Citation
P Sorption Characteristics on Amorphous Al-Fe Co-Precipitated Hydr(oxides) as Determined by Batch Experiments and Flow Calorimetry: The Energetics and the Effect of Time, pH and Al Content

Material Information

Title:
P Sorption Characteristics on Amorphous Al-Fe Co-Precipitated Hydr(oxides) as Determined by Batch Experiments and Flow Calorimetry: The Energetics and the Effect of Time, pH and Al Content
Creator:
HARVEY, OMAR RICHARD ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Adsorption ( jstor )
Anions ( jstor )
Electric potential ( jstor )
Oxides ( jstor )
pH ( jstor )
Phosphates ( jstor )
Precipitation ( jstor )
Signals ( jstor )
Soils ( jstor )
Sorption ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Omar Richard Harvey. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2009
Resource Identifier:
439083750 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

P SORPTION CHARACTERISTICS ON AMORPHOUS AL-FE CO-PRECIPITATED HYDR(OXIDES) AS DETERMINED BY BATCH EXPERIMENTS AND FLOW CALORIMETRY: THE ENERGETICS AND THE EFFECT OF TIME, PH AND AL CONTENT By OMAR RICHARD HARVEY A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Omar Richard Harvey

PAGE 3

To my entire family especially my mom Angella and my wife, Keisha

PAGE 4

ACKNOWLEDGMENTS I would like to acknowledge all the persons who have made this experience a successful and enjoyable one. First and foremost I am grateful to the Lord Jesus Christ, who above all else has guided me through this research in perfect health and with amazing revelations; I am grateful to my wife and earthly joy, Keisha Rose-Harvey, for her understanding and support throughout this project; and I am grateful to my professor and mentor Dr. R.D. Rhue who throughout all this has taught me more about science than I had ever learned or will ever learn in a formal class room setting. It was through his enthusiasm and love for science coupled with his desire to see students succeed that I have developed a great love and dedication towards research. My gratitude also goes out to my other advisors Dr. Vimala Nair, Dr. Willie Harris and Dr. Nick Comerford for their input in making the completion of this project a possibility, particularly to Dr. Harris for always being available to listen to my concerns and thoughts as well as his immeasurable contribution in the characterizing study. Additionally I would like to thank Bill Reve, Keith Hollien (X-ray), Brad Willenberg (SEM) and Gill Brubaker (PSD) for assistance in laboratory. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................xi CHAPTER 1 LITERATURE REVIEW.............................................................................................1 Introduction...................................................................................................................1 P Sorption in Soils and Hydr(oxides)...........................................................................1 Al and Fe Fractions in Soil and Their Importance to P Sorption..........................1 Mechanism of P Sorption in Soils and Hydr(oxides)............................................4 Factors Affecting P Sorption........................................................................................6 Effect of Surface Species.......................................................................................6 Effect of P Species.................................................................................................8 Effect of Competing Anion...................................................................................8 Energetics of P Sorption.............................................................................................10 2 SYNTHESIS AND CHARACTERIZATION OF MIXED METAL AL-FE HYDR(OXIDES)........................................................................................................12 Introduction.................................................................................................................12 Materials and Methods...............................................................................................12 Synthesis..............................................................................................................12 Scanning Electron Microscopy (SEM)................................................................14 X-ray Diffraction.................................................................................................16 Particle Size Distribution.....................................................................................17 Chemical Composition........................................................................................17 Results and Discussion...............................................................................................18 Qualitative Observations.....................................................................................18 SEM Analysis......................................................................................................18 X-ray Diffraction Analysis..................................................................................39 Metal Content......................................................................................................41 Particle Size Distribution.....................................................................................43 v

PAGE 6

Conclusions.................................................................................................................45 3 BATCH P SORPTION ON MIXED-METAL HYDR(OXIDES).............................48 Introduction.................................................................................................................48 Material and Methods.................................................................................................49 P Sorption as a Function of Time........................................................................49 P Sorption as a Function of pH and Al:Fe Content.............................................50 P Analysis............................................................................................................50 Sorbed P Determination......................................................................................51 Results and Discussions..............................................................................................52 P Sorption as a Function of Time at Fixed pH....................................................52 P Sorption as a Function pH on Different Hydr(oxides).....................................55 P Sorption as a Function of Al Content at Different pH.....................................59 Conclusion..................................................................................................................62 4 P SORPTION EFFECTS ON MIXED-METAL HYDR(OXIDE) SURFACES AS DETERMINED BY FLOW CALORIMETRY..........................................................63 Introduction.................................................................................................................63 Materials and Methods...............................................................................................64 Instrumentation....................................................................................................64 Ion Exchange and P Sorption..............................................................................65 Results and Discussion...............................................................................................67 P Sorption............................................................................................................67 Ion Exchange and the Effect of P Sorption.........................................................69 Cation exchange...........................................................................................69 Anion exchange............................................................................................70 Conclusions.................................................................................................................79 5 SUMMARY................................................................................................................80 LIST OF REFERENCES...................................................................................................83 BIOGRAPHICAL SKETCH.............................................................................................89 vi

PAGE 7

LIST OF TABLES Table page 2-1 Mass of salt used and expected Al content in hydr(oxides)....................................14 4-1 Heats of P sorption on mixed-metal hydroxides......................................................68 4-2 Heats of anion exchange on mixed-metal hydroxides.............................................78 vii

PAGE 8

LIST OF FIGURES Figure page 1-1 P species in solution...................................................................................................9 2-1 Images (x1000) of (a) Al-hydr(oxide) (b) Fe-hydr(oxide) at 15kV.........................22 2-2 Distribution of Al and Fe in 1:10 hydr(oxide) as determined by SEM at 15kV......23 2-3 Distribution of Al and Fe in 1:5 hydr(oxide) as determined by SEM at 15kV........23 2-4 Distribution of Al and Fe in 1:2 hydr(oxide) as determined by SEM at 15kV........24 2-5 Distribution of Al and Fe in 1:1 hydr(oxide) as determined by SEM at 15kV........25 2-6 Distribution of Al and Fe in 2:1 hydr(oxide) as determined by SEM at 15kV........26 2-7 Distribution of Al and Fe in 5:1 hydr(oxide) as determined by SEM at 15kV........27 2-8 Distribution of Al and Fe in 10:1 hydr(oxide) as determined by SEM at 15kV......28 2-9 Distribution of Al and Fe in suspected Al rich area for 10:1 hydr(oxide) as determined by SEM at 15kV....................................................................................29 2-10 Metal distribution in Fe-hydr(oxide) as a function of accelerating voltage.............30 2-11 Metal distribution in Al-hydr(oxide) as a function of accelerating voltage.............31 2-12 Metal distribution in 1:10 hydr(oxide) as a function of accelerating voltage..........32 2-13 Metal distribution in 1:5 hydr(oxide) as a function of accelerating voltage............33 2-14 Metal distribution in 1:2 hydr(oxide) as a function of accelerating voltage............34 2-15 Metal distribution in 1:1 hydr(oxide) as a function of accelerating voltage............35 2-16 Metal distribution in 2:1 hydr(oxide) as a function of accelerating voltage............36 2-17 Metal distribution in 5:1 hydr(oxide) as a function of accelerating voltage............37 2-18 Metal distribution in 10:1 hydr(oxide) as a function of accelerating voltage..........38 viii

PAGE 9

2-19 X-ray diffractograms of freshly precipitated mixed-metal hydr(oxides).................40 2-20 X-ray diffractograms of mixed-metal hydr(oxides) three months after synthesis...41 2-21 Aluminum content of hydr(oxides)..........................................................................42 2-22 Total metal content of mixed-metal hydr(oxides)....................................................42 2-23 Particle size distribution in mixed-metal hydr(oxides)............................................44 2-24 Variation in particle size fraction with Al content...................................................44 3-1 Standard Curve for P standards................................................................................51 3-2 Batch P Sorption with time on mixed-metal hydr(oxides).......................................54 3-3 Changes in rate of P sorption on hydr(oxides) over time.........................................54 3-4 Fractional sorption with time (based on 24 hr sorption maximum).........................55 3-5 P sorption on Al-Fe hydr(oxides) as a function of pH.............................................58 3-6 Loss of sorption with Al content at different pH.....................................................59 3-7 P sorption as a function of Al content......................................................................61 3-8 Al-Fe interactions accounting for differences in sorption........................................62 4-1 Chloride standard curve...........................................................................................66 4-2 Exotherms for P sorption on mixed-metal hydr(oxides)..........................................68 4-3 Relationship between peak area and quantity of P sorbed.......................................69 4-4 Heats of anion exchange on Fe (0%) hydr(oxide)...................................................72 4-5 Heats of anion exchange on 1:10 (9.1%) hydr(oxide).............................................73 4-6 Heats of anion exchange on 1:5 (16.7%) hydr(oxide).............................................73 4-7 Heats of anion exchange on 1:2 (33.3%) hydr(oxide).............................................74 4-8 Heats of anion exchange on 1:1 (50%) hydr(oxide)................................................74 4-9 Heats of anion exchange on 2:1 (66.7%) hydr(oxide).............................................75 4-10 Heats of anion exchange on 5:1 (83.3%) hydr(oxide).............................................75 4-11 Heats of anion exchange on 10:1 (90.9%) hydr(oxide)...........................................76 ix

PAGE 10

4-12 Heats of anion exchange on Al (100%) hydr(oxide)...............................................76 4-13 Relationship between peak area for NO 3 / Cl exchange and Al content...................77 4-14 Change in heat signal peak area per unit of Cl sorbed.............................................77 4-15 Anion exchange capacity of hydr(oxides)................................................................78 4-16 Effect of P sorption on anion peak area...................................................................79 x

PAGE 11

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science P SORPTION CHARACTERISTICS ON AMORPHOUS AL-FE CO-PRECIPITATED HYDR(OXIDES) AS DETERMINED BY BATCH EXPERIMENTS AND FLOW CALORIMETRY: THE ENERGETICS AND THE EFFECT OF TIME, PH AND AL CONTENT By Omar Richard Harvey August 2004 Chair: R.D. Rhue Major Department: Soil and Water Sciences Al and Fe hydr(oxides) are significant adsorbers of P in soil and aquatic systems. They are commonly found existing as mixed rather than pure phase hydr(oxides). This study was conducted to determine the effect of Al content on P sorption in co-precipitated amorphous Al-Fe systems. Al-Fe hydr(oxides) containing between 0 and 100 mol% Al was synthesized by co-precipitation from Al-Fe chloride solutions and characterized by XRD, SEM, PSD and acid digestion. P sorption as a function of time, pH and Al content were determined by batch experiments; heats of adsorption and the effect of P sorption on surface charge were determined by flow calorimetry. SEM results suggested that surface precipitation of Al onto Fe had occurred, at least above 10 mol% Al, though this was not confirmed by optical observations. The results also showed the formation of a dual phase of mixed metal hydr(oxide) and pure Al hydr(oxide) at Al content greater than 80 mol%. P adsorption rate in batch experiments was not affected by Al content but the amount of P xi

PAGE 12

adsorbed increased with Al content. P sorption decreased with pH at all Al contents but the decrease tended to be greater for the lower Al containing hydr(oxides). P adsorption was exothermic on all hydroxides with an average heat of adsorption of 33 5 KJ/mol. P adsorption resulted in reduced heats of Cl/NO 3 exchange but did not result in any detectable cation exchange. Heats of anion exchange on the hydr(oxides) averaged (2.8.4 KJ/mol) at ratios greater than 1:2. Heat signal peak areas for Cl/NO 3 exchange on pure Fe hydroxide were about those on pure Al hydr(oxide) and were attributed to increased AEC with Al content at least above 20 mol% Al. Although the amount of P adsorbed in the calorimetric study was similar on all the hydr(oxides), the heat of Cl/NO 3 exchange following P sorption on the Fe-rich hydr(oxides) was about that on the Al-rich hydr(oxides). Additional work is needed to determine the extent to which Fe affects the heat of exchange and the cause of this effect. xii

PAGE 13

CHAPTER 1 LITERATURE REVIEW Introduction Phosphorus (P) has been implicated as a one of the nutrients responsible for eutrophication of many water bodies through leaching or surface run-off from agricultural farms. The retention of P in the soil is therefore of critical importance to limiting the quantity of P loss form these systems to water bodies. P retention (sorption) capacity of acid soils (Freese et al. 1995) as well as wetland systems (Reddy et al. 1995) has been found to be strongly correlated with Al and Fe hydr(oxide) content. These hydr(oxides) exist more commonly in nature as mixed Al-Fe hydr(oxides) rather than pure phase Al or Fe hydr(oxides) and maybe formed through co-precipitation, sequential precipitation or agglomeration (Anderson and Benjamin 1990a). Al content in these hydr(oxides) may vary from 0-100 mol% with > 40 mol % being more common in sandy soils. The properties of the hydr(oxides) formed are largely dependent on the formation condition such as Al and Fe solution concentration (Mani and Rao 1982), pH (Blangenois et al. 2004) and temperature (El-Sharkawy et al. 2000). P Sorption in Soils and Hydr(oxides) Al and Fe Fractions in Soil and Their Importance to P Sorption Aluminum and Fe are ubiquitous in soil systems. In addition to existing in soil solution as hydrated ions, they can be found in soils as components of primary minerals such as biotite and or, secondary minerals such the clays and metal hydr(oxides). Al and Fe may also be found in the organic fraction of the soil where they are often bound to 1

