument. Samples exposed to water or aqueous solutions were briefly rinsed with distilled water a nd then dried by touching with a dry kimwipe before placing in the vacuum chamber of the XPS machine. Rinsing with water will not remove all chemisorbed cesium as cesium adsorption appears to be partially irreversible (Grütter et al. 1986). The spectra were collected with an effective spot size of 47mm and ten successive scans were performed at a pass energy of 192 eV. High resolution scans were performed on selected peaks using a pass energy of 48 eV.

To more accurately predict the binding site of cesium ions with chlorite in solution, a computer simulation was performed. The CERIUS molecular simulations package was used with the sorption module. This package uses a Metropolis Monte Carlo statisti cal mechanics method (Metropolis et al 1953) to simulate binding locations and energies of host molecules or ions in a crystal lattice. The algorithm generates a completely random position and orientation of the sorbate (CsCl) in a framework (chlorite). The configuration is accepted or rejected based upon the Metropolis algorithm which evaluates the energy of the system. These steps are repeated until enough configurations are generated to measure properties with statistical accuracy.

In our simulation we used an extended chlorite cell with dimensions of 11, 18, and 30 Å in the a, b, and c cell directions respectively. One CsCl group was used in the cell with the constant loading or canonical ensemble. The Burchant universal force field was used to predict interaction amongst all the atoms. Independent simulation in a separate lab was done to verify the computer output.

Results and Discussion:

The {001} surface of the talc-like layer under water appears as a well ordered siloxane sheet (figure [4-1]). Images are reproducible and the individual SiO4 tetrahedra making up the six-membered siloxane ring are resolved. Adding a Cs2CO3(aq) soluti on to the AFM fluid cell shows the appearance of fluxional islands of cesium atoms in the siloxane cavity with a c(1x1)R60º unit mesh (figure [4-2]). The start of surface precipitation may also have been imaged with Cs2CO3 precipitating in an island fash ion where the carbonate group is not resolved in the images. These islands of cesium atoms appear stable over a few scans, but will disappear and then reappear, probably due to the cantilever tip removing the cesium atoms when the surface binding energy of cesium is surpassed by the force of the tip (Cs removal will occur if the contact force is on the order of 10-7 N).

Imaging the surface of chlorite in a CsCl (aq) solution also shows the appearance of cesium atoms inside the siloxane cavity (figure [4-3]). The images are lower in resolution than those with Cs2CO3(aq) solution and the presence of cesium inside the s iloxane cavity as specks could easily be dismissed as noise unless compared to images of the surface in water (figure [4-1]). No island clustering of atoms can be distinguished from the background noise. The possibility of surface cesium chloride precip itation is unlikely at this lower cesium-salt concentration.

While the images of cesium adsorbed to the chlorite surface indicate that it adsorbs into the siloxane cavity as expected due to the localization of negative charge from Si4+/Al3+ substitution, some of the images (figures [4-2c] and [4-3b]) show prefer ential alignment of the cesium atom to one side of the siloxane cavity. This is probably an artifact created by the tip pushing the atom to one side of the cavity during the data collection, however, more work is needed before any conclusion can be made. In addition, the cesium ions appear to be adsorbed directly into the cavity rather than above the cavity as would be expected from crystal chemical reasoning. This binding of cesium into the cavity may be due to the positive cesium atom moving closer t o the source of negative charge (Si4+/Al3+ substitution) or to exchange of a hydrogen atom in the structure with Cs+. Alternatively, the force of the tip scanning above the siloxane surface may be 'pushing' the cesium atom into the cavity giving a false representation of the cesium binding site. However, the observed Cs+ binding site at the negatively charged siloxane cavity of the talc-like layer agrees well with what is expected from theory (Sposito 1990). Collapse of the hydrogen bonded interlayer c reates a better adsorption site for cesium (Grütter et al 1986) and may 'trap' the ion explaining the binding of cesium low in the cavity. However, this is unlikely as large flat areas of chlorite were viewed (highly exchangeable) as opposed to edges whe re layer collapse is likely to occur.

