UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A model for the incongruent dissolution of pyrite Irwin, Keith Wayne 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1996-0060.pdf [ 5.28MB ]
JSON: 831-1.0086964.json
JSON-LD: 831-1.0086964-ld.json
RDF/XML (Pretty): 831-1.0086964-rdf.xml
RDF/JSON: 831-1.0086964-rdf.json
Turtle: 831-1.0086964-turtle.txt
N-Triples: 831-1.0086964-rdf-ntriples.txt
Original Record: 831-1.0086964-source.json
Full Text

Full Text

A M O D E L FOR THE INCONGRUENT DISSOLUTION OF PYRITE. by K E I T H W A Y N E IRWIN B.Sc , The University of British Columbia, 1991 B.Ed., The University of British Columbia, 1992 A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Soil Science) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A December 1995 © Keith Wayne Irwin, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract To understand the incongrucnt dissolution phase of natural pyrite, polished and cleaned samples were subjected to oxidative dissolution at low pH from 0 to 72 hours. All samples were operationally defined as pure based on XRD, XRF, SEM/EDXA and XPS analyses. At 12 hour intervals, one pyrite sample and 10 mLs of solution were extracted from the reaction vessel for analysis. The air-sensitive pyrite samples were introduced into the UHV of an x-ray photoelectron spectrometer. Iron (Fe 2p) and sulfur (S 2p) oxidation states were measured and S/Fe mole ratios were calculated. All samples were compared to a series of controls which indicated the Fe 2p and S 2p binding energies to be 706.7 eV and 162.4 eV respectively. The reference S/Fe ratio was found to be 3.6. This high ratio is believed to be indicative of a natural weathering rind. Curve fitting indicated the presence of elemental sulfur in small amounts (7%), whereas, most sulfur retained its pyritic form. Throughout the dissolution experiment, both photoemission peaks showed no significant signs of oxidation despite the presence of oxygen at the surface. The S/Fe ratio increased from 3.6 to 5.6 (t = 48 h) after which point the ratio decreased rapidly to its original value. The 10 mL aliquot of solution was partitioned into two sub-aliquots - one for the determination of total iron by atomic absorption, the other for total sulfate by nephelometry, a method which proved itself to be superior to the more conventional turbidimetric method. Solution phase data indicated the relatively constant production rate of iron until 48 hours at which point, there is a noticeable increase in this rate. Sulfate analysis indicated a similar trend to that of iron though the increased rate at 48 hours was dramatic. Based on a correlation of the surface data with that of the solution phase, the following mechanism was proposed. Iron diffuses along a S/Fe mole ratio gradient. Eventually, a critical ratio is achieved at which time the surface begins to crumble. Flakes of elemental sulfur and sulfur-rich pyrite detach from the surface. The elemental sulfur is thought to remain as such based on the findings cited in the literature review. Given the significant rise in the production rate of sulfate, the sulfur-rich pyrite n flakes are believed lo convert quickly to sulfate while simultaneously releasing iron into solution. ill Table of Contents Abstract ii Tabic of Contents iv List of Tables vii List of Figures viii Acknowledgements x Introduction 1 Chapter One: Literature Review. A) Structure of Pyrite . 1 introduction 4 2)Bulk Features of Pyrite 4 a) Composition and Crystallography 4 b) The Disulfide Environment 5 c) The Iron Environment 6 d) Qualitative Description of the Bonding in Pyrite 6 a) Surface Topography 8 b) Surface Composition 10 B) Mechanisms for the Oxidation of Pyrite by Dissolved Oxygen 1) Introduction 10 2) Chemical Oxidation 11 3) Electrochemical Oxidation 16 4) Summary 23 Chapter Two: Thesis objectives 29 Chapter Three: Materials and Methods A) Materials 1) Reagents 31 2) The Pyrite Samples 32 iv 3) The Reaction Vessel and Supporting Equipment 32 B) Methods 1) Glassware Cleaning Procedure 34 2) Various Pyrite Cleaning Procedures 34 3) Monitoring the Conditions of the Experiment 36 4) Sample Preparation for Characterization and Instrument Settings 37 5) Sample Initiation, Collection and Storage 39 6) .Analysis of the Solid Phase 40 7) Analysis of the Liquid Phase 42 a) Low-Level Sulfate Determination 42 b) Nephelometry 43 c) Turbidimetry 46 d) Total Iron Determination 47 Chapter Four: Results and Discussion 50 A) Operational Definitions 50 B) Assumptions and Their Justifications 51 C) The Cleaning Procedures 52 1) Chemical State of Iron and Sulfur 52 2) Sulfur to Iron Mole Ratio 53 3) O/S and O/Fe Mole Ratios 54 4) Summary 55 D) Characterization of the Pyrite Samples 55 E) The Dissolution Experiment - Results 1) The Solid Phase 56 2) The Liquid Phase 57 3) A comparison Between Nephelometry and Turbidimetry 59 v F) A Model for the Incongruent Dissolution of Pyrite 60 1) The Original Surface 60 2) The Surface Under Dissolution 61 Chapter Five: Conclusion 76 Bibliography 80 Appendix. 1. Photograph of Experimental Apparatus 85 Appendix 2. Data on Characterization of Pyrite Samples 86 Appendix 3. Data on Nephelometry, Turbidimetry and A. A. Experiments 87 vi List of Tables. Table 1. Various published S-S distances in Pyrite 5 Table 2. XPS results of the various cleaning procedures 54 Table 3. XPS results indicating the ratio of elemental sulfur to pyritic-S (S°/S) 55 Table 4. XPS results of the dissolution experiment from 0 to ~2 hours 57 Table 5. Total sulfate and iron in the liquid phase as determined from nephelometry and atomic absorption respectively 58 vii List of Figures. Figure 1. The unit cell of pvrile 25 Figure 2. A qualitative molecular orbital diagram for pyrite 26 Figure 3. A (100) view of pyrite's unit cell 27 Figure 4. An example of a stepped and unreconstructed (010) surface 28 Figure 5. A schematic diagram of the reaction vessel and supporting equipment used for the dissolution experiment 49 Figure 6. XPS narrow scan spectra for the Fe 2p and S 2p photoemission peaks of a cleaned an uncleaned pyrite sample 65 Figure 7. XPS narrow scan spectra for the Fe 2p and S 2p peaks showing the effects of the various cleaning procedures on chemical state 66 Figure 8. Results of curve fitting of the S 2p peak indicating the presence df elemental sulfur on a cleaned sample 67 Figure 9. Typical XRD spectrum of powdered pyrite samples 68 Figure 10. Typical XPS wide scan spectrum of a cleaned pyrite sample indicating the presence of oxygen and carbon contaminants 68 Figure 11. Three SEM photographs of inclusions 69 Figure 12. EDXA analysis of inclusions 70 Figure 13. XPS narrow scan spectra of Fe 2p and S 2p peaks during the course of dissolution 71 Figure 14. A schematic of a surface undergoing dissolution 72 Figure 15. A plot of total iron and sulfate during dissolution 73 Figure 16. A plot of the S/Fe mole ratio at the surface of pyrite during dissolution 73 Figure 17. A plot of five adjusted standard curves from the nephelometry experiment for the determination of total sulfate 74 Figure 18. A plot of the adjusted and averaged standard curve from the ncphclomctry experiment for the determination of total sulfate 74 Figure 19. A plot of five sulfate standard curves from turbidimetry 75 Figure 20. Plot of the averaged sulfate standard curve from turbidimetry 75 ix Acknowledgments. I would like to extend unreserved appreciation to each of my committee members. Dr. L.M. Lavkulich showed extreme sensitivity throughout the study and an acute sense of direction which he was able to communicate well to me. Dr. P. Wong's acicular knowledge of surface science and his meticulous approach to adhering to the scientific method paralleled his commitment to assisting me throughout this sudy. Dr. K. Mitchell's ability to explain complex concepts in simple terms certainly matched his impressionability. Also I would like to bestow my thanks to all the technicians and other graduate students who took a delight in assisting me at various times during the thesis. And last but not least, I would like to thank and dedicate this work to my two children, Liam and Sarah, who exhibited love and patience at all times. Finally, I would like to offer my thankfulness and indebtedness to my God whithout whose grace and infinite mercy this work would not have have been possible. x Introduction One of the more important geochemical processes to be modeled is pyrite oxidation in aqueous systems. Pyrite (FeS2) which is a non-economical yet ubiquitous mineral will, upon exposure to atmospheric oxygen, yield acidic conditions in soils (acid-sulfate soils) and in rock dumps at mine sites (acid mine drainage, AMD). Acidifying products of largely unknown reactions are released into the environment causing havoc with nature. Two examples are given in the following discussion. Of the estimated 10 -12 million hectares of acid-sulfate soils in the world (Van Breemen, 1976; Beek et al., 1980) most occur in recent: coastal marine sediments which are suitable for agriculture (Van Breemen, 1976). The pyritizatipn of ferric iron occurs in intertidal environments with mangroves or reeds and the highest pyrite contents are found where tidal flushing is strongest (Pons et al., 1982). Furthermore, pyritic sediments of the Pleistocene and Tertiary ages are common in many inland areas (Pons et al., 1982). Temperate and upland regions are not exempt from the woes of pyrite oxidation either (Gosling and Baker, 1980; Prokopovich, 1988; Wagner et al., 1982). Subsequent to the reclamation and drainage of acid-sulfate soils, the pyrite is oxidized by atmospheric oxygen thus producing sulfuric acid (Van Breemen, 1978). Within 50 cm of the surface, acid-sulfate soils have a pH below 4. At these low pH values, metals such as Al reach toxic levels rendering otherwise productive land practically useless for food production unless remediation programs are implemented to raise the pH. Moreover, toxic elements can leach into the surrounding surface and groundwaters thus affecting more distant soils and waterways. In the tropics, ecologically valuable wetlands located downstream from a reclamation site are seriously affected by such acidic waters (Bronswijk and Groenenberg, 1993). Probably the largest single environmental problem currently facing the mining industry is AMD (Gould et al., 1982). In Canada, AMD is generated at Ni, Zn, Cu, Pb, Au, U and coal mines either in waste rock dumps or in the tailings ponds produced during mineral processing (Gould etal., 1982). 1 In a rock dump, the most common and noticeable manifestation of AMD is the reddish brown staining due to the presence of ferric salts (Ritchie, 1994). The salts themselves pose no environmental hazard. Rather the pellucid drainage waters which contain 1 - 100 ppm of trace metals such as Cu, Cd, Zn etc. have detrimental effects on the aquatic flora and fauna (Ritchie, 1994). Acid neutralizing minerals adjacent to oxidation sites are able to, at least initially, suppress the effects of AMD. However, as these acid-neutralizing minerals are consumed, acids produced thereafter are transported away by the movement of pore water in the rock dump. Eventually, and in a similar way to that of acid-sulfate soils, these acidic products infiltrate surface and sub-surface waters thus affecting larger areas of land (Beek et al., 1980). Since the 1970's the generation of AMD has been intimately linked to the potential development of a mine (Jambor, 1994). Unlike the past where minimal capital was spent on site rehabilitation, nowadays and throughout North America, stringent environmental controls have been implemented largely as the result of the combination of legislated regulation and the involvement of special-interest groups. In fact, environmental concerns may be the dominating factor that could prevent a potential ore body from being mined (Jambor, 1994). Nicholson (1994) outlines three purposes for conducting laboratory studies of sulfide mineral oxidation (SMO): (1) to explore the mechanisms responsible including the rate-determining step, (2) to understand the role of emironmental factors in SMO and (3) to develop predictive models which describe the oxidation processes. This thesis wishes to contribute to the extensive library of knowledge of SMO by considering the incongruent dissolution phase of p\Tite oxidation. Regarding the general concept of incongruent dissolution, it is believed that most, if not all, minerals initially undergo an incongruent dissolution phase. Then, because of the formation of various product layers on the surface of the mineral, congruent dissolution occurs where reactants and products must diffuse across these newly formed weathered layer(s). It is conceivable that samples used in past studies already possessed a natural yet unsuspected weathering rind. That is the surface of pyrite may not be stoichiometric to begin with. 2 Also, as Ihc mineral undergoes dissolution, it may simply cycle through periods of incongrucnt dissolution. That is, during the course of pyrite weathering, the ratio of sulfur atoms to iron atoms released may simply fluctuate between some fixed values. In part, this thesis will concern itself with these ideas. 3 Chapter One: Literature Review. A) The Structure of Pyrite. 1) Introduction. A prerequisite to elucidating any aqueous oxidation mechanism of pyrite is a knowledge of the mineral's structure. This section is a synthesis of some of the details of the bulk features of pyrite followed by a more general description of some of its surface features. 2) Bulk Features of Pyrite. a) Composition and Crystallography. Semiconductor pyrite ( FeS2 ) actually has variable sulfur content in the range FeS2.n1 to FeSi.93 (Lowson, 1982). The iron exists as Fe 2 + while the sulfur exists as S2 "2 (Lowson, 1982; Finklea et al., 1976; Klein and Hurlbut, 1985, p. 285). Pyrite belongs to the isometric crystal system (2/m bar 3 ; Pa3 space group ) and crystallizes in a modified sodium chloride structure where the ferrous ions occupy the Na + sites (the corners and face centers of the unit cell) while the centre of the disulfide ions occupy the CI" sites. All six faces of cubic p>Tite are equally reactive (Klein and Hurlbut, 1985, p. 285). There exists ample data on the geometric features of pyrite (Finklea et al., 1976; Lowson, 1982; Mellor, 1935; Kjekshus, 1971; Panzner and Egert, 1984): unit cell length = 0.54175 nm distance between 2 Fe atoms on the (110) face = 0.382 nm distance between Fe and center of an S-S pair on the (001) face = 0.270 nm S-S distance « 0.21 nm Fe-S distance = 0.2259 nm S-Fe-S bond angle = 85.66 and 94.34 degrees ( a departure from the ideal 90 degrees) Fe-S-Fe bond angle = 115.5 degrees S-S-Fe bond angle = 102.4 degrees cell volume = 159.000 x 10 " 2 4 cm 3 x-ray density = 5.0116 g cm(high due to approximate close packing of atoms) 4 molar volume = 23.940 cm / mol FcS2-b) The Disulfide Environment. The disulfide ions lie parallel to the four space diagonals of the cubic unit cell (Deer et al., 1992, p.583) and are located at the edges, corners and centre of the unit cell (Figure la, p.25). The sulfur atoms are arranged in layers with approximate hexagonal stacking. Each sulfur atom has four nonequivalent nearest neighbors - one close S atom and three slightly further Fe atoms which form a planar triangular group - resulting in a distorted tetrahedral arrangement (Figure lb, p.25). The S-S pairs coordinate around Fe centers in a distorted octahedron. The S-S Lewis bases are strong field ligandS: A number of values for the distance between the two S atoms of the disulfide ion have been published, see Table 1. Table 1. Various published S-S distances in Pyrite. S-S Distance (nm) Reference 0.206 a 0.2135 b 0.2153 c (a) Lowson, 1982; (b) Van Der Heide et al, 1980; (c) Finklea et al, 1976, and Panzner and Egert, 1984. Although the S-S distance is short due to the positive overlap of the 3s (a l g , a g ) orbitals, it is not this covalent bond which is responsible for the actual bonding in the S-S pair because the antibonding 3s* (a2U, °"u *) orbital is equally populated (Figure 2. p.26). The sigma bonding in the S-S pair is due to the positive overlap of the 3pz (a2g )orbitals ( where z is the internuclear axis) giving the S-S ion an overall bond order of 1. The inner core levels (Is, 2s, 2p) of the S-S pair do not contribute to the overall bonding due to restrictions of the small radial extension of these orbitals. The molecular orbital (MO) diagram for the disulfide ion has been proposed by 5 some workers (Luther, 1987; Finklea ct al., 1976; Mishra and Osseo - Asarc. 1988; Ennaoui et al., 1986: Van Dcr Hcide el al., 1980). In these papers, both the hybridized and unhybridized MO diagrams are presented. Quantum mechanics reminds us that it is not necessary to invoke hybridization in order to construct the molecular orbitals of the ion. As mentioned above, the 3pz (a2g, CTg ) MO is responsible for the bonding in the S-S ion. The 3pz* (a2U , CTu* ) MO remains empty. The remaining twelve valence electrons of the ion are located in the 3s (aig ? a g ) [2], 3s* (a2u , °" u * ) 3Px,y (eiu . 7 1 u ) [4], and 3p x y* (eg*, n g * ) [4] molecular orbitals (Figure 2, p.26). It is these 12 electrons that are donated to the nearest 6 Fe atoms, that is, S-S ligands form ligand-group orbitals which are. shared equally with all neighbouring iron atomic orbitals. The resulting disulfide ion is diamagnetic. c) The Iron Environment. Figure lc, p.25, illustrates the spatial arrangement of the iron atoms in pyrite's unit cell. As noted earlier, the Fe-S distance is 0.2259 nm. This is much shorter than the distance one would obtain by simply adding the ionic radii of Fe 2 +and S1" (0.262 nm ) suggesting that Fe-S bonding is essentially covalent. Each ferrous ion ( d 6 ) is located at the center of a distorted octahedron . The ground state for these low spin d 6 ions is 5 D which transforms, in an octahedral field, as 5 E g and 5 T2 g states (Huhhey, 1983, p.442). Trigonal distortion splits the lower t2g level into non bonding ai g and eg levels wherein all 6 d-electrons reside, resulting in (1) the non ideal S-Fe-S bond angle (Finklea et al., 1975) and (2) the diamagnetism of the ferrous atom. The aj g and eg levels are nonbonding because they and their matching S-S levels are filled and therefore satisfied. The resulting d-orbital splitting pattern is similar to that of trigonal prismatic coordination. d) Qualitative Description of the Bonding in Pyrite. The sigma bond'between sulfur and iron can be understood using a simple valence bond picture. In this case, each lone pair of electrons from the S-S pair is donated to a hybridized d s p orbital on iron. However, this approach fails to account for most bonding between iron and sulfur. The valence bond theory (VBT) portion of the molecular orbital diagram is restricted to the 6 lower segment of the diagram only (Figure 2, p.26). Moreover, hybridization of the 3d, 4s and 4p orbitals on iron is not necessary. Figure 2 combines features of VBT and crystal field theory (CFT) to obtain a qualitative MO diagram for pyrite. Seven points to keep in mind when considering the MO diagram are (1) even though d orbitals appear to contribute qualitatively well to bonding, quantitatively their overlap with other orbitals is poor because d orbitals are diffuse, (2) the eg upper valence band is filled . (3) the upper antibonding orbitals, which form the lower conduction band, are more iron - like in character; whereas, the lower bonding orbitals, which form the upper valence band, are more sulfur - like in character, (4) because of symmetry constraints, the sulfur a2g orbital does not bond to.any iron orbitals, (5).bonding is strongest between sets of orbitals which are closest in energy, (6) the degree of orbital overlap is dependent on the angle between the orbitals and (7) no electrons reside in pyrite's antibonding molecular orbitals thus, in part, accounting for this mineral's hardness (~ 6). Salmeron et al. (1976) find it convenient to think of the interaction between sulfur and iron in terms of tunneling. Electrons in the iron band may tunnel toward sulfur atoms thus increasing their electron density. It is well known that pyrite has no preferred cleavage plane. It is also known that the inherent strength of a crystalline material is a function of structural defects and bonding mechanism. Regarding structural defects, the natural presence of randomly distributed trace elements in pyrite are an inherent source of defects in the crystal. Pyrite may contain small amounts of Ni and Co, in fact, there exist a complete solid solution series between bravoite, (Fe, Ni)S2, and pyrite (Klein and HurlbuL 1985, p. 285). It is thought by some workers that the Ni and Co are responsible for the ability of pyrite to behave as both n- and p-type semiconductors; in fact, n- and p-type semiconduction is possible within the same crystal (Lowson, 1982). The ionic radii of N i 2 + and Co 2 + are greater than that of Fe 2 - ; therefore, these impurities serve to increase the cell edge length. Also, there may be small quantities of Cu, Au, Zr, Pb, Zn present as trace impurities (Klein and Hurlbut, 1985, p. 285). It is conceivable that these impurities are one reason for the lack of cleavage plane in pyrite. Regarding bonding, consider the Z = 0 and Z = 0.5 planes 7 of Figure 3 (p.27). Note the octahedra are not all aligned along the same axis. This non-alignment of trigonally distorted octahedra throughout the lattice is a direct function of the positions of the S-S ligands. Obviously, these ligands have positioned themselves in such a way as to minimize the lattice energy. Consequently, no unique plane of weakness occurs, and again, no preferred cleavage plane exists. A mention should be made regarding the S2 eg and eg* orbitals. These orbitals arise from the linear combination of 3p orbiuals on the S atoms and, consequently, have n symmetry. However, consider, for example, the Tl ee orbital. It is reasonable that one lobe of this orbital 2 ' 2 2 interacts with an Fe dz - orbital (c ) while the other lobe interacts with the Fe d x - v (P ) orbital of a different neighboring Fe atom. Overall, sigma bonding is acheived. e) The Band Gap. There exists a range of band gap values ( E G ) in the literature, the range being due to the various methods of determination: 0.77 10 1.2 eV (Ennaoui et al., 1986. Lowson, 1982) Simplified schematic band gap diagrams have been published (Mishra and Osseo - Asare, 1988; Luther, 1987; Ennaoui et al., 1986). In these diagrams, the lower and upper boundaries of the band gap are the iron's t 2 g valence band and eg* conduction bands respectively. More precisely, because the Fe u° energy level has undergone a perturbation from trigonal distortion, die upper valence band is composed of the iron's nonbonding ajE and ee bands. Electrons are excited from the upper valence band across the small band gap to the conduction band (Mishra and Osseo - Asare, 1988). The zero to slightly positive and temperature dependent magnetic susceptibility of pyrite was considered to be the result of the small band gap (Lowson, 1982). Moreover, the small band gap may suggest that covalent bonding between the sulfur and iron orbitals is appreciable. 3) Surface Features of Pyrite. a) Surface Topography. The surface of any solid may be defined as a two dimensional plane in which exists atoms and bonds that have preserved tfoeir bulk-like features ( positions and energies ). However, in reality, surfaces are three dimensional spaces (composed of terraces, adatoms. ledges, kinks 8 etc.) that undergo structural and energetic changes, thus distinguishing them from the bulk. Three types of changes are discussed by Cox (1989, p.232-240) and Harrison (1989, p.232-249). Firstly, there exist surface relaxation mechanisms wherein surface atoms may simply move inward or outward in an effort to minimize their energies. Earlier, mention was made regarding the effect trigonal distortion has on the S-Fe-S bond angle. At the surface, we may expect surface atoms to relax in such a way that this angle may approach 90 degrees more closely. Surface relaxation does not affect symmetry. Secondly, surface reconstruction affects surface geometry and energies. Surface reconstruction may be viewed as a natural result of the Jahn -Teller theorem which states that a system with a degenerate ground state will spontaneously reconstruct in an effort to remove the degeneracy ; consequently, symmetry is always lowered. If reconstruction is to occur, then new bonds would form between partially filled orbitals or between filled and empty orbitals. Many reconstruction scenarios are probable. Consider the following two examples: (1) dangling S-S pairs could conceivably regroup to form elemental sulfur, (2) because covalent bonding between two sulfur atoms has been suggested to be stronger than that between sulfur and iron atoms, iron-sulfur bonds are more likely to cleave than sulfur-sulfur bonds; therefore, those iron atoms which have lost a surface ligand could form a bond with a sulfur atom of a neighbouring S-S pair. Surface reconstruction is common in covalent solids but not so in transition metal solids. Thirdly, adsorbates affect surface structure and energy. Obviously, the type of adsorbate will influence the manner in which the surface reconstructs. An example is the work performed by Ibach and Rowe (1974) who studied the adsorption of hydrogen onto a silicon (111) surface. In this case, the hydrogen atoms form bonds with dangling hybrid orbitals thus removing the 2x1 reconstruction. Obtaining a clean surface is a challenging task. Water vapor, hydrocarbons and oxygen-containing carbon species may be present even under the ultra high vaccuum of an x-ray photoelectron spectrometer. b) Surface Composition. Surface composition is operationally defined here as the sulfur to iron mole ratio at the surface of pyrite. In their studies on the detection of sulfides and polysulfides on electrochemically 9 oxidized surfaces of pyrite, Mycroft et al. (1990) have found that XPS measurements consistently gave anomalously high S / Fe mole ratios of 3.2. This ratio was taken as their reference ratio. This experimental ratio may not be an anomaly at all; in fact, it may very well be that the ratio is naturally high at the surface relative to the bulk indicative of some type of weathering rind. If this assumption is true, an S-rich surface layer would tend to lower pyrite's reactivity in solution because of sulfur's hydrophobicity. c) Surface Electronic States. In addition to speaking of pyrite's surface in terms of its topography and chemical composition, we may also speak of its surface in terms of its electronic states. Surface states may be defined as dangling bonds wherein electronic states are restricted to the surface. Figure 3 (p.27) suggests filled and empty iron 3d, 4s and 4p-orbitals as well as S2 2 molecular orbitals extend from the surface. Filled orbitals could conceivably behave as sites of reduction whereas empty orbitals, sites of oxidation.The coordination number of surface atoms is lower than their bulk counterparts; consequently, surface, atoms become more atomic-like and have narrower bands associated with them. These same band narrowing and reduction in coordination number phenomena occur as the density of steps increases on a surface. Figure 4 (p.28) illustrates a cross section of a hypothetical stepped surface, and suggests that the surface of pyrite could reconstruct to form a sulfur rich layer as a result of dangling S-S orbitals. 4) Conclusion. All chemical reactions involving solids with liquids occur at the solid/liquid interface. With respect to the solid, an understanding of its surface is crucial in the development of any reaction mechanism. In the following section, it should become apparent that all reaction mechanisms for the oxidation of pyrite begin with the assumption that pyrite's surface structure is ideal and stoichiometric. That is, the surface of pyrite has maintained its bulk-like features. Making this assumption is convenient, however, it can hinder any realistic understanding of what processes are actually occurring at the solid/liquid interface. B) Mechanisms for the Oxidation of Pyrite by Dissolved Oxygen. 10 1) Introduction. This section reviews the various mechanisms that have been put forth for the aqueous oxidaton of pyrite by dissolved oxygen ( DO) primarily under acidic conditions. There exist two major abiotic oxidation mechanisms for the oxidation of pyrite by DO: (1) chemical oxidation and (2) electrochemical oxidation. A major difference between these two mechanisms is that in the chemical pathway, the oxygen in the products is derived from molecular O2 ; whereas, in the electrochemical pathway, the oxygen in the products is derived from H 2 0, and the DO is consumed in a separate cathodic reaction (Lowson, 1982; McKibben and Barnes, 1986; Ritchie, 1994; Mathews and Robins, 1974; McKay and Halpern, 1959). Before discussing the work done on the chemical oxidation of pyrite, two general statements may be made: (1) The reaction between DO and the heterogeneous pyrite surface should not be limited by O2 solubility provided the mass transfer of 0 2 gas to the dissolved state is not limiting. The solubility of O2 at room temperature is approximately 2.65 x 10 " 4 M (2) Pyrite oxidation is usually far from equilibrium (Moses and Herman, 1991). 2) Chemical Oxidation. The salient features in the chemical oxidation are an adsorption step followed by a chemical reaction at the solid / liquid interface. Type CI: McKay and Halpern (1959) proposed the following mechanism. Initially, there is the fast chemisorption of a monolayer of DO: FeS2 + 02(aq) -> FeS2 * 0 2 (1) followed by a second adsorption step (the rate determining step) : FeS2 * 0 2 + 0 2 (aq) -> FeS2 * 2 0 2 (2) Desorption occurs according to : FeS2 * 2 0 2 - » FeS04 + S° (3) 11 The distribution of Fe 2 \ Fe 3', S° (elemental sulfur), S0 4 ~2. can be derived from the following reactions: 4 FeS04 + 0 2 + 2 H 2 S0 4 2 Fc2 (S0 4) 3 + 2 H 2 0 (4) 2 S° + 3 0 2 + 2 H 2 0 -> 2 H 2 S0 4 (5) The apparent entropy of activation is - 75 J M"1 K"1 which is reasonable for a reaction occurring at a solid / liquid interface and whose rate determining step may involve an adsorption step (see p. 16). The experiment of McKay and Halpern (1959) was conducted in the temperature range 100- 130 °C andPo2 range of 1-4 aim.. The authors point out that the rate of oxidation is proportional to the surface area and partial pressure.of oxygen, and independent of ferrous and ferric iron concentrations up to 0.1 M and 0.55 M respectively. Elemental sulfur was found in solution, as an intermediate, in large amounts under conditions of high acidity and at the lower temperatures (~100 °C). It is known that the production of S° in solution is minimal at room temperature, reaches a maximum at 100 - 150 °C (see Type C3 (p. 13) and Type E3 (p. 19)) and reduces at T > 150 °C. As a result, minimal to nil S° is observed, in solution, in low (room) temperature experiments. Type C2: In addition to the above mechanism, McKay and Halpern (1959) also proposed an alternative mechanism wherein elemental sulfur and sulfate are actual end products produced via a common intermediate ( FeS 2*0 2) formed in the slow adsorption step : FeS2 + 0 2 -> FeS 2*0 2 (6) then desorption occurs according to: FeS 2*0 2 -> FeS04 + S° (7) FeS 2*0 2 + H 2 0 -> FeS0 4 + H 2 S0 4 (8) Reactions (7) and (8) are parallel reactions. The distribution of S° and SO4"2 , in solution, will depend upon the rates of reactions (7) and (8) respectively. At least initially, the distribution of S° and SO4 "2 may be ~ 50 : 50. 12 The fact that S° may be an end product is supported by the leaching studies of Bieglcr and Swift (1979). Once produced, S° was very stable even at elevated temperatures. S° was found to be very stable during the anodic polarization of H2S in 0.1 M H2SC>4 between 1.2 and 1.7 V (relative to the Standard Hydrogen Electrode, SHE). In fact, Mishra and Osseo-Asare (1988) found that S° remains the only product up to potentials of 1.45 V. Non d-band sulfides such as CdS, ZnS and PbS also react in aqueous solutions to yield S° rather than sulfate. Now that Fe 2 + and SO4" 2 are formed, Fe + 2 must be oxidized by 0 2 to form Fe~3 . The pyrite can be further oxidized by this ferric iron. Below pH - 3, oxidation by DO or Fe3" has little to no dependence on pH: (Moses et al., 1987; Ritchie, 1994; Lowson, 1982). Also, below pH ~ 3 and under abiotic conditions, oxidation by ferric iron is 10 to 100 times faster than by DO arid a significant amount of ferric iron can remain in solution to react with the pyrite surface. (Ritchie, 1994; Singer and Stumm, 1970; Nordstrom, 1982; Moses et al., 1987; Nicholson, 1994; McKibben and Barnes, 1986). However, the conversion of ferrous to ferric iron is a process which decreases rapidly with decreasing pH. In their experiments, Singer and Stumm (1970) could not detect any ferric iron even after several days of oxidation. The oxidation of pyrite by Fe + 3 may be according to FeS2 + 14Fe + 3 + 8 H 2 0 -> 15 Fe"2 + 2 S04"2 + 16 H" (9) Here, pyrite provides sites for the reduction of ferric to ferrous iron and water acts as a source of oxygen in the sulfate. In terms of acid mine drainage, because the pyrite can reduce Fe"3 much more quickly than the Fe + 2 can be regenerated to the Fe"3 form, the latter becomes the rate determining step. (Goldhaber, 1983; Singer and Stumm, 1970). TypeC3: In an acid leaching experiment conducted at T > 100 C, Majima and Peters (Lowson, 1982) proposed a modified adsorption - chemical reaction mechanism involving the following adsorption step: 02(aq) -> 2 0 (ads) (10) This is followed by the rate determining step: 13 3 O (ads) + FcS2 - » FcS203(ads) (11) involving adsorbed thiosulfatc. The thiosulfate then detaches and disproportionates to sulfite and elemental sulfur, in solution, according to the reaction: FeS203(ads) + 2 H + - » F e 2 + + H 2 S0 3 + S° (12) then , H 2 S0 3 + 1/2 0 2 - » H S O 4 " 1 + H + . (13) Invariably, elemental sulfur was an end product but never exceeding yields of 50 %. This chemical mechanism is thought to be one of two parallel mechanisms involved in the oxidation of pyrite, the other being electrochemical. The elementary step (11).implies that the migration of adsorbed oxygen to favorable sites in preparation for the subsequent formation of FeS 20 3 (ads) is rate limiting. This in contrast to the idea that 0 2 adsorption itself may be rate limiting. This migration phenomena is similar to that proposed by Mishra and Osseo-Asare (1988) for the migration of hydroxyls (see Type E6, p.22). Type C4: Goldhaber (1983) provides more insight into the oxidation of pyrite by DO in the pH range 6-9 and T = 30 °C. First, the pH of zero point charge (pHzpc) for pyrite is approximately 7.3 . Oxidation of pyrite seems fastest at ~ 7.3 where there exists a net negative charge on the S-S pairs extending from the surface. One atom of each S-S pair extending from the surface becomes protonated. It is these protonated sulfur atoms which are thought to be sites for the fast adsorption of DO. The rate determining step is thought to be the subsequent adsorption of water. After some rearranging, the ferrous iron is oxidised to the ferric form by a second adsorption step of DO. No elemental sulfur is formed in this mechanism. Furthermore, Goldhaber assumes an ideal surface. Type C5: McKibben and Barnes (1986) oxidised crushed and pretreated pyrite using a number of oxidants ( DO, H 2 0 2 , Fe + 3 ) in the pH range 1 - 4. The specific initial rate law was found to be Rsp, DO = - 10 6 7 7 f O2 ]° 5 f moles pyrite cm "2 min ] (14) 14 Samples were periodically removed, over the 1.5 day experiment, for the determination of total iron and total sulfur concentrations. Despite the presence of polythionates. sulfate was found to be the dominant form of sulfur in agreement with the trends observed by Goldhaber (1983) at higher pH values. The surface area to volume ratio was assumed not to change during the course of the reaction (see p. 16). The activation energy (Ea) was 56.9 + 7.5 kJ per mol. which is indicative of a reaction controlled mechanism. Diffusion controlled mechanisms tend to have Ea values < 16 kJ per mol. The authors point out that the adsorption of a DO molecule onto the pyrite surface and it's subsequent dissociation is suggested by the square root rate law. Often, the overall rate of a solid / liquid reaction is controlled by the rate of the. adsorption or desorption step. Hence, the chemical oxidation rate of pyrite is thought to be reaction controlled and not diffusion controlled (Moses et al., 1987; Lowson, 1982; Goldhaber, 1983). Lasaga (1981) has shown analytically that the surface coverage of adsorbed species onto available sites is dependent on the square root concentration of the adsorbed species. Williamson and Rimstidt (1992) argue that the fractional order is strong supporting evidence for 0 2 sorption. In their study, they use a Freundlich isotherm that seems to indicate non-specific sorption and possibly multi-layers of oxygen at the surface. Furthermore, McKibben and Barnes (1986) mention that any mechanism occurring at the surface could involve reactions such as : 0 2 -> 2 0 (15) or 0 2 + 2 H 2 0 -> 2 H 2 0 2 (16) However, these reactions need to be reconciled with the electrochemical mechanism inferred from potentiometric studies. In this case, oxygen may be involved in the cathodic reaction: 0 2 + 2 H + + 2 e' - » H 2 0 2 (17) The balancing anodic reaction may be : FeS2 + 8 H 2 0 - » Fe + 2 + 2 HS04" + 14 H + + 14 e" (18) No unique mechanism reconciling both chemical and electrochemical steps was proposed. Using cyclic voltammetry, Biegler et al. (1975) studied 0 2 reduction on the pyrite 15 surface. As a result of their investigations, they believe they have identified peroxide as a soluble intermediate during pyrite oxidation in acidic solutions. Adsorbed 0 2 is thought to convert to adsorbed peroxide which can (1) convert to water and (2) desorb and diffuse into the surrounding medium. As noted above, McKibbens and Barnes (1986) pretreated their pyrite. The pretreatment of pyrite is a common procedure intended to circumvent the problems associated with the initial incongruent dissolution phase of oxidation. Different workers use different pretreatment procedures but the purpose in each is the same (Moses, 1982; Nicholson, 1994; Fornasiero, 1992; Moses etal., 1987). ... Reactive surface area is that fraction of the total area which is characterised with high surface energy due to the presence of grain edges, inclusion pits, corners ect.. Regarding the assumption that the surface area does not change significantly during the course of a reaction, Moses and Herman (1991) showed that the reactive surface area is small and changes negligibly during the course of dissolution. Through the use of SEM photographs, they were able to show that large areas of their mineral surfaces remained smooth and relatively unaltered before and after oxidation in contrast to smaller areas of high surface energy which appeared ragged and etched after oxidation. From their compilation of much data, White and Peterson (1990) have concluded that the reactive surface area is one to three orders of magnitude less than the BET surface area. Regarding the adsorption step, Luther (1987), Moses et al. (1987) and McKibbens and Barnes (1986) have suggested the following explanation for the slow adsorption of oxygen onto the pyritic surface. Oxygen, in its ground state, is P with two unpaired electrons in FI * orbitals; consequently, 0 2 is paramagnetic. However, because pyrite is diamagnetic, it's reactivity with 0 2 is spin restricted. 