PAGE 14

2 organic functional groups to form highly stable organo-metal complexes (McBride 1994). Despite their ubiquity only the Al and Fe fractions in clays and hydr(oxides) have been shown to be significant in P sorption. The available pool of P in soils is the most important factor from an agronomic or environmental standpoint. It represents the fraction of soil P in solution and solid phase that is available for plant uptake and the P more likely to leach into groundwater sources with subsequent movement laterally to surface water bodies. Phosphorus mobility is greatly restricted by sorption on soil constituents. In alkaline soils Ca has the greatest effect on P mobility (McDowell and Condron 2001) through the formation of CaP precipitates while in acidic soils the Al and Fe fractions have the greatest effect on P mobility (McDowell and Condron 2001, Zhou et al. 1997) through sorption to surfaces of oxides, precipitation as metal phosphates or formation of organo-metal complexes. Oxides, hydroxides and oxyhydroxides collectively referred to hydr(oxides) are the most important of the Al and Fe fraction. Mcdowell and Condron (2001) using experiments to remove the organic, Al and Fe oxide, and acid soluble fraction of soil found that removal of Al and Fe oxides had the greatest effect in decreasing P sorption and increasing desorption. The quantity and properties of Al and Fe hydr(oxides) are largely dependent upon soil genesis conditions such as parent material and climate. Although crystalline oxides of Al and Fe may influence P sorption, the amorphous hydr(oxides) have been shown to be more important in sorption of P. Freese et al. (1992) reported that total P sorption in topsoil and subsoil samples from eastern Germany was predominantly related to the amounts of amorphous Fe and Al oxides with no significant

PAGE 15

3 relationship between crystalline Fe and Al oxide and total P sorption. In addition, in evaluating models for P sorption, Freese et al. (1992) showed that sorption on soils where crystalline oxides were prevalent could be adequately predicted by amorphous oxides only. A recent report by Agbenin (2003), however, indicated the opposite. In comparing dithioniteand -oxalate extractable Fe and Al effect on P sorption in several savanna Alfisols he found a greater correlation between dithionite extractable Al (Al-d)and Fe (Fe-d) than oxalate extractable (Al-o and Fe-o). He reported that 73% of variability in P sorption could be accounted for by dithionite extractable Al and Fe. Further he reported a difference in the effect of Al and Fe with Fe-d > Al-d > Al-o and no relation with Fe-o. Agbenin (2003), though attributing greater significance to crystalline oxides did report low sorption capacities from soils studied. Zhang et al. (2001), using sequential extraction methods to investigate the influence of various fractions of Fe and Al on phosphate solubility and reactions in citrus sandy soils, reported a significant correlation between P, crystalline and amorphous Fe as well as amorphous Al but not with crystalline Al. The importance of an Al or Fe fraction (amorphous or crystalline) in P sorption is likely dependent on the amounts of those fraction. However, long term sorption and the quantity of P adsorbed is predominantly a function of crystallinity and surface area (including porosity) rather than quantity (Borggaard 1983, Parfitt 1989, Freese et al. 1992). The large surface area and low crystallinity of amorphous oxides makes them more reactive towards P. Total surface area of crystalline oxides is less, due to lack of internal surfaces and terminal functional OH groups (Freese et al. 1992). This lack of internal surfaces prevents significant penetration of phosphate into the crystal lattice

PAGE 16

4 (Torrent et al. 1990, Barrow 1983). Torrent et al. (1990) attributed the slow reaction phase observed in P sorption on goethite to the presence of amorphous ferrihydrite. Parfitt (1989) observed no slow reaction phase in studying phosphate sorption on highly crystallized goethite and concluded that solid-state diffusion of phosphate does not occur in these crystals. Mechanism of P Sorption in Soils and Hydr(oxides) Phosphorus sorption occurs mainly on the surfaces of variable charge minerals such as allophanes, hydr(oxides) of Al, Fe and Mn as well as on the edges of silicate clays via ligand exchange. In ligand exchange, H 2 0 or valence unsatisfied OH ligands attached to a single metal atom (terminal OH ) or two metal atoms (bridging OH ) are replaced by the P anion resulting in the formation of a bi-nuclear, inner sphere metal-phosphate complex (Bleam et al. 1991, Persson et al. 1996). P is also known to replace other ligands such as sulphate or silicate (Pardo and Guadalix, 1990). In addition to ligand exchange, P may also be adsorbed to the variable charged surfaces through anion exchange. This occurs via electrostatic attractions at pH values below the point of zero charge of the surface. P sorption by anion exchange is rapid, reversible and non-specific and is responsible for only a small fraction of the total sorption observed. Ligand exchange on the other hand, is specific and tends towards irreversibility. Additionally, ligand exchange is characterized by the release of OHinto solution and a change in surface charge to a more negative value (Rhue and Harris 1999). Phosphorus sorption on Al and Fe hydroxides in soils occurs via two time-dependent reaction mechanisms. The first is thought to be a quick ligand or ion exchange reaction occurring on anion exchange sites. These exchange reactions can last from a few minutes to a few days. The second mechanism is not as clear or widely agreed upon

PAGE 17

5 (Rhue and Harris 1999). Suggested mechanisms reflect the fact that this reaction is much longer than the exchange reactions of the first mechanism (lasting up to months) and account for the hysteretic behavior observed in P adsorption/ desorption studies. What is not clear is the actual mechanism by which P is sorbed to the soil during this phase. The two most common theories are precipitation of P as metal phosphate on surfaces, and diffusion of P into micropores or through the amorphous hydr(oxide) coatings. The theory of formation of metal phosphate on surfaces is supported by spectroscopic evidence (Bleam et al., 1991 and Lookman et al., 1994), and potentiometric titration evidence (Li and Stanforth, 2000). Recent research however, using the more accurate X-ray Adsorption Near Edge Spectroscopy (Arai and Sparks, 2001) to look at phosphate adsorption on soils, hydrous ferric oxides and Al hydroxide showed no evidence of a metal phosphate phase. Precipitation is further disputed because, except in highly fertilized soils, solutions are often undersaturated with respect to most crystalline phosphorus compounds. Pierzynski et al. (1990) found P existing as amorphous Al-P precipitates in highly fertilized soils. These were found as discrete particles or coatings on other particles. The formation of these were however attributed to P induced weathering of aluminum silicate minerals and subsequent co-precipitation of Al-Si-P, rather than P precipitation onto the Al hydr(oxide) or silicate surface. Van Riemsdijk (Van Riemsdijk and Haan, 1981 and Van Riemsdijk et al., 1984) suggested that the irreversibility of phosphate sorption observed in adsorption/ desorption studies was due to P diffusing through the poorly structured Al and Fe hydr(oxide) coating on soil surfaces and being adsorbed on these internal surfaces. He suggested that

PAGE 18

6 this adsorbed phosphorus is often “fixed” because of the resistance of the amorphous hydroxides to weathering rendering internal sites inaccessible. Other work (Madrid and de Arambarri 1985 and Willet et al 1988) has attributed the slow phase of P sorption to accessibility of the phosphate to surface mesoporosity. They suggested that as P “diffuses” into the hydr(oxide) structure the pores become smaller, and smaller hence the accessibility to sorption sites is reduced. Both author working on Fe-hydr(oxides) found no evidence of Fe phosphate formation. Willet et al.(1988) found that all the P sorbed could be desorbed with 0.1 M NaOH presenting evidence that no precipitates were formed. Factors Affecting P Sorption Effect of Surface Species The surface species present affects P sorption by influencing the number and reactivity of the surfaces. Speciation of metal hydr(oxides) is largely pH dependent and can be sufficiently modeled by observing the behavior of the metal in solution (McBride 1994). Al and Fe exist as hydrolyzed ions in solution, particularly above pH 4, resembling an Al-OH surface site. The degree of hydrolysis determines the number of ligands present. Easily hydrolyzable metal such as Al and Fe, with their high charge density, tend to have more ligands associated with them than other metals. The charge on this site is dependent on the pH, being positively charged at lower pH (4.7-6.5), having no charge at pH 6.5-8.0 and negatively charged at pH values >8.0 (Nilsson et al. 1992). The positive charge and negative charge arises from protonation and deprotonation respectively. At low pH the hydroxyl groups accept a proton resulting in the formation of OH 2 + (water) groups attached to the metal atom. Ligands of this form tend to dissociate more readily than OH and therefore more ligand exchange is expected and is observed at

PAGE 19

7 lower pH where the M-OH sites (M is the coordinating metal) are protonated to form M-OH 2 . At high pH, above their pK a values the M-OH group deprotonates forming negatively charged species which repels the negatively charged P, resulting in less sorption. Ligands coordinated to two metal atoms (bi-dentate ligands) tend to dissociate less readily than those coordinated to one metal atom (mono-dentate ligands) and hence are less reactive. Based on co-ordination and overall charge three types of hydroxyl groups have been identified on metal oxide surfaces. These are the ol, hydroxo and aquo type hydroxyl groups (Rhue and Harris 1999). The ol groups are bi-dentate with no charge, while the hydroxo and aquo are mono-dentate with a charge of -1/2 and +1/2 respectively. Assuming the OH group is in octahedral co-ordination with Al and Fe the aquo group is dominant under acidic condition and is the most reactive of towards P. Although mono-dentate ligand exchange and hence the formation of a mono-dentate metal-P complex would be more thermodynamically favorable, infrared (IR) spectroscopic evidence (Tejedor-Tejedor and Anderson 1990) suggests that P may be largely sorbed through bi-dentate complexation and only a small amount of P is adsorbed by mono-dentate complexation. Persson et. al. (1996), suggested that this discrepancy between thermodynamics and spectroscopic evidence is because the samples used in the spectroscopy studies were not prepared in accordance to distribution models for surface species derived from potentiometry or adsorption studies. By using adsorption samples of P on goethite which represented different points in a distribution diagram obtained from thermodynamic data in FTIR studies they concluded that P is mostly adsorbed by mono-dentate complexation.

PAGE 20

8 Effect of P Species Like surface speciation, P speciation is greatly dependent on pH. In solution, P can exist as PO 4 3 or the protonated species HPO 4 2, H 2 PO 4 and H 3 PO 4 . As pH increases, the change in dominance of each species coincides with the pK a1 , pK a2 and pK a3 of H 3 PO 4 (Figure 1-1) indicating increasing de-protonation with pH. For pH values less than 2 the tri-protic, uncharged H 3 PO 4 species dominates; between pH 2 and 6 the diprotic, mono-valent species H 2 PO 4 dominates; between 7 and 12 the mono-protic, divalent species, H 2 PO 4 dominates; and above pH 12 the dominant species is the un-protonated, trivalent PO 4 3. Since anions associate more easily with protons and given that the surfaces of hydr(oxides) are protonated only at low pH values, P is expected to adsorb more readily at low pH and less readily at higher pH especially above the pKa of the surface sites. Also, P adsorption is expected to be impeded at higher pH due to competition from OHgroups in solution (Lijklema 1980). Therefore it is unlikely that significant quantities of monoprotic or unprotonated species of P would be found adsorbed to soils under normal conditions. Effect of Competing Anion Most anions of environmental importance, such as arsenate, phosphate and chromate, exist as oxyanions in solution. They adsorb very little in humus and their sorption to mineral surfaces account for their retention in soils. When in solution together these anions will compete for sorption sites. The competitiveness of an anion is measured by its selectivity for a particular surface site (McBride 1994). Selectivity is determined by anion properties such as shared charge and electronegativity. For anions with smaller shared charge the effective negative charge residing on each O is greater, allowing a

PAGE 21

9 greater affinity for the surface. Additionally, the surface-oxyanion bond formed for smaller shared charge anions tends to be greater for the same reason. For anions with similar shared charge, electronegativity is used to determine their relative competitiveness. Anions with smaller electronegativity tend to be more competitive than anions with similar shared charge but a larger electronegativity. Chromate despite having similar shared charge, of 1.5, as selenate has been shown to bind more strongly to Fe-hydr(oxide) at any particular pH. Figure 1-1. P species in solution. Based on shared charge and electronegativity only borate (B(OH) 4), silicate (SiO 4 4 ), and OH would be expected to out compete P for sorption sites, but the weak acidic nature of these groups means that they dissociate at higher pH values and therefore would only compete with P at higher pH values. Other anions such as NO 3 , ClO 4 and the halides (except F ) only bind to variable charge surfaces via outer-sphere electrostatic

PAGE 22

10 bonds and therefore do not inhibit P sorption. Tanada et al. 2003 reported selectivity values for P 1000 times greater than those for Cl and NO 3 and more than 100 times greater than those for SO 24 and CO 23 . Energetics of P Sorption The energetics of P sorption is not clearly understood and is surrounded by conflicting reports. Malati (Malati et al. 1993) in studying P sorption on silicate clays and a titanium oxide at pH values below their point of zero charge (PZC) reported that P sorption on these surfaces was endothermic. Hundal (1988) also reported P sorption on medium clay loam soils to be endothermic. Endothermic P sorption has also been reported by Mustafa et al. (1990) for sorption onto OH and Cl forms of the anion exchanger, Amberlite IRA-400 and by Chien et al. (1982) on soils. In all cases, sorption was observed to increase with increasing temperature a trend consistent with endothermic reactions. The opposite effect of temperature was observed by Froelich (1988) who reported that P sorption decreased with increasing temperature for suspended sediments that had reached equilibrium with P solution. Barrow (1983) found that when neither sorption nor desorption was occurring, on soils incubated for several days at different temperatures, P concentration in solution increased with temperature. He concluded that the reverse reaction, sorption, was therefore exothermic. These contradictory results suggest that equilibration time may be significant in determining whether P sorption is endothermic or exothermic. Initial sorption may be endothermic but as time increases the sorption mechanism may become exothermic as suggested by Rhue and Harris (1999). Contradictions in P sorption energetics may also have arisen from the method of determination used to describe P sorption. Many times indirect methods, such as Langmuir based derivations (Mustafa et al. 1990, Malati et al. 1993) or the Clapeyron

PAGE 23

11 equation (Hundal 1988), rather than direct methods were used to determine energies of sorption. The problem with indirect methods is that standard conditions often used vary and may cause discrepancies in the results obtained. Rhue et al. (2002) used flow calorimetry to measure the direct enthalpy associated, they found that P sorption on soils and Al-hydr(oxide) was exothermic while precipitation was endothermic. Hundal (1988) in explaining his results suggested that precipitation may be responsible for the endothermic P sorption observed. The major objective of this research was to investigate the effect of Al content on P sorption characteristics in amorphous co-precipitated Al-Fe hydr(oxide) systems. This was done through several specific tasks or objectives. Task 1 was to synthesize Al-Fe hydr(oxides) with Al content ranging between 0-100 mol% via co-precipitation from solution and then characterizing the hydr(oxides) based on physical and chemical properties. Task 2 was to determine the P sorption characteristics on each hydr(oxide) using batch experiments and flow calorimetry. Batch experiments were used to determine P sorption as a function of time, pH and Al content while flow calorimetry was used to determine the heats of adsorption and the effect of P sorption on surface charge. The final task was to use differences in the chemical and physical properties of the hydr(oxides) to explain differences in P sorption characteristics observed (if any).