XPS analysis provides surface information on chemical composition and local electronic structure. X-ray photoelectron spectra of chlorite exposed to water and then exposed to Cs2CO3(aq) and CsCl(aq) solutions are shown in figure [4-4]. The chemical c omposition of the surface is similar to that predicted from the bulk chemical composition. However, carbon contamination is present along with a slight excess of silicon and aluminum. The excess silicon and aluminum probably arise from a partial leached layer of silanol and aluminol groups at the surface. The shift of each individual peak in the spectra is indicative of the type of bonding at the surface. Peak shifts for each element at the surface generally indicate the type of bonding expected from the chlorite structure. Evidence of Si-O, Al-0-Al, Al-OH, Fe-O, Fe-OH, Mg-OH, and O-H bonding is found in the high resolution spectra as predicted for the chlorite structure. Adsorption of cesium from the Cs2CO3(aq) and CsCl(aq) solutions is verifie d in the spectra by the appearance of cesium 3d5/2 and 4d5/2 peaks shown in figures [4-4]b,c and d. The cesium 3d peak shift is characteristic of a Cs-OH type of bond. We interpret the cesium peaks as resulting from surface adsorption rather than bulk s urface precipitation because the intensity of the cesium 3d peak indicates only a small amount of cesium at the surface. The cesium at the surface is comparable to the amount of iron at the surface, rather than a bulk coating of cesium carbonate. A slig ht increase in the intensity of the carbon 1s peak for this sample may indicate a slight amount of precipitation at the surface and edges which cannot be ruled out entirely. Chlorite exposed 24 hours to a saturated Cs2CO3(aq) solution yielded an X-ray ph otoelectron spectrum with greatly increased cesium and carbon peak intensity masking much of the chlorite spectra due to the bulk precipitate.

The output of the computer simulation is shown in figure [4-5] with the position of Cs over the siloxane cavity with Cl coordinated by the basal oxygens of an SiO4 tetrahedron (part of the structure has been omitted for clarity). This verifies our pre dicted behaviour of cesium with the siloxane cavity, but further simulations are necessary to simulate the actual surface as opposed to an enlarged unit cell. Ideally interaction of the AFM tip with such a group could be modeled and the results compared with output from the microscope.

Conclusions:

We have imaged cesium adsorption at the mineral-water interface using the AFM. X-ray photoelectron spectra indicate that the cesium adsorbs to the surface via a Cs-OH type of bond. Careful inspection of the images reveals tip-sample interaction effec ts such as the aligning of cesium with one side of the siloxane cavity and the possible 'pushing' of the cesium atom into the cavity by the tip. Island-like adsorption of cesium seen in some images may result from clustering of a excess negative charge. Computer simulation of our system confirms that the binding site for cesium is in the siloxane ring as observed in our AFM images.

Removal of cesium atoms could be accomplished by scanning the surface at higher tracking force. The relationship between contact force to the binding energy of the surface active species could be investigated by choosing various collectors and then co mparing their bond strength at the surface to the force required for removal with the AFM. More work needs to be done to characterize the surface-tip interaction and to improve the quality of AFM images so that the majority of images are not subjected to excessive artifacts/drift/noise/etc. In this way, adsorption of species from solution could be more readily identified and scanning force/bond energy comparisons made.

Acknowledgments:

This work was funded by a Natural Science and Engineering Research Council of Canada research and equipment grants to G.S.H.

Figure [4-1]: a) AFM image (distance units in all AFM images are in nanometers) of chlorite viewed in distilled water. An autoflattening routine has been applied to the image to remove image bow and a lowpass filter applied to remove high frequency noi se. b) 2D-Fast fourier transformed image of figure a to remove noise and point out the periodic structure of chlorite. The image has been enlarged (zoomed) to show structural features more clearly. An overlay of the (001) siloxane surface of chlorite has been added to show the agreement with the calculated structure.

Figure [4-2]: a) Raw AFM image of chlorite in Cs2CO3 solution. No filtering has been applied. b) Treatment of figure a with a highpass filter to exaggerate differences in the height aspect of the image reveals the appearance of cesium atoms inside th e siloxane cavity. c) 2D-FFT filtered and zoomed image of figure a shows the presence of cesium atoms inside the siloxane cavity.

Figure [4-3]: a) AFM image of chlorite surface from figure [4-1] viewed in CsCl solution. Autoflattening and lowpass filtering has been used. b): 2D-FFT filtered and zoomed image of figure [4-3a]. The inside of the siloxane cavity, as shown in the s tructural overlay, appears distorted; or has a new species present.

Figure [4-4]: a) X-ray photoelectron survey spectra of chlorite exposed to water for 48 hours. Peaks characteristic for each element are annoted on the spectra along with the orbitals from which they arise. b) X-ray photoelectron survey spectra of ch lorite exposed to a 0.07030.0006 M Cs2CO3 solution of pH 10.9 for 48 hours. Note the appearance of a new cesium peak indicating adsorption of cesium by chlorite at the surface. Bulk precipitation of Cs2CO3 is unlikely as a large cesium peak would mask t he rest of the spectrum.

