MINERAL PHASES IN RAW AND PROCESSED MUNICIPAL WASTE INCINERATION RESIDUES-TOWARDS A CHEMICAL STABILISATION AND FIXATION OF HEAVY METALS

Dissertation zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften an der Fakultät für Geowissenschaften der Ruhr-Universität Bochum

vorgelegt von Athanasius Priharyoto Bayuseno aus Ponorogo, Indonesien

Bochum 2006

Die vorliegende Arbeit wurde von der Fakultät für Geowissenschaften der Ruhr- Universität Bochum 2006 als Dissertation angenommen.

Erster Gutachter : Prof. Dr. Wolfgang W. Schmahl, LMU München Zweiter Gutachter : Prof. Dr. Hermann Gies Fachfremder Gutachter : Prof. Dr. Stefan Wohnlich Tag der mündlichen Prüfung : 4. Mai 2006

ii DANKSAGUNG An dieser Stelle möchte ich mich bei Herrn Prof. Dr. Wolfgang Schmahl herzlichst bedanken. Er hat zur Bearbeitung des Themas angeregt und durch seine stetige Unterstützung die vorliegende Arbeit ermöglicht. Durch seine Freude und seinen Enthusiasmus an der Kristallographie und Mineralogie hat er auch mich für dieses Fachgebiet begeistert. Über die Wissenschaft hinaus ist er in den drei Jahren zu einem richtigen Doktorvater für mich geworden. Mein besonderer Dank gilt Herrn Dr. Thomas Reinecke, der durch seine Förderung und fachliche Hilfe erheblich zum Gelingen dieser Arbeit beigetragen hat. Für die sehr gute Zusammenarbeit bedanke ich mich bei Herrn Dr. Heinrich Geiger, Frau Karin Bialas und vielen anderen Leuten am KAAD ( Katholischer Akademischer Austausch Dienst ) Bonn. Ohne ihre Hilfe und die finanzielle Unterstützung wäre diese Arbeit kaum möglich gewesen. Frau Dr. Anna Mielniczuk- Pastoors und Frau Ursula Röttsches beim Katholischen Hochschulzentrum Bochum danke ich recht herzlich für die Förderung zum Gelingen des KAAD Stipendiums. Für die finanzielle Unterstützung durch das Institut für Geologie, Mineralogie und Geophysik der Ruhr-Universität Bochum bedanke ich mich auch. Herrn Prof. Eko Budihardjo, Rektor der Diponegoro-Universität (UNDIP), Semarang, Indonesien, und seinen Mitarbeitern danke ich recht herzlich für die Förderung meines Studiums an der Ruhr-Universität Bochum. Mein besonderer Dank gilt Herrn Dr. Neuser und Herrn Dr. Bernhardt, die durch ihre fachliche Hilfe meine Arbeit am Rasterelektronenmikroskop und der Elektronenstrahlmikrosonde ermöglicht haben. Herrn Hendrik Narjes, Frau Astrid Michele, Frau Sandra Grabowski und Herrn Udo Trombach danke ich für chemische Analysen und Hilfe bei Laborexperimenten. Ich möchte hier Frau Antoinette für ihre praktische Hilfe danken, sowie auch allen Mitarbeitern des Mineralogischen Institutes der Ruhr-Universität Bochum und meinem Kollegen Mahamudul Hasan. Ebenfalls bedanken möchte ich mich bei Herrn Dr. Schlüter am Mineralogischen Museum der Universität Hamburg für die Beschaffung von Gordaite, und bei dem MHKW Essen und Herrn Dr. Müllejans vom MHKW Iserlohn für die Beschaffung der Materialproben. Von ganzem Herzen danke ich meiner Frau Nunuk, meiner Tochter Yustina und meinem Sohn Andreas, die die Arbeit jederzeit voll unterstützt haben.