PAGE 24

CHAPTER 2 SYNTHESIS AND CHARACTERIZATION OF MIXED METAL AL-FE HYDR(OXIDES) Introduction Compounds commonly found in soils are often formed through co-precipitation, sequential precipitation or agglomeration (Anderson and Benjamin 1990). In addition to the mechanism by which a compound is formed, the properties of that compound are often largely dependent upon formation condition such as pH (Blangenois et al. 2004), temperature (El-Sharkawy et al. 2000) and solution composition. El-Sharkawy (2000) in his work found that the pore size and pore size variability increases with increase temperature of formation in Al-Fe co-precipitated system. There is also evidence that the Al and Fe content in the precipitating solution may influence the particle size distribution, surface area, the structure as well as the metal distribution of the mixed hydr(oxide) formed (Anderson and Benjamin 1990, Rodic et al. 2001 and Wolska et al. 1994). Additionally surface charge and sorption characteristics may be influenced. Materials and Methods Synthesis Mixed-metal hydroxides containing Al: Fe in ratios ranging between 0:1 and 1:0 (0-100 mol% Al) were synthesized through the titration of their respective mixed metal salt solution with a base. (Throughout this thesis Al:Fe and mol% Al will be used interchangeably. For metal ratios the first number will always be Al and the second Fe). Varying quantities of reagent grade Al and Fe chloride salt (Table 2-1) were added 12

PAGE 25

13 together in 250mL beakers and dissolved with DDW water to yield Al-Fe mixed-metal chloride solutions. Each metal chloride solution was then titrated, at room temperature, against 5M NaOH. This was done through the drop wise addition of the NaOH to the metal chloride solution, until the pH of the resulting suspension was between 6 and 7. A magnetic stirrer was used to ensure constant and uniform stirring throughout the titration. The suspension was then left overnight to allow settling of the precipitate. After settling, a decanting technique was used to carefully separate the resulting NaCl solution from the precipitate. The precipitate was then washed several times to remove any remaining salt in the precipitate. In each washing step, double distilled water (DDW) was added to the precipitate, the suspension was stirred, centrifuged at 2000 RPM, and decanted. Following washing, the precipitate was dried for 24 hours at 70 C. The dried precipitate was then crushed using a mortar and pestle and sieved through a 150 micron sieve. The resulting powder was stored in 20 mL plastic vials at room temperature for use in characterization, batch, and calorimetric studies. The quantity of each metal salt used to make the salt solution was dependent on the dominant metal in the hydr(oxide) and the mass of powder targeted. For the pure Al (1:0) and Fe(0:1) hydr(oxides), where a mass of 1g of hydr(oxide) was targeted, an equivalent mass of the respective metal chloride salt sufficient for 1g of hydr(oxide) was used. For the mixed hydr(oxides), where target mass was between 1 and 2g the dominant metal dictated the total quantity of each salt used. For an Al dominated hydr(oxide) a mass of Al chloride salt containing at least 1g of Al is added. The quantity of iron is then proportioned accordingly to ensure that the desired Al: Fe ratio and mol% Al is achieved at pH 6-7, since maximum precipitation is expected at these pH values. The opposite was

PAGE 26

14 done for Fe dominated hydr(oxides). For the 1:1 (50 mol% Al), where neither of the metals was dominant, the mass of Al and Fe chloride salt used was calculated to give an equal Al and Fe molar composition while ensuring a mass above 1g. Table 2-1. Mass of salt used and expected Al content in hydr(oxides) Sample ID (molar ratio) Expected Al content (mol%) AlCl 3 .6H 2 0 (g) FeCl 3 .6H 2 0 (g) Al content (mol) Fe content (mol) 0:1 0 4.83 0 0.018 1:10 9.1 0.44 4.83 0.0018 0.018 1:5 16.7 0.87 4.83 0.0036 0.018 1:2 33.3 2.16 4.83 0.009 0.018 1:1 50.0 4.47 5.00 0.0185 0.0185 2:1 66.7 8.94 5.00 0.037 0.0185 5:1 83.3 8.94 2.00 0.037 0.0074 10:1 90.9 8.94 1.00 0.037 0.0037 1:0 100 8.94 0.037 0 Scanning Electron Microscopy (SEM) Scanning electron microscopy (SEM) uses a finely focused beam of electrons to irradiate the sample being observed. When the beam is radiated onto the sample, characteristic signals (eg. x-rays) are released. The intensity of the signals released will depend on the shape, chemical composition and crystal orientation of the irradiated volume. When analyzed, these signals can be used to give structural and elemental information about the sample being examined (Goldstein et al. 2003). The SEM is often fitted with accessories that can analyze different signals. When fitted with an energy dispersive spectrometer (EDS) and an electron probe micro-analyzer, SEM can provide a rapid evaluation of the elemental constituents of a sample. The EDS and electron probe micro-analyzer detects x-rays of elements above atomic number 4 at typical beam currents used for secondary electron imaging. The output information from the EDS is both qualitative and quantitative, but the quantitative data is

PAGE 27

15 more safely regarded as semi-quantitative. The detection limit is usually between 1000-3000 mg/L. With SEM, the sample is analyzed non-destructively, under vacuum with a spatial resolution of 1m. Lateral resolution ranges between 10 and 50 nm and depth resolution from 1-1000 nm depending on the type of signal and the mode of operation. A low voltage mode of operation uses an accelerating voltage 5kV to irradiate the sample and is used in close-to-surface analyses (1-5 nm). For depths greater than 5nm, accelerating voltages ranging between 10 and 30kV are usually required. Accelerating voltages of these magnitudes are characteristic of high voltage modes of operation and produces output that is of higher resolution than those obtained in low voltage mode. The quality of the output is also largely dependent on the time of analysis used for a sample. The longer the time of analysis, the better the quality and resolution of the output. For qualitative analyses a time of 10-100s is required. For quantitative analyses and mapping the required time is 100-500 and 1000-10000 s, respectively. SEM was employed in this research to determine the morphology of the hydr(oxides) and the distribution of Al and Fe throughout each sample. Before being analyzed, hydr(oxide) samples were mounted onto carbon mounts with a maximum of 5 samples per mount. Samples were mounted as drops of suspensions. The liquid was dried leaving the powder particles fixed to the surface of the mount. In mounting the samples, 2-3 mg of hydr(oxide) powder were suspended in approximately 2 mL of DDW and shaken to maintain even distribution of particles. A drop of the resulting suspension was then placed on the carbon mount and allowed to air-dry for 24 h. During this time the mounts were partially covered to prevent contamination of the samples. A carbon-based

PAGE 28

16 coating of vaporized graphite was then placed over the mounted samples to reduce charging effects and increase conductivity during analysis. Samples were examined using a Jeol Scanning electron microscope (JSM 6400) at accelerating voltages of 10, 15 and 20kV. Analyses were initially done using an accelerating voltage of 15kV, but results were consistent with the formation of solid solution or surface precipitation. To get an idea of the process that is more likely occurring a thin film on a substrate method (p 454-465, Goldstein et al. 2003) was employed. Accelerating voltages of 10kV and 20kV were used. Because sample penetration depth varies with accelerating voltage it would be expected that if surface precipitation dominated, the relative concentration of one the metals would vary disproportionately with accelerating voltage. If a solid solution is formed however there would be no relative change in concentration relative to depth or accelerating voltages, assuming signals are kept proportional. An image, EDS and Al-Fe dotmap were generated for each sample, at each accelerating voltage. The image was used for qualitative observations (morphology) while the EDS and dotmap were used to determine the relative distribution of Al and Fe at the surface and interior of the particle. Images were done at magnifications of x1400 (for 10kV and 20kV) and x1000 (for 15kV), while EDS and dotmaps were done at x10000. The area for scanning was chosen so as to ensure a representative distribution of particle sizes and minimal charging effects. X-ray Diffraction X-ray diffraction (XRD) analysis, like EDS, is semi-quantitative and is used in determining the structure of a sample. Additionally XRD is used in the identification of crystalline phases in solid (powder) samples. The technique is based on the diffraction of x-rays on a crystal lattice and assumes that every crystalline phase has a characteristic x

PAGE 29

17 ray diffraction (fingerprint), which can be used in its identification. X-ray diffraction was used to determine the structure of the different mixed-metal hydr(oxides) and to identify any Al or Fe crystalline phases present. Because the samples were needed for batch and flow sorption studies non-destructive dry powder XRD was used. Samples were mounted as dry powder onto aluminum powder mounts and analyzed using a Nicolet x-ray diffractometer with a CuK beam at 35kV and 20 mA. The resulting diffractograms were then observed for crystalline phases of Al or Fe attributable to the hydr(oxides). Particle Size Distribution Particle size distribution in the hydr(oxides) was determined with a Beckman Coulter particle size analyzer (LS13320) at a standard obscuration of 8-12% and 40-50% polarization intensity differential scattering (PIDS) and laser loading rate respectively. An alumina in water optical model was used with an analysis time of 90 seconds per sample for two replicates, rinsing between each run. Before analysis 10-20 mg of hydr(oxide) was suspended in DDW water and sonicated for two minutes to ensure particles are not agglomerated. The sonicated suspension was then loaded into the instrument with a pipette to the desired obscuration and analyzed. A pump speed of 75% percent was used to prevent settling of larger particles during analysis. Chemical Composition Al and Fe content were quantitatively determined through acid digestion of hydr(oxide) samples and analysis of the digest for the respective metal using atomic absorption spectrometry. Twothree mg of each sample were placed into 2 mL glass vials and 3-4 drops of hydrochloric acid added to dissolve the hydr(oxide) which is insoluble in water but soluble at low pH. The total volume in each flask was then made

PAGE 30

18 up to about 1.5mL with DDW water, capped and placed in boiling water for 20-30 minutes to complete digestion. After complete digestion, flasks were allowed to cool to room temperature and the solution transferred to 25 mL volumetric flasks. The solution was subsequently brought to volume (25mL) with DDW water and then analyzed for Al and Fe concentration using a Varian atomic absorption spectrometer (SpectraAA 220FS). Results and Discussion Qualitative Observations When uncrushed, the mixed metal hydr(oxides) all had a characteristic brown color (inherited from the iron) and were indistinguishable from each other. The only exception was the 10:1 mixture for which, in addition to the brown color, streaks of lighter colored material could be seen layered between the darker material. Separation and acid digestion of the two phases showed that the dark material was an Al-Fe phase with Al:Fe ratio of 5:1, while the light colored material was almost purely Al with an Al:Fe ratio of 20:1. This suggests that for ratios above 5:1, the solid consists of a mixed Al-Fe phase as well as a pure Al phase. Crushed samples were easily distinguishable by color and texture. As the Al:Fe ratio increased the color of the hydr(oxide) became paler. Also the texture of the hydr(oxides) had a more silt-like powder feel with increased Al:Fe. Low Al:Fe containing hydr(oxides) were more granular and coarse textured. SEM Analysis Elemental dotmaps (Figures 2-1 to 2-18) showed a close intra particle association between Al and Fe for all hydr(oxides). This was indicative of the formation of a solid solution or the surface precipitation of one metal onto the other. Solid solution formation in mixed metal Al-Fe oxides has been reported by Mani and Rao (1982) and Wolska et

PAGE 31

19 al. (1994) but only at Al content of 10 mol% (Al:Fe of 1:9) and or high temperatures much greater than the 70C maximum used in the synthesis of these hydr(oxides). Except for the 1:10 mixture (9.1 mol%), the Al:Fe content of all the hydr(oxides) exceeded the limit for solid solution formation as suggested by Wolska et al. Solid solution formation in Al-Fe hydr(oxides) is also unlikely based on chemical and solubility criteria. Although Al and Fe have very similar atomic radii (0.05nm and 0.06nm respectively) and could be easily substituted for each other into the hydr(oxide) structure, as in isomorphous substitution, it is unlikely that they would form a solid solution as they have very different solubility constants (K so ) in hydr(oxides) (Fe = 10 -39 and Al= 10 -31 ) (McBride, 1994). This difference in solubility constants would mean that the distribution co-effiecient for these metals in a solid would vary significantly from 1 and therefore they should not form a homogeneous compound together. The different pHs at which these metals precipitate also casts doubt on the formation of a solid solution as Fe tends to precipitate at lower pH values and therefore would precipitate before the maximum precipitation pH ( 6-7) for Al is reached. The lower precipitation pH as well as the lower K so for Fe hydr(oxides) would suggest that if surface precipitation is the major process involved in the formation of these mixed metal hydr(oxides) the most likely scenario would be a precipitation of Al onto the surface of the Fe. This precipitation would therefore concentrate the Fe toward the interior of the particle and the Al to the exterior. The comparison of dotmaps and EDSs (Figure 2-12 to 2-18 ) obtained at 10kV and 20kV indicates that surface precipitation may well be the process by which these mixed metal hydr(oxides) are formed, especially above 1:10 Al:Fe ratio. At the lower

PAGE 32

20 accelerating voltage (Figure 2-16 b and d), 10kV, all samples showed a greater intensity and relative amounts of Al compared to Fe in the dotmaps and EDSs respectively. With increased accelerating voltage (Figure 2-16 c and e) however, and hence increase depth of penetration into the sample, a disproportionate increase in Fe intensity (from dotmap) and relative amounts (from EDSs) compared to Al were observed suggesting that there is a greater concentration of Fe in the interior of the particle relative to the surface. The dotmaps and EDSs also showed that at 10kV (Figure 2-14b,d to 2-18 b,d), as Al:Fe ratio increases in the hydr(oxides) the intensity and relative amounts of Fe detected decreases. This lowering of the intensity and relative amounts of Fe detected at constant accelerating voltage, and hence constant depth of penetration, suggests an increase in thickness of the Al coating as Al content is increased. Although the SEM results suggested a surface precipitation of Al onto Fe it is worth mentioning that there are several things to be considered before a definitive conclusion can be made. The first is based on the energetics of Al relative to Fe. The fact that Fe characteristic radiation is more energetic than that of Al means that at higher accelerating voltages (deeper penetration depths) Fe is more likely to emerge from the particle relative to Al radiation, thereby resulting in a disproportionate increase in the EDS peaks of the Fe relative to the Al. Whether this effect was significant or accounted for the increase in Fe relative to Al seen in the hydr(oxides) with increasing accelerating voltage is unclear but is worth taking into consideration. The second thing to consider is the fact that it is not clear at what scale (nanometer or micrometer) Al precipitation onto an Fe-rich core is likely occurring since the particles are likely to be aggregates rather than discrete particles since the hydr(oxide) powders were obtained by crushing the dried

PAGE 33

21 precipitate. Additionally, based on optics, if the particles had a discrete inner Fe core surrounded by Al, observation of the particles under transmitted light (assuming precipitation of Al was occurring at the micrometer level) should reveal an opaque inner core with bright edges since the Fe would absorb the light and Al would transmit it. This was however not seen upon observation of 1:1 hydr(oxide) particles under a light microscope. Instead the particles looked uniform. Although this was largely scale dependent it is worth considering and may suggest that precipitation may be occurring at the nanometer scale. The use of a more definitive method such as x-ray photoelectron spectroscopy would be useful in confirming the SEM results. Observing the EDS results for Al and Fe in a crystalline material such as biotite under the same SEM conditions used may also be helpful. The presence of an Al phase detected at the 10:1 ratio (90.9 mol% Al) (Figure 2-9), probably corresponded to the light colored streaks observed in the uncrushed samples, supporting the suggested phase change above Al:Fe ratio of 5:1. In addition, these results suggest an increase in thickness of the coating of Fe by Al between 1:1 and 5:1 Al:Fe content. A closer look at the dotmaps in Figures 2-4 and 2-5 suggests that full coating of the Fe by Al is not reached until the Al:Fe ratio is 1:1. Above 1:1 there may be a precipitation of Al onto Al resulting in the increase in the thickness of the Al coating on Fe detected in Figures 2-15b to 2-18 b. The Al coating for 5:1 Al:Fe maybe 5 times as thick as that of 1:1 since for a single layer precipitate it would be expected that the phase change occur at an Al:Fe ratio of 1:1 rather than 5:1. Although the suggestion of a multilayered type coating process of Fe by Al might be new, the presence of a pure Al phase in the study of mixed-metal Al-Fe oxides has

PAGE 34

22 been observed by Korecz (1972) and Wolska (1994). Wolska while using infra-red spectroscopy to study the mechanism of Al for Fe substitution in mixed Al-Fe oxides with a maximum 12 mol% Al observed that as Al content increases above 10 mol%, increased amounts of bayerite could be detected. Korecz in using mossabauer spectroscopy to study the incorporation of iron into the structure of corundum reported that for 9:1 Al-Fe oxides only lines characteristic of corundum could be detected. From there he suggested an incorporation limit for Fe into corundum of 5:1 Al:Fe. (a) (b) Figure 2-1. Images (x1000) of (a) Al-hydr(oxide) (b) Fe-hydr(oxide) at 15kV (a) (b)

PAGE 35

23 (c) Figure 2-2. Distribution of Al and Fe in 1:10 hydr(oxide) as determined by SEM at 15kV. (a) image (x1000) (b) EDS (c) Al-Fe dotmap. (a) (b) (c) Figure 2-3. Distribution of Al and Fe in 1:5 hydr(oxide) as determined by SEM at 15kV. (a) image (x1000) (b) EDS (c) Al-Fe dotmap. Nb. NaCl coating masking hydr(oxide) causing distortion in map and image.

PAGE 36

24 (b) (a) (c) Figure 2-4. Distribution of Al and Fe in 1:2 hydr(oxide) as determined by SEM at 15kV. (a) Image (x1000) (b) EDS (c) Al-Fe dotmap.

PAGE 37

25 (a) (b) (c) Figure 2-5. Distribution of Al and Fe in 1:1 hydr(oxide) as determined by SEM at 15kV. (a) Image (x1000) (b) EDS (c) Al-Fe dotmap.

PAGE 38

26 (a) (b) (c) Figure 2-6. Distribution of Al and Fe in 2:1 hydr(oxide) as determined by SEM at 15kV. (a) Image (x1000) (b) EDS (c) Al-Fe dotmap.

PAGE 39

27 (a) (b) (c) Figure 2-7. Distribution of Al and Fe in 5:1 hydr(oxide) as determined by SEM at 15kV. (a) Image (x1000) (b) EDS (c) Al-Fe dotmap.

PAGE 40

28 (a) (b) (c) Figure 2-8. Distribution of Al and Fe in 10:1 hydr(oxide) as determined by SEM at 15kV. (a) image (x1000) (b) EDS (c) Al-Fe dotmap.

PAGE 41

29 (a) (b) (c) Figure 2-9. Distribution of Al and Fe in suspected Al rich area for 10:1 hydr(oxide) as determined by SEM at 15kV. (a) Image (x1000) (b) EDS (c) Al-Fe dotmap.

PAGE 42

30 (a) (b) (c) (d) Figure 2-10. Metal distribution in Fe-hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) dotmap (c) EDS at 10kV (d) EDS 20kV.

PAGE 43

31 (a) (b) (a (c) (d) Figure 2-11 . Metal distribution in Al-hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) dotmap (c) EDS at 10kV (d) EDS 20kV.

PAGE 44

32 (a) (d (b) (c) (d) (e) Figure 2-12. Metal distribution in 1:10 hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV.

PAGE 45

33 (a) (b) (c) (d) (e) Figure 2-13. Metal distribution in 1:5 hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV.

PAGE 46

34 (a) (b) (c) (d) (e) Figure 2-14. Metal distribution in 1:2 hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV.

PAGE 47

35 (a) (b) (c) (d) (e) Figure 2-15. Metal distribution in 1:1 hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV.

PAGE 48

36 (a) (b) (c) Figure 2-16 . Metal distribution in 2:1 hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV. (d) (e)

PAGE 49

37 (a) (b) (c) (d) (e) Figure 2-17 . Metal distribution in 5:1 hydr(oxide) as a function of accelerating voltage. (a) image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV.

PAGE 50

38 (a) (b) (c) (d) (e) Figure 2-18 . Metal distribution in 10:1 hydr(oxide) as a function of accelerating voltage. (a) Image(x1400) (b) Al-Fe dotmap at 10kV (c) Al-Fe dotmap at 20kV (d) EDS at 10kV (e) EDS 20kV.

PAGE 51

39 X-ray Diffraction Analysis X-ray diffractograms (Figure 2-19) showed no diffraction peaks characteristic of Fe or Al hydr(oxide) compounds. The peaks observed were attributable to the presence of NaCl from synthesis and aluminum from the mount used. The absence of any Al or Fe hydr(oxide) peaks indicate that all the synthesized hydr(oxides) were amorphous in nature. A region of higher intensity material (between 2 values of 5 and 14 degrees) was observed in the diffractogram of the 10:1 hydr(oxide) (Figure 2-19). This was only observed in 5:1 and 1:0 samples that had been aged for three months (Figure 2-20). The higher intensity suggests that the material in this region is more ordered (has more structure), compared to the lower intensity material. The fact that this was only observed in the higher Al containing hydr(oxides) indicates that it is related to the structure of Al in these materials. Carim et. al.(1997), in studying the conversion of diaspore to corundum observed that both these phases contained only octahedrally co-ordinated Al. Nuclear magnetic resonance data of an intermediate transitional compound however, showed that Al was occupying both octahedral and tetrahedral positions. This was correlated to higher intensities observed in the x-ray reflections of the same compound. Rodic et. al (2001), in studying the cation distribution in a range of sintered mixed-metal hydroxides with varying Al:Fe ratio, concluded that Fe preferentially occupies the octahedral positions, while the smaller Al ions preferentially occupy the tetrahedral positions. The higher intensity material observed in the freshly precipitated 10:1 hydroxide material may, therefore, be attributed to tetrahedrally co-ordinated aluminum. In addition, since it is observed in freshly precipitated hydroxides, it suggests that in these mixed

PAGE 52

40 metal systems an Al:Fe ratio of 10:1 and greater may enhance crystallization of Al minerals. Figure 2-19. X-ray diffractograms of freshly precipitated mixed-metal hydr(oxides).

PAGE 53

41 Figure 2-20. X-ray diffractograms of mixed-metal hydr(oxides) three months after synthesis. Notice the development of higher intensity regions in 5:1, 10:1 and Al. Metal Content Measured Al content in synthesized hydr(oxides) were consistent with targeted Al content (Figure 2-21) indicating that the method used to prepare the hydr(oxides) was efficient. Total metal content (Figure 2-22) showed a linear decrease with increasing Al content, from 65% for 0:1 (0 mol% Al) to 30% for 1:0 (100 mol% Al). This decrease may be attributed to increase in water associated with the structure of the hydroxide, or increased hydroxylation as Al content increases. Mani et. al.(1982), suggested that due to

PAGE 54

42 R2 = 0.999301020304050607080901000102030405060708090100target Al (mol%)measured Al (mol%) Figure 2-21. Aluminum content of hydr(oxides). R2 = 0.899501020304050607080900102030405060708090100Al (mol%)metal content(%) Figure. 2-22. Total metal content of mixed-metal hydr(oxides) capillary condensation and the colloidal nature of the hydr(oxides) high temperatures are required to remove water from the interior of these materials. Using DTA he recommended that a temperature of 170C was sufficient to remove the bound water in the pure iron oxide. Higher temperatures are, however, required to remove water from

PAGE 55

43 Al-hydr(oxides). The different types of water associated with Al are removed at different temperatures. Bound and constitutional waters are removed at 120C and 260C respectively. El-Sharkawy et. al. (2000), ascribed endothermic peaks observed on the DTA of co-precipitated Al-Fe hydr(oxides) to the sequential removal of physisorbed, bounded and constitutional water. These peaks were observed at 100, 134 and 177C respectively. The drying temperature 70C used for the hydr(oxides) in this study, was probably only sufficient to remove the physisorbed and could therefore, account for the variability in metal content observed in the hydr(oxides). Particle Size Distribution Particle sizes of up to 60 m were observed in the 1:0 (100 mol% Al) and >140m for all other samples. The distribution of particles over the size ranges (Figure 2-23) indicate that most of the particles are less than 100 m (100% for Al hydr(oxide) and > 70% for the others). The similarities in the shape of the particle size distribution for Fe and the mixed hydr(oxides) indicate that of the two metals Fe is controlling the particle size. Figure 2-24 shows that as Al content is increased there is a reduction in the <40 m particle size fraction from 60% 45% and an increase in the > 80 m fraction. The 60-80 m fraction remained constant at 15% irrespective of Al content. The increase in the larger particle size fraction is possibly due to precipitation of Al onto Fe hydr(oxide) particles.

PAGE 56

44 010203040506070809010009.116.733.35066.783.390.9100Al content (mol%)volume % <20 20-40 40-60 60-80 80-100 100-120 120-140 >140 Figure 2-23. Particle size distribution in mixed-metal hydr(oxides). 051015202530354045500102030405060708090100Al content (mol%)volume % <20 20-40 40-60 60-80 80-100 100-120 120-140 >140 Figure 2-24. Variation in particle size fraction with Al content

PAGE 57

45 Conclusions Co-precipitation of Al and Fe from solution resulted in structural as well as textural differences in the resulting mixed-metal hydr(oxides) relative to pure Al or Fe hydr(oxides). The process of formation of these mixed-metal hydr(oxides) is consistent with a surface precipitation of Al-rich hydr(oxide) material onto Fe-rich hydr(oxide) material. The morphology and particle size distribution of these mixed metal hydr(oxides) was dictated by Fe. The degree of Al coverage is likely dependent on the Al:Fe molar ratio in the mixed hydr(oxide) with partial coverage occurring below 1:1, full coverage at 1:1, and increase in thickness of the layer thereafter up to 5:1 above which there is a separation of phases, resulting in the formation of a pure Al phase in addition to the typical Al-coated Fe-rich phase. The separation of phases was accompanied by the appearance of structure in the 5-14 2 region of the x-ray diffractograms which was interpreted as a shift in co-ordination of the Al atoms from octahedral to tetrahedral. This is indicative of an increase in structural development preliminary to crystallization of the Al-Fe hydr(oxide).

PAGE 58

CHAPTER 3 BATCH P SORPTION ON MIXED-METAL HYDR(OXIDES) Introduction Batch experiments are the most commonly used method for studying sorption reaction kinetics on soils and their components. In this technique a known mass of soil (or soil component), from hereon referred to as the adsorbent, is placed in containers with a known volume of adsorptive solution. The resulting adsorbent-adsorptive suspension is then allowed to equilibrate at constant pressure and temperature for different time intervals. During each time interval the suspension is kept constantly mixed or stirred to achieve maximum adsorbent-adsorptive interaction. At the end of a time interval the suspension is filtered or centrifuged and the supernatant analyzed for the concentration of adsorptive. The difference between the input adsorptive concentration and the supernatant adsorptive concentration is taken to be the sorbed concentration. A plot of sorbed concentration over time is then generated for different time intervals. This plot can be used to determine kinetic parameters for a reaction. In addition to determining kinetic parameters, batch experiments are often used to describe the interaction of an adsorptive with an adsorbent (Sparks 1995). The procedures are essentially the same as described above except that only one equilibration time is used with varying input adsorptive concentration. The resulting plot is an adsorption isotherm of sorbed concentration as a function of input adsorptive concentration. The shape of the plot is indicative of the affinity between the absorbent and absorptive. 48

PAGE 59

49 Traditionally there are some concerns about batch experiments (Sparks 1995). There may be changes in solid: solution ratio particularly if one container is used for several equilibration times with aliquots of the suspension being removed at each interval. For quick reactions there is also concern that the reaction might be completed before measurements can be made. Too much or too little mixing may cause changes in the surface area of adsorbent or influence mass transfer and transport processes. Another major concern is that released species are often not removed and may cause interferences in the reactions. Batch experiments were used in this research to investigate and compare P sorption characteristics on Al-Fe mixed metal hydr(oxides) with varying Al:Fe ratios at room temperature (25C). P sorption as a function of time, pH and Al: Fe ratio was determined for each hydr(oxide). All experiments were carried out at a constant concentration of P and solid: solution ratio of 100 mg/L and 1:1000 respectively. The P solutions were made by dissolving 0.439 g of K 2 HPO 4 in 1L of 50mM KCl. Material and Methods P Sorption as a Function of Time P sorption was observed as a function of time at seven equilibration intervals over a 24 h period. Ten mg of each hydr(oxide) were placed in seven 20 mL plastic scintillation vials (one per time interval) and 10 mL of 100 mg/L P solution (pH 4.83) added. The scintillation vials were then capped and allowed to equilibrate on a reciprocal stirrer. At intervals of 1, 2, 3, 8, 10, 20 and 24h a scintillation vial for each hydr(oxide) was removed from the stirrer. The hydr(oxide)-P suspension was then immediately filtered through 0.45 m syringe filters and 100 l of the supernatant transferred to a 30 mL Pyrex tube. The transferred supernatant was then dried in the oven at 70C brought to a

PAGE 60

50 volume with 5 ml molybdate solution and analyzed for P. Sorbed P concentrations were subsequently determined and plotted against time. P Sorption as a Function of pH and Al:Fe Content The effect of pH on P sorption on each hydr(oxide) sample was investigated by varying the pH of the P solution equilibrated with the hydr(oxides). The pH values of P solutions used were 3.3, 4.8, 6.1, 7.2 and 8.5. Solution pH values were obtained by adjusting the pH of the P solution (pH 4.8) through the drop wise addition of NaOH (to increase pH), or HCl (to decrease pH). Ten mg of each of the hydr(oxides) were then placed in each of five 20 mL plastic vials (one vial per pH value) and 10mL of P solution added. The resulting suspension was allowed to equilibrate at a constant stirring rate on a reciprocal stirrer for 24 h after which it was removed, filtered, and a 100l aliquot of the supernatant transferred to a 30 mL Pyrex tube, dried and analyzed for P. P Analysis All P analysis throughout the research was done using the molybdate blue method of Murphy and Riley (1962). Molybdate reagent solution (reagent B) was made by dissolving 1.5 g of L-ascorbic acid in 100mL of an acid-molybdate solution (reagent A)and brought to a volume of 1 L with DDW. 5mL of the reagent B was then added to each of the pyrex tubes containing the dried P supernatant, shaken and left to allow full development of the blue color. The intensity of the blue color developed in each tube was then compared to that developed in standards that were derived from 0.5 mL of 0, 1, 2, 3, 4, 5 mg/L P. Tubes with a more intense blue color than that seen in the 5 mg/L standard were diluted by factors of 5, 10, 20 or 1000 (for extremely blue colors) with reagent B until the intensity of the color was less than that seen in the 2.5 standard.

PAGE 61

51 Following color development and dilution of supernatant samples (where necessary) analysis was carried out on a Bausch and Lamb (Spectronic 100) spectrometer at a wavelength of 880 nm. Absorbance values obtained were subsequently converted to concentration values using the equation of the standard curve (Figure 3-1). y = 16.482x 1.264R2 = 0.999012345600.10.20.30.40.5AbsorbanceConcentration (ug/ml) Figure 3-1. Standard Curve for P standards Sorbed P Determination To determine the amount of P sorbed (ug P/mg ) to each hydr(oxide), the concentration of P in the supernatant values were converted to mass of P in the supernatant, by multiplying the concentration by the dilution factor used in P analysis. This resulting mass (ug) was subsequently subtracted from 1000 ug, the mass of P in 10 mL of the 100 mg/L P solution, to yield the mass of P sorbed (ug). By further dividing the mass of P sorbed by the mass (mg) of hydr(oxide) used the amount of P sorbed was obtained. To account for the water in the structure of the hydr(oxides), the amount of P sorbed was expressed based on the metal content ie. ug P/mg of metal. This was done in

PAGE 62

52 all sorption experiments except the sorption as a function of time study which kept the units of ug P/mg of hydr (oxide). Conversion to ug P/mg of metal was done by dividing ug P/mg of hydr (oxide) by the percent metal content. Results and Discussions P Sorption as a Function of Time at Fixed pH The amount of P sorbed showed a relative increase over time for all hydr(oxides) (Figure 3-2) over the 24 h time period. The rate of increase (Figure 3-3) however, began to decline exponentially after about 1h becoming constant after about 10 h. This type of rate change suggests a change in sorption mechanism from a quick sorption process in the first hour of equilibration to a slow diffusion type mechanism thereafter. This is typical of P sorption on hydr(oxide) surfaces and is thought to be due to a quick ligand or ion exchange sorption process (in the first hour) which eventually slows down as pore size and therefore accessibility to sorption sites decreases, requiring P to move deeper into the interior of the hydr(oxide) particle. Despite, the rapid decrease in P sorption rate there was no point at which the rate of sorption was observed to be zero, indicating equilibrium or sorption max. Instead an asymptotic decrease towards zero was observed (Figure 3-3) indicating that sorption was continuing at a slow but constant rate. This suggests that conditions of solid: solution ratio, P concentration and equilibration times used in the experiment were insufficient to cause saturation of the P sorption sites and subsequently equilibrium on the hydr(oxides). The presence of numerous increasingly smaller pores (Willet et al., 1988) in the amorphous hydr(oxides), providing additional surface sites for P sorption, are likely the reason for the length of the slow phase. As these pores decrease in size and accessibility, the time taken for the P molecule to reach the sorption sites increases and hence, the rate

PAGE 63

53 of P sorption decreases towards zero, accounting for the more or less constant P sorption rate observed after 10h. From a practical standpoint, despite being important in understanding the trend which P follows when being sorbed to hydr(oxides) and long-term evaluations, the slow Psorption phase may be unimportant on a rate of P sorption (removal) basis. For example, on average, 80% of the total P (Figure 3-4) that was sorbed in 24 h on all hydr(oxides) was sorbed in the first hour of equilibration. For the next 23 hours only 20% of the total was sorbed. In fact by the tenth hour of equilibration 98% of the 24 h max was already sorbed. P sorption also showed a general increase with Al:Fe ratio. The similarity in the shape of the curves (Figure 3-2), particularly during the slow phase of P sorption, indicate that P has similar access to sorption sites on all hydr(oxides) and suggests that the distribution of pore sizes and hence internal surface area are also similar irrespective of Al:Fe ratio. It is therefore unlikely that diffusion effects are a significant contributor to the differences in sorption observed in these hydr(oxides). The increased sorption with Al:Fe ratio was therefore more likely due to the type of sorption site present and or the external surface area of the hydr(oxides). The type of sorption site is unlikely to account for the difference seen because as discussed in chapter 1, the surface of these hydr(oxides) tend to be Al rich and therefore sorption will likely occur on Al-OH type sites initially and later Fe-OH type site. The most probable explanation is an increase in external surface area with Al:Fe ratio arising from an increasing proportion of smaller sized particles. This is supported by particle size data and believed to be due to dispersing effect of soluble Al species on Fe during co-precipitation (Anderson et. al 1990 a).

PAGE 64

54 R2 = 0.927R2 = 0.926R2 = 0.927R2 = 0.938R2 = 0.861R2 = 0.96105101520253035404550556065024681012141618202224Time (hrs) (ug P /mg of hydr(oxide)) Fe 1:05 1:02 1:01 2:01 5:01 10:01 Al Figure 3-2. Batch P sorption with time on mixed-metal hydr(oxides) R2 = 0.8270510152025303540455003691215182124Time (hrs)Rate of P sorption (ug/hr) Figure 3-3. Changes in rate of P sorption on hydr(oxides) over time.

PAGE 65

55 00.20.40.60.811.203691215182124Time (hrs)Fractional sorption Figure 3-4. Fractional sorption with time (based on 24 h sorption maximum). P Sorption as a Function pH on Different Hydr(oxides) It is important to note that P sorption in this section is presented as g/ mg of metal rather than g /mg of hydr(oxide). This was done to eliminate the effect of variable water content in the solid. P sorption on all hydr(oxides) decreased with increasing pH(Figure 3-5 a, b). The decrease was more or less linear over the pH range studied with inflection points occurring at pH 6 and 7 for Al:Fe ratio below and above 1:1 respectively. P sorption decreased significantly at pHs above these points, as shown by the sharp change in slope thereafter (Figure 3-5). The points also corresponded well with reported pK a1 values for Fe-OH (6.5) and Al-OH (7.5) (McBride, 1994) and is believed to represent a shift in the type of the sorption sites on the solid’s surface towards less reactive hydroxyl groups. Under the experimental conditions of this study the shift in sorption would most probably reflect a shift from the more reactive aquo (M-OH 2 + ) to the

PAGE 66

56 less reactive hydroxy (M-OH) group, where M is either Al or Fe. Additionally, changes in surface charge may at least account for a portion of the reduction in sorption observed over the pH range particularly between pH 6 and 9. At these pH values the system is essentially at the PZC values for the Fe-OH (7.5) and Al-OH (9.0) (McBride, 1994) and hence the surface charge on the solids is close to neutrality or negatively charged resulting in repulsion of the negatively charged phosphate ions. There is also indication that sorption remains essentially constant between pH 5 and 7 on the 2:1, 5:1 and 10:1 hydr(oxides) (Figure 3-5 b). The reason for this is unclear but is possibly due to increase structural organization in these hydr(oxides) compared to the others. As suggested earlier based on XRD data there was an increase in structural development with Al content in the mixed hydr(oxides). This structural development may increase the solids ability to resist pH change effects at least within this pH range. Work done by Dominik et al. (2002) and Wells et al. (2001) on the dissolution of mixed Al-Fe oxides showed that the rate of dissolution was negatively linearly related to Al content indicating a decrease in dissolution with increase Al content. Wells ( 2001) found that the rate was directly related to the surface area. Additionally they concluded that the results were influenced directly by metal-oxygen bond energy and crystallinity, and indirectly by crystal size and shape resulting from association between Al and Fe. Figure 3-6 shows the percent loss in sorption with Al content at each pH value. These values were calculated by taking sorption at pH 3 on all hydr(oxides) to be O% loss. Difference in sorption, calculated as percent, between pH 3 and each of the other pH values were then taken to be the loss in sorption up to that pH value. The results indicate that at a given pH, loss of sorption on the mixed hydr(oxides) decreases with increasing

PAGE 67

57 Al content. Additionally, up to 80% of the sorption was shown to have been lost over the entire pH range for the hydr(oxides) with less than 50 mol% Al compared to around 50-60% losses for hydr(oxides) with a greater Al content. Interestingly, 25-30% of the overall loss was found to occur around the pH values where the inflection points occurred (Figure 3-5). This corresponded to pH 6-7 for an Al content less than 50 mol% and pH 7-8 for greater than 50 mol% Al. This also indicates that sorption characteristics were largely controlled by the properties of the hydr(oxides). For hydr(oxides) below 50 mol% Al there is only partial coverage of Fe by Al, hence the Fe-OH sites are the dominant type of sites on these surfaces. Sorption characteristics in these hydr(oxides) under the experimental conditions used are therefore influenced by the properties of Fe-OH 2 + which has a pK a1 of 6.5, hence accounting for the great loss in sorption between pH 6 and 7 for these hydr(oxides). For hydr(oxides) containing more than 50 mol% Al full coating of the Fe by Al is achieved. The surface is therefore dominated by Al type sites, in this case AlOH 2 + which has a pK a1 of 7.5 and accounting for the loss in sorption between pH 7 and 9.

PAGE 68

58 0204060801001201401603456789pHP sorbed (ug/mg of metal) Fe(0%) 1:10(9.1%) 1:5(16.7%) 1:2(33.3%) 1:1(50.0%) 4060801001201401601802003456789pHP sorbed (ug/mg of metal) 2:1(66.7%) 5:1(80.1%) 10:1(90.1%) Al(100%) Al content (a) Al content (b) Figure 3-5. P sorption on Al-Fe hydr(oxides) as a function of pH.(a) low Al containing hydr(oxides) (b) higher Al containing hydroxides.

PAGE 69

59 R2 = 0.844R2 = 0.945R2 = 0.859R2 = 0.763010203040506070809010002040608010012Al content (mol%)loss of sorption(%) 0 4.83 6.08 7.17 8.53 Figure 3-6. Loss of sorption with Al content at different pH P Sorption as a Function of Al Content at Different pH To account for variability in water content P sorption in this section was also expressed on a per mg of metal basis. P sorption showed a general increase in sorption with increasing Al content at all pH values. Sorption of up to 180 g/mg was obtained (Figure 3-7) for the pure Al hydr(oxide) (100 mol%) which was twice that observed for the highest mixed hydr(oxide) (90.9 mol%) and four times that observed in the pure Fe-hydr(oxide) (0 mol%). This trend was the same at all pH values and is most likely due to differences in the number of sorption sites arising from differences in surface area and particle size distribution. The increase in sorption with Al content was strongly non-linear, particularly at pH values below 8.5, and was best described by a 4 factor polynomial model (Figure 3-7). The model had 3 points of local maxima and minima occurring at about 10 mol% (1:9),

PAGE 70

60 50 mol%( 1:1) and 80 mol %( 5:1) Al content. Each point can be related to a shift in the relationship between P sorption and the Al content of the hydr(oxide) arising from changes in Al-Fe interaction in the hydr(oxides). Up to 10 mol %, Al content has no effect on P sorption. Above 10 mol%, P sorption increased linearly with Al content up to 50 mol% after which no additional increase was observed up to about 80 mol% Al. P sorption then increased exponentially thereafter up to 100% Al content. The local minima at 10 mol%, is proposed to represent a change from Al-Fe solid solution to precipitation of Al-hydr(oxide) onto, and partial coverage of, the surface of Fe-hydr(oxide) particles. Al-hydr(oxide) precipitation continues with the Fe surface becoming completely covered by Al-hydr(oxide) at 50% Al content. There is an increase in the thickness of Al-hydr(oxide) coating on the Fe between 50-80 mol% at which point the quantity of Al hydr(oxide) that can be supported by the Fe-hydr(oxide) core is apparently reached . Further increase in Al content above 80 mol% results in a complex formation of a mixed Al coated Fehydr(oxide) and a pure Al-hydr(oxide) phases. Figure 3-8 shows a proposed scheme for Al-Fe interaction accounting for differences seen in P sorption. The results also suggest that as pH increases the effect of Al: Fe interactions are less pronounced eventually becoming negligible at pH values above the pK a of Al-OH. At pH 8.5, sorption can be sufficiently modeled as a linear function of Al content (R 2 =0.945). This linear relationship essentially represents the minimum sorption capacity of the hydr(oxides) under the experimental conditions.

PAGE 71

61 R2 = 0.960R2 = 0.973020406080100120140160180200020406080100120Al content (mol%)sorption(ug/mg of metal) 3.28 4.83 R2 = 0.979R2 = 0.964R2 = 0.95602040608010012014002040608010012Al content (mol%)sorption(ug/mg of metal) 0 6.08 7.17 8.53 pH pH (b) Figure 3-7. P sorption as a function of Al content at (a) pH 3.3 and 4.8 (b) pH 6.1, 7.2 and 8.5.

PAGE 72

62 IV. Phase separation mixed + Al-hydr(oxide) III. Full coating of Fe by Al II. Partial coating of Fe by Al I. Solid solution formation III IV II I Figure 3-8. Al-Fe interactions accounting for differences in sorption. Conclusion P sorption on all hydr(oxides) was characterized by an initial rapid phase followed by a very slow phase that was still active after 24 h of equilibration. The rate of uptake during the slow phase was similar indicating that the difference in P sorption among the hydr(oxides) occurred in the rapid phase and was as a result of difference in surface properties rather than diffusion effects. P sorption decreased with increasing pH for all hydr(oxides) studied, with the magnitude of change being greater at the lower Al contents. The relationship between P sorption was strongly non-linear with P sorption being about 4 times on the Al-hydr(oxide) compared to that on the Fe-hydr(oxide). Changes in P sorption were related to changes in morphology and composition of the hydr(oxides) as Al content varied.

PAGE 73

CHAPTER 4 P SORPTION EFFECTS ON MIXED-METAL HYDR(OXIDE) SURFACES AS DETERMINED BY FLOW CALORIMETRY Introduction Flow calorimetry provides a direct, quantitative measure of the heat involved in a reaction (Rhue et al. 2002). This measured heat is related to the change in enthalpy. The method is ideally suited for measuring interactions occurring at the liquid/ solid interface and has been widely used to study the surface chemistry of many types of solids such as organics (Schneider et al. 1997 and Taraba 1990) and synthetic inorganics (Meziani et al 1997). Until recently, however little use had been made of flow calorimetry in the study of surface reactions on soils (Rhue et al. 2002). Coupled with other macroscopic and spectroscopic techniques flow calorimetry can be used to yield information about surface chemical reactions, such as how binding energies vary with surface coverage, that could not be obtained by other methods (Appel et al. 2002). It has several advantages over batch calorimetry (Steinberg, 1981) in that: i) it can resolve complex series of reaction that occur more or less simultaneously but at different rates; ii) multiple adsorption/desorption cycles can applied to the same sample, allowing reversible and irreversible processes to be distinguished; iii) changes in surface properties associated with specific treatment or aging effects can be quantified; and iv) when both the amount of sorption and its associated heat are measured information about surface heterogeneity can be obtained. Flow calorimetry may also be used to resolve the contradictions in 63

PAGE 74

64 enthalpies derived from the temperature dependence of sorption isotherms (Rhue et al. 2002). In this study, flow calorimetry was used to determine the energies associated with K/Ca and NO 3 /Cl exchange as well as P sorption on the mixed metal hydr(oxides). In addition, surface homogeneity between samples was assessed with the aim of better understanding formation of the mixed hydr(oxides). Materials and Methods Instrumentation The flow calorimeter used in this study was built in our lab by Dr. Dean Rhue. A detailed description of how it works is outlined by Rhue et al. (2002). For this study 10-20 mg of hydr(oxide) sample was packed into a column flanked by two thermistors, one at the column inlet (reference thermistor) and the other at the column outlet (column thermistor). Solutions containing different reactants were then forced through the sample column by using a 100cm water column to pressurize the solution containers. Flow through the column was maintained between 0.35 and 0.40 mL min -1 by a precision needle valve located at the outlet of the calorimeter. The temperature of the incoming solution was sensed by the reference thermistor. Any heat given off as a result of either physical or chemical interaction between the solution components and the hydr(oxide) in the column was sensed by the column thermistor. The differences between reference and column thermistor readings were transmitted to a computer where they were recorded and displayed as a graph of signal (volts) over time (minutes). The end of the reaction was taken to be the point at which the graph returned to baseline (difference in signals is zero). The resulting graph was integrated as flow rate-averaged peak area and compared with 5 second heat pulse calibration peaks. Differences in peak characteristics for each

PAGE 75

65 hydr(oxide) were used to make inferences about the surface properties of the hydr(oxides). Ion Exchange and P Sorption 50 mM KNO 3 and 50mM KCl were used as the reacting solutions for NO 3 /Cl anion exchange while 50 mM KCl and 25 mM CaCl 2 were used for K/Ca cation exchange. For P sorption the reacting solution was a 1mM P solution made from KH 2 PO 4 salt. Ionic strength and pH of all the solutions used in the flow calorimetry experiments were maintained at 50 mM and 4.8 respectively. To obtain an ionic strength of 50mM for the P solution 50 mM KCl was used as the background electrolyte. Prior to ion exchange and P sorption treatment cycles the hydr(oxide) was allowed to equilibrate with 50 mM KCl solution. Each hydr(oxide) sample was subsequently treated with several anion (NO 3 /Cl) and cation (K/Ca) exchange cycles (pre-P sorption) followed by a single P sorption treatment cycle using a 1 mM P solution. The hydr(oxide) was exposed to P solution after which the solution was changed to KCl. Cl was unable to replace P on the hydr(oxide) surface and therefore no heat signal was observed in going from the P solution back to KCl. Essentially its purpose was to give the hydr(oxide) limited exposure to P and then stop the reaction by putting it back into KCl. After P treatment several anion and cation (post-P) exchange cycles were done. A single anion exchange cycle comprised NO 3 replacing Cl followed by Cl replacing NO 3 and lasted for a total of between 40-60 minutes. Cation exchange treatment cycles were between K and Ca and lasted 20-30 minutes. Following post-P ion exchange the columns were again allowed to equilibrate in 50 mM KCl. The hydr(oxide) was then removed from the column, digested in concentrated HNO 3 acid and the digest was subsequently analyzed for P, Al, Fe, K and Cl. P analysis was done by the molybdate blue method and the

PAGE 76

66 metals by atomic adsorption. Initially chloride analysis was done using high performance liquid chromatography (HPLC) but no chloride was detected by the method . A silver nitrate test however showed that chloride was present in the digest indicating that there may have been some interference in the HPLC method. A method was developed to estimate the Cl concentration in the digest based on the turbidity of the AgCl precipitate formed when AgNO 3 was added to the digest. A mL of digest was brought to a volume of 5 mL using DDW and 100 l of 15mM AgNO 3 was then added. The resulting mixture was allowed to stand for 50 minutes, during which time it was periodically shaken to ensure proper mixing. Following the 50 minute reaction time the mixture was vigorous shaken and the turbidity of the AgCl precipitate formed was measured using a Bausch and Lamb spectrometer (Spectronic 100) at a wavelength of 440 nm. The turbidity of the digests was then compared to 0, 0.1, 0.2, 0.3 and 0.5 mM Cl standards. The standard curve (Figure 4-1) was linear over the range (R2= 0.996) y = 0.9575x 0.0439R2 = 0.99600.10.20.30.40.50.600.10.20.30.40.50.6turbidityconc. (umol/ml) Figure 4-1. Chloride standard curve.

PAGE 77

67 Results and Discussion P Sorption Figure 4-2 shows the graphs of signal over time for the P treatment cycle on the hydr(oxides). The fact that these plots were exotherms indicated that P sorption on the hydr(oxides) was exothermic. The similarity in shape of the graphs indicates that the P sorption mechanism was the same and was independent of the Al content. No return to baseline during the 20 minute P treatment indicates that equilibrium was not reached in this time period and hence the surface was not saturated by P. The quantity of P sorbed (Figure 4-3) averaged 5.48 .59 g/mg with no significant correlation with increasing Al content, indicating that Al content did not affect the magnitude of P sorption under the experimental conditions. The quantity of P adsorbed was linearly related to the normalized flow-averaged peak area (Figure 4-4) indicating that as more P was adsorbed there was a proportional increase in total quantity of energy released. Despite an increase in total energy with total P sorbed, the energy of adsorption per unit mass of P sorbed was similar, 33 kJ/mol (Table 4-1) irrespective of Al content. That no significant differences were observed in the heats of adsorption indicated that the species involved in P sorption were the same for all the hydr(oxides). That meant that the surface hydroxyl groups involved were the same irrespective of Al content and the energy required to break the metal-hydr(oxide) bond in these groups were essentially the same whether the metal was Al or Fe. Additionally the results suggest that the same P species was being adsorbed. Based on P speciation model (Figure 1-1) the P species expected to dominate the solution at the experimental pH (4.8) is H 2 PO 4 and is most likely to be the species involved in the sorption process.

PAGE 78

68 -0.200.20.40.60.811.21.405101520253035404550time(min)si g nal ( volts ) 0.811.21.41.61.822.205101520253035404550time (min)si g nal ( volts ) Figure 4-2. Exotherms for P sorption on mixed-metal hydr(oxides) . (a) 0-15 mol% Al and (b) 66.7-100 mol% Al. Table 4-1. Heats of P sorption on mixed-metal hydroxides. Sample ID Column wt (mg) peak area (v.mL) v.mL/mg Al content (mol%) P sorbed (ug) P ug/mg of hydr(oxide) Heat of sorption KJ/mol Fe 16.1 256 15.90 0.00 89.05 5.53 39.7 1:10 15.1 232 15.36 10.20 75.90 5.03 42.2 1:5 19.8 156 7.88 17.92 60.58 3.06 35.6 1:2 16.9 255 15.09 33.77 109.75 6.49 32.1 1:1 17.3 275 15.90 49.83 118.03 6.82 32.2 2:1 16.6 157 9.46 65.82 86.45 5.21 25.1 5:1 14.9 128 8.59 81.80 55.90 3.75 31.6 10:1 14.7 273 18.57 89.65 122.18 8.31 30.9 Al 10.2 152 14.90 100.00 52.80 5.18 39.8 (b) (a)

PAGE 79

69 R2 = 0.87305101520250123456789P sorbed (ug/mg)peak area (v.ml/mg) Figure 4-3. Relationship between peak area and quantity of P sorbed. Ion Exchange and the Effect of P Sorption Cation exchange No heat signal was obtained for K/Ca exchange on any of the hydr(oxides) indicating that the surfaces of these hydr(oxides) had little or no cation exchange capacity at the pH of experiments. Even for post-P cation exchange when the adsorbed P might be expected to impart its negative charge to the surface, thus increasing the surface cation exchange capacity, no heat signal was obtained. These results are consistent with what was expected as the pH of the reacting solutions were below the PZC for the surfaces and thus the surfaces would be dominated by positive charge, generating far more AEC than CEC. The quantity of P sorbed (0.17 0.05mol/mg) to the hydr(oxides) as well as the equilibration time may have been too low to cause any measurable increase in CEC and hence detection by flow calorimetry was limited.

PAGE 80

70 Anion exchange Unlike cation exchange, anion exchange gave integrateable signal peaks indicating that the surfaces of the hydr(oxides) were largely positively charged and hence had a greater anion exchange than cation exchange capacity. That the surfaces were largely positively charged was expected because as mentioned earlier, the experimental pH was below the PZC for the hydr(oxide) and therefore the surface would be protonated. Figure 4-4 to 4-12 show the heat signals and their resulting peak area over time for anion exchange before (pre-P) and after (post-P) P treatment on the hydr(oxides). Nitrate replacing Cl on all the hydr(oxides) was exothermic while Cl replacing NO 3 was endothermic. Peak areas for the endotherms and exotherms for a given hydr(oxide) were essentially equal indicating that Cl/NO 3 was completely reversible. When normalized to the mass of hydr(oxide) in the column, peak areas(Figure 4-13) for both pre-P and post-P anion exchange increased with Al content suggesting either an increase in the AEC or the heat of exchange. A plot of heat signal against exchangeable Cl (Figure 4-14) suggested that the heat of exchange was independent of Al content except for the high Fe containing hydr(oxides) (Table 4-2). Calculated heats of exchange for Al: Fe ratios greater than 1:5 averaged 2.8.4 KJ/mol. Anion exchange capacity (Figure 4-15) increased with Al content. This increase in AEC with Al followed the same trend as that seen in Figure 4-13, particularly above 20 mol% Al, indicating that differences in AEC were largely responsible for some of the differences in anion peak areas. Below 35-40 mol% Al, AEC was constant (Figure 4-15) but the calorimetric peak areas decreased (Figure 4-13) indicating that the heats of exchange was lower for the Fe-rich hydr(oxides). Although interferences prevented the measuring of anion exchange on the Fe sample, the fact that the heat of exchange on the 1:10 hydr(oxide) (which behaved

PAGE 81

71 similarly to Fe throughout the research) was about that seen on the other hydr(oxides) is consistent with a lower heat of exchange at high Fe content. The increase in AEC with Al content could be partly related to the change in PZC with Al content. The fact that the experimental pH was closer to the PZC of Fe-hydr(oxide) (~7.5) meant that surfaces dominated by these sites were more likely to have less positive charge compared to those dominated by Al type sites (PZC~9.5). Other factors that could be involved are increasing surface area and or increasing charge density with Al content. Flow rate-averaged peak areas were greater for pre-P than post-P anion exchange on all hydr(oxides) indicating that there was a reduction in the number of positively charged sites due to P sorption. The percentage change (Figure 4-16) however varied with Al content, decreasing from 50% to 25% at an Al content of 50 mol% with no significant change with increasing Al up to 100 mol%. This points to the heterogeneity of the surface up to 50% Al content and the homogeneity of the surface thereafter and suggests that as Al content increases the surface of the mixed hydr(oxides) became more and more like pure Al hydr(oxide) which is consistent with the earlier proposal of partial surface coverage up to 50% Al and complete surface coverage thereafter. Since no significant difference in the quantity of P sorbed was observed and given that the length of P-treatment and the concentration of P in the treatment solution was kept constant the difference in pre-P and post-P peak areas with Al content is most likely due to the number of sorption sites on the surface. As previously discussed higher Al containing hydr(oxides) may have a greater number of sites compared to higher iron containing hydr(oxides) at the experimental pH. It is therefore expected that since a fixed quantity of P was adsorbed a larger proportion of sorption sites would be occupied on the Fe

PAGE 82

72 dominated surfaces leading to greater percent reduction in the number of sites available for post-P anion exchange which is manifested as a lower post-P peak area compared to pre-P peak area. The reduction in anion exchange peak areas following P sorption also indicate that P sorption and anion exchange occur on the same surface sites, at least under these experimental conditions. These sites are most likely OH 2 + (aquo ligand) sites since anion exchange occur only on positively charged sites. Water ligand sites in addition to being more prevalent at low pH are much more easily removed by P in ligand exchange than other surface OH groups and may have been the primary site of P sorption. The fact that the anion exchange capacity could not be regenerated during subsequent Cl treatment points to the irreversibility of P sorption at constant pH. -0.2-0.100.10.20.30.40510152025Time(min)si g nal ( volts ) Post-P NO3 Post-P Cl Pre-P NO3 Pre-P Cl -60-40-2002040608015202530Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl 10 (b) (a) Figure 4-4. Heats of anion exchange on Fe (0%) hydr(oxide). (a) heat signal (b) peak areas.

PAGE 83

73 -0.4-0.3-0.2-0.100.10.20.30.4051015202530Time(min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P NO3 Post-P Cl -100-80-60-40-200204060801001015202530Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl (a) (b) Figure 4-5. Heats of anion exchange on 1:10 (9.1%) hydr(oxide). (a) heat signal (b) peak areas. -0.3-0.2-0.100.10.20.30.40.505101520253035Time(min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P Cl Post-P NO3 -160-120-80-400408012016020010152025303540Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl Figure 4-6. Heats of anion exchange on 1:5 (16.7%) hydr(oxide). (a) heat signal (b) peak areas.

PAGE 84

74 -0.5-0.4-0.3-0.2-0.100.10.20.30.40.505101520253035Time (min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P NO3 Post-P Cl -160-120-80-40040801201601520253035Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl 10 (b) (a) Figure 4-7. Heats of anion exchange on 1:2 (33.3%) hydr(oxide). (a) heat signal (b) peak areas. -0.6-0.4-0.200.20.40.60.805101520253035Time (min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P NO3 Post-P Cl" -200-160-120-80-4004080120160200101520253035Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 p ost-P C l (a) (b) Figure 4-8. Heats of anion exchange on 1:1 (50%) hydr(oxide). (a) heat signal (b) peak areas.

PAGE 85

75 -0.8-0.6-0.4-0.200.20.40.60.8051015202530Time (min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P NO3 Post-P Cl -200-160-120-80-400408012016020024051015202Time (min) 5 p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl (a) (b) Figure 4-9. Heats of anion exchange on 2:1 (66.7%) hydr(oxide). (a) heat signal (b) peak areas. -0.6-0.4-0.200.20.40510152025303540Time (min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P Cl Post-P NO3 -200-160-120-80-400408012016020010152025303540Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl (a) (b) Figure 4-10. Heats of anion exchange on 5:1 (83.3%) hydr(oxide). (a) heat signal (b) peak areas.

PAGE 86

76 -1-0.8-0.6-0.4-0.200.20.40.60.8051015202530Time (min)si g nal ( volts ) Pre-P NO3 Pre-P Cl Post-P NO3 Post-P Cl -200-160-120-80-40040801201602002401015202530Time (min) p eak area ( v.ml ) pre-P NO3 pre-P Cl post-P NO3 post-P Cl (a) (b) Figure 4-11. Heats of anion exchange on 10:1 (90.9%) hydr(oxide). (a) heat signal (b) peak areas. -0.3-0.2-0.100.10.20.30.40.5051015202530354045Time (min)si g nal ( volts ) Pre-P Cl Pre-P NO3 Post-P NO3 Post-P Cl -160-120-80-40040801201602001015202530354045Time (min) p eak area ( v.ml ) post-P NO3 post-P Cl pre-p Cl pre-P NO3 (a) (b) Figure 4-12. Heats of anion exchange on Al (100%) hydr(oxide). (a) heat signal (b) peak areas.

PAGE 87

77 R2 = 0.982R2 = 0.98402468101214160102030405060708090100Al content (mol%)peak area (v.ml/mg) Pre-P AEC Post-P AEC Figure 4-13. Relationship between peak area for NO 3 / Cl exchange and Al content y = 5.928xR2 = 0.8680246810120.00.51.01.52.0exchangeable Cl (umol/mg)v.ml/mg Figure 4-14. Change in heat signal peak area per unit of Cl sorbed

PAGE 88

78 Table 4-2. Heats of anion exchange on mixed-metal hydroxides. Sample ID Al content (mol%) Column wt (mg) Peak area (v.mL/mg) Cl sorbed (umol) Cl (umol/mg hydr(oxide)) mJ/v.mL Heat of exchange (KJ/mol) Fe 0.00 16.1 0.445 1:10 10.20 15.1 2.82 14.98 0.99 0.444 1.27 1:5 17.92 19.8 5.13 0.414 1:2 33.77 16.9 5.09 14.24 0.96 0.378 2.55 1:1 49.83 17.3 6.62 19.14 1.11 0.453 2.27 2:1 65.82 16.6 7.45 17.59 1.06 0.474 3.22 5:1 81.80 14.9 8.72 21.79 1.46 0.473 2.83 10:1 89.65 14.7 9.55 21.22 1.44 0.427 3.16 Al 100.00 10.2 10.98 18.13 1.78 0.445 2.64 80901001101201301401501601701801900102030405060708090100Al content (mol%)AEC (cmol/Kg) Figure 4-15. Post-P Anion exchange capacity of hydr(oxides)

PAGE 89

79 R2 = 0.91201020304050600102030405060708090100 Al content (mol%) peak area change (%) Figure 4-16. Effect of P sorption on anion peak area. Conclusions P sorption was exothermic and showed no effect of Al content. The similar heats of P adsorption and the fact that the shape of the exotherms was similar for all the hydr(oxides) indicated that P sorption mechanism was the same regardless of Al content. P sorption resulted in decreased AEC with no apparent change in CEC. Anion exchange capacity increased with Al content for Al:Fe ratios greater than about 1:2. The data suggest that the heat of exchange between Cl and NO 3 decreased for the Fe-rich hydr(oxides).

PAGE 90

CHAPTER 5 SUMMARY Co-precipitation of mixed Al-Fe hydr(oxides) can be easily done through the titration of mixed Al-Fe chloride solution against NaOH. The Al and Fe content in the hydr(oxide) is dependent on the initial Al:Fe molar ratio in the initial solution as well as the final precipitation pH. Under conditions of room temperature (25 C) and final precipitation pH of 6.5-7 Al and Fe content (in mol%) was found to be equal to that in the initial solution. For Al content at least greater than 10 mol% surface precipitation of Al onto Fe is believed to be the dominant mechanism by which the mixed hydr(oxides) were formed, particularly if precipitation was done under conditions of constant stirring and gradually increasing pH. Partial coverage of Fe by Al is believed to occur at Al content below 50 mol%, complete coverage of Fe by Al between 50 and 80 mol% Al and the formation of a dual phase of mixed Al-Fe and pure Al hydr(oxide) above 80 mol% Al. The thickness of the Al layer is also believed to increase with Al content between 50 and 80 mol%. Freshly co-precipitated hydr(oxides) are largely amorphous in nature and structural development increased with Al content and aging time. Additionally the particle size distribution as well as morphology of the mixed hydr(oxide) particles are controlled by Fe particularly below 80 mol% Al. Differences in P sorption characteristics were largely influenced by the properties of the hydr(oxides). P sorption occurred via an initial quick sorption phase followed by a slow phase. The affinity of the hydr(oxide) surface for P at a constant pH was the same irrespective of Al content. At a solid: solution ratio of 1:1000 and input P concentration 80

PAGE 91

81 of 100 mg/L, it was found that the quick phase lasted for about 3 h and accounted for 90% of the P sorbed over a 24 h period. The rate of P sorption during the slow phase was the same irrespective of Al content and differences in the quantity of P sorbed were due to differences in surface properties rather than diffusion effects. P sorption decreased over the pH range 3-9. Twenty to forty percent of total loss in sorption over the pH range was due to change in P species and 25-30 % was due to change in surface species. Change in surface species occurred at pH values corresponding to the theoretical pKa 1 value of the dominant surface M-OH group in the hydr(oxide). For hydr(oxides) containing less than 50 mol% Al where Fe dominated the surface this species change occurred between pH 6 and 7 while for greater than 50 mol% Al where Al dominated the surface this species change occurred between pH 7 and 9. Total loss in P sorption over the entire pH range studied ranged between 65-80% for pure Al and Fe as well as mixed hydr(oxides) containing less than 50 mol% Al compared to less than 50 % loss for hydroxides containing greater than 50 mol% Al. This is believed to be due to increased structural development in the higher Al containing mixed hydr(oxides). P sorption increased non-linearly with Al content, particularly below pH 8. A four factor polynomial provided the best logical fit and each local maxima or minima could be correlated with a change in hydr(oxide) phase from solid solution formation partial coverage of Fe by Al complete coverage of Fe by Al a dual phase of mixed Al-Fe hydr(oxide) and pure Al hydr(oxide). At a given pH, P sorption on the pure Al-hydr(oxide) was about 4 times that observed in the pure Fe and mixed hydr(oxides) respectively. This is believed to be due to difference in surface area and structural development with Al content.

PAGE 92

82 At pH 5, P adsorption was exothermic, irreversible and the mechanism was the same irrespective of Al content. The quantity of P sorbed within a 20 minute exposure to P in the calorimetric experiments was unrelated to the Al content and linearly related to the energy released. The heats of adsorption were equal for all Al contents averaging 33 KJ/mol. There was no detectable CEC before or after P adsorption. P adsorption and anion exchange occurred on protonated aquo ligand groups resulting in decreased AEC after P treatment. Anion exchange was completely reversible with heats of exchange of 2.8.4 KJ/mol fro Al contents greater than about 35-40 mol%. Nitrate for chloride exchange was exothermic and Cl/ NO 3 exchange endothermic. The AEC increased for Al contents greater than 35-40 mol%. There was an indication that the heat of anion exchange decreased for the Fe-rich hydr(oxides), i.e < 35-40 mol% Al.

PAGE 93

LIST OF REFERENCES Agbenin, J.O. 2003. Extractable iron and aluminum effects on phosphate sorption in a savanna alfisol. Soil Sci. Soc. Am. J. 67: 589-595. Anderson, P.R. and M. M Benjamin. 1990 (a). Surface and bulk characteristics of binary oxide suspensions. Environ. Sci. Technol. 24: 692-698. Anderson, P.R. and M. M Benjamin. 1990 (b).Modelling sorption in aluminum-iron binary oxide suspensions. Environ. Sci. Technol. 24: 1586-1592. Appel C., D. Rhue., L. Ma., and B. Reve. 2002. Heats of K/Ca and K/Pb exchange in two tropical soils as measured by flow calorimetry. Soil Sci. 167: 773-781. Arai, E., and D.L. Sparks. 2001. ATR-FTIR spectroscopic investigation on phosphate adsorption mechanisms at the Ferrihydrite-water interface. J. Coll. Interf. Sci. 241: 317-326. Barrow, N.J. 1979. Three effects of temperature on the reactions between inorganic phosphate and soil. J. Soil Sc. 30: 271-279. Barrow, N.J. 1983. A mechanistic model for describing the sorption and desorption of phosphate by soil. J. Soil Sc. 34: 733-750. Bastin, O., F. Janssens, J. Dufey and A. Peeters. 1999. Phosphorus removal by a synthetic iron oxide-gypsum compound. Ecol. Eng. 12: 339-351. Blangenois, N., M. Florea, P. Grange, R.P. Silvy, S.P. Chenakin, J.M. Bastin, N. Kruse, B.P. Barbero and L. Cadius. 2004. Influence of the co-precipitation pH on the physico-chemical and catalytic properties of vanadium aluminum oxide catalyst. Appl. Catalysis (in press). Bleam, W.F. 1991. Soil Science application of nuclear magnetic resonance spectroscopy. Adv. Agron. 46: 91-148. Boers, P.C.M., W. van Raaphorst and D.T. van der Molen. 1998. Phosphorus removal retention in sediments. Water Sci. Technol. 37: 31-39. Borggard, O.K. 1983. The influence of iron oxides on phosphate sorption by soil. J. Soil Sci. 34: 333341. 83

PAGE 94

84 Cabrera, F., P. Aranbarri, L.Madrid, and G.G. Toca. 1981. Desorption of phosphorus from iron oxide in relation to pH and porosity. Geoderma 26: 203-216. Carim, A.H., G.S. Rohrer, N.R. Dando, S. Tzeng, C. L. Rohrer and A.J. Perotta. 1997. conversion of diaspore to corundum: A new -alumina transformation sequence. J. Am. Ceram. Soc. 80: 2677-2680. Chien, S.H., N.K. Savant, and U. Mokwunye. 1982. Effect of temperature on P sorption and desorption in two acid soils. Soil Sci.133: 160-166. Deist, J., and O. Talibudeen. 1967. Thermodynamics of K-Ca exchange in soils. J. Soil Sci. 18: 138-148. Dominik, P., H.N. Pohl, N. Bousserrhine, J. Berthelin and M. Kaupenjohann. 2002. Limitations to the reductive dissolution of Al-substituted goethites by Clostridium butyricum. Soil Biol. and Biochem. 34: 1147-1155. El-Sharkawy, E. A., S.A. El-Hakam and S.E. Samra. 2000. Effect of thermal treatment on the various properties of iron(III)-aluminum(III) co-precipitated system. Materials letter 42: 331-338. Freese D., S.E.A.T.M. van der Zee, and W.H. van Riemsdijk. 1992. Comparison of different models for phosphate sorption as a function of the iron and aluminum oxides of soils. J. Soil Sci. 43: 729-738. Freese D., S.E.A.T.M. van der Zee, , and W.H. van Riemsdijk. 1995. Modelling phosphate sorption kinetics in acid soils. Eur. J. Soil Sci. 46: 239-245. Froelich, P.N. 1988. Kinetic control of dissolved phosphate in natural rivers and estuaries: A primer on the phosphate buffer mechanism. Limnol. Oceanogr. 33: 649-668. Frossard, E., L.M. Condron and A. Oberson. 2000. Processes governing phosphorus availability in temperate soils. J. Env. Qual. 29: 15-23. Galerneau, E. and R. Gehr. 1997. Phosphorus removal from wastewaters: experimental and theoretical support for alternative mechanisms. Water Res. 31: 328-338. Goldstein, J., D. Newbury, D. Joy, C. Lyman, P Echlin, E. Lifshin, L. Sawyer and J. Michael. 2003. Scanning electron microscopy and x-ray micro-analysis. 3 rd ed. Klumer Academic/Plenum Publishers. New York. Groszek, A.J., and S. Partyka. 1993. Measurements of hydrophobic and hydrophilic surface sites by flow microcalorimetry. Langmuir 9: 2721-2725. Hundal, H.S. 1988. A mechanism of phosphate adsorption on Narrabi medium clay loam soil. J. Agric. Sci. 11: 155-159.

PAGE 95

85 Korecz, L., I. Kurucz, G. Menczel, E. Papp-Molnar, E. Pungor and K. Burger. 1972. Mossbauer investigation of iron-aluminum mixed oxides. Talanta 19: 1599-1604. Li, L., and R. Stanforth. 2000. Distinguishing adsorption and surface precipitation of phosphate on goethite. J. Colloid and Interface Sci. 230: 12-21. Lijklema, L. 1980. Interaction of orthophosphate with iron(III) and aluminum hydroxides. Am. Chem. Soc. 14: 537-541. Lookman, R., P. Grobet, R. Merckx and K Vlassak. 1994. Phosphate sorption by synthetic amorphous aluminum hydroxides: A 27Al and 31P solid-state MAS NMR spectroscopy study. Eur. J. Soil Sci. 45: 37-44. Lukasik, J., Farrah, S.R., Truesdail, S. and Shah, D.O. 1996. Adsorption of micro-organisms to sand and diatomaceous earth particles coated with metallic hydroxides. Kona 14: 87-91. Madrid, L. and P. deArambarri. 1985. Adsorption of phosphate by two iron oxides in relation to their porosity. J. Soil Sci. 36: 523-530. Maguire, R.O., R.H. Foy , J.S. Bailey and J.T. Sims. 2001. Estimation of the phosphorus sorption capacity of acidic soils in Ireland. Eur. J. Soil. Sci. 52: 479-487. Malati, A.M., R.A. Fassam, and I.R. Henderson. 1993. Mechanism of phosphate interaction with two reference clay and an anatase pigment. J. Chem. Technol. Biotechnol. 58: 387-389. Mani. R. and V. Sitakara Rao. 1982. Structural and phase changes in iron(III)Al(III) mixed hydroxides-oxide system: A thermoanalytical study. Thermo. Acta. 53: 175-182. McBride, M.B. 1994. Environmental Chemistry of Soils. Oxford University Press, New York, NY. McDowell, R. and L. Condron. 2001. Influence of soil constituents on soil phosphorus sorption and desorption. Comm. Soil Sci. and Plant Anal. 32: 2531-2547. Meziani, M.J., J. Zajac, D. J. Jones, J. Roziere and S. Partyka. 1997. Surface characterization of mesoporous silicoaluminates of the MCM-41 type: evaluation of polar surface sites using flow calorimetry, adsorption of cationic surfactant as a function of pore size and aluminum content. Langmuir 13: 5409-5417. Murphy, J. and J.P. Riley. 1962. A modified single solution method for the determination of phosphorus in natural waters. Anal. Chim. Acta 27: 31-36. Mustafa, S., S.Y. Hussain, and R. Ahmad. 1990. Phosphate/hydroxide exchange studies on amberlite IRA 400. Solv. extraction and ion exch. 8: 325-340.

PAGE 96

86 Mustafa, S., S.Y. Hussain, N. Rehana. 1989. Temperature effect on ion exchange sorption of phosphate. Solv. extraction and ion exch. 7: 705-720. Nanzyo, M. 1984. Diffuse reflectance infrared spectra of phosphate sorbed on alumina gel. J. Soil Sci. 35: 63-69. Nilsson, L., L. Lovgren and S. Sjorberg. 1992. Phosphate complexation at the surface of goethite. Chem. Spec. Bio. 4: 121-130. Noll, L.A. 1987. Adsorption calorimetry of surfactant interactions with minerals. Colloids surf. 26: 43-54. OwusuBennoah, E., C. Szilas, H.C.B. Hansen and O.K. Borggaard. 1997. Phosphate sorption in relation to aluminum and iron oxides of oxisols from Ghana. Comm. Soil Sci. and Plant Anal. 28: 685-697. Pardo, M.T. and M.E. Guadalix. 1990. Phosphate sorption in allophonic soils and release of sulphate silicate and hydroxyl. J. Soil Sci. 41: 607-612. Parfitt, R.L. 1989. Phosphate reactions with natural allophone, ferrihydrite, and goethite. J. Soil Sci. 40: 359-369. Persson, P., N. Nilsson, and S. Sjoberg. 1996. Structure and bonding of orthophosphate ions at the iron oxide-aqueous interface. J. Colloid and Interface Sci. 177: 263-275. Pierzynski, G.M., T.J. Logan, and J.M. Bingham. 1990. Phosphorus chemistry and mineralogy in excessively fertilized soils:a) Quantitative analysis of phosphorus-rich particles, b) description of phosphorus-rich particles, and Solubility equilibria. Soil Sci. Soc. Am. J. 54: 1576-1595. Reddy, K.R., O.A. Diaz, J Scinto and M Agami. 1995. Phosphorus dynamics in selected wetlands and streams of the Lake Okeechobee Basin. Ecol. Eng. 5: 183-207. Rhue, R.D., C. Appel, and N. Kabenji. 2002 .Measuring surface chemical properties of soil using flow calorimetry. Soil Sci. 167: 782-790. Rhue, R.D. and W.G. Harris.1999. Phosphorus sorption/desorption reactions in soils and sediments. P. 187-206. in K.R. Reddy(ed.) Biogeochemistry of subtropical ecosystems. CRC Press. Rodic, D., M. Mitric, R. Tellgren and H. Rundlof. 2001. The cation distribution and magnetic structure of Y 3 Fe (5-x) Al x O 12 . J. Magnet. and Magnet. Mat. 232: 1-8. Scheidegger, A.M., G.M. Lamble and D.L. Sparks. 1997. Spectroscopic evidence fro the formation of mixed-cation hydroxide phases upon metal sorption on clays and aluminum oxides. J. Coll. Inter. Sci. 186: 118-128.

PAGE 97

87 Schneider, S., F Simon, D. Pleul, and H.J. Jacobasch. 1997. Flow sorption calorimetry,a powerful tool to investigate the acid base character of organic polymer surfaces. Fresenius’ J. Anal. Chem. 358: 244-247. Sharpley, A.N. 1995. Soil phosphorus dynamics: agronomic and environmental impacts. Ecol. Eng. 5: 261-279. Sparks, D.L. 1995. Environmental soil chemistry. Academic Press Inc. San Diego, California. Stanjek, H. and U. Schwertmann. 1992. The influence of aluminum on iron-oxides: hydroxyl and aluminum substitution in synthetic hematites. Clays and Clay Min. 40: 347-354. Steinberg, G. 1981. What you can do with surface calorimetry. Chemtech. 11: 730-737. Tanada, S., M. Kabayama, N. Kawasaki, T. Sakiyama, T. Nakamura, M. Araki, and T. Tamura. 2003. Removal of phosphate by aluminum oxide hydroxide. J. Colloid and Interface Sci. 257: 135-140. Taneja, S.P. and D. Raj. 1993. Identification of iron oxide and hydroxide in soil clays. Nucl. Instr. Meth. Physics Res. Sect. B: Beam Interact. Materials and Atoms 76: 230-232. Taraba, B. 1990. Reversible and irreversible interaction of oxygen with coal using pulse flow calorimetry. Fuel. 69: 1191-1199. Tartaj, P. and J Tartaj. 2002. Preparation, characterization and sintering behavior of spherical iron oxide doped alumina particles. Acta Materialia 50: 5-12. Tejedor-Tejedor, M. I. and M.A. Anderson. 1990. The protonation of phosphate on the surface of goethite as studied by CIR-FTIR and electrophoretic mobility. Langmuir 6:602. Torrent, J., V. Barron and U. Schwertmann. 1990. Phosphate adsorption and desorption by goethites differing in crystal morphology. Soil Sci. Soc. Am. J. 54: 1007-1012. Tzou Y.M., M.K. Wang, R.H. Loeppert. 2003. Sorption of phosphate and Cr(VI) by Fe(III) and Cr(III) hydroxides. Arch. Environ. Contam. Toxicol. 44: 445-453. van Riemsdijk, W.H., and F.A.M. De Haan. 1981. Reaction of orthophosphate with a sandy soil at constant supersaturation. Soil Sci. Soc. Am. J.45: 261-266. van Riemsdijk, W.H., L.J.M. Boumans and F.A.M. De Haan. 1981. Phosphate sorption by soils: I. A model for phosphate reaction with metal-oxides in soil. Soil Sci. Soc. Am. J.48: 537-541.

PAGE 98

88 Wells, M.A., R.J. Gilkes and R.W. Fitzpatrick. 2001. Properties and acid dissolution of metal substituted hematites. Clays and Clay Min. 49(1): 60-72. Willet, I.R., C.J. Chartres, and T.T. Nguyen. 1988. Migration of phosphate into aggregrated particles of ferrihydrite. J. Soil. Sci. 39: 275-282. Wolska, E., W. Szajda and P. Piszora. 1992. Determination of solid-solution limits based on the thermal behavior of aluminum substituted iron hydroxides and oxides. J. Therm. Anal. 9: 2115-2122. Wolska, E., W. Szajda and P. Piszora. 1994. Mechanism for Alfor Fe-substitution during the -(Fe, Al)OOH -(Fe, Al) 2 O 3 transformation. Solid State Ionics 70/71: 537-541. Xie, R.J., A.F. MacKenzie, J.W. Fyles, and I.P. O’Halloran. 1993. Assessing energy of phosphate adsorption and desorption using an integrated Gibbs equation. Geoderma 59: 289-310. Zhang, M.. A.K. Alva and Y.C. Li. 2001. Aluminum and iron fractions affecting phosphorus solubility and reactions in selected sandy soils. Soil Sci. 166: 940-948. Zhou, M., R.D. Rhue and W.G. Harris. 1997. Phosphorus sorption characteristics of Bh and Bt horizon in horn sandy coastal plain soils. Soil Sci. Soc. Am. J. 61: 1364-1369.

PAGE 99

BIOGRAPHICAL SKETCH Omar Richard Harvey was born in St. Mary, Jamaica, on the 29 th day of November 1978 to Angella Frazer and Clifton Harvey. He was raised under the watchful eyes of his mother and step-father, Stafford Murray. After completing basic school in 1984, Omar attended the Port Maria Primary School for 6 years. Throughout these years he was nutured into becoming focused on his education. In 1990 at the age of 11 he passed his high school entrance exam on the very first attempt, thanks to the stewardship of his mother and his grade six teacher, Mrs. Kennedy. He then attended St. Mary High school where over the next 5 years he developed a passion for the natural sciences, particulary chemistry and agricultural science. By the end of his fifth year in high school, he had decided that he wanted to become an agronomist and was successful in passing 7 subjects in his exit exams. He became the first male student in his high school’s history to receive a grade 1 in agricultural sciences. Omar’s extra-curricular student life was equally exciting. He represented his high school in cricket and table tennis as well as science competitions and was a member of the student government. At age 17 Omar left his native Jamaica to pursue a degree in agronomy at the University of the West Indies in Trinidad and Tobago. He continued to excel and like previous times was always tipped to succeed. Under his advisor Dr. Isaac Bekele Omar received numerous academic awards. In his first year at UWI he received the Faculty of Natural Science award for best overall student. That same year he was the recipient of the Ernest Johnston memorial bursary, the UWI academic bursary and the Jamaica Agricultural Development Foundation 89

PAGE 100

90 scholarship for being the Jamaican student studying agriculture in Trinidad with the highest average grade. For the next two years, Omar was awarded the best agronomy student. In 2000, he was awarded the Cyrus McCormick academic scholarship as an exchange student at Virginia Tech. In December of that same year he completed his degree with first class honors becoming the youngest and first Jamaican to do so. He returned to Jamaica and worked as a teacher and later as a customer service representative with Grace Kennedy Limited, the nation’s largest company, until July 2002. During that same year Omar married his high school sweet heart of 7 years and was awarded an assistantship to pursue a master of science degree at the University of Florida. At UF, Omar was a student of Dr. R.D. Rhue and focused on research in environmental soil chemistry, specifically P interaction with Al and Fe hydr(oxides).