Figure [4-4]: c) Low resolution x-ray photoelectron spectra of same sample shown analyzed in figure [4-4b]. Note the second cesium peak appearing overlapping with aluminum. d) Survey spectrum of chlorite exposed to (5.9560.006)10-3 M CsCl solution at pH 7 for 48 hours. A cesium peak is present, but with much lower intensity than in 4b due to a lower concentration of cesium in the solution.

Figure [4-5] Output of the computer simulation of cesium adsorbed to chlorite. Cesium is shown over the siloxane cavity of chlorite as seen in the AFM images.

V. Summary:

In this thesis a study of layer silicate surface chemistry has been done using the phyllosilicate mineral clinochlore. The atomic structure of the {001} clinochlore surface was first obtained in air utilizing the AFM. Some of the microscope images sh owed a slight relaxation effect in the talc-like surface in the form of shortening and lengthening of the a and b unit cell distances. The surface was also characterized using the AFM in water and oil. This was done to see the effect hydration has on th e surface structure and a slight expansion of a and b unit cell distances was seen.

The aqueous reaction of cesium ions with the clinochlore surface was studied. Cation exchange on a layer-silicate was seen for the first time with the AFM. Cesium cations bind to the negatively charged siloxane ring of the talc-like surface, balancin g the charge arising from isomorphic substitution. Support for the AFM results are given by XPS of the clinochlore surface before and after reaction, providing quantitative chemical and bond environment information. Metropolis Monte Carlo computer simul ation of the reaction showed the same binding site for cesium sorption as seen in the AFM images.

This work is an investigation into only one aspect of surface chemistry, but is applicable to a wide variety of geological surface chemical problems. Various unknown binding sites for surface active species in soils can be determined directly and the strength of the surface-adsorbate bond may eventually be determined utilizing AFM. Visualization of contaminant adsorption onto clay minerals and the likelihood of fixation from groundwater can be investigated with the AFM. The same procedures could be used to study the surface chemistry of other layer silicate minerals in other geological settings such as the weathering of various minerals in the development of soils.

This work provides an expansion in AFM use, making direct observation of adsorption mechanisms possible. Further AFM investigations into other aspects of surface chemistry (such as mineral dissolution, crystal growth, and colloid-surface interactions) are also possible. One unique advantage in AFM is the direct atomic scale visualization of surface processes. As the operation of the instrument is difficult, especially for atomic resolution imaging, a detailed procedure is provided to aid in reproduc ing the experiments.

List of References:

Anderson, M.A., and Rubin, A.J. (1981) Adsorption of Inorganics at Solid-Liquid Interfaces. Ann Arbor, Michigan, Ann Arbor Science,.357pp.

Hans Christian von Baeyer (1992) Taming the Atom: the emergence of the visible microworld. 1st ed. New York, Random House, 223pp.

Bailey, S.W. (1975) Chlorites, In J.E. Gieseking Ed., Soil Components, p., 191-263. Springer Verlag, New York.

Bailey, S.W. (1988) Chlorites: Structures and crystal chemistry. In S.W. Bailey Ed., Hydrous Phyllosilicates (exclusive of micas), p. 347-403. Mineralogical Society of America, Reviews in Mineralogy Vol. 19, Washington, DC.

Berresheim, K., M. Mattern-Klossen, and Wilmers, M. (1991) A standard form of spectra for quantitative ESCA-analysis. Fresnius J Anal Chem, 341, 121-124.

Binnig, G., Quate, C.F., and Gerber, C. (1986) Atomic Force Microscope. Physical Review Letters, 56, 930-933.

Binnig, G., Rohrer, H., Gerber, Ch., and Weibel, E. (1982) Surface Studies by Scanning Tunneling Microscopy. Physical Review Letters, 49, 57-61.

Brown, B.E., and Bailey, S.W. (1963) Chlorite polytypism: II Crystal structure of a one-layer Cr--chlorite. American Mineralogist, 48, 42-61.

Christmann, K. (1991) Introduction to Surface Physical Chemistry. New York, Springer-Verlag, 274pp.

Comans, R.N.J., Haller, M., and De Preter, P. (1991) Sorption of cesium on illite: Non-equilibrium behaviour and reversibility. Geochimica et Cosmochimica Acta, 55, 433-440.

Drake, B., and Hellmann, R. (1991) Atomic force microscopy imaging of the albite (010) surface. American Mineralogist, 76, 1773-1776.

Drake B., Prater, C.B., Weisenhorn, A.L., Gould, S.A.C., Albrecht, T.R., Quate, C.F., Connell, D.S., Hansma, H.G., and Hansma, P.K. (1989) Imaging crystals, polymers, and processes in water with the atomic force microscope. Science, 243, 1586-1589.

Gould, S.A.C., Burke, K., and Hansma P.K. (1989) Simple theory for the atomic-force microscope with a comparison of theoretical and experimental images of graphite. Physical Review, 40, 5363-5366.

Grütter, A., von Gunten, H.R., and Rössler, E. (1986) Sorption, desorption, and isotope exchange of cesium (10-9-10-3 M) on chlorite. Clays and Clay Minerals, 34 677-680.

Hochella Jr., M.F., and White, A.F., (1990) Mineral-water interface geochemistry: An overview. In M.F. Hochella Jr. and A.F. White, Eds., Mineral-Water Interface Geochemistry, p. 1-16, Mineralogical Society of America, Reviews in Mineralogy Vol. 23, Wa shington, DC.

Lindgreen, H., Gaenees, J., Hansen, P.L., Besenbacker, F., Laesgaard, E., Stensgaard, I., Gould, S.A.C., and Hansma, P.K. (1991) Ultrafine particles of North Sea illite/smectite clay minerals investigated by STM and AFM. American Mineralogist, 76, 1218-12 22.

Metropolis, N., Rosenbluth, A.W., Rosenbluth, M.N., Teller, A.H., and Teller, E. (1953) Equation of state calculations by fast computing machines. Journal of Chemical Physics, 21, 1087-1092.

Myers, D. (1991) Surfaces, Interfaces, and Colloids Principles and Applications. NewYork, VCH Publishers, 433pp.

Ohnesorge, F. and Binnig, G. (1993) True atomic resolution by atomic force microscopy through repulsive and attractive forces. Science, 260, 1451-1456.

Rachlin, A.L. (1992) An Atomic Force Microscope and Electron Microscope Study of Calcite and Apophyllite. unpublished M.Sc. thesis, University of Toronto.

Rachlin, A.L., Henderson, G.S., and Goh, C.M. (1992) An atomic force microscope (AFM) study of the calcite cleavage plane: Image averaging in Fourier space. American Mineralogist, 77, 904-910.

Sheiko, S.S., Möller, M., Reuvekamp, E.M.C.M., and Zandbergen, H.W. (1993) Calibration and evaluation of scanning-force-microscopy probes. Physical Review B, 48, 5675-5678.

Sposito, G. (1989) The Chemistry of Soils. New York, Oxford University Press, 277pp.

(1984) The Surface Chemistry of Soils. New York, Oxford University Press, 234pp.

Stumm, W. (1992) Chemistry of the Solid-Water Interface. Toronto, John Wiley & Sons, 428pp.

Stumm, W. and Morgan, J.J. (1981) Aquatic Chemistry. Toronto, John Wiley & Sons, 780pp.

van Attenkum, P.M.T., and Trooster, J.M. (1977) Removal of X-ray satellites from Mg K excited photoelectron spectra. J Electron Spectrosc Relat Phenom, 11, p. 363-370.

Vrdoljak, G.A., Henderson, G.S., Fawcett, J.J., and Wicks, F.J. (1993) Structural relaxation of the chlorite surface imaged by the atomic force microscope. American Mineralogist, 79, 6957-6962.

Weisenhorn, A.L., MacDougall, J.E., Gould, S.A.C., Cox, S.D., Wise, W.S., Massie, J., Maivald, P., Elings, V.B., Stucki, G.D., and Hansma, P.K. (1990) Imaging and manipulating molecules on a zeolite surface with an atomic force microscope. Science, 247, 1330-1333.

Wicks, F.J., Kjoller, K., and Henderson, G.S. (1992) Imaging of the hydroxyl surface of lizardite at atomic resolution with the atomic force microscope. Canadian Mineralogist, 30, 83-91.

Yong, R.N., Mohamed, A.M.O., and Warkentin, B.P. (1992) Principles of Contaminant Transport in Soils. In Developments in Geotechnical Engineering, Volume 73, New York, Elsevier Science, 327pp. v 61