0 2 has no vacant sigma orbitals of proper energy and symmetry to accept electrons from the S-S pair (a surface state). Therefore the transition state complex would lack a strong sigma bond and be unstable. Moreover, upon electron transfer from a sulfur higher occupied molecular orbital (HOMO) to an 0 2 lower unoccupied molecular orbital (LUMO), the 16 bond order of the O2" species would lower from 2.0 to 1.5 offsetting the gain in energy of the S-S bond. O2 is known to chemisorb slowly to many metals and their oxides making the adsorption step essentially irreversible and likely rate-determining. The aquo - Fe 3 complex, although it is also paramagnetic, is able to transfer a hydroxyl radical from a water ligand to the diamagnetic pyrite surface. Consequently, it's reaction with pyrite is approximately one to two orders of magnitude greater than that of 0 2 (Luther, 1987; Moses et al., 1987; McKibben and Barnes. 1986). 3) Electrochemical Oxidation. Electrochemical and oxidative dissolution are synonymous. As mentioned earlier, in the electrochemical mechanism, the oxygen in the products is derived from water while the DO is consumed in a separate cathodic reaction at the pyrite surface. In isotopic studies, Taylor et al. (1984 a,b) found that water is a more important source of sulfate-oxygen than DO, at least at low pH. Unlike the chemical approach, there is no need for a 50 : 50 distribution of S° and sulfate because these are thought to arise from separate independent reactions occurring at different rates (Lowson, 1982) Type E l : From pressure leaching experiments, Biegler and Swift (1979) have proposed that the initial step in the oxidation process is the slow adsorption of water: FeS2 + H 2 0 = F e S 2 ( O H ) a d s + H + + e" (19) This is followed by deprotonation according to: FeS2 ( OH )ads = FeS2 ( O ) a d s + H + + e" (20) Hence, reactions (19) and (20) provide some insight into the way in which sulfate-oxygen may be derived from water. Type E2: Garrels and Thompson (1960) observed reaction (21) to limit pyrite oxidation at low potentials (0.25-0.4 V): 17 FcS2 —> Fe"2 + 2 S° ( s u r f a c c ) + 2 c" (21) S°(suiface) represents a layer of S° formed at the surface and is often termed a passive film. The reason (21) may have a limiting effect on pyrite oxidation is as follows. Recall that minimal S° is produced at room temperature. However, once formed, S°( S U I f a c c ) may remain as S°(surfacc) o r ' l m a Y convert to sulfate depending on the conditions of the experiment. There exists much evidence that shows S° does not easily convert to sulfate in low pH solutions or at temperatures below the melting point of sulfur (112.8 °C) (Biegler and Swift, 1979; Najdeker and Bishop, 1973; Lowson, 1982; Mishra and Osseo-Asare, 1988; Brocket al., 1976; Clark, 1966). The rate of conversion of sulfide-S to SU( s u rf a c <») is slightly greater than that of S°( s u r f a c e ) to SO4" 2 (Nordstrom, 1983). Using, cyclic voitammetry, Hamilton and Woods (1981) have observed multilayers of sulfur on pyrite surfaces oxidised in an acidic solution ( 0.5 M acetic acid and 0.5 M sodium acetate with pH = 4.6 ). At higher pH values (9.2 and 13 ), they found the sulfur to be restricted to approximately one monolayer. From work on newly polished surfaces of pyrite, it was concluded that high acidity experiments under reducing conditions do favor the formation of elemental sulfur on pyrite (see type E5, p.21) (Lowson, 1982; Mathews and Robins, 1974). The fact that pyrite does not actively corrode has been attributed to the formation of a passive film (Lowson, 1982). Pyrite has a molar volume less than that of its sulfur content when in the S°form ( 24 vs 32.7 cm3). Therefore, if S° forms at the surface, then it is able to protect pyrite from oxidation given the S° does not oxidize to sulfate readily. Moreover, S° is hydrophobic, and therefore, would resist attack from water. According to Lowson (1982), a perfect fitting layer remains intact if less than 27 % of the pyrite decomposes simultaneously to form S° and any excess S° is oxidized to sulfate. At values greater than 27 % , S° begins to crack because of strain thus affording less protection. Unless the formation of S° is site specific, there can be accumulation over the entire surface. As long as the S° remains at the surface , it will evade detection in the solution phase. A protective layer is not necessarily expected from a chemical oxidation mechanism. 18 According to (21), electrons arc produced at the pyrite surface. These electrons must now be conducted to a cathodic site where 0 2 is reduced : 0 2 + 2 H' + 2e" - » H 2 0 2 . (22) Overall, the electrochemical reaction is FeS2 + 0 2 + 2H* - » Fe2" + S° + H 2 0 2 (23) The cathodic reaction (22) above is thought to be the sum of (24) and (25) 0 2 + e" -> 0 2 " (24) 0 2 ' * 211" + e" -> H 2 0 2 (25) where (24) may be rate limiting at pH < 4 (Biegler and Swift, 1979; Biegler et al., 1975). It is interesting that the formation of peroxide from the reduction of oxygen in acid media is not thermodynamically favorable, thus limiting the role of reactions (16), (17) and (22). Peroxide may be an intermediate only in the series of elementary steps that eventually lead to reaction (26) below (Lowson, 1982). Regardless of pH, Biegler et al. (1975) have found pyrite to be a good electrocatalyst for 0 2 reduction. However, Biegler (1976 b) has found that pyrite specimens from different sources show no direct correlation between their semiconducting properties and the rate at which they can reduce 0 2 . This observation was attributed to the nature and amount of trace impurities which serve to perturb the pyrite crystals ideal structure. In their work on the characterization of semiconducting properties of naturally occurring pyrite, Lalvani et al. (1990) concluded that impurities can impart metal-like behaviour to pyrite. Also, these impurities are thought to be responsible for the lack of consistency in pyrite's physical and chemical properties. There exists evidence that reaction (21) might not occur as written (see Type E5, p.21) though this evidence would not discredit the formation of S° by some other route. Type E3: According to Lowson (1982), there exists a number of possible cathodic reactions. Based on a systematic thermodynamic argument; however, Lowson points out that the principle cathodic reaction occurring at the pyrite surface should be: 19 0 2 + 4 H + + 4 e" -> 2 H 2 0 (26) This half cell reaction is actually the sum of four elementary steps, of which reaction (24) is limiting. Bailey and Peters (1976) have shown through the use of O that all reactant oxygen becomes water in acidic solutions at H O C and between 4 and 7 MPa. Under these conditions, the mechanism seems entirely electrochemical and produces an S° layer at the pyrite surface. The principle anodic reaction suggested is: FeS2 + 8H 2 0 - » Fe 3 + + 2 S0 4 2" + 16 H+ + 15 e" . (27) If we consider the oxidation of ferrous iron; Fe 2 + + 1/4 0 2 + H + - » Fe 3 + + 1/2 H 2 0 (28) then, by combining reactions (26), (27) and (28) according to: 1 (27) + 15/4 (26) - (28) = (31) (29) 1 (27) + 15/4 (26) - 15 (28) = (32) (30) we get, FeS2 + 7/2 0 2 + H 2 0 Fe 2 + + 2 S0 4 " + 2 H + (31) FeS2 + 14 Fe3" + 8 H 2 0 15 Fe2" 2 S0 4 " + 16 H + (32) where reactions (31) and (32) may represent the overall stoichiometry of aqueous abiotic pyrite oxidation (Moses and Herman, 1991; Lowson, 1982). The actual potential exhibited by a mineral is called the open circuit or rest potential, which is the sum of all reversible potentials of the individual reactions and which is independent of the semiconducting properties of the mineral (Lowson, 1982). Moses and Herman (1991) point out that the open circuit potential for pyrite is midway between the potentials of the dominant anodic and cathodic reactions. Hence, pyrite oxidation may be under mixed control. This is supported by the fact that pyrite has a small band gap and high surface concentration of electrons and positive holes thus making pyrite conducive to cathodic and anodic reactions at its surface. According to Lowson (1982). the potential for the dominant anodic reaction is thought to be dependent upon pH, ferric iron and sulfate; whereas, the potential for the dominant cathodic 20 reaction is believed to be dependent upon pH and the partial pressure of oxygen. Majima and Peters (Lowson, 1982) surmise that adsorbed 0 2 appears to be responsible for the high rest potential of pyrite ( ~ 0.62 V), which is in contrast to pyrite's lower rest potential of ~ 0.25 V in the presence of He (g). Consequently, under anodic conditions, irreversibility is so extreme that the pyrite surface may be termed passive in the presence of 0 2 . Upon closer investigation of reaction (27), Meyer (1979) points out that the anodic reaction may proceed via two parallel reactions. The first is the oxidation of ferrous iron (28), the 2 ' 2 other being the oxidation of S 2 " through a number, of thio intermediates such as S 2 0 2 "and S 20 3 2". Type E4: An electrochemical mechanism may involve a galvanic effect as a result of passive and active sites on the pyrite surface (Mathews and Robins, 1974). Those sites relatively deficient in sorbed 0 2 form the anodic sites where the pyrite is dissolved. Cathodic sites are where 0 2 is reversibly reduced to hydroxyl ions according to the rate determining step: 1/2 0 2 + H 2 0 + 2 e" = 2 OH" (33) In more recent electrode studies, only sulfate, and not S°, has been found to form at the anode suggesting reaction (26) to be the cathodic half cell reaction (Mathews and Robins, 1974) TypeE5: Fornasiero et al. (1992) have conducted an electrokinetic study on pyrite oxidation. They have proposed the following mechanism for the incohgruent oxidation of pyrite: i) the formation of an iron-deficient layer according to: FeS2 = x Fe 2 + + Fe(i.X) S2 (34) ii) the oxidation of the ferrous iron: Fe 2 + = Fe 3 + + e" (35) iii) the hydrolysis of the ferric iron to various ferric hydroxide species whose concentrations are pH dependent, iv) the preciptation of iron hydroxide species onto the pyrite surface. 21 The balancing cathodic reaction is (26). Reaction (34) docs not implicitly allow for the formation of S° although disulfides could conceivably regroup to form S° or S° "2 species. Chandcrs and Briccno (1987) offer more insight into reaction (34) They suggest that the migrating iron atoms could oxidize to form some layer of iron hydroxides depending on pH. The resulting iron oxide layer may or may not be uniform or porous. In a later study, Pang et al. (1990) proposed that the underlying S-rich layer could develop cracks and microfissurcs as this layer begins to oxidize to various oxysulfide forms. Under heavy oxidation, the porous sulfur layer, could disappear. The overlying oxidized iron layer can and most likely acts as a diffusion barrier to both incoming oxidants and outgoing reaction products. If smooth, this outermost iron layer decreases surface area, thus reducing the rate of any subsequent reactions. Other workers propose that the sulfur rich layer may be either an S 8 (Sato, 1960) or some type of non-Sg layer (Peters, 1984). Alternatively, Goldhaber (1983) suggested there may be the preferential release of sulfur in which case, the sulfur would oxidize first then detach from the surface as a sulfoxy anion. The remaining surface is thought to be a layer of iron hydroxides. Type E6: Mishra and Osseo - Asare (1988) have studied the electrochemical dissolution of pyrite and have proposed the following mechanism: FeS2 + H 2 0 + h" = Fe(OH)S2 + F f (36) Fe(OH)S2 = FeS2(OH) (37) FeS2(OH) + 3 H 2 0 + 3 h + -> Fe(OH)2*S2(OH)2 +3 H + (38) Fe(OH)2*S2(OH)2 + 2 h" Fe 2 + + S 20 3 2" + 2 f f + H 2 0 (39) Overall, we get, FeS2 + 3 H 2 0 + 6 h + -> Fe 2 + + 6 H + + S 20 3 2" (40) where, h is a semiconductor hole in the valence band. (Recall that the saturation region of pyrite is satisfied at room temperature.) Mishra and Osseo-Asare (1988) explain this mechanism as follows. Initially, there is the electroadsorption of OH" which can undergo reversible and irreversible electroadsorption / electrodesorption at the pyrite surface. This forms a Helmholtz 22 double layer. The adsorption / desorption phenomena is attributed to the presence of Fe d electrons in the upper valence band (eg ) which capture and transfer the OH" groups to sulfur sites (S22) on the surface. In so doing, the d states play an important catalytic role in the oxidation of pyrite. Sulfur p states are thought to play a similar catalytic role during the oxidation of PbS. Mishra and Osseo-Asare (1988) further investigated the fate of the thiosulfate formed in reaction (40), and proposed the following mechanism: S 20 3 2" —1-> Sn062" + HSO3" - » S0 4 2 ' (41) S 20 3 2" + H + —.2 -> S° + HS03" — 3-> S0 4 2" (42) Step I is anodic oxidation, the rate of which is proportional to the anodic potential. Step 2 does nOt depend on the anodic potential, and is inhibited by high pH (6-9) leaving thiosulfate to react with 0 2 and H 2 0 to become polythionates and sulfites, both of which will convert to sulfate. The polythionates are relatively stable in acidic solutions. At pH 9, thiosulfate is the main product because of its kinetic stability in alkaline media. S° was formed only in acidic and high temperature experiments. Also, step. 2 shows that, at most,, only 50% of pyrite - sulfur can become S°. In step 3, it is the bisulfite that is further oxidized to sulfate. Steps 1 and 3 are mass transfer controlled. In a later study, Mishra and Osseo - Asare (1992) proposed a mechanism to account for the current - potential behaviour for the reduction of oxidants based on their results of cyclic voltammetry experiments. They put forth an isoenergetic charge transfer model wherein electrons are transferred directly from the conduction band edge to empty levels of the oxidant. The relative positions of the energy states on both sides of the interface are constant. The concentration of electrons at the conduction band edge depends on the concentration of electrons in the bulk of the, solid phase and on the potential difference across the space charge layer. The current produced depends on the concentration of oxidant. This model is quite similar to that described by Luther who used molecular orbital theory to explain charge transfer from the pyrite surface to oxygen (see p. 16). A second model is described wherein charge transfer is afforded via surface states. This model is actually borrowed from the combined efforts of Vandermolan et al. (1980) and Dcsilvcstro et al. (1986). Vandcrmolan et al. proposed that an electron can be trapped from the conduction band edge to the surface states in the band gap region (step 1). Then electron exchange occurs between the surface states and the oxidani (step 2). A third step is suggested by Desilvestro et al.. Here, mass transfer of the reduced species through the electrohie is considered. Now, if step 1 is rate determining, the current will be independent of the stirring rate and the concentration of the oxidant. If step 2 is rate determining, then the current depends on the concentration of oxidant. But if step 3 limits the rate of oxidation- then the current will depend on the stirring rate and the concentration of oxidant. 4) Summary. Despite the vast amount of research dooe on the aqueous oxidation of p>iite, there exists very little agreement as to mechanism or products of reaction. Nevertheless, some common features of the oxidation process, whether chemical or electrochemical, are as follows: 1) All suggested mechanisms begin with the convenient assumption that pyrite :s surface is ideal. 2) Oxygen is involved in the oxidation process. Most mechanisms suggest oxygen adsorbs to the surface. Its adsorption may or may not be rate limiting and may be at a cathodic site on the pyrite surface. The product of this cathodic reaction may be H2O2 (p.15, 18), H2O (p.19) or OH" (p.21). Water is involved in the oxidation and does adsorb to the surface. Its adsorption may be rate limiting. 3) Sulfate-oxygen is most likely derived from waier (electrochemical method) rather than from DO (chemical method) in low pH solutions.. 4) Under some conditions, S° is believed to be an intermediate; whereas, under other conditions, S° is thought to be an end-product. The production of elemental sulfur in solution is minimal at room temperature but is favored in high acidity solutions under reducing conditions. Also, the rate of conversion of S° to sulfate is very slow at room temperature and in low pH solutions. Finally, the conversion rate of sulfide-S to S° is only slight!) greater than that of S° to sulfate. S° never exceeds yields of 50 %. 24 5) Some type of sulfur layer is formed at the pyritic surface and plays a protective role in the oxidation mechanism. 6) Most elementary steps reviewed do not provide any mechanistic details on the oxidation process. Most workers provide reactions only. 7) No studies, to my knowledge, combine spectroscopic data of the pyrite surface with that from the solution phase to better understand the incongruent dissolution of pyrite. 25 Kb) 1 (a) Figure 1. The unit cell of pyrite. (a) A schematic of the S-S positions in the unit cell. The line segments drawn represent the S-S internuclear axes, (b) A representation of the immediate distorted tetrahedral environment around the S2 2 moiety. The large spheres represent the sulfur atoms, (c) Positions of the iron atoms in the unit cell. Z = 0 refers to the XY plane at Z = 0. In both (a) and (c) the dimension along the Z-axis has been enlarged four times for clarity. All drawings are not to scale. 26 Pyrite MO's Core Levels Figure 2. A qualitative molecular orbital (MO) diagram for pyrite. AO and MO refer to atomic and molecular orbitals respectively. Superscripts indicate the number of electrons present at a given energy level. E\; refers to the Fermi energy level. A schematic band diagram is drawn alongside the molecular orbital diagram. For more explanations, see the text. 27 0- 0. a. 3(b) Figure 3. A (100) view of pyrite's unit cell, (a) Z = 0 plane, (b) Z = 0.5 plane. In both (a) and (b), ellipses represent sulfur atoms; whereas, the center of each octahedron represents an iron atom. Dotted lines represent the S-S internuclear axes. Drawings are not to scale. 28 Figure 4. An example of a stepped and unreconstructed (010) surface. Line segment ABCDE represents the steps. The coordination number of some surface iron atoms has decreased, and therefore, these iron atoms are more atomic-like than their counterparts located further from the surface. Also, the potential for dangling sulfur atoms to regroup into elemental sulfur is evident. 29 Chapter Two: Thesis Objectives Although much work has been carried out on the oxidation of pyrite (FeS2), very little is known regarding the incongruent dissolution phase of this noble sulfide. Moreover, most oxidation experiments that have been carried out have been performed on crushed samples of the sulfide. The common practice of crushing a mineral serves to introduce large amounts of defects in the crystal structure. It is common knowledge that the oxidation rate of any mineral is greatly affected by the type and extent of defects present in the crystal structure. Defects, whether natural or induced, form sites of high surface energy, and therefore, can be extremely labile. Though methods of sample preparation apparently exist for the removal of defects introduced by the mechanical crushing of samples, if a method could be developed that does not employ crushed samples but rather leaves a sample in its natural state then this methodology would certainly be favorable. To my knowledge, little definitive work has been performed on the oxidation of pyrite which links surface data with that of the liquid phase. In this context, surface data is defined as information on the changes in the S/Fe ratio and changes in the chemical states of both sulfur and iron at the solid surface as the mineral undergoes oxidation; liquid phase data is defined as the information on the changes in the sulfate and total iron concentrations during the course of oxidation. Correlating surface data to that of the solution phase can only serve to improve our understanding of the oxidation mechanism. Finally, the main weathering product identified in the weathering of pyrite is the mineral jarosite with a general formula of M+Fe3(S04)2(OH)6 where M is K, Na or H 3 0 (Lindsay, 1979). Below pH 3, a value not uncommon for acid-sulfate soils and acid mine drainage (Moses et al., 1987), jarosite has a Ksp value of approximately 10~^u (Lindsay, 1979). Therefore, to avoid contamination of the pyritic surface, a pH of 1 was selected for this study. Given the above comments, the objectives for this thesis are as follows: a) To subject pyrite to oxidative dissolution for 3 days under abiotic conditions and at pH = 1, T = 25 °C, P = 101.3 kPa, P 0 2 = 20 % and P C 0 2 = 0%. 30 b) To correlate surface data with solution data in an effort to propose a general mechanism for the incongruent dissolution of pyrite. Surface data, as defined above, will be gathered from x-ray photoclcctron spectroscopy; whereas, in the solution phase, total sulfur concentrations will be determined from analysis of barium sulfate precipitate using the method of nephelometry, while total iron concentrations will be determined from atomic absorption. Solid and liquid phase samples will be collected at twelve hour intervals. c) To develop a method for studying mineral dissolution wherein the mineral is not crushed but rather maintains. its natural crystal state, as much as possible, throughout the course of the dissolution experiment. Finally, though not a direct objective, but a necessity, is the development of a method for the determination of low-level sulfate. This thesis is semi-quantitative. The specific surface area of the pyrite samples may be assumed to be quite low, and therefore, cannot be readily determined by conventional techniques. Consequently, a normalised rate law cannot be calculated. Furthermore, concentration of the products of the dissolution reaction are expected to be very low. Therefore, speciation of sulfur and iron is not warranted. In fact, it is assumed that total sulfur and iron concentrations will be extremely low themselves. 31 Chapter Three: Materials and Methods. A. Materials 1) Reagents. a) Reagents used in the dissolution solvent: - Distilled water. This water is derived from a Fisher Scientific Milli-Q pure water polishing system which consists of a 4-bowl organic extraction system (CDOF 012 05). The filter cartridges include: 1 highly activated carbon filter which effectively removes organics and chlorine, 2 ion-exchange resins each with 1076 cc of acid - base nuclear grade mixed-bed resin, 1 organic extraction filter and 1 hollow fibre (0.2 um) which is radiation sterilized. Distilled water will be referred to simply as water. - Hydrochloric acid (Fisher Scientific, ACS. ) . After ignition, the maximum levels of, sulfate, sulfite and iron are 1, 1 and 0.2 ppm respectively. b) Reagents used for iron determination: The only reagent used, in addition to water, was the iron reference solution used in preparing the standards. This solution is a certified Fisher Scientific product (CSI 124-500) suitable for atomic absorption spectroscopy. The solute in the solution is iron powder in a 2-5% nitric acid solvent. c) Reagents used for sulfate determination: Two methods for sulfate determination were employed in this experiment - turbidimetry and nephelometry. Each method uses different reagents, i) Turbidimetry. - hydrochloric acid (6 M) (as above) - barium jello consisting of (a) water, (b) 225 bloom calf skin gelatin (Aldrich 27,162-4), (c) polyvinyl alcohol (98% hydrolysed) (Aldrich 18,967-7), (d) certified A.C.S. barium chloride dihydrate from Fisher Scientific. - sulfate reference solution, for standards, which consists of water and Analar grade anhydrous sodium sulfate from BDH. 32 ii) Ncphelometry. - conditioning reagent consisting of (a) glycerol, (b) concentrated HC1, (c) water, (d) denatured alcohol, (e) certified A.C.S. sodium chloride from B.D.H. ( containing 0.004% sulfate) - barium chloride dihydrate (as above) - sulfate reference solution for standards (as above) 2) The Pyrite Samples. Fifteen specimens of striated cubic pyrite crystals (~lcm3) were purchased from the geological museum located at the University of British Columbia. These crystals were reported to have been collected from Navajun, Spain. The exact location is not known. Recall that each face of cubic pyrite is equally reactive (Klein and Hurlbut, T985, p.285), therefore, any face could be used in the present study. All faces of each crystal were cut to a thickness of ~ 1 mm with a water-cooled diamond saw then polished with a 1 micron diamond paste. All work was performed by the technical staff of the Department of Geological Sciences at the University of British Columbia. It was these cut and polished faces that were used in the dissolution experiment. 3) The Reaction Vessel and Supporting Equipment. The equipment used in the dissolution experiment will be listed in the same order as presented in Figure 5 (p.50). 1) Ultra high purity nitrogen gas cylinder fitted with a two-stage regulator. The nitrogen gas was used to purge the glove box. 2) OxiClear disposable gas purifier. This gas purifier reduces the oxygen and water content in the nitrogen gas to below 50 ppb. 3) Omega flow meter (# 32301-32400) with a glass float. This flow meter was used to track air flow into the glove box during its purging thus reducing the possibility of over pressurising the glove box. 4) Cole Parmer one way gas valve (# 6393-60). This valve was used to ensure unidirectional flow of nitrogen gas into the glove box during purging. 5) Glove box. This plexiglass box (20 in. x 24 in. x 18 in.; volume ~ 143 L) was designed and constructed by the author. Joints were scaled initially with dichloroethane then with clear marine silicone. A removable lid was secured to the rest of the box by twenty bolts (4 at each end and 6 along each length). Heavy black neoprene gloves were fitted to the front of the box and sealed with plexiglass rings each fastened to the box by eight small bolts and nuts. Silicone was then applied around each ring. The glove box was always kept under a slight positive pressure of nitrogen gas. This positive pressure could be easily verified by simply observing the gloves which would protrude outward slightly from the face of the box. -6) Reaction vessel. The reaction vessel consists of a 500 mL Erlennieyer flask tightly fitted with a #9 black stopper! Tw'o holes were bored through the stopper and fitted with glass rods (1/16 inch inside diameter) - one for gas inflow, the other for gas outflow. Silicone was used to seal the glass rods to the stopper to prevent any gas leakage. The tip of the air inflow rod had been heated over an open flame in order to reduce its inside diameter to that of the size of a capillar)'. Consequently, as air was forced through the capillary, a jet stream was formed thus producing a continuous and effective swirling action, therefore, no stir bar was required for mixing. 7) Omega flow meter (same as above). This meter was used to regulate the air flow into the reaction vessel. 8) ThomasR Ascarite II (8 - 20 mesh). This ascarite consists of a coating of solid sodium hydroxide (CAS 1310-73-2) on a non-fibrous silicate carrier (CAS 1318-00-9). The ascarite would undergo a color change from light brown to white upon adsorption of CO2 and formation of sodium carbonate. 9) Drierite (#23005). This material consists of 97% anhydrous calcium sulfate and 3% cobalt (II) chloride. Because the compressed laboratory air was known to contain moisture, it was felt that the air should be dried prior to entering the reaction vessel. 10) Compressed laboratory air. 11) Cole Parmer one way gas valve (same as above). This valve was used to ensure that no gases from outside the glove box would enter the reaction vessel. 34 12) Cole Parmer one-way gas valve (same as above). This valve was used as an outflow valve for the nitrogen gas that was used in purging the glove box. This valve was controlled by the vaccuum system afforded by the water tap. Therefore, if the flow of nitrogen gas into the glove box was held constant, then varying the flow of water would help regulate the pressure of gas in the glOve box. 13) Omega flow meter (same as above). This meter was used to help regulate the gas outflow from the glove box. 14) Vaccuum. 15) Constant temperature room. B) Methods. 1) Glassware Cleaning Procedure All glassware was cleaned according to the following procedure: 1) acid washed in 4 M HC1 overnight, 2) rinsed with water, 3) rinsed in 6 M HC1 (twice) 4) rinsed with water (four times). 2) Various Pyrite Cleaning Procedures. To establish control samples, three separate cleaning methods were employed in trial runs. These are labelled simply as methods A, B and C. Method A is a synthesis of several published methods commonly used in pyrite oxidation studies (Goldhaber, 1983; McKibben and Barnes, 1986; Moses et al., 1987; Moses and Herman, 1991). Methods B and C were my modifications of method A. The cleaning procedures were considered acceptable if they met the following two criteria: (1) removed any oxides from the pyrite surfaces and (2) did not alter the sulfur to iron mole ratio from that of an uncleaned sample. To know whether these criteria have been met will be discussed later. Method A. - In a fume hood, boil 4-5 pyrite samples in 400 mL of 6 M HC1 (4-6 hours). 35 - Rinse with 300 mL of boiling 6 M HC1 twice or until no yellow color appears in the rinse solution. - Rinse three times with 200 mL of hot acetone to dehydrate the surfaces. - Clean ultrasonically with 300 mL of denatured alcohol (85% ethanol and 15% methanol) for 90 seconds to remove any adhering powder. - Rinse with 300 mL of 1 M HNO3 (1-2 minutes) to dissolve any remaining sharp edges or damaged surfaces. - Rinse twice with 300 mL of room temperature water. - Rinse once with 200 mL of denatured alcohol to dehydrate the surfaces. - Under a heavy stream of nitrogen gas, transfer the samples into a plastic bottle. Seal the bottle with Teflon tape. The bottle contains Drierite to absorb any moisture. The samples were not in contact with the Drierite . - For a final cleaning, transfer samples to glove box which is being continuously purged with nitrogen gas. - In glove box, transfer samples from the plastic bottle to a 400 mL beaker. - Rinse three times with 200 mL of hot 6 M HC1. - Rinse three times with 200 mL of hot acetone. - Store samples into a bottle as above. - Using a glove bag which is sealed to the XPS transfer chamber and which is also being continuously purged with a heavy stream of nitrogen gas, samples were mounted onto the sample holder then transferred into the pre-vaccuum chamber of the spectrometer. Method B. This method is essentially the same as that described above. The only changes were that concentrated HC1 was used and neither HNO3 or water were used. The reason I chose to use concentrated HC1 instead of 6 M HC1 was that I wished to reduce the amount of water used in the cleaning process. Water is thought to be an important reactant in the electrochemical oxidation of pyrite. For this same reason, the water rinses were omitted in this cleaning method. The reason I 36 chose nol to use the HNO3 was simply because this acid is often used for crushed samples to remove any sharp edges. My samples were not crushed, and therefore, sharp edges were of no concern. Method C. - In a fume hood, boil 4-5 samples in 400 mL of 6 M HC1 (1 hour). - Rinse with 300 mL of hot 6 M HCI twice or until no yellow colour appears in the rinse solution. - Rinse twice with 300 mL of room temperature water. Decant. - Under a heavy stream of nitrogen gas, transfer samples from the beaker onto a KimwipeR and dry samples with the nitrogen gas. - Under the stream of nitrogen gas, store samples in a plastic bottle. Seal the bottle with TeflonR tape. This bottle has no Drierite in it. - Transfer samples into the pre-vaccuum chamber of the x-ray photoelectron spectrometer as above. The.reason water was reintroduced into the cleaning procedure will be discussed later. Moreover, this third method was further modified. Instead of boiling the samples in the acid for 1 hour, some samples were boiled for 1, 15 and 30 minutes only to see if the two criteria mentioned above could still be met by these shorter time periods. Results will be discussed later. 3) Monitoring the Conditions of the Experiment. This section describes the operating conditions of the dissolution experiment, ie., the conditions of the dissolution solvent, the oxidizing gas, the sampling times, and the instuments used to measure these conditions. a) Conditions of the Dissolution Solvent. The pyrite samples were introduced into a solution which had the following properties: - volume = 500 mLs - pH a 1.00 + 0.02. This solution pH was obtained by acidifying water with HCI. The pH was found not to vary more than + 0.1 pH units during the course of the experiment. The pH values were obtained with an Orion pH meter (420A) which was equipped with an Orion pH Triodo™ 37 electrode (# 91-57BN; Ag/AgCl internal reference electrode with a built in thermistor for automatic temperature compensation). - temperature = 25 °C. The temperature could be obtained from the pH meter. Recall that the glove box was housed in a constant temperature room which was set to 25 +1 °C. - pressure = 1 atmosphere. - ionic strength = 0.1 molal - oxidizing gas = dry CCVfree air. The solubility of the oxygen in the solution can be calculated to be ~ 8.5 mg/L given a water vapour correction of 0.0313 aim. at T = 25 °C. Eh measurements were difficult to obtain; in fact, a number of attempts to measure redox potentials failed. Results showed no consistency; in fact, readings would drift continuously. It is well known that Eh measurements are difficult to obtain in natural systems. Sulfides can poison a platinum electrode in reducing environments; whereas, the electrode may behave as an oxide electrode and respond to pH in oxidising environments. Furthermore, if redox species are too low in concentration, the electrode may very well have difficulty in responding to such species. I suspect the latter condition is the reason for the lack of consistency in the Eh readings. Eh readings were attempted with the same Orion meter as described above equipped with a platinum electrode # 96-78-00. Sampling times were 0, 12, 24, 36, 48, 63 and 72 hours. 4) Sample Preparation for Characterization and Instrument Settings. Characterization of the samples is operationally defined as the determination of the purity of the samples. Pyrite is known to contain several trace impurities and is often associated with a number of other mineral phases present as inclusions. With respect to trace impurities, these have an undefined effect on the chemical and physical properties of pyrite. Therefore, though obtaining an inventory of the impurities and inclusions allows us to determine the chemical status of the samples, this knowledge would be difficult to assess as to understanding the dissolution process. To characterize the samples, four instrumental techniques were employed: (1) 38 x-ray diffraction (XRD), (2) x-ray fluorescence (XRF), (3) scanning electron microscopy with an energy dispersive x-ray analyser (SEM/EDXA), and (4) x-ray photoclcclron spectroscopy (XPS). Recall that the original cubes of pyrite had had their faces cut and polished. Several of the remaining portions of the crystals were crushed in an agate mortar. The size of the powder was not determined because this information was not considered to be relevant. This powder was used for the XRD and XRF analysis, and therefore, neither of these techniques are surface-sensitive. SEM/EDXA, on the other hand, will be used to qualitatively identify inclusions that are visible with the naked eye. Only in this sense is SEM/EDXA considered surface-sensitive. Otherwise; in general, EDXA is not surface-sensitive having a. sampling depth of approximately 1 u.fn. XPS is the only truly surface-sensitive technique employed throughout this study. Although its sampling depth is approximately 100 A, only the elastically scattered photoelectrons will yield the desired photoemission peaks. The calculation for the inelastic mean free path for pyrite will be discussed later (p.51). No attempt to clean the samples were made except in the case of the XPS measurements. (a) X-Ray Diffraction. This method is qualitative in nature and was used to determine the presence of mineral phases other than pyrite. If gangue mineral phases are present in small quantities only, their peaks may not be easily discernible by this method. A 1.5 cm X 1.5 cm peice of double-sided sticky tape was placed on the centre of a glass slide. Powdered pyrite was sprinkled onto the tape and pressed to yield a flat and uniform layer. The glass slide was then placed into the goniometer. Using Cu (Ka) radiation (A. = 1.542 A), three samples were analysed. The x-ray diffractometer is a Philips product (PW 1050/25), and was made available by the department of Soil Science at the University of British Columbia. (b) X-Ray Fluorescence. X-ray fluorescence is quantitative in nature and was used to detect the type and amounts (ppm ± 5%) of minor elements in three powdered samples. Minor elements include Nb, Zr, Y, Sr, Rb, Pb, Zn, Ni, Cu, Co, Mn, V, Cr, Ba and Na. Sample preparation involved simply pressing 4.00 39 grams of powdered sample onto flat 32 mm diameter boric acid discs using a hydraulic press. The discs were then loaded into the XRF chamber for analysis. Calibration of the instrument was obtained by means of a series of international geochemical reference standards (Abbey, 1983). All calibration curves were linear over the full working range. The instrument is a computer-controlled Philips 1400 fully automatic X-ray spectrometer, and was made available by the department of Oceanography at the University of British Columbia. (c) Scanning Electron Microscopy with an Energy Dispersive X-Ray Analyser. This method was used in a qualitative mode to detect the type of mineral phases,.other than pyrite, that "were initially visible through a Zeiss (MC 63A) high powered compound microscope (500X magnification). Though the purpose of using this method in the qualitative mode is the same as that for XRD, SEM/EDXA is more sensitive because any small inclusions could be analysed individually. Six pyrite faces were mounted onto a glass slide. All six samples were then carbon coated to improve electrical conductivity between the samples and the instument. Carbon coating contaminates the samples making them useless for the dissolution experiment. The samples were then transferred to the instrument's chamber for analysis. The instrument is a Philips XL 30 and was made available by the department of Geology1 at the University of British Columbia. The accelerating voltage was set to 15.0 kV at a base pressure of 10 mbar. Incident electrons were produced by a LaBg thermal emission gun. (d) X-Ray Photoelectron Spectroscopy. This method is primarily intended to characterize the surface of the pyrite samples. Survey scans (0 - 1500 eV) were obtained using incident non-monochromatic Al K a x-rays (kinetic energy = 1486.6 eV) operating at 10 kV and 20 mA with an analyser pass energy of 192 eV (AE = constant) to collect the photoelectrons. The area of analysis was 2mm X 4mm. The 9 o operating pressure and temperature of the UHV chamber was 2x10 mbar and 25 C respectively. 40 Sample preparation for this method has been discussed earlier. The instrument is a Lcybold Max 200 and was made available by the department of Chemistry at the University of British Columbia. 5) Sample Initiation, Collection and Storage. Sample initiation, collection and storage refers to the method in which samples were introduced into the reaction vessel, after having being cleaned, removed from the reaction vessel, after having undergone dissolution, and stored in preparation for XPS analysis respectively. a) Sample Initiation. Immediately after the final rinse, 65 samples were transferred immediately, under a heavy stream of nitrogen gas, into the reaction vessel using tweezers coated with teflon tape. Of these 65 samples, six were viewed under the Zeiss microscope to ensure the absence of inclusions and then set aside in the reaction vessel for the XPS analysis. The reaction vessel was filled with 500 mLs of the dissolution solvent. The reaction vessel was then transferred to the glove box. The reaction vessel was tightly stoppered and the oxidizing gas was applied. The glove box was then sealed and purged for no less than 90 minutes at a flow rate of 2.5 L/minute. b) Sample Collection and Storage. At the time of collection, the glove box was opened and the reaction vessel unstoppered. Ten mLs of the solution was removed using a clean syringe and transferred to a plastic bottle. Clean stainless steel tweezers were used to remove one pyrite sample from the reaction vessel to the same plastic bottle which contained the solution. The reaction vessel was reinserted into the glove box and restoppered. Air flow was applied to the reaction vessel. At this point, however, the glove box would not be sealed and purged because I wished to deal with the newly removed sample immediately. Under a heavy stream of nitrogen gas, the sample in the plastic bottle was removed from its solution using another set of smaller and clean stainless steel tweezers. The sample was laid onto a Kimwipe under a stream of nitrogen gas, thoroughly dried, then stored in another small clean plastic bottle while it also was being purged with the nitrogen gas. This bottle was then scaled with TeflonR tape, and labelled appropriately. The solution remained in the 41 first plastic bottle which was then scaled. This solution was stored in a refrigerator (approximately 4 °C) to quench any reactions that may occur prior to analysis. Once all samples had been collected, each was separated into two 5 mL aliquots and analysed for total sulfate and total iron. 6) Analysis of the Solid Phase. This section will detail the various aspects of XPS that are pertinent to the experiment. X-ray photoelectron spectroscopy is also often termed Electron Spectroscopy for Chemical Analysis (ESCA). With respect to this experiment, XPS was used in both qualitative and quantitative modes, that is, the chemical; state.and mole ratios of surface species were determined. XPS provides useful information to aid in understanding the nature of atoms and molecules at a solid surface. In this XPS experiment, a beam of non-monochromatic Al Ka photons is incident on the pyrite surfaces. Core level electrons are ejected from the surface of the sample and collected in a hemispherical analyser before reaching the detector. The kinetic energy (KE) of the photqelectrons are measured from: KE = Ei - BEj - Osp (43) where, Ei = incident photon energy = 1486.6 eV BEj = binding energy of core level electrons for element i Osp = work function of the spectrometer. This is determined with reference to the Au photoemission peak whose binding energy is 84.0 eV. Because iron is known to exist in the sample, the Fe 2p3/2 peak of control samples was chosen as the reference peak. The experimental binding energy of this peak is 706.7 eV, which is in agreement with that of van der Heide et al. (1980). Once survey scans (0 - 1500 eV) had been obtained, narrow scans for the Fe 2p3/2, 1/2, 5 2p3/2, 1/2 and O Is photoemission peaks were collected for chemical state and S/Fe mole ratio determination. The lineshape analysis of a spectra includes (1) the binding energy of a photoemmission peak, and (2) the width of a peak at the half maximum intensity value. Earlier, two criteria had been listed to help determine whether or not a cleaning procedure was acceptable in providing control samples. It is the lineshape of a given spectrum that would serve as a guide in knowing whether or not a cleaning procedure had met the first criterium.For a given element, its lineshape will be affected by the chemical state of that element. Also, however, a lineshape can be affected by any differential charging that may occur in the solid. Differential charging results from poor conductivity of the sample and poor electrical contact between the sample and the spectrometer, a problem more inherent with powdered samples than with single flat crystal faces as used in this study. No samples used in this study exhibited differential charging. Before mole ratios could be calculated, background was subtracted using the Shirley non-linear background subtraction subroutine. Peaks were then integrated, then normalized relative to the Fe 2P 3/2 photoemission peak. The spectrophotometer measures absolute intensity (I) values for the iron, sulfur and oxygen photoemission peaks. From these values, absolute atomic density values (n) are calculated from the equation: n = I / S [atoms/cm3] (44) where S is a Leybold sensitivity factor for each individual photoemission peak of interest (SFe2 P = 3.01, Ss2P = 0.76, S o i s = 0.75) relative to that of F l s which is given a value of 1.0. Because the exact nature and history of the samples is unknown, and because samples do differ slightly from each other, absolute atomic density values should not be attempted because no reference point exists. For this reason, ratios are calculated as given in the following example for iron and sulfur: n (S) / n (Fe) = I (S) S (Fe) / S (S) I (Fe) (45) It is this type of calculation which is performed to determine whether or not a cleaning procedure has met the second criterium mentioned earlier. 7) Analysis of the Liquid Phase. a) Low-Level Sulfate Determination. In this study, the only sulfur species assumed to be in solution is sulfate. This assumption is justified later in the discussion. The conditions of the incongruent dissolution experiment are very restrictive in terms of the amount of sulfate expected to form considering the low surface area 43 of the pyrite samples, the relatively short sampling times and the relatively slow kinetics of dissolution. Therefore, a method for the determination of low-level sulfate had to be employed. In order to determine the upper limit of sulfate concentration to be expected from a three day dissolution experiment, preliminary turbidimetry tests were conducted in the Pedology laboratory using the automated Lachat QuickChemR AE method. Two very important discoveries were made during the course of these tests. First, results were not satisfactorily reproducible, and second, despite the lack of reproducibility, the maximum amount of sulfate to be expected was less than 2 ppm. These results confirmed the need.for a more precise methodology wherein low levels of sulfate could be measured with reasonable confidence ie., a method in which (1) the probability of losing any of the analyte during the course of its determination was kept to an absolute minimum, and (2) a reproducible standard curve could be obtained for the range of sulfate from 0 to 4 ppm. This section compares two techniques for low-level sulfate determination - nephelometry and turbidimetry. Both methods are based on radiative scattering, however, in nephelometry, the detector is placed out of the path of the incident radiation (usually at 90°); whereas, in turbidimetry, the detector is placed directly behind the cell and thus aligned with the incident radiation. Therefore, while nephelometry measures scattered radiation, turbidimetry measures a decrease in incident radiation (Braun, 1987, p. 427). b) Nephelometry. As polarised light of intensity Io passes through a solution, each molecule essentially becomes a source of light of the same wavelength as Io and scatters this light in all directions with intensity Is. The ratio (Is / Io ) is defined by the Rayleigh ratio, R Q , given by: R Q = i s r 2 / l o c o s 2 0 = (16 7 i 4 a 2 ) / ( X04) (46) where: i s is the distance from the scattering centre to the observer, 9 is the angle between the incident beam and the scattered beam, a is a constant (the polarizability of the particle), X0 is the wavelength of the incident light (546 nm). 44 The scattered beam is expected to be the result of Tyndall scattering given the large diameters of the barium sulfate colloids relative to the wavelength of the incident beam. When 0 = 90°, (4) becomes R 9 0 = [ T D / T T ( 1.049) h J n 2 F(Gs/Gw) (47) where: TD is the diffuse transmittance of the working standard filter. For our instrument. TD = 0.0172, h is the width of the diaphragm (1.20 cm), F is the product of the transmittances of the filters used in determining the scatter ratio, Gs is the meter reading of the scattered light measured at 90 degrees to the incident beam, Gw is the meter reading of the transmitted light measured directly into the incident beam. For colorless solutions, this is approximately equal to a meter reading of the incident light. n is the refractive index of the solution. For the purpose of this experiment, it is sufficient to use the latter portion of equation (47), that is, F (Gs / Gw), because the primary1 interest is in the ratio of scattered light to incident light. In practice, the intensity of the transmitted beam is several orders of magnitude greater than that of the scattered beam and as such it is not feasible to measure such a wide range with a single photomultiplier tube. Therefore, in addition to a calibrated working standard filter, there are four neutral density filters which all serve to attenuate the transmitted light. Each of these four density filters has a known transmittance value which is tabulated in Table 5 of Appendix 3 (p. 91). The following procedure is a modification of that found in Methods of Soil Analysis. Part 2 - Chemical and Microbiological Methods. 2 n d edition (1982) p. 175: 1) Conditioning reagent (colloid stabilizing reagent): While stirring, mix 50 mL glycerol with a solution containing 30 mL concentrated HC1, 300 mL denatured alcohol (85 % ethanol and 15 % methanol), 300 mL deionized water and 70 g sodium chloride. Stir for ten minutes. Not all the sodium chloride will have dissolved, therefore, decant and store the conditioning reagent into a one litre plastic container. In the referenced procedure, only 100 mLs of alcohol is suggested; however, because barium sulfate is known to have reduced solubility in alcohol, 300 mLs of the 45 alcohol was chosen for the sake of improved precipitation especially at the lower concentration levels. 2) Stock sulfate solution (1000 ppm): In a 1 L volumetric flask, place 800 mLs water. Dissolve 1.479 g anhydrous sodium sulfate. Make to volume and shake. 3) Working sulfate solution 1 (100 ppm): Volumetrically pipette 10 mL of the 1000 ppm stock sulfate solution into a 100 mL volumetric flask. Make to volume and shake. 4) Working sulfate solution 2 (10 ppm): 'Volumetrically pipette 10 mL of the working sulfate solution 2 into a 100 mL volumetric flask. Make to volume and shake. 5) Sulfate standards: Volumetrically pipene appropriate aliquots of working sulfate solution 2 to 100 mL volumetric flasks to prepare standards according to: 0, 0.5, 1.0, 1.5,.2:0, 2.5 ppm. 6) Transfer all standards to plastic 150 mL containers. 7) Add 10 mL of the conditioning reagent to samples and standards. Shake well. This can be done a day before readings are taken- Again, to improve precipitate formation and stability, 10 mLs of the conditioning reagent was employed instead of the suggested 5 mLs in the referenced procedure. The ratio of conditioning reagent to sample volume becomes 1:10. 8) Regarding the 5 mL aliquot set aside for total sulfate determination in the dissolution experiment, add 0.5 mL of conditioning reagent. 9) Add the barium chloride to the standards and samples according to the ratio: 0.3g / 100 mL solution. Shake for one minute, then pour into a spectrophotometer cell. Let stand for five minutes before taking any readings. 10) Obtain F, Gs and Gw values for each standard and sample. Each standard is measured five times randomly for quality control. The only equipment used throughout the nephelometry experiment was a light scattering photometer made available by the department of Chemistry at the University of British Columbia. Materials included a spectrophotometer cell with a 3 cm path length and a timer. For all standards, the following calculation were performed: (a) F(Gs/Gw) (48) 46 (b) F(Gs/Gw) a v e , b I i i n U = F(Gs/G\v) b l ! i n k / n (49) where n = number of replicates. (c) F(Gs/Gw)!ld j = [ F(Gs/G\v) - F(Gs/Gw)av<,. b I i i n k ] (50) where F(Gs/Gw) includes the blanks. (d) For each set of five replicates F(Gs/Gw) a d j > a v g = F(Gs/Gw)adj / n (51) (e) Plot F(Gs/Gw)adj,aVg vs. standards. (f) Perform linear regressions on the standards'data. c) Turbidimetry. At low concentrations, the ratio of incident radiation of intensity Io to unscattered (transmitted) radiation of intensity I can be determined from (Braun, 1987, p.426): S = - log (I / Io) = kbC (52) where, S is termed turbidance (to distinguish it from absorbance in Beer's law) k is the turbidity coefficient b is the path length (as in Beer's law) C is the concentration of the scattering particles. In this procedure, as well as in that of nephelometry, the dominant form of scattering particles is considered to be barium sulfate colloids. Turbidimetric calculations are most precise when the ratio (I/Io) are not near 0 or 1. For this reason, solutions which are either very dilute or very concentrated do not typically yield reproducible results. The method used in this study is completely automated. The only work required is the preparation of the reagents and standards. The following procedure has been taken from Lachat's Methods Manual for the QuickChem- Automated Ion Analyzer (method # 10-116-10-1-C). Preparation of Reagents: Make all solutions with deionized water (10 megohm). 47 1) Hydrochloric acid Slock, 6 M: In a fume hood, pour 50 mL of concentrated HCI into 50 mL water. Stir. 2) Working hydrochloric acid, 0.015 M. In a fume hood, dilute 2.5 mL of the stock acid to 1 liter of water. Stir. 3) Barium Jello: Boil ~ 800 mL water in a 1 liter beaker with graduations. Remove from the heat, add a stir bar and place on a magnetic stirrer. In a weighing boat, pour 4.0 g of 225 bloom calf skin gelatin, 1.2 g polyvinyl alcohol and 50 g barium chloride dihydrate. Add the contents of the weighing boat to the boiled water while stirring (30 minutes). Make to volume (1000 mLs). Filter through rinsed glass wool into a 1 L container. Allow to cool before using. The barium jello, which is slightly turbid, is actually barium sulfate formed from sulfate naturally occurring in the gelatin. This jello acts as a seed for the rapid analytical precipitation of barium sulfate from the analyte sulfate. Preparation of Standards: 1) Stock sulfate standard (1000 ppm): Into a 1 L volumetric flask, dissolve 1.479 g of anhydrous sodium sulfate in approximately 800 mL water. Make to volume and shake. 2) Working stock standard 1 (100 ppm): Into a 1 L volumetric flask, place approximately 800 mL water. Pipet exactly 100 mL of the stock solution. Make to volume and shake. 3) Working stock standard 2 (10 ppm): Into a 1 L volumetric flask, place approximately 800 mL water. Pipet exactly 100 mL of the working stock standard 1. Make to volume and shake. 4) Working standards (set of 5): Into five 250 mL volumetric flasks, pipet exactly 0.00, 25.0, 75.0, 125.0 and 250.00 mL respectively of the working stock standard 2. Make each to volume and shake. This makes standards of 0.00, 1.00, 3.00, 5.00, 10.00 ppm S O 4 2 " respectively. For quality control, 5 randon measurements of each standard were taken . The only instrument used was the Lachat Instruments' QuickChem Automated Ion Analyzer made available by the department of Soil Science at the University of British Columbia. 48 d) Total Iron Determination. Analysis of total iron was made with the use of an atomic absorption spectrophotometer supplied by the department of Soil Science at the University of British Columbia. Total iron can be assumed to be an indication of the ferrous iron concentration. This assumption is discussed later. The only procedure involved is the preparation of standards. There is no need to plot a standard curse because this is done internally by the spectrophotometer's software. Based on this internal standard curve, analyte absorption readings are displayed automatically as concentration values (ppm). Standards were prepared according to the following protocol: 1) Stock iron standard (1000 ppm ± 1% solution - a Fisher Scientific product as described earlier). 2) Working stock standard 1 (100 ppm Fe): Into a 1 L volumetric flask, place approximately 800 mL water. Pipet exactly 100 mL of the stock iron standard. Make to volume and shake. 3) Working stock standard 2 (10 ppm Fe): Into a 1 L volumetric flask, place approximately 800 mL water. Pipet exactly 100 mL of the working stock standard 1. Make to volume and shake. 4) Working standards (set of 5): Into five 250 mL volumetric flasks, pipet exactly 0.00, 25.0, 75.0, 125.0 and 250.00 mL respectively of the working stock standard 2. Make each to volume and shake. This makes standards of 0.00, 1.00, 3.00, 5.00, 10.00 ppm Fe respectively. With respect to the 5 mL aliquots set aside for total iron determination in the dissolution experiment, four absorption readings per aliquot were taken to check reproducibility. From these four readings, average iron concentrations were calculated. The spectrophotometer used was a Perkin Elmer model 306 made available by the department of Soil Science at the University of British Columbia. 49 Figure 5. A schematic diagram of the reaction vessel and supporting equipment used for the dissolution experiment. (1) nitrogen gas, (2) OxiClearR, (3, 7, 13) flow meters, (4, 11, 12) one R R way gas valves, (5) glove box. (6) reaction vessel, (8) Ascarite , (9) Drierite , (10) compressed air, (14) vaccuum, (15) constant temperature room. 50 Chapter Four: Results and Discussion. Prior to discussing the results of the dissolution experiment, several operational definitions, assumptions and their justifications are addressed. The purpose of this section is to bridge the gap between pure and applied science. That is, the samples used in the dissolution experiment are not well defined in terms of knowing their exact history and physico/chemical properties prior to subjecting them to the oxidation environment. Each sample differs slightly from the other and no attempt has been made to study the surface details of these differences. Yet, the samples are very good geologic samples representative of iron sulfides in natural systems. Therefore, in an effort to maintain internal consistence within the boundaries of the experiment, reasonable definitions and assumptions applicable to thisstudy are appropriate. A) Operational Definitions. 1) Incongruent Dissolution: This is the non-stoichiometric dissolution of a mineral. In general, this dissolution process is dependent upon the conditions of the experiment (temperature, pressure, pH, pE, ionic strength, P02 and Pco2) a n a upon the conditions of the samples (face of crystal, type and extent of impurities and defects, distribution of grain sizes (for crushed samples) and any sample pretreatment protocols). 2) Surface: This is a three dimensional structure impregnated with trace impurities and natural defects. Its depth can be approximated using an empirical approach devised by Seah and Dench (1979) for calculating the inelastic mean free path (Xe) of inorganic compounds: te = (2170/E2) + 0.72 (a E ) 0 - 5 (53) where, E = 1090 eV, and a is the monolayer thickness (the distance from the centre of the S-S ion to a near iron atom (- 0.2376 nm). This calculation yields a value of 11.6 monolayers or 2.75 nm. 3) Original Surface: This is the sample whose surface is physically and chemically unaltered from that of the original sample except that it has been polished in order to obtain reasonable XPS spectra. 4) Reference sample: This is a control sample whose surface will be defined as clean and to which all oxidised samples will be compared. The reference sample is not one particular sample 51 but rather an average sample in terms of the sulfur to iron mole ratio and chemical states of iron and sulfur. 5) Clean Surface: This is a surface which differs from that of the original sample in the following ways as determined from x-ray photoelectron spectra: - no significant iron or other sulfoxy oxide peaks observed, - binding energy for the Fe 2p3/2 _ S2p 1 / 2 , S2p 3 / 2 peaks are 106.1 eV, 163.4 and 162.4 eV respectively (Van der Heide et al., 1980). - sulfur to iron mole ratio is not significantly different from that of the original sample. B) Assumptions and Their Justifications. Assumption #1: The oxidizing medium is sterile, and therefore, there should be no concern about the presence of sulfur oxidizing bacteria (eg. Thiobacillus). This is justified on the basis that the water used throughout the experiment has been specially purified and filtered prior to use (p.32) and pH adjusted to unity. Also, all glassware has been carefully cleaned (p.35). Finally, all samples used in the dissolution experiment have been boiled in 6 M HC1 for no less than 10 minutes. Assumption #2: The only sulfur species present in solution is sulfate. Although this is not completely true, many studies show that sulfate is certainly the most dominant form of sulfur produced during the oxidation of pyrite at room temperature (pp. 12-13). As for the presence of elemental sulfur, this will be discussed later. Assumption #3: Ferrous iron is the dominant form of iron. Recall from the literature review that Singer and Stumm (1970) could not detect any ferric iron in their solutions even after several days of aqueous pyrite oxidation (p. 14). Assumption #4: The pyrite samples are pure yet they do not behave identically to one another. Pyrite is known to have variable sulfur content and to contain trace amounts of various impurities. However, although some work has showii that trace impurities can have some undefined effects on the semiconducting properties of pyrite (Beigler, 1976b and Lalvani et al., 1990) no studies, to my knowledge, have accurately assessed the effects that impurities or gangue 52 minerals have on the chemical or physical properties of pyrite during dissolution. Therefore, although results from the characterization of the pyrite samples will indicate the presence of trace impurities and gangue minerals, these will be ignored in terms of having any significant or adverse effects on the dissolution process. This is a general assumption and has been used by other workers (Buckley and Woods, 1987). C) The Cleaning Procedures. The purpose of this section is to determine whether or not the three cleaning procedures have met the two criteria mentioned earlier (p.34). 1) Chemical State of Iron and Sulfur. Figure 6 (p.65) compares narrow scan spectra of an original sample (229) with that of a cleaned sample (205). The iron peak (229) indicates the presence of iron oxides. No curve fitting (peak synthesis) was performed to determine the exact nature of the iron oxide(s) because (1) this was not an objective of the study number and (2) two possible oxides or combinations thereof are possible: (a) Fe-0 (709.3 eV), and (b) Fe203 (hematite; 711 eV), where the binding energies listed refer to the Fe 2p 3 / 2 photoemission peaks (Wagner et al., 1979). For this reason, any iron oxides are simply referred to as FeOx. Both FeOOH (710.8 eV) and Fe304 (magnetite; 711.2 eV) are definitely not possible oxides at the surface because both yield lineshapes that are distincly different from those obtained in this study. The Fe 2p3/2 spectrum (205) of Figure 6 does definitely illustrate that iron oxides have been significantly, but not completely, removed. The gently sloping tail on the high binding energy side of the Fe 2p 3 / 2 spectrum is a result of shake-up effects - a natural feature in XPS spectra. Figure 6 also illustrates the sulfur photoemission peaks for uncleaned (229) and cleaned (205) samples. The uncleaned sample does not show significant amounts of any oxides. Sulfate does have a binding energy at 168.25 eV but this peak is not clearly evident from Figure 6. From both sets of photoemission peaks of Figure 6, it appears that oxygen is preferentially bound to iron rather than to sulfur. This is despite the fact that the Fe-O bond 53 dissociation energy- (410 kJ/mol) is much less than that of S-O (522 kJ/mol). Bond dissociation energies are determined experimentally from heats of formation in the gaseous phase. In such a phase, no matrix effects exist, unlike in solids. Therefore, bond dissociation energies cannot be used to rationalize the preferential bonding of oxygen to iron. Figure 7 (p.67) illustrates the results of the three cleaning procedures as described earlier. In terms of chemical state, all three methods yield similar results. That is, all three methods remove oxides from the surface. The iron 2P 3/2 and sulfur 2P 3/2 binding energies are listed on p,51. Spectra for the 15 and 30 minute trials of method C have been purposely omitted because they were essentially identical to those illustrated in Figure 1. 2) Sulfur to IrOn Mole Ratio. Table 2 (p.55) tabulates the S/Fe mole ratios as calculated from equation (45) for the various cleaning procedures. Results from Table 2 indicate that the original sample (229) has S/Fe = 3.58. Methods A and C (1, 15 and 60 minutes) all yield similar ratios. Method B yielded an anomalously high ratio. The reason for this is uncertain. Perhaps the hot concentrated HCI. used in the cleaning procedure (p. 3 5) lowers the stability of surface iron atoms thus facilitating their diffusion from the surface crystal lattice into the cleaning solution. Method C (15 minutes) yielded a high ratio also, although, it is conceivable that this sample simply had a naturally high ratio. The reference S/Fe mole ratio can be taken to be ~ 3.6. This is an average of those samples cleaned by methods A (200), C (205) and of the uncleaned sample (229) as listed in Table 2. This ratio is significantly higher than one would expect from pyrite's empirical formula, but agrees reasonably well with that of Mycroft et al. (1990). Excess sulfur at the surface could regroup into elemental sulfur rings. Figure 8 (p.68) illustrates an XPS spectrum of a sample cleaned by method C (60 minutes). Curve fitting, using the Leybold Max 200 data system software package, was performed on this spectrum and indicates the presence of elemental sulfur. Table 3 (p. 56) tabulates the atomic density ratio S° / pyritic-S based upon the 2p3/2 peaks. Pyritic-S refers to the S22 moieties. From this data, approximately 7 % of sulfur is in the elemental form. Under 54 the UHV of the x-ray spectrometer and at room temperature, elemental sulfur is known to be volatile. In their XPS study on pyrite oxidation, Buckley and Woods (1987) Table 2. XPS results indicating the effect of the cleaning procedures on S'Fe, O S and O/Fe atomic ratios relative to those of an uncleaned sample. Sample Atomic Ratios (n) number S/Fe O/S O/Fe 229 3.58 1.30 4.64 200 3.71- 0.87 3.22 202 4.06 0.85 3.47 244 3.48 0.45 1.56 246 3.64 0.99 3.61 248 3.81 0.38 1.46 205 3.55 0.46 1.64 Each sample and the method by which it was cleaned is as follows: (229) uncleaned; (200) method A; (202) method B; (244) method C, 1 minute; (246) method C, 15 minutes; (248) method C, 30 minutes; (205) method C, 60 minutes. lowered the temperature of their UHV chamber to less than 200 K to prevent losses of elemental sulfur. Therefore, in this study the actual percentage of elemental sulfur at the crystal surface should be higher than 7% prior to placement into the UHV chamber of the spectrometer, but most likely not much higher given that the production rate of elemental sulfur in room temperature experiments is known to be low. Given that all samples had been placed into the UHV chamber for varying amounts of time, no reference point for quantifying the elemental sulfur exists. Therefore, no trends in S° formation can be determined. For some reason sulfur is able to maintain this disulfide form despite the apparent deficiency of iron atoms in the crystal lattice. 3) O/S and O/Fe Mole Ratios. 55 Tabic 2 tabulates the 0/S and O/Fe mole ratios obtained from XPS work. Relative to the uncleancd sample, only method C (1, 15 and 60 minutes) shows a substantial reduction in the O/S ratios. However, in all three methods, any oxygen remaining at the surface appears to be either physisorbed or chemisorbed only. This is evident from the fact that oxygen does exist at the surface yet no change in the chemical states of sulfur or iron has occurred. Therefore, methods A and B are just as effective as method C at significantly removing oxidized species from the pyrite surface. Because physisorption involves van der Waal interactions with low binding energies and low activation energies, any physisorbed oxygen should desorb quickly, and therefore, not interfere with any dissolution mechanism. Method C (30 minutes) shows a high O/S mole ratio. This is thought to be solely the result of contamination which may have occurred at some point in the overall preparation of the sample (cleaning, storing and/or transfer to the UHV chamber). The above statements also apply to the O/Fe mole ratio results. Table 3. XPS results from analysis of a cleaned sample (method C, 60 minutes) indicating the ratio of elemental sulfur to pyritic-sulfur (S°/S). Sulfur Species Binding Energy (2p 3/2) ( e V ± 0 . 3 eV) n(S°) / n(S) elemental (S°) 164.8 0.07 pyritic-S (S"1) 162.4 4) Summary. In terms of chemical state, each cleaning method is equally effective at removing oxides from the crystal surface. However, with regards to the S/Fe mole ratio, method B seems to alter this ratio. Consequently, this method is not used in this study. Methods A and C do not appear to alter the S/Fe ratio, and therefore, both methods are deemed to yield clean surfaces. Because 56 method C is much less cumbersome, it was chosen for the preparation of samples to be used in the dissolution experiment. D) Characterization of the Pyrite Samples. Figure 9 (p.69) illustrates a typical XRD spectrum obtained from a powdered pyrite sample. Part A of Appendix 2 (p.87) lists experimental, published and extraneous d-spacings. There is very good agreement between the experimental and published d-spacings. Figure 10 (p.69) is an example of a typical XPS wide scan spectrum obtained from analysis of an uncrushed pyrite sample exhibiting the absence of inclusions under the Zeiss microscope. Carbon and oxygen are the only clearly evident contaminant peaks. Typically, no other mineral phases or impurities were discernible frOm these spectra. The usual background due to inelastic scattering is well exhibited in Figure 10. Figures 11 (p.70) and 12 (p.71) illustrate typical SEM/EDXA results obtained from uncrushed pyrite samples exhibiting inclusions. EDXA results indicate the presence of silica, two forms of aluminosilicates and calcium carbonate. The difficult)' in locating these mineral phases suggests that these and other possible phases are not abundant. Part B of Appendix 2 (p.87) tabulates XRF data obtained from three crushed samples. A number of minor elements were detected but none of these appeared in significant quantities. Consequently, their presence was not analysed for in the oxidizing solution. E) The Dissolution Experiment - Results. 1) The Solid Phase. Figure 13 (p.72) illustrates XPS narrow scan spectra for two of the seven sampling times. The remaining five sets of spectra were essentially identical to those shown in Figure 13, and therefore, are not shown. In terms of chemical state, there appears to be no significant change in the oxidation states of iron and sulfur. Though there does exist oxidant (0 2 or H 20) in the surface region, it is either physisorbed or chemisorbed only. This may imply that the adsorption rate of any oxidant that could result in a change in chemical state of iron or sulfur is either very fast followed by a quick etching process or that the adsorption rate is very slow. Part B of the 57 literature review (p. 16) suggests that the oxygen (O2) adsorption step may be quite slow and perhaps rate-limiting. Table 4 lists the O/S and O/Fe atomic ratios. In both cases, these ratios seem to follow the same trend as that exhibited by the S/Fe ratio. Table 4. XPS results of the dissolution experiment from 0 to 72 hours. Sampling Time Atomic Ratios (n) (hours) S/Fe O/S O/Fe 0 (average) 3.64 0;43 1.55 " ; " 12 3.40 . 0.52 1.76 24 • 3.79 0.72 2.72 36 3.57 1.15 4.11 48 5.60 1.00 5.61 63 4.24 0.83 3.51 72 3.52 0.61 2.15 Note. For sampling time t = 0 hours, the S/Fe atomic ratio is an a\>erage of controls cleaned by methods A (200), C (205) and the original sample (229) as listed in Table 2 (p. 54); whereas, the O/S and O/Fe ratios are an average of controls cleaned by method C (1, 30 and 60 minutes) only. XPS results of S/Fe mole ratios during the course of the dissolution experiment are listed in Table 4. These results are plotted in Figure 16 (p.74). The plot suggests that for approximately the first 36 hours, there is very little change in the ratio. From 36 to 48 hours, there is a significant increase in the ratio which seems to reach a maximum value of 5.6 at 48 hours. Then, at some point between 48 and 63 hours, this ratio drops significantly. At 72 hours, the ratio reassumes its original value (~ 3.6). Again, throughout the course of the entire dissolution experiment, the majority of sulfur in the sulfur-rich surface layer is in the pyritic form. 2) The Liquid Phase. 58 Each time a pyrite sample was removed from the reaction vessel, a ten mL aliquot of the solution was also removed for total sulfur (sulfate) and iron analysis. Total sulfur and iron results arc tabulated in Table 5 along with the liquid S/Fe ratio. Total iron and sulfate data are plotted in Figure 15 (p.73). Raw data for total sulfur and iron can be found in Tables 1 to 6 of Appendix 3 (P-88). Except for an initial rapid rate of iron production in solution, Figure 15 (p.74) illustrates the iron production rate to be a relatively constant until about 48 hours at which point the rate increases till 63 hours. From 63 to 72 hours, the rate appears to approach the same value as that from 12 - 48 hours. Total sulfur follows a similar pattern except that at 48 hours, there is a significant rise in its rate of production. Table 5. Total concentrations of sulfate and iron in liquid phase as determined from nephelometry and atomic absorption respectively. Sampling Time Sulfate Iron S/Fe . (hours) (ppm) (ppm) 0 0.00 0.0 12 0.21 0.3 0.7 24 0.23 0.4 0.58 36 0.33 0.5 0.66 48 0.55 0.6 0.92 63 1.51 0.84 1.8 72 1.59 0.9 1.8 The iron production in solution may be simply the result of diffusion. The iron concentration gradient which exists between the bulk of the mineral and the surrounding solution, 59 though complex, is steep considering pyrite's molar volume and the very low concentration of iron in solution. This gradient provides the driving force for the diffusion process. Regarding the liquid S/Fe ratio presented in Table 5, dissolution is obviously incongruent for the first 48 hours. However, from 48 to 72 hours, the liquid S/Fe ratio approaches a value of two thus reflecting the stoichiometry of pyrite. Also, solid phase data suggests that the surface is in the process of being regenerated. This implies that had sampling been carried out at intervals much greater than 12 hours the incongruency of the dissolution would have been missed and a different mechanism would have been proposed. 3) A Comparison Between Nephelometry and Turbidimetry. Figures 17 and 18 (p.74) are plots of standard curves derived from the nephelometry experiment. Data for these plots are listed in Tables 1, 2 and 3 of Appendix 3 (p.##). Figure 17 illustrates that there is very good reproducibility in data from 0 to 4.5 ppm sulfate amongst the five random sets of measurements. The adjusted and averaged sets of measurements are plotted in Figure 18. This plot has been fitted to two linear regressions, at the 95% confidence interval, using a standard statistical software package, Microsoft™ Excel (Version 4.0). The linear regressions are arranged such that the concentration of sulfate (ppm) in the samples from the 72 hour dissolution experiment can be determined readily. These regressions, their domains and R 2 values are as follows: [sulfate] = 1.997 Z + 0.009 (0 < Z < 0.69, R 2 = 0.9991) (58) [sulfate] = 0.0367 Z + 1.38 (0.69 < Z < 87.87, R 2 = 0.9998) (59) where, Z = F(Gs/Gw)E-03 (adjusted and averaged). Two worthwhile points should be made regarding the linear regressions. Firstly, no exponential nor polynomial function could provide fits as good as the two linear regressions. Polynomials up to and including order 6 were tested, yet all gave unsatisfactory results especially between 0 and 1.0 ppm. Secondly, the standard curve could have been fit to three linear regressions (0-1, 1-1.5 and 1.5-4.5 ppm); however, no advantage was observed. 60 Figures 19 and 20 (p. 75) are plots of standards curves derived from the turbidimetry experiment. Data for these plots can be found in Table 7 of Appendix 3 (p.90). Figure 19 demonstrates the inability of the conventional turbidimetric method to produce replicable results from 0 to 10 ppm sulfate relative to that of nephelometry. Figure 20 illustrates an averaged standard curve along with its associated linear regression also derived from the same statistical package mentioned above. Although the averaged standard curve from nephelometry shows a more pronounced non-linear response than that from turbidimetry at low concentrations, the fact that nephelometry is able to produce more consistent results gives it an operational advantage in this study. Being able to track the changes in sulfate concentration with reasonable confidence is of outmost importance in order to correlate liquid phase data with that of the solid phase. At a single glance, the averaged standard curve from nephelometry is obviously more reliable than that from turbidimetry. This is due to the simple fact that nephelometry compares scattered light and not transmitted light to the incident beam. F) A model for the Incongruent Dissolution of Pyrite. Before correlating liquid data to that of the solid phase, two experimental constraints must be addressed. Firstly, it must be remembered that at each sampling time, solid phase data is gathered from a single pyrite sample. Liquid phase data, however, represents concentrations of iron and sulfur which are derived from all samples in the reaction vessel. Secondly, analysis of the various solid samples from the cleaning procedures reminds us that there exists natural variability in the S/Fe ratios. These two constraints force us to relax the need to compare S/Fe ratios of both phases too rigidly. For example, the relatively constant rate of iron diffusion into the solution is not well reflected in the S/Fe mole ratio at the solid surface. Therefore, simply observing trends may do more justice to proposing a mechanism for the incongruent dissolution of pyrite. 1) The Original Surface. This and other data (Mycroft et al., 1990) suggests a high S/Fe mole ratio. Data from this study also reveals the presence of elemental sulfur. The physical distribution of the elemental 61 sulfur is not certain. Figure 14 (p.73) attempts to illustrate the distribution of elemental sulfur along a section of a surface which is littered with features such as steps, adatoms, kinks and ledges. Although this study contains no specific data that suggests the presence for these surface features, they are assumed to exist. Pyrite is known to be in equilibrium with elemental sulfur which is able to segregate to corners and fracture lines (Lowson, 1982). Figure 14 is meant to suggest that the percentage of elemental sulfur at the surface is greater than the calculated 7% but that the majority of sulfur remains in the pyritic form. Figure 14 also depicts a natural S/Fe gradient which increases from 2:1 in the bulk to 3.6:1 at the surface. The original gradient may have been stimulated by trace impurities acting as point defects. These defects may, in turn, have stimulated the release of iron atoms from the surface region of the crystal. Alternatively, the sulfur-rich surface may have arisen from migration of sulfur towards the surface. In any event, the surface is deemed to be naturally sulfur-rich. Although the exact nature of any surface reconstruction is beyond the scope of this study. Figure 14 does not assume the pyrite surface to be unaffected by surface reconstruction. Each cube in the diagram simply represents the location of the unit cells in space. A further note regarding the sulfur-rich surface is that because the chemistry of this surface is unknown, the extent and type of surface states is also unknown. Consequently, Figure 2 (p.26) simply represents the surface states as a "black box". 2) The Surface Under Dissolution. The mechanism to be proposed for the incongruent dissolution of pyrite is based upon (1) the original surface as described above, (2) a correlation of liquid and solid phase data and (3) current knowledge of the various pyrite oxidation mechanisms as presented in the literature review. The mechanism favors an electrochemical pathway rather than chemical because in the latter no protective layer is expected to form (p. 18). As ferrous iron diffuses along the S/Fe gradient, this gradient steepens as seen in Figure 14 between 36 and 48 hours. Once the S/Fe mole ratio gradient approaches some critical value, the stability of the sulfur-rich layer is reduced. Cracks and microfissures may occur resulting in 62 the desorption of sulfur-rich pyrite flakes (S-pyrite). This desorption may explain the sudden decrease in the solid S/Fe mole ratio between 48 and 72 hours. This desorption process may or may not increase the surface area of the pyrite. The average density of surface features (eg. kinks per unit area) may maintain some equilibrium value. This statement should be apparent from Figure 14. Before continuing with the mechanism, two points need mentioning regarding the critical S/Fe mole ratio. First, this critical ratio may be related to the chemical and physical properties of pyrite. Because each sample varies from one another in these properties, there may be variation in the value of this critical ratio from sample to sample. Second, this critical ratio may or may not be related to Lowsorfs (1982) description of the passive film which is able to form: at the surface of pyrite (p. 18). Recall that this film is considered stable if less than 27% of pyrite decomposes simultaneously and any excess sulfur is converted to sulfate. In the present mechanism, it may not be necessary for a certain percentage of the surface to decompose simultaneously but rather it may be that, over time, the accumulation of sulfur reaches an unstable level. The S-pyrite flakes would conceivably have an extremely large surface area. Upon their detachment from the pyrite surface, fresh surfaces would become exposed to the solution thus allowing for an increased production rate of sulfate. This would explain the sudden increase in sulfate concentration occurring immediately after the sudden decrease in the solid S/Fe mole ratio from 48 to 72 hours (compare Figures 15 and 16 (p. 73)). From Figure 15, between 63 and 72 hours, the rate of sulfate production seems to have slowed down. This may imply that all flakes of S-pyrite have been converted to sulfate thus yielding a liquid S/Fe ratio of approximately two. Also, given there is the desorption of elemental sulfur flakes from the surface, this sulfur would essentially remain as is given the extremely slow conversion rate to sulfate compounded with the slow production rate of S° itself. Conceivably, the flakes of S-pyrite and S° could contain ferrous iron which could diffuse outwardly explaining the rise in the rate of iron production in solution during the same time period (48 - 63 hours, Figure 15, p.73). The integrity of the two types of flakes may be such that they themselves crumble even further into smaller size fractions over time thus ultimately further increasing the rate of sulfate and iron production. Upon detachment of both types of flakes, the solid S/Fe mole ratio data seems to suggest that the original surface is continuously being regenerated. This, in turn, may suggest that the mechanism described above may simply repeat itself indefinitely. That is, Figures 15 and 16 may represent one incongruent cycle of the total congruent dissolution process. The following reactions are meant to support the mechanism described above. FeS3.6 -> x Fe 2 + + Fe ( 1 . x ) S 3 . 6 (56) Fe (i- x )S 3 . 6 -> xFe 2 + + Fe(i_2x)S3.6 (57) Fe (i_2x)S3.6 ->' x F e 2 + + Fe (i_3 x )S 3. 6 (58) Fe(i_3X)S3.6 —>• FeS 3. 6 + S-pyrite + nS° (59) where, FeS3.6 represents the pyrite surface which includes elemental sulfur, Fe(i-3X)S3.6 represents pyrite at its critical S/Fe mole ratio, and n S° represents flakes of elemental sulfur which, in total, does not exceed yields greater than 50% of the total available sulfur. Decomposition of the S-pyrite occurs according to the following anodic reaction: y H 2 0 + S-pyrite ->• (0.25y) S0 4 2 " + (2y) H + + (1.5y) e" (60) Overall, FeS 3. 6 + y H 2 0 -> 3x Fe 2 + + Fe (i_ 3 x )S 3 . 6 + n S° + 0.25y S04"2 + 2y H + + 1.5y e" (61) Given a long period of time, dissolution will be congruent. The mechanism implies that elemental sulfur and sulfate occur via independent reactions occurring, most likely, at different rates.The balancing cathodic reaction could be (26), 0 2 + 4 H + + 4 e" - » 2 H 2 0 . No data in this study supports reaction (60). This reaction is based only on the work of Lowson (1982), McKibbcn and Barnes (1986), Ritchie (1994), Mathews and Robins (1974) and 64 McKay and Halpern (1959) wherein the oxygen in the sulfate is thought to be derived from water. Reactions (56) to (58) represent the steady diffusion of iron into the liquid phase thus increasing the S/Fe mole ratio gradient within the pyrite crystal. These three steps parrallel (34) but unlike (34) the initial surface is sulfur-rich. All three steps could have been summarized into one. however, three steps were used to emphasize (1) that the diffusion of ferrous ions is slow and (2) the critical S/Fe mole ratio will vary from sample to sample. Steps (56) to (58) also parallel (21) in that there is the formation of elemental sulfur at the pyrite surface. Once the critical S/Fe mole ratio is reached, the structural integrity of the sulfur-rich regions is destabilized and flakes of S-pyrite and S° detach from the surface as depicted by (59). Consequently, the original surface is regenerated. The detachment of both types of flakes, however, is not necessarily simultaneous. The diffusion of reaction products away from the mineral surface needs some consideration. In this study, stirring was achieved by a jet stream of dry C02-free air (p.37). Consequently, reaction products were not allowed to accumulate near the solid/liquid interface. However, in a static environment where there exists no mechanism for the constant removal of products away from the solid surface, conceivably, the diffusion rate of ferrous ions from the crystal lattice could be significantly reduced. This, in turn, would delay the onset of a critical S/Fe mole ratio and the subsequent detachment of flakes from the surface. Furthermore, a coating of both types of flakes could develop on the mineral surface thus further impeding the diffusion rate of ferrous ions from the crystal lattice. 65 Figure 6. XPS narrow scan spectra comparing the chemical states for (a) Fe 2p and (b) S 2p photoemission peaks. 229 ( uncleaned sample), 205 ( Method C, 60 minutes). Only the Fe 2p spectrum (229) indicates the significant presence and removal of iron oxides. 66 Figure 7. XPS narrow scan spectra for (a) Fe 2p and (b) S 2p photoemission peaks showing the effects of cleaning on the chemical states. 200 ( Method A), 202 ( Method B), 244 ( Method C, 1 minute), 205 ( Method C, 60 minutes). 67 Counts (arb. units) Figure 8. Curve fitting for an XPS narrow scan spectrum of the sulfur photoemission peak for a cleaned pyrite sample (method C, 60 minutes). Fitted peaks are labelled as follows: (1) pyritic-S 2p 3/ 2, (2) pyritic-S 2pj/2, (3) elemental S 2p3/2, (4) elemental S 2pj/ 2. 68 Figure 9. Typical XRD spectrum of powdered pyrite samples. Peaks are tabulated in table 1 of Appendix 3. + S CAES) — i 1 1 1 i • 1200 1000 BOO 600 /OO 200 0 Binding energy (eV) Figure 10. Typical XPS wide scan ploc ( 0 - 1500 eV ) of pyrite samples indicating various XPS and Auger (AES) peaks. Impurities include oxygen and carbon. 69 Figure 11. Three SEM photographs of inclusions in pyrite samples. All photographs are at the same scale. Inclusions 1, 2, 3 and 4 are typical examples of gangue minerals found in the pyrite samples and correspond to spectra (a), (b), (c) and (d) respectively in Figure 12. 7 0 S I Figure 12. Typical EDXA results for the characterization of the pyrite samples. Spectra (a) and (b) represent silica and some form of aluminosilicate respectively; spectrum (c) represents some other form of aluminosilicate; spectrum (d) represents calcium carbonate. 71 1 1 1 1 1 , , , 712 711 710 709 708 707 706 705 Binding energy (eV) i 1 1 , 1 1 — 170 168 166 164 162 160 ( M Binding energy (eV) Figure 13. XPS narrow scan spectra for (a) Fe 2p and (b) S 2p photoemission peaks indicating no significant changes in chemical states during the dissolution experiment 258 (36 h), 263 (72 h). 72 t Figure 14. A schematic drawing of a section of a hypothetical pyrite surface (1,2) undergoing incongruent dissolution. Each cube represents a unit cell. Iron atoms are drawn as spheres. Elemental sulfur has been drawn free-hand at corners. Flakes of elemental sulfur (3) and S-pyrite (4) have been sketched as products of dissolution. The surface region is depicted as being sulfur-rich; whereas, the bulk is stoichiometric. For more information, see the text. 73 Status of Liquid Phase During Pyrite Dissolution o 'is 3 Of Sulfur to Iron Mole Ratio During Pyrite Dissolution 0 (Av) 12 24 ....... . , . „ • • • 36 Time (hours) 46 63 n(S)/n(Fe) | 72 Figure 16. Plot of the sulfur to iron mole ratio at the surface of pyrite during dissolution. 74 Adjusted Standard Curves for Sulfate Determination by Nephelometry 90.00 80.00 -70.00 --60.00 ^ 50.00 -55 40.00 + o fa 30.00 --20.00 -10.00 0.00 -10:00, I *< o in -cf- T-" " T- IM IN "" e>—rr-Standards (ppm) — 0 — Z (adj.l) -A - Z (adj.2) - • - Z (adj.5) -— Z (adj.5) ^ - z Figure 17. Plot of the five adjusted standard curves from the nephelometry experiment. Adjusted and Averaged Standard Curve for Sulfate Determination by Nephelometry Standards (ppm) Figure 18. Plot of the adjusted and averaged standard curve based upon the five standard curves of Figure 16 above. 75 Average Standard Curve for Sulfate Determination by Turbidimetry 10 -t , Standard (ppm) Figure 20. Plot of the averaged standard curve for the determination of sulfate by turbidimetry. Also included is the standard curve as determined from the linear regression of the averaged standard curve. 76 Chapter Five: Conclusion. In relation to the objectives of this study, the following summary and conclusions appear warranted. A trend in the liquid and solid S/Fe ratios has been observed upon which a mechanism for the incongruent dissolution of pyrite has been proposed. The sudden rise and subsequent fall in the solid S/Fe mole ratio followed by the sudden increase in sulfate production in solution strongly suggests evidence for a critical S/Fe mole ratio at the pyrite surface. Moreover, the claim that pyrite's original surface is sulfur-rich deserves special attention. Although this study has not rigorously defined the physical and chemical nature of the surface, it nevertheless holds to its claim based on the principle that a surface is a pertubation of a crystal's bulk structure. Hence, one should not expect the surface stoichiometry to be identical to that of the bulk. In fact, this surface may behave as a passive film under some conditions. To my knowledge, no published work has investigated the incongruent dissolution of pyrite from a correlation of solid and liquid phase data. Combining the strengths of XPS with those of nephelometry has proven to be worthwhile and instructional. A key feature of XPS is its ability to provide information on the chemical state of surface species. Nephelometry, on the other hand, has shown to be sensitive and consistent in detecting low concentrations of sulfate, thus proving to be more reliable than the traditional turbidimetric procedure. Together, changes in the chemical status of both solid and liquid phases have shown to be correlated. Unlike many published dissolution experiments, the pyrite samples in this study were not crushed prior to dissolution. Rather, they were only polished to obtain reasonable XPS spectra. Moreover, from a hardware point of view, the design and construction of the apparatus used in this experiment is noteworthy. Because pyrite is known to be extremely air-sensitive, much care and attention had to be given to protect each sample from exposure to air prior to analysis. Therefore, although much time was devoted to designing, acquiring, building and testing the various peices of hardware, once developed, the setup allowed for the quick and straightforward preparation of samples. 77 This thesis contributes to the existing body of knowledge on pyrite dissolution. The fulfillment of each objective itself is a contribution. In addition, this study provides (1) an up-to-date, and concise review of the various aqueous oxidation mechanisms of pyrite by dissolved oxygen and (2) an alternative qualitative molecular orbital diagram for pyrite based upon the author's description of chemical bonding in the iron sulfide. Because XPS did not reveal any changes in chemical states of either iron or sulfur, no insight into the roles of dissolved oxygen and water as reactants is available. Water is only suspected to act as a reactant in the conversion of S-pyrite to sulfate. This study suggests that the frequency of sampling can largely influence the type of dissolution mechanism one proposes. Had the dissolution experiment been carried over a longer period of time, it is possible that a series of incongruent cycles would have been observed. Consequently, at the end of each cycle, both solid and liquid S/Fe ratios would show essentially no change. Therefore, sampling interval is crucial to any mechanism. The pretreatment of pyrite as a means of circumventing the problems associated with the incongruent dissolution phase was mentioned earlier (p. 16). This study, however, suggests that the incongruent phase may always be at work. Most oxidation mechanisms that have been reviewed are based on short term data (< 1 day). According to this study, 1 day would not be sufficient to witness any incongruency, and therefore, only congruency would be assumed. The mechanism proposed in this study finds commonalities with a number of mechanisms reviewed earlier. Types E2 (p. 17) and E5 (p.21) are especially closely related to the proposed mechanism." Type E2 suggests the production of ferrous iron and the formation of a passive S° film on the pyrite surface; whereas, type E5 suggests the formation of an iron-deficient layer. Also, the proposed mechanism allows for the production of elemental sulfur but not in yields greater than 50% of total available sulfur thus paralleling mechanisms CI (p. 11), C2 (p. 12), C3 (p.l3),El(p.l7) andE6(p.22). Despite the common ground that this mechanism shares with other mechanisms, two essential differences exist. Firstly, the mechanism begins with the fundamental assertion that 78 pyrite's surface is naturally sulfur-rich. As dissolution progresses, there appears to be an enrichment of sulfur at this surface thus steepening the S/Fe mole ratio gradient along which ferrous ions may diffuse. Most other mechanisms suggest pyrite's surface to be, and remain, stoichiometric throughout the oxidative dissolution process. Secondly, the mechanism invokes, in general terms, a critical S/Fe mole ratio. Once exceeded, the surface becomes unstable and flakes of S-pyrite and elemental sulfur begin to detach from the surface yielding higher rates of sulfate and ferrous iron production. This critical S/Fe ratio may be related to the passive film often referred to in other studies. It should be apparent that like many other mechanisms, this one provides no insight into any electron transfer mechanisms that are at work: Until more comprehensive and fundamental surface science is initiated, such surface reactions will not be uncovered. Very few references have been made regarding other XPS studies on the aqueous oxidation of pyrite. This is simply because not much work has been done under conditions similar to those in this study. What was crucial to this experiment, in terms of XPS, was to ensure reasonable spectra and a good reference binding energy for the Fe 2P3/2 photoemission peak. Once ensured, internal consistency throughout the dissolution experiment could be established. For example, because iron was known to exist in the samples, the Fe 2P3/2 photoemission peak was chosen as a reference. Although this procedure is not unique, it is not widely practiced. Instead, for example, many workers use the Cls peak as a reference. During the course of this study, many questions came to mind. Five such questions are as follows: (1) What is the physical and chemical nature of pyrite's surface in terms of composition, and structure ? The answer to this question will depend largely on the technology available to study a complex binary system such as pyrite. (2) Can we define the critical S/Fe mole ratio more precisely. That is, what value or range of values is considered critical, and what makes this ratio(s) critical ? (3) What molecular mechanisms exist to allow for the diffusion of iron ? 79 (4) What role do trace impurities and guangue minerals play in the dissolution process. (5) Can we better partition the distribution of elemental sulfur and S-pyrite during the course of dissolution. These and other questions will need to be answered in order to more fully, understand the nature of pyrite disolution. Only then can we more stategically abate the problems associated with sulfide mineral dissolution in the contexts of acid mine drainage and acid-sulfate soils as mentioned in the introduction. 80 Bibliography Abbey S.. Studies in "Standard Samples" of Silicate Rocks and Minerals 1969 - 1982. Geological Survey of Canada. (1983) Pap 83-15, 114 pp. Bailey L. K. and Peters E., Decomposition of Pyrite in Acids by Pressure Leaching and Anodization: The Case of an Electrochemical Mechanism. Canadian Metallurgical Quarterly. (1976) 15, no.4, 333 - 344. Beek K. J., Blokhuis W. A., Driessen P. M., Van Breemen N. and Pons L. J., Problem Soils: Their Reclamation and Management in Land Recalamation and Water Management. Institute for Land Reclamation and Improvement. (1980) 27, 43 - 72. Biegler T., Oxygen reduction on Sulfide Minerals. II. Relation between Activity and Semiconducting Properties of Pyrite Electrodes. Journal of Electroanalytical Chemistry and Interfacial Electrochemistry. (1.976 b) 70, 265 -275. Biegler T:, Rand D. A. J. and Woods R,, Oxygen Reduction on Sulfide Minerals. I. Kinetics and Mechanism at Rotated Pyrite Electrodes, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry. (1975) 60, 151 - 162. Biegler T., Rand D. A. J. and Woods R., Oxygen Reduction on Sulfide Minerals. Trends in Electrochemistry. Eds.: Bockris J. O'M., Rand D.A.J, and Welch J.J., (1976 a), 291 - 302. Biegler T. and Swift D. A., Anodic Behaviour of Pyrite in Acid Solutions. Electrochimica Acta. (1979) 24, 415 -420. Bierens de Haan S., A Review of the Rate of Pyrite Oxidation in Aqueous Systems at Low Temperature. Earth-Science Reviews. (1991) 31, 1 - 10. Braun R. D., Introduction to Instrumental Analysis. McGraw - Hill Book Company, New York. (1987), 1004 pp. Brock T. D., Cook S., Petersen S. and Mosser J.L., Biogeochemistry and Bacteriology of Ferrous Iron Oxidation in Geothermal Habitats. Geochimica Cosmochimica Acta. (1976) 40, 493 - 500. Bronswijk J. J. B. and Groenenberg J. E., A simulation Model for Acid Sulfate Soils. I. Basic Principles. 4th International Symposium on Acid Sulfate Soils, Wageningen, the Netherlands. (1993), 173 - 182: Buckley A. N. and Woods R., X-Ray Photoelectron Spectroscopy of Oxidized Pyrrhotite Surfaces. I. Exposure to air. Applications of Surface Science. (1985) 22/23, 280 - 287. Buckley A. N. and Woods R., X-Ray Photoelectron Spectroscopy of Oxidized Pyrrhotite Surfaces. II. Exposure to aqueous solutions. Applications of Surface Science. (1985) 20, 472 - 480. Buckley A. N. and Woods R., The Surface Oxidation of Pyrite. Applied Surface Science. (1987) 27,437 -452. Chander S. and Briceno A., Kinetics of Pyrite Oxidation. Minerals and Metallurgical Processing. (1987) August, 171 - 176. 81 Clark C. S., Oxidation of Coal Aline Pyrite. Journal of Sanitation Engineering Division of American Society of Civil Engineering. 92nd Proceedings Paper 4802. (1966), 127 - 145. Cox P. A., The Electronic Structure and Chemistry of Solids. Oxford Science Publications, Oxford University Press. (1989), 259 pp. Deer W. A., Howie R. A. and Zussnian J., An Introduction to the Rock-Forming Minerals. 2n d ed. Longman Scientific & Technical, England. (1992). 696 pp. Dcsilvestro J., Gratzel M. and Pajkossi T., Electron Transfer at the WOj - Electrolyte Interface Under Controlled Mass Transfer Conditions. Journal of Electrochemical Society. (1986) 133, 331 - 336. Eisenberg D. and Crothers D., Physical Chemistry with Applications to the Life Sciences. The Benjamin / Cummings Publishing Company, Inc., Ontario. (1979), 868 pp. Ennaoui A., Fiechter S., Jaegermahn W, and Tributsch H., Photoelectrochemistry of Highly Quantum Efficient Single-Crystalline ri-FeS2. (Pyrite). Journal of Electrochemical Society. (1986) 133, 97 - 106. Finklea S. L. Ill, Cathey L. and Amma E. L., Investigation of the Bonding Mechanism in Pyrite using Mossbauer Effect and X-ray Crystallography. Acta Crystallography. (1976) A 32, 529 -537. Fornasiero D., Eijt V. and Ralston J., An Electrokinetic Study of Pyrite Oxidation. Colloids and Surfaces. (1992) 62, 63 - 73. Garrels R. M. and Thompson M. E., Oxidation of Pyrite by Iron Sulfate Solutions. American Journal of Science. (1960) Bradley 258-A, 57 - 67. Goldhaber M. B., Experimental Study of MetasTable Sulfur Oxyanion Formation During Pyrite Oxidation at pH 6 - 9 and 30° C. American Journal ofScience. (1983) 283, 193 -217. Gosling L. M. and Baker S. J., Acidity Fluctuations at a Broadland Site in Norfolk. Journal of Applied Ecology. (1980) 17, 479 - 490. Gould W. D., Bechard G. and Lortie L., Chapter 7. The Nature and Role of Microorganisms in the Tailings Environment in Environmental Geochemistry7 of Sulfide Mine Wastes, Eds.: Jambor J.L. and Blowes D.W.. Short Course Handbook, Mineralogical Association of Canada. (1994) 22, 185 - 199. Hamilton I. C. and Woods R., An Investigation of Surface Oxidation of Pyrite by Linear Sweep Voltammetry. Journal of Electroanalytical Chemistry. (1981) 118, 327 - 343. Harrison W. A., Electronic Structure and the Properties of Solids. The Physics of the Chemical Bond. Dover Publications, Inc., New York. (1989), 586 pp. Huheey J. E., Inorganic Chemistry. Principles of structure and reactivity. 3 r d ed. Harper and Row, New York. (1983), 936 pp. Ibach, H and Rowe J. E., Hydrogen Adsorption and Surface Structures of Silicon. Surface Science. (1974) 43, 481 - 492. 82 Jambor J. L., Chapter 3. Mineralogy of Sulfide-Rich Tailings and Their Oxidation Products in Environmental Geochemistry of Sulfide Mine Wastes, Eds.: Jambor J.L. and Blowes D.W.. Short Course Handbook, Mineralogical Association of Canada. (1994) 22, 59 - 102. Joint Committee on Powder Diffraction Standards, Selected Powder Diffraction Data for Minerals. Ed.: Berry L.G., lstcd. (1974), 833 pp. Klein C. and Hurlbut C. S. Jr., Manual of Mineralogy. 20 t h ed. John Wiley & Sons, New York. (1985), 596 pp. Lachat Instruments, Methods Manual for the QuickChemR Automated Ion Analyzer. 1992. Lalvani S. B., Weston A. and Masden J. T., Characterization of Semiconducting Properties of Naturally Occurring Polycrystalline Iron Disulfide (Pyrite). Journal of Material Science. (1990) 25, 107- 112. Lasaga A . C , Chapter .1: Rate Laws of Chemical Reactions. Reviews in Mineralogy. Ed. Ribbe PH. . (1981) 8 c.2, 1 -68. Lindsay W.'L., Chemical Equilibria in Soils. John Wiley & Sons, New York. (1979), 449 pp. Lowson R. T., Aqueous Oxidation of Pyrite by Molecular Oxygen. Chemical Reviews. (1982) 82 no. 5, 461 -497. Luther G. W. Ill, Pyrite Oxidation and Reduction : Molecular orbital theory considerations. Geochimica et Cosmochimica Acta. (1987) 51, 3193 - 3199. Mathews C. T. and Robins R. G., Aqueous Oxidation of Iron Disulfide by Molecular Oxygen. Australian Chemical Engineering. (1974) Nov. / Dec, 19 - 24. McKay D. R. and Halpern J., A Kinetic Study of the Oxidation of Pyrite in Aqueous Suspension. American Institute of Mining, Metallurgy and Petroleum Engineers. (1959) 212, 301 - 309. McKibben M. A. and Barnes H. L., Oxidation of Pyrite in low temperature acidic solutions: Rate laws and surface textures. Geochimica et Cosmochimica Acta. (1986) 50, 1509 - 1520. Mellor J. W., A Comprehensive Treatise on Inorganic and Theoretical Chemistry: Fe (Part 3)-Cobalt. Longmans, Green and Company, New York. (1935) 14, 187 pp. Meyer R. E., Electrochemistry ofFeS2 . Journal of Electroanalytical Chemistry. (1979) 101, 59 -71. Mishra K. K. and Osseo - Asare K., Aspects of the Interfacial Electrochemistry of Semiconductor Pyrite (FeS2 )• Journal of Electrochemical Society. (1988) 135, 2502 - 2509. Mishra K. K. and Osseo - Asare K., Electroreduction of Fe3+, 02 , and Fe(CN)^S at the n-pyrite (FeS2) Surface. Journal of Electrochemical Society. (1992) 139, 3116 - 3120. Moses C. O., Kinetic Investigation of Sulfide Mineral Oxidatio in Sterile Aqueous Media. Msc. paper at University of Virginia, Department of Environmental Sciences. (1982), 100 pp. Moses C. O. and Herman J. S., Pyrite oxidation at circumneutral pH. Geochimica et Cosmochimica Acta. (1991) 55, 471 - 482. 83 Moses C. O., Nordstrom D. K., Herman J. S. and Mills A. L., Aqueous Pyrite Oxidation bv Dissolved Oxygen and by Ferric Iron. Geochimica et Cosmochimica Acta. (1987) 51, 1561 -1571. Mycroft J. R., Bancroft G. M., Mclntyre N. S., Lorimcr J. W. and Hill I. R.. Detection of Sulfur and Polysulfidcs on Electrochemically Oxidized Pyrite Surfaces by X-Ray Photoelectron Spectroscopy and Raman Spectroscopy. Journal of Electroanalytical Chemistry. (1990) 292, 139 -.152. Najdeker E. and Bishop E., The Formation and Behaviour of Platinum Sulfide on Platinum Electrodes. Journal of Electroanalytical Chemistry. (1973) 41, 79 - 87. Narayan, P. B. V., Anderegg J. W. and Chen C. W., An ESCA Study of Iron Sulfidation in H2 S. Journal of Electron Spectroscopy and Related Phenomena. (1982) 27, 233 -242. Nicholson R. V., Chapter 6 : Iron - Sulfide Oxidation Mechanisms: Laboratory Studies in . Environmental Geochemistry of Sulfide Mine Wastes, Eds.: Jambor J.L. and BlowesD.W, Short Course Handbook, Mineralogical Association o/Cawaaa. (1994) 22, 163 - 183. Nordstrom D. K., Chapter 3: Aqueous Pyrite Oxidation and the Consequent Formation of Secondary Iron Minerals in Acid Sulfate Weathering, Eds.: HosSoer L.R., Kittrick J.A. and Fanning D.F., Soil Science Society of America. (1982), 37 - 56. Oshawa A., Yamamoto H. and Watanabe H., X-Ray Photoelectron Spectra of Valence Electrons in FeS2, CoS2 and NiS2. Journal of the Physical Society of Japan. (1974) 37, 568. Page A. L., Miller R. H. and Keeney D. R. Eds., Methods of Soil Analysis Part 2 - Chemical and Microbiological Properties. 2nd edition. American Society of Agronomy Inc. and Soil Science Society of America Inc. (1982), 1159 pp. Pang J., Briceno A. and Chander S. A Study of Pyrite / Solution Interface by Impedance Spectroscopy. Journal of Electrochemical Society. (1990) 137, 3447 - 3455. Panzner G. and Egert B., The Bonding State of Sulfur Segregated to a-Iron Surfaces and on Iron Sulfide Surfaces Studied by XPS, AES and ELS. Surface Science. (1984) 144, 651 -664. Peters E., Electrochemical Mechanisms for Decomposing Sulfide Minerals. Proceedings, International Symposium on Electrochemistry in Mineral and Metal Processing, Eds.: Richardson PR., Srinivasan S.S. and Woods R.. The Electrochemical Society, Pennington, N.J.. (1984) 84-10,343 - 361. Pons L. J., Van Breemen N. and Driessen P. M., Chapter 1. Physiography of Coastal Sediments and Development of Potential Soil Acidity in Acid Sulfate Weathering. Soil Science Society of America Special Publication. (1982) 10, 1 - 18. Prokopovich N , Engineering Impacts of Acid Sulfate Deposits in California in Selected Papers of the Symposium on Acid Sulfate Soils, Wageningen, the Netherlands. (1988) 44, 49 - 58. Ritchie A. I. M., Chapter 8: Sulfide Oxidation Mechanisms : Controls and Rates of Oxygen Transport in Environmental Geochemistry of Sulfide Mine Wastes, Eds.: Jambor J.L. and Blovves D.W.. Short Course Handbook, Mineralogical Association of Canada. (1994) 22, 201 - 245. 84 Salmcron M., Baro A. M. and Rojo J. M , Interatomic Transitions and Relaxation Effects in Auger Spectra of Several Gas Adsorbates on Transition Metals. Physical Review. (1976) 13:4. 4348 -4363. Sato M., Oxidation of Sulfide Ore Bodies. II. Oxidation Mechanisms of Sulfide Minerals at 25 C. Economic Geology. (1960) 55, 1202 - 1231. Seah M. P. and Dench W. A. Quantitative Electron Spectroscopy of Surfaces: A Standard Data Base for Inelastic Mean Free Paths in Solids. Surface and Interface Analysis. (1979) 1 no. 1, 2 -11. Singer P. C. and Stumm W., Acid Mine Drainage - The Rate Limiting Step. Science. (1970) 167. 1121-1123. Taylor B. E., Wheeler M. C. and Nordstrom D. K., Oxygen and Sulfur Compositions of Sulfate in Acid Mine Drainage: Evidence for Oxidation Mechanisms. Nature. (1984 a) 308, 538 - 541. Taylor B. E., Wheeler M: C. and Nordstrom D. K., STable Isotope Geochemistiy of Acid Mine Drainage: Experimental Oxidation of Pyrite. Geochimica Cosmochimica Acta. (1984 b) 48. 2669 -2678. Van Breemen N , Genesis and Solution Chemistry of Acid Sulfate Soils in Thailand. Ph.D. dissertation. Wageningen Agricultural University, Wageningen, the Netherlands. (1976), 239 pp. Van Der Heide H., Hemmel R., Van Bruggen C. F. and Haas C , Photoelectron Spectra of 3d Metal Pyrites. Journal of Solid State Chemistry. (1980) 33, 17 - 25. Vandermolan J., Gomes W. P. and Cardon F., Investigation on the Kinetics of Electroreduction Processes at Dark Ti02 and SrTiO^ Single Crystal Semiconductor Electrodes. Journal of Electrochemical Society. (1980) 127, 324 - 328. Wagner D. P., Fanning D. S., Foss J. E., Patterson M. S. and Soow P. A., Chapter 7. Morphological and Mineralogical Features Related to Sulfide Oxidation Under Natural and Disturbed Land Surfaces in Maryland in Acid Sulfate Weathering. Soil Science Society1 of America Special Publication. (1982) 10, 109 - 125. White A. F. and Peterson M. L., Chapter 35: Role of Reactive-Surface-Area Characterization in Geochemical Kinetic Models in Chemical Modelling of Aqueous Systems II ACS Symposium Series 416, Eds: Melchior D.C. and Bassett R.L., American Chemical Society, Washington DC, (1990) 461 -475. Williamson M. and Rimstidt, J. D., The Kinetics of Aqueous Pyrite Oxidation by Ferric Iron and Dissolved Oxygen. Poster, American Chemical Society. 204th National Meeting, Washington, D . C , (1992) August, 23 -28. 85 Appendix 1. Photograph of Dissolution Apparatus. Appendix 2. Characterization of Pyrite Samples. Part A. Results from XRD analysis listing experimental, published and extraneous d-spacings for three powdered samples. Experimental d-spacings 3.13 2.719 2.43 2.22 1.92 1.634 Published d-spacings 3.13a 2.71a 2.719b 2.42a 2.424b 2.212b 1.916b 1.63a 1.633b. Extraneous d-spacings 4.45 1.629 (a) Joint Committee on Powder Diffraction Standards, 1974, p:92. (b) Klein and Hurlbut, 1985, p.285. Part B. Results from XRF listing minor elements (ppm +/- 5%) in three powdered samples. Sample Nb Zr Y Sr Rb Pb Zn Cu 1 2.2 103.2 9.4 9.4 2 20.4 27.5 23.8 2 2 102.4 7.7 8.4 2.3 21 29.5 24.7 3 2 103 8.2 8.7 2.5 20.1 29.1 23.1 Sample Mn V Cr Ba Na Co « Ni 1 21.8 5.7 30 2 0.03 171.8 163.6 2 19.1 3.5 26.4 2 0.03 177.8 160.5 3 20.5 5.2 30.5 2 0.03 174.2 167.3 Note. The SEM/EDXA and XPS results are provided in the text. 87 Appendix 3. Raw Data from Nephelometry, Turbidimetry and Atomic Absorption Experiments. Table 1. Raw data from nephelometry experiment for the determination of low-level sulfate. Z= F(Gs/Gw)E-03. Standard (ppm) Filters F value Gs Gw Z (expt.) Z (adj.) 0.0 BGR 0.00333 35.8 114.5 1.04 0.00 0.0 BGR 0.00333 35.7 114.4 1.04 0.00 0.0 BGR 0.00333 35.8 114.3 1.04 0.00 0.0 : BGR 0.00333 36.0 114.5 1.05 0.01 . 0.0 BGR 0.00333 35.9 . 114.4 1.04 0.00 0.5 BGR . .0.00333 42.8 . 111.4. 1.28 0.24 . 0.5 BGR 0.00333 42.8 111.5 1.28 0.24. 0.5 BGR 0.00333 42.9 11.1.6 • 1.28 . . 0.24 ' 0.5 BGR 0.00333 43.0 111.5 1.28 0.24 0.5 BGR 0.00333 42.9 111.5 1.28 0.24 1.0 BGR 0.00333 53.8 116.0 1.54 0.50 1.0 BGR 0.00333 53.8 116.2 1.54 0.50 1.0 BGR 0.00333 53.9 116.2 1.54 0.50 1.0 BGR 0.00333 54.0 116.2 1.55 0.51 1.0 BGR 0.00333 53.9 116.1 1.54 0.50 1.5 BGR 0.00333 119.8 84.6 4.71 3.67 1.5 BGR 0.00333 120.0 84.7 4.71 3.67 1.5 BGR 0.00333 120.1 84.7 4.72 3.68 1.5 BGR 0.00333 120.0 84.9 4.70 3.66 1.5 BGR 0.00333 119.9 84.7 4.71 3.67 2.0 YGR 0.01357 109.8 81.8 18.22 17.18 2.0 YGR 0.01357 109.9 82.0 18.19 17.15 2.0 YGR 0.01357 110.1 82.1 18.20 17.16 2.0 YGR 0.01357 110.0 • 82.0 18.21 17.17 2.0 YGR 0.01357 110.1 82.1 18.20 17.16 2.5 BR 0.02971 101.3 90.9 33.11 32.07 2.5 BR 0.02971 101.5 90.7 33.25 32.21 2.5 BR 0.02971 101.6 90.9 33.21 32.17 2.5 BR 0.02971 101.6 91.0 33.17 32.13 2.5 BR 0.02971 101.5 90.8 33.21 32.17 3.0 BR 0.02971 129.8 83.4 46.24 45.20 3.0 BR 0.02971 130.1 83.6 46.24 45.20 3.0 BR 0.02971 130.1 83.4 46.35 45.31 3.0 BR 0.02971 129.8 83.4 46.24 45.20 3.0 BR 0.02971 129.9 83.5 46.22 45.18 3.5 B 0.0586 83.7 81.0 60.55 59.51 3.5 B 0.0586 83.6 81.3 60.26 59.22 3.5 B 0.0586 83.8 80.9 60.70 59.66 3.5 B 0.0586 83.8 81.1 60.55 59.51 3.5 B 0.0586 83.7 81.2 60.40 59.36 88 Table 1. Continued. Standard (ppm) Filters F value Gs Gw Z (expt.) Z (adj.) 4.0 B 0.0586 101.2 79.1 74.97 73.93 4.0 B 0.0586 101.1 79.2 74.80 73.76 4.0 B 0.0586 101.1 79.3 74.71 73.67 4.0 B 0.0586 101.2 79.2 74.88 73.84 4.0 B 0.0586 101.1 79.2 74.80 73.76 4.5 B 0.0586 . 120.0 78.9 89.13 88.09 4.5 • B , 0:0586 119:8 78.8 89.09 88.05 4.5 :•• ' B: 0.0586 119.9 78.7 .89.28 88.24 4.5 B 0.0586 120.1 78.8 89.31 88.27 4.5 .. B 0.0586 . 120.0 78.9 89.13 88.09 Table 2. Adjusted Z values for each of the five replicates. Standard Z Z Z Z Z (ppm) (adj. 1) (adj. 2) (adj. 3) (adj. 4) (adj. 5) 0.0 0.00 0.00 0.00 0.01 0.00 0.5 0.24 0.24 0.24 0.24 0.24 1.0 0.50 0.50 0.50 0.51 0.50 . 1.5 3.67 3.67 3.68 3.66 3.67 2.0 17.18 17.15 17.16 17.17 17.16 2.5 32.07 32.21 32.17 32.13 32.17 3.0 45.20 45.20 45.31 45.20 45.18 3.5 59.51 59.22 59.66 59.51 59.36 4.0 73.93 73.76 73.67 73.84 73.76 4.5 88.09 88.05 88.24 88.27 88.09 Table 3. Adjusted and averaged Z values for each of the standards. Standard Z (ppm) (adj. and avg) 0.0 0.00 0.5 0.24 1.0 0.50 1.5 3.67 2.0 17.16 2.5 32.15 3.0 45.22 3.5 59.45 4.0 73.79 4.5 88.15 89 Table 4. Results from nephelometry for the determination of sulfate in the liquid phase at various sampling times. Sample (hours) Filters F value Gs Gw Z (expt.) Z (expt. adj.) 0 BGR 0.00333 35.7 114.4 1.04 0.00 12 BGR 0.00333 46.0 134.2 1.14 0.10 24 YBG 0.00157 49.0 66.6 1.15 0.11 36 BGR 0.00333 47.5 131.9 1.20 0.16 48 BGR 0.00333 44.1. 112.0 . 1.31 0.27 63 BGR 0:00333 136.8 99:5 4.58 • 3.54 72 BG 0.00656 . 117.6 113.3 6.81 5.77 Table 5. Filter transmittance values used to calculate the F value. Red 0.507 Yellow 0.239 Green 0.112 Blue 0.0586 Table 6. Results from atomic absorption for iron determination in the liquid phase. All readings in ppm. Iron 1, 2, 3 and 4 refer to four replicates. Sampling time (hours) Iron 1 Iron 2 Iron 3 Iron 4 Iron (avg.) 0 0.0 0.0 0.0 0.0 0.0 12 0.3 0.35 0.3 0.3 0.3 24 0.4 0.4 0.4 0.4 0.4 36 0.5 0.5 0.5 0.5 0.5 48 0.6 0.6 0.6 0.6 0.6 63 0.85 0.83 0.85 0.84 0.84 72 0.9 0.9 0.9 0.9 0.9 Table 7. Absorbance results from the turbidimetry experiment for the determination of sulfate in several standards. All absorbance readings in ppm. Abs. 1. 2, 3. 4 and 5 refer to five random replicates. Standard (ppm) Abs. 1 Abs. 2 Abs. 3 Abs. 4 Abs. 5 Abs. (avg.) 0 0.091 0.096 0.093 0.184 0.092 0.11 1 0.958 0.944 0.825 0.667 0.678 0.814 3 3.052 3.036 3.149 2.95 2.989 3.04 5 4.916 4.915 4.652 4.527 4.384 4.679 10 9.64 9.427 9.186 8.973 8.914 9.23 90 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items