iii ZUSAMMENFASSUNG Die Eigenschaften von Filterstäuben und Schlacken aus deutschen Müllverbrennungsanlagen (MVA) und der chemisch-mineralogische Umsatz von Filterstäuben wurden mit dem Ziel erforscht, die enthaltenen Schwermetalle durch kristallchemische Substitution in beständigen Mineralphasen zu immobilisieren. Chemische Pauschalanalysen zeigen, dass die Schlacken Al, Si und Fe als Hauptkomponenten enthalten. Im Gegensatz dazu bestehen Filterstäube vor allem aus Ca, Na, K, Cl und S. Zusätzlich finden sich in den Filterstäuben hohe Konzentrationen toxischer Schwermetalle (z.B. Zink, Cadmium und Blei). Die Hauptgemengteile (>1 Gew.-%) der frischen Schlacken, die mit der Rietveldmethode quantitativ analysiert wurden, sind Silikate (z.B. Gehlenit, Augit, Diopsid, Quarz), Oxide (z.B. Magnetit, Spinel, Hämatit) und Karbonate (z.B. Calcit), außerdem ist ein hoher Gehalt an Glas (>30 Gew.-%) vorhanden. Als untergeordnete Phasen kommen Baryt, Rutil und Wüstit vor. Die frischen Filterstäube erhalten eine erhebliche Menge Glas (> 40 Gew.-%) und unterschiedliche kristalline Phasen wie Kalziumtetrachlorozinkat (K 2ZnCl 4), Gehlenit, Halit, Quarz und Feldspat. Andere identifizierte Phasen sind Magnetit, Hämatit, Kalk und verschiedene Ca-Sulfate. Die magnetische Fraktion der Filterstäube, die Magnetit, Hematit und weitere untergeordnete Phasen enthält, konnte abgetrennt werden. Der Einfluß der natürlichen Alterung auf die Mineralstabilität in Schlacken und Filterstäuben wurde mit Röntgenbeugungsmethoden untersucht. Der Alterungsprozeß in den Schlacken führt zur Bildung neuer Minerale wie Ettringit und Hydrocalumit. Die Alterung der Minerale in den Filterstäuben führt zur Neubildung von Syngenit, Gips und Hydrocalumit. Gordaite wurde in gealterten Proben gefunden, die zuvor mit Wasser vermischt worden waren. Es ist erkennbar, dass die Alterung von Schlacken und Filterstäuben eine Abnahme des pH-Wertes bewirkt. Einzelne Partikel von ausgewählten Schlacken und Filterstäuben wurden mit der Elektronenstrahlmikrosonde analysiert. Die groben Partikel der Schlacken enthalten vor allem CaO, SiO 2, Al 2O3 und Fe 2O3, und bestehen aus den Mineralen Quarz, Spinel, Melilit, und Glasphase mit Si-oder Fe-reichen aluminosilikatischen Zusammensetzungen. Dagegen enthalten die feinen Partikel der Filterstäube vor allem

SiO 2, Al 2O3 und Fe 2O3, außerdem erhöhte Konzentrationen toxischer Schwermetalle (z.B. Zn, Pb und Cd). Die Partikel der Filterstäube bestehen aus sehr heterogenen

iv Anteilen von Glas, Metall und anderen kristallinen Phasen, in denen die Schwermetalle verteilt sind. Das Glas hat eine Ca-reiche Aluminosilikat-Zusammensetzung. Außerdem wurde die mineralogische Alterung von Schlacken und Filterstäuben mit Röntgenbeugungsmethoden über einen Zeitraum von 6 Monaten untersucht. Portlandit und Ettringit bildeten sich sofort in der abgeschreckten Schlacke. Ettringit wurde nachfolgend zum stabileren Gips und Hydrocalumit abgebaut. In Abhängigkeit von der chemischen Zusammensetzung der Filterstäube führte ihre Alterung zu einer beträchtlichen Neubildung von Syngenit, Gips, Hydrocalumit und Gordait. Zur Feststellung der Extrahierbarkeit wasserlöslicher Phasen und Schwermetalle aus Filterstäuben dienten zwei Arten von Versuchen. Einerseits wurden die Filterstäube ausgewaschen, um lösliche Salze abzutrennen. Andererseits wurde das Soxhlet- Verfahren mit heißem Wasser als Lösungsmittel angewandt. Das Löslichkeitsprodukt der Alkalichloride (NaCl und KCl) ist verantwortlich für die Freisetzung von Na, K und Cl aus den Filterstäuben. Lösungs-Fällungsgleichgewichte zwischen Ca-, K-, Al-, und

2− SO 4 -haltigen Mineralen führten zur Bildung der Hydratphasen Gips, Syngenit, und Ettringit. Sowohl die Auswaschung als auch das Soxhlet-Verfahren sind offenbar keine wirksamen Methoden zur Extraktion von Schwermetallen aus Filterstäuben. Die pozzolanische Verfestigung der Filterstäube wurde mit der Rietveld Methode in Abhängigkeit von der Reaktionszeit untersucht. Das Massenverhältnis von Lösung zu Feststoff betrug in den Experimenten 3 bzw. 10. Syngenit und Gips bildeten sich in den verfestigten Original-Filterstäuben. Zusätzlich entstanden nur geringe Mengen Ettringit, Hydrocalumit und CSH Phasen. Auch die gewaschenen Filterstäube weisen zementartige Eigenschaften auf. Bei der Hydratation scheint die Glasphase zu Kristalhydratphasen zu reagieren. Der hydrothermale Umsatz der ungewaschenen Filterstäube mit wässrigen NaOH- und KOH-Lösungen unterschiedlicher Molarität bei verschiedenen Temperaturen (90 O -180 OC) und Reaktionszeiten wurde mit der Rietveld Methode untersucht. Bei 180 OC, 48 h entstand in Gegenwart von 0.5 M NaOH-Lösung eine erhebliche Menge Al-substituierten 11Å-Tobermorits und Katoits. Bei ähnlichen Bedingungen bildete sich in Anwesenheit von KOH-Lösung eine nur geringe Menge Al-substituierten 11Å-Tobermorits. Eine Kristallisation von Zeolithen (Analcim und Hydroxylcancrinit) ließ sich nicht erreichen. Ein erheblicher Anteil der

v Aluminosilikatglas-Matrix wandelte sich beim hydrothermalen Umsatz zu kristallinen Phasen um. Zudem war eine dramatische Abnahme des Quarzanteils im Verhältnis zu den unbehandelten Filterstäuben beobachten. Der hydrothermale Umsatz gewaschener Filterstäube mit 0.5 M NaOH-Lösung bei 180 OC erzeugte in 48 h in gleicher Weise wie beim ungewaschenen Edukt das gemischte Produkt Al-Tobermorit und Katoit. Das Waschen der Filteraschen erwies sich im hydrothermalen Umsatz als wichtige Voraussetzung für die Bildung von Zeolithen (Analcim, Hydrocrancrinit). Die Löslichkeit der Minerale von Filterstäuben in sauren wässrigen Lösungen wurde ebenfalls untersucht. Minerale wie NaCl, KCl und K2ZnCl 4 in den unbehandelten Filterstäuben lösten sich gut, ebenso Syngenit, Ettringit, Hydrocalumit und Gordaite in den gewaschenen Filterstäuben. Auch in den verfestigten Produkten gewaschener Filterstäube ging Ettringit und CSH-Phase in Lösung. Dagegen lösten sich die Produktphasen des hydrothermalen Umsatzes, nämlich 11Å-Tobermorit, Analcim und Hydroxylcancrinit nicht in sauren Lösungen. Diese Ergebnisse zeigen, dass der hydrothermale Umsatz von Filterstäuben eine stabile Produktparagenese erzeugt. Schließlich wurden Auswaschversuche nach dem TCLP-Verfahren ( toxicity characteristic leaching procedure test ) zur Untersuchung der potentiell toxischen Materialien durchgeführt. Aus den unbehandelten und gewaschenen Filterstäuben und ihren verfestigten Folgeprodukten ließen sich die Schwermetalle, insbesondere Zn, Pb und Cd, in Konzentrationen herauslösen, die über den jeweiligen TCLP Grenzwerten liegen. Dagegen ließen sich die Schwermetalle aus den hydrothermalen Reaktionsprodukten der Filterstäube wesentlich schlechter extrahieren. Damit eröffnet das hydrothermale Verfahren einen viel versprechenden Ansatz, das von Filterstäuben ausgehende potentielle Risiko der Umweltgefährdung zu reduzieren und die hydrothermal umgesetzten Filterstäube als Resource zu nutzen.

Schlagworte : MVA Schlacken, Filterstäube, Rietveld Methode, Hydratation und Hydrothermale Verfahren.

vi ABSTRACT The characteristics of German MSWI residues (bottom ash and fly ash) and the use of fly ash for the development of stable mineral phases with the intention of immobilising heavy metals were explored. Bulk chemical analyses indicated that the bottom ash had major components of Al, Si and Fe, while fly ash was composed largely of Ca, Na, K, Cl, S, in addition to high concentrations of toxic elements ( e.g ., Zn, Pb and Cd). The major crystalline phases (> 1 wt.%) of the fresh bottom ash determined by the XRD Rietveld method are silicates ( e.g ., gehlenite, augite, diopside, quartz), oxides (e.g ., magnetite, spinel, hematite) and carbonates ( e.g ., calcite), in addition to a significant glassy content (>30 wt.%). The minor phases included barite, rutile and wüstite. Likewise, the fresh fly ash contained a high amount of amorphous phase (> 40 wt.%) and various crystalline phases such as potassium tetrachlorozincate (K 2ZnCl 4), gehlenite, halite, quartz, and feldspar. Other phases are magnetite, hematite, lime and Ca-sulfates. The magnetic fraction of fly ash could be extracted using the magnetic stirring procedure. Here, the magnetic fraction contained mainly magnetite, hematite and other minor minerals. The effect of natural aging on the mineral stability of bottom and fly ash was evaluated by XRD. The aging process modified mineralogy of the bottom ash, leading to the formation of new minerals ( e.g ., ettringite and hydrocalumite). Similarly, aging of fly ash resulted in the formation of syngenite, gypsum and hydrocalumite. In contrast, gordaite was found in the aged sample mixed with water. Evidently, aging of the bottom ash and the fly ash resulted in the reduction of the pH values. Individual particles of the selected bottom and fly ash samples were examined by electron probe x-ray microanalysis (EPMA). The coarse particles of the bottom ash contained primarily CaO, SiO 2, Al 2O3 and Fe 2O3, corresponding to minerals, quartz, spinel, melilite, and amorphous material with Si-rich aluminosilicate compositions. In contrast, the fly ash particles were composed largely of SiO 2, Al 2O3 and Fe 2O3, and enriched in many toxic elements ( e.g ., Zn, Pb, and Cd). Additionally, the fly ash particles consisted of a very heterogeneous assemblage of glasses, metals, and other crystals in which heavy metals were distributed. The glasses were composed of Ca-rich aluminosilicate.

vii Further, aging of bottom and fly ash was specifically investigated by XRD for a period of 6 months. Portlandite and ettringite were formed immediately in the quenched bottom ash, but ettringite was further altered to form the stable minerals gypsum and hydrocalumite. Aging of fly ash obviously yielded the significant formation of syngenite, gypsum, hydrocalumite and gordaite, depending on the chemical compositions of the fresh fly ash. Extractability of water-soluble compounds and heavy metals from the fly ash was investigated in two experiments. First, washing was applied to remove soluble salts. Second, a Soxhlet hot water-extraction was employed. The dissolution product of chloride minerals ( i.e., NaCl and KCl) was responsible for the release of Na, K and Cl.

2− Moreover, the dissolution-precipitation of Ca, Al, Si- and SO 4 containing minerals resulted in the formation of hydrate phases such as gypsum, syngenite, and ettringite. Both washing and Soxhlet-extraction were obviously not effective methods for extracting the heavy metals from the fly ash. Pozzolanic solidification of fly ash at various times has been investigated by the XRD Rietveld method. Liquid to solid ratios (L/S) equal to 3 and 10, respectively were selected. Syngenite and gypsum were formed in the solidified raw fly ash. Only small amounts of ettringite, hydrocalumite and CSH phase were produced. Likewise, the washed fly ash exhibited the cementitious properties, yielding a high quantity of gypsum and ettringite. Here the amorphous phase seemed to be transformed to the crystalline hydrate phases during pozzolanic reaction. Hydrothermal treatment of fly ash at different molarity of NaOH and KOH, temperatures (90 O-180 OC) and times has been investigated by the XRD Rietveld method. Significant amounts of Al-substituted 11Å tobermorite and katoite were produced at 180 OC for 48 h in the presence of 0.5 M NaOH. Similar treatment regimes in the presence of KOH did not yield significant quantities of Al-substituted 11Å tobermorite. A significant formation of zeolitic materials (analcime and hydroxylcancrinite) could not be gained. Amounts of alumino-silicate glassy matrix were also converted to minerals during the hydrothermal reaction. A dramatic reduction of the quartz level in the raw fly ash was noted. Likewise, the hydrothermal treatment of the washed fly ash with 0.5 M NaOH at 180 OC for 48 h yielded the mixed product of Al-substituted 11Å tobermorite and katoite. Additionally, the water washing was found

viii to be an important process to produce zeolitic compounds such as analcime and hydroxylcancrinite. The stability of minerals in the acidic solution was examined by XRD. Certain soluble minerals ( i.e., NaCl, KCl and K 2ZnCl 4) of the raw fly ash were clearly leached out, whereas syngenite, ettringite, hydrocalumite and gordaite formed in the washed fly ash also disappeared. Likewise, ettringite and CSH phase in the solidified products of the washed fly ash were dissolved in the leaching solution. In contrast, the main hydrothermal products (namely tobermorite-11Å, analcime and hydroxylcancrnite) remained stable under acidic conditions. Importantly, the findings showed that hydrothermal treatment of the MSWI fly ash produce the stabilised products, which can be proven to be non-hazardous. Finally, the TCLP ( toxicity characteristic leaching procedure ) test was employed to investigate the potential toxicity of the materials. The raw and washed fly ashes and their solidified products showed the leachability of heavy metals such as Zn, Pb and Cd beyond the respective regulatory limits. By contrast, the hydrothermal products of fly ash exhibited less leachability of heavy metals relative to the parent materials. Thus, the hydrothermal method provides a promising solution to the reduction of environmental hazards of fly ash deposits and it opens opportunities for giving value to the fly ash as a resource.

ix TABLE OF CONTENTS Page DANKSAGUNG iii ZUSAMMENFASSUNG iv ABSTRACT vii TABLE OF CONTENTS x LIST OF TABLES xiii LIST OF FIGURES xv LIST OF ABBREVIATIONS AND ACRONYMS xxiii CHAPTER 1. INTRODUCTION 1.1 Background to Study 1 1.2 Characteristics of MSWI Residues 4 1.3 Treatment Strategy of MSWI Residues 6 1.4 Objectives 10 1.5 Research Plan 10 1.6 Structure of Doctoral Thesis 11 CHAPTER 2. STABILISATION TREATMENTS OF MSWI FLY ASH 2.1 Stabilisation of MSWI Fly Ash 13 2.2 Characteristics of MSWI Residues 16 2.3 Stabilisation of MSWI Fly Ash in a Cement-Based System 22 2.3.1 Pozzolanic Solidification of MSWI Fly Ash 23 2.3.2 Microstructural Model of C-S-H Phase 24 2.3.3 Stabilisation Mechanism in a Cement-Based System 26 2.4 Hydrothermal Treatments of MSWI fly ash 30 2.5 Leaching Behaviour of MSWI Residues 35 2.5.1 Leaching Mechanism 36 2.5.2 Factors Controlling Leaching 37 2.5.3 Evaluation of Leaching Resistance 44 2.6 Summary 44 CHAPTER 3. METHODOLOGY OF RESEARCH 3.1 Experimental Design 46 3.2 Powder Processing 50

x 3.3 Pozzolanic Solidification Experiment 54 3.4 Hydrothermal Processing 54 3.5 Characterisation 55 3.5.1 Chemical Analysis 55 3.5.2 Optical and SEM Microscopy 55 3.5.3 Electron Probe X-ray Microanalysis (EPMA) 56 3.5.4 Analytical Method 56 3.6 Leaching Experiment 62 CHAPTER 4. CHARACTERISTICS OF MSWI RESIDUES 4.1 Characterisation of MSWI Residues 64 4.1.1 Bulk-Chemistry 64 4.1.2 Particle Chemistry and Morphology 68 4.1.3 Mineralogy of MSWI Residues 72 4.1.4 Summary 89 4.2 Microscopic Characteristics of MSWI Residues 90 4.2.1 MSWI Bottom Ash 90 4.2.2 MSWI Fly Ash 96 4.2.3 Summary 103 4.3 Mineralogical Alteration of the MSWI Residues 104 4.3.1 Mineralogical Changes of Bottom Ash 104 4.3.2 Mineralogical Changes of Fly Ash 108 4.3.3 Summary 115 4.4 Water-Extraction Processes of Fly Ash 115 4.4.1 Water-Washing Process 115 4.4.2 Soxhlet-Water Extraction 121 4.4.3 Summary 126 4.5 Discussion 126 CHAPTER 5. SYNTHESIS AND LEACHING ANALYSIS OF STABILISED MATERIALS FROM RAW AND WASHED MSWI FLY ASHES

xi 5.1. Pozzolanic Solidification of MSWI Fly Ash 132 5.1.1. Raw Fly Ash 132 5.1.2. Washed Fly Ash 142 5.1.3. Summary 149 5.2. Hydrothermal Treatments of MSWI Fly Ash 149 5.2.1. Raw Fly Ash 149 5.2.2. Washed Fly Ash 164 5.2.3. Summary 170 5.3. Stability and Toxic Potential of Stabilised Materials 170 5.3.1. Stability of Raw Fly Ash and Stabilised Materials 170 under Acidic Conditions 5.3.2. Stability of Washed Fly Ash and Stabilised Materials 179 under Acidic Conditions 5.3.3. Summary 186 5.4. Discussion 187 CHAPTER 6. CONCLUSION 6.1. Characteristics of MSWI Residues 194 6.2. Pozzolanic and Hydrothermal Activations of the Raw and Washed 198 Fly Ashes 6.3. Leaching Resistance of the Raw and Washed Fly Ashes and the 200 Stabilised Materials 6.4. Suggestions for Further Research 201 REFERENCES 203 APPENDICES I. Lebenslauf 216 II. Publications from Thesis 217

xii LIST OF TABLES Page Table 2.1 Ranges of chemical elements in MSWI residues (Hjelmar, 1996). 18 Table 2.2 Physical properties of metals and their compounds 19 (Lide, 1997; Youcai et al. , 2002). Table 2.3 Mineralogy of MSWI residues. 20 Table 2.4 Joint Committee of Powder Diffraction Society (JCPDS) card 32 number for some zeolites (Querol et al ., 2002). Table 2.5 Zeolites and other neomorphic phases synthesised from coal fly 33 ash as a function of activation agent (NaOH or KOH), temperature and activation solution/fly ash ratio (ml/g) (Querol et al ., 2002). Table 2.6 Maximum concentration levels of contaminants in leachates from 36 various MSWI residues (Hjelmar, 1996). Table 4.1 Chemical composition of the MSWI solid residues. 65 Table 4.2 XRD-based mineralogy of the BA-A1 sample. The minerals are 77 arranged in the following sequence (1) major minerals in the fresh bottom ash, (2) major new mineral after aging or water treatment, (3) minor phases (<3 wt.% each) in alphabetical order. Table 4.3 XRD-based mineralogy of bottom ash samples. 80 Table 4.4 XRD-based mineralogy of the FA-A1 Sample. 85 Table 4.5 XRD-based mineralogy of fly ash samples. 88 Table 4.6 Chemical composition of phases in bottom ash particles analysed 94 by EPMA as illustrated in figures 4.9 and 4.10. Composition is given in wt. % of oxides except for the alloy (spot 2), where the numbers refer to element wt.%. Table 4.7 Mineralogy of bottom ash particles from EPMA analysis 95 (composition given in wt. % of oxides, except for the metallic phases in wt.% element) Table 4.8 Chemical composition (wt.%) of phases in fly ash particles 100 analysed by EPMA spot analysis as illustrated in figures 4.11 and 4.12. Table 4.9 Chemical composition of phases in fly ash particles analysed by 101 EPMA (composition given in wt. % of oxides, except for the alloy given in wt.% elements). Table 4.10 Chemical composition of phases in fly ash particles analysed by 102 EPMA (composition given in wt. % of oxides). Table 4.11 The quantity of chemical elements in the raw and washed fly ash 117 samples. Table 4.12 XRD-based mineralogy of water-washed fly ash samples. 120 Table 4.13 Bulk chemical composition of solid residues obtained from the 122 Soxhlet extraction process of sample FA-A1. Table 4.14 XRD based mineralogy of residues from the Soxhlet experiment 125 of FA-A1. Table 5.1 XRD-based mineralogy of the FA-A1 sample treated with 0.5 M 156 NaOH. Table 5.2 XRD-based mineralogy of the FA-A1 sample treated with 0.5 M 159 KOH.

xiii Table 5.3 Concentrations of heavy metals in the leaching solution after 18-h 177 reaction of leaching process for the raw fly ash samples and their stabilised materials. Table 5.4 Concentrations of heavy metals in the leaching solution after 18-h 184 reaction of the leaching process for the washed fly ash samples and their stabilised materials.

xiv LIST OF FIGURES Page Figure 2.1 Idealised structure of zeolite framework (after Querol et al ., 15 2002). Figure 2.2 (A) and (B) structure of a single layer of 1.4-nm tobermorite in 25 bc -and ac -projections, respectively. In (B), the chains are seen end on. (C) Suggested structure for a single layer of jennite, in ac projection; the chains are seen end on and the Ca-O sheets edge on, parallel to their corrugations, and circled ‘H’s represent hydroxyl groups. In (A), (B) and (C), full circles represent calcium atoms, P and B denote paired and bridging tetrahedra, respectively. Jennite axes relate to a monoclinic pseudocell with a=1.00nm, b=0.36nm, c=2.14nm, β=101.9 O (after Taylor, 1990). Figure 2.3 A sorosilicate model of C-S-H Phases: (a) fully stoichiometric 26 sorosilicates (Ca/Si=2.0), (b) sorosilicate with all Ca(OH) 2 removed (Ca/Si=1.5), (c) sorosilicate with half of its calcium removed (Ca/Si=1.0); (d) the sorosilicate has ”unzipped” to form dreierketten (Ca/Si =1.0) (after Grutzeck et al ., 1999). Figure 2.4 Schematic sorption isotherms of metal ion (Me) at a mineral 30 surface at constant pH for different cases (after Stumm, 1992). Figure 2.5 Schematic of leaching mechanisms (after Corner, 1990). 37 Figure 2.6 Solubility of (a) Cd, (b) Al, and (c) B concentration in leaching 39 solution from samples of fresh and aged MSWI fly ash (after Sabbas et al ., 2003). Figure 2.7 Leaching behaviour of Na, K, Cl from weathered MSWI bottom 41 ash as function of the L/S ratio (after Sabbas et al ., 2003). Figure 2.8 Point of zero charge and effect of pH on the surface charge for 43 some common minerals (after Stumm and Morgan, 1996) . Figure 3.1 A Soxhlet of water-extraction apparatus. A: Liebig cooler, B: 53 paper filter capsule, C: glass flask, and D: glass extractor. Figure 3.2 Geometry of the Bragg-Brentano diffractometer (after Siemens, 57 1986). Figure 3.3 Quality of the Rietveld pattern-fitting results for the aged BA-A1 61 bottom ash where ( ······ ) and ( −−−−) denote the observed and calculated patterns, respectively, along with the associated difference plot. Figure 4.1 Major chemical elements of bottom ash (BA-A1) in comparison 66 to the element abundance in the Earth’s crust (Rudnick and Fountain, 1995). Figure 4.2 SEM micrographs of particles for (a) bottom ash (BA-A1) and (b) 70 fly ash (FA-A1) collected from incinerator A. Note the difference in scale for the two micrographs. Figure 4.3 (a) SEM micrograph of the FA-B1 sample collected from 71 incinerator B: a spherical particle is visible; on its surface some small aggregated particles are present, and (b) EDX spectrum of the spherical particle shown in figure 4.3a. Figure 4.4 XRD patterns of the fresh (a) bottom ash and (b) fly ash. BA-An, 73 FA-An and BA-Bn, FA-Bn refer to the bottom ash and fly ash

xv collected from the incinerator plants A and B respectively (see section 3.2). The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Bo (boehmite), C (calcite), Co (corundum), Cr (cristobalite), Ct (caracolite), E (ettringite), Fe (iron), G (gehlenite), He (hematite), Hc (hydrocalumite), Hl (halite), L(lime), Lz (lazurite), M (magnetite), Mi (minium), Pz (K 2ZnCl 4), Q (quartz), R (rutile), S (sylvite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns. Figure 4.5a Diffractograms of (a) fresh, (b) 6-months storage, (c) water- 76 treated and (d) magnetically separated bottom ash of BA-A1. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Co (corundum), Cq (coquimbite), E (ettringite), G (gehlenite), Gy (gypsum), He (hematite), Hc (hydrocalumite), Lp (lepidocrocite), M (magnetite), Q (quartz), R (rutile) and W (wüstite). Figure 4.5b Quality of the Rietveld pattern-fitting results for the 6-months 76 storage of the BA-A1 sample measured without internal standard. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern. Figure 4.6a XRD patterns of (a) fresh, (b) aged (c) water- treated, and (d) 81 magnetically separated fly ash. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite) An (anhydrite), Ba (bassanite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), L (lime), M (magnetite), Po (portlandite) Pz (K 2ZnCl 4), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Sy (syngenite) and Us (ulvöspinel). Figure 4.6b Quality of the Rietveld pattern-fitting results for the water-treated 81 FA-A1 sample without internal standard. The curves are observed (······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern. Figure 4.7a XRD pattern for the fresh FA-A1 sample. The principal 82 diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), G (gehlenite), Hl (halite), He (hematite), L (lime), Pz (potassium tetrachlorozincate), Q (quartz), Po (portlandite), R (rutile) and Us (ulvöspinel). Figure 4.7b Quality of the Rietveld pattern-fitting results for the fresh FA-A1 82 sample without internal standard. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern. Figure 4.8a XRD patterns of the fresh fly ash samples from incinerator A and 87 B. FA-An and FA-Bn stand for samples of fly ash from the

xvi incinerator facilities A and B respectively. The peaks are labelled An (anhydrite), Bo (boehmite), C (calcite), Cr (cristobalite), Ct (caracolite), Fe (iron), G (gehlenite), Gd (gordaite), Hl (halite), Hc (hydrocalumite), He (hematite), L (lime), Lz (lazurite), Po (portlandite), Q (quartz), R (rutile) and S (sylvite). Figure 4.8b Quality of the Rietveld pattern-fitting results for the fresh sample 87 of FA-A3. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern. Figure 4.9 BSE images of bottom ash polished thin sections illustrating the 92 location of EPMA spot analyses. The numbers refer to analyses in table 4.6. Figure 4.10 BSE images of bottom ash polished thin sections illustrating the 93 location of EPMA spot analyses. The numbers refer to analyses in table 4.6. Figure 4.11 BSE images of fly ash polished thin sections illustrating the 97 location of EPMA spot analysis. The numbers refer to analyses in table 4.8. Figure 4.12 BSE images of fly ash polished thin sections illustrating the 98 location of EPMA spot analysis. The numbers refer to analyses in table 4.8. Figure 4.13 Diffractograms of BA-A2 aged at room temperature over times; 105 (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Cr (cristobalite), D (diopside), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), M (magnetite), Q (quartz) and R (rutile). Note: SQRT refers to the square root of the intensity scale for the XRD patterns. Figure 4.14 Diffractograms of BA-B1 aged at room temperature over times; 107 (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Co (corundum) Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Po (portlandite), Q (quartz) and R (rutile). Note: SQRT refers to the square root of the intensity scale for the XRD patterns. Figure 4.15 Diffractograms of FA-A2 aged at room temperature as a function 109 of time; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Ba (bassanite), Bo (boehmite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gd (gordaite), Hc (hydrocalumite), He (hematite), Hl (halite), Lz (lazurite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite) and Sy (syngenite). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

xvii Figure 4.16 Diffractograms of FA-A3 aged at room temperature as a function 111 of time; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), Ba (bassanite), Bo (boehmite), C (calcite), Cr (cristobalite), G (gehlenite), Hc (hydrocalumite), He (hematite), Hl (halite), Lz (lazurite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite) and Sy (syngenite). Note: SQRT refers to the square root of the intensity scale for the XRD patterns. Figure 4.17 Diffractograms of FA-A4 aged at room temperature as a function 113 of time; (a) from fresh to 6 months, and (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), Ct (caracolite), Fe (iron), G (gehlenite), Gd (gordaite), Hl (halite), Lz (lazurite), M (magnetite), Q (quartz), R (rutile), Sy (syngenite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns. Figure 4.18 Diffractograms of FA-B1 aged at room temperature as a function 114 of time; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Ba (bassanite), Bo (boehmite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), L (lime), Lp (Lepidocrocite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns. Figure 4.19 Diffractograms of (a) washed (WFA-An and WFA-Bn) fly ashes 119 and (b) the raw and washed FA-A3 samples. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel). Figure 4.20 XRD patterns of (a) solid residues extracted from sample FA-A1 124 by the Soxhlet device and (b) raw and extracted fly ash. The principal diffraction peaks for some of the more predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Ac (arcanite), Ba (bassanite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), Kt (katoite), M (magnetite), Ms (monosulfate), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

xviii Figure 5.1 XRD patterns of FA-A1 solidified with the saturated solution of 134 Ca(OH) 2 using liquid/solid ratios of (a) L/S = 3, and (b) L/S =10 at various times. The peaks are labelled Al (alunite), An (anhydrite), Ba (bassanite), Be (bernalite), C (calcite), Cr (cristobalite), E (ettringite), Fe (iron), Gy (gypsum), G (gehlenite), Hc (hydrocalumite), He (hematite), Hl (halite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Sy (syngenite), Us (ulvöspinel), Tb (tobermorite-14Å-CSH phase) and Th (thenardite). Figure 5.2 (a) SEM image of prismatic crystals of syngenite, and (b) EDX 136 spectrum obtained from FA-A1 solidified with the saturated solution of Ca(OH) 2 at the L/S ratio of 10 for 1 month. Figure 5.3a Diffractogram of FA-A1 solidified with the saturated solution of 137 Ca(OH) 2 at L/S ratio of 3 for 14 days. The peaks are labelled An (anhydrite), C (calcite), E (ettringite), Gy (gypsum), G (gehlenite), Hc (hydrocalumite), Hl (halite), M (magnetite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel). Figure 5.3b Quality of the Rietveld pattern-fitting results for FA-A1 solidified 137 with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 14 days where (·········) and ( ) denote the observed and calculated patterns, respectively. Figure 5.4 Main hydrate phases as a result of FA-A1 solidified with the 138 saturated solution of Ca(OH) 2 using ratios of: (a) L/S = 3, and (b) L/S = 10 at various times. Note that the amorphous phase in the solidified FA-A1 sample at L/S ratio of 3 and various times was not detected by the quantitative XRD analysis. Figure 5.5a Diffractograms of FA-An and FA-Bn samples solidified with the 141 saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days. The peaks are labelled An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), Hl (halite), Hz (hydrozincite), Pa (potassium alum), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel). Figure 5.6a Diffractograms of WFA-A1 solidified with the saturated solution 143 of Ca(OH) 2 at L/S ratio of 3 and various times. The peaks are labelled An (anhydrite), C (calcite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), M (magnetite), Po (portlandite), Q (quartz), R (rutile) and Sp (sphalerite). Figure 5.6b Main hydrate phases as a result of pozzolanic solidification of 143 WFA-A1 with the saturated solution of Ca(OH) 2 at L/S ratio of 3 and various times. Note that the amorphous phase in the solidified WFA-A1 sample at L/S ratio of 3 and various times was not detected by the quantitative XRD analysis. Figure 5.7a Diffractograms of WFA-A1 solidified with the saturated solution 145 of Ca(OH) 2 at L/S ratio of 10 and various times. The peaks are labelled An (anhydrite), C (calcite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite),

xix M (magnetite), Q (quartz), R (rutile) and Tb (tobermorite-14Å- CSH phase). Figure 5.7b Main hydrate phases as a result of pozzolanic solidification of 145 sample WFA-A1 with the saturated solution of Ca(OH) 2 at L/S ratio of 10 and various times. Figure 5.8 (a) SEM image of a prismatic morphology of gypsum, and (b) 146 EDX spectrum obtained from WFA-A1 solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 10 for 2 months. Figure 5.9a Diffractograms of WFA-An and WFA-Bn samples solidified with 148 the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cs (cerussite-PbCO 3), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), M (magnetite), Pa (potassium alum), Q (quartz), R (rutile), Sp (sphalerite) and Tb (tobermorite-14Å-CSH phase) . Figure 5.9b Main hydrate phases as a result of pozzolanic solidification of 148 WFA-An and WFA-Bn samples with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days. Figure 5.10 XRD patterns of FA-A1 hydrothermally treated at various 151 temperatures for 48 h in (a) 0.5 M NaOH, and (b) 0.5 M KOH. The peaks are labelled Ac (analcime), Al (alunite), An (anhydrite), Ba (bassanite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), Hcr (Hydroxylcancrinite), He (hematite), Hl (halite), Ilt (illite), Kt (katoite), Lp (lepidocrocite) M (magnetite), Ms (monosulfate), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Us (ulvöspinel) and Tb (tobermorite- 11Å). Figure 5.11 (a) SEM image of a platy network Al-substituted tobermorite- 153 11Å crystal morphology, and (b) EDX spectrum obtained from FA-A1 hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h. Figure 5.12 (a) SEM image of a platy network Al-substituted tobermorite- 154 11Å crystal morphology, and (b) EDX spectrum obtained from FA-A1 hydrothermally treated at 180 OC in 0.5 M KOH for 48 h. Figure 5.13a Diffractogram of FA-A1 hydrothermally treated at 180 OC in 155 0.5 M NaOH for 48 h. The peaks are labelled C (calcite), Ce (internal standard of CeO 2), G (gehlenite), Gy (gypsum), Ilt (illite), Kt (katoite), M (magnetite) and Tb (tobermorite-11Å). Figure 5.13b Quality of the Rietveld pattern-fitting results for FA-A1 155 hydrothermally treated with 0.5 M NaOH at 180 OC for 48 h where (······) and ( ) denote the observed and calculated patterns, respectively. Figure 5.14 Minerals produced from hydrothermal conversion of sample FA- 160 A1 in (a) 0.5 M NaOH, and (b) 0.5 M KOH at different reaction temperatures and fixed time of 48 h. Note H-Cancrinite : hydroxylcancrinite. Figure 5.15 XRD patterns of FA-A1 hydrothermally treated (a) at 180 OC in 162 0.5, 1 and 2.5 M NaOH for 48 h, and (b) at 180 OC in 0.5 M NaOH for 36, 48 and 168 h. The peaks are labelled Ac (analcime), Al (alunite), An (anhydrite), Ba (bassanite), C (calcite), Cr

xx (cristobalite), Fe (iron), Gy (gypsum), G (gehlenite), Hc (hydrocalumite), Hcr (hydroxylcancrinite), He (hematite), Hl (halite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Us (ulvöspinel) and Tb (tobermorite- 11Å). Figure 5.16a XRD patterns of fly ash samples (FA-A2, FA-A3, FA-A4 and FA- 165 B1) hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h. The peaks are labelled Ac (analcime), C (calcite), Fe (iron), G (gehlenite), Gy (gypsum), Hcr (hydroxylcancrinite), He (hematite), Hl (halite), Hs (metahalloysite), Ilt (illite), Kt (katoite), Lp (lepidocrocite), M (magnetite), Po (portlandite), Tb (tobermorite- 11Å) and V (vuagnatite). Figure 5.16b Minerals produced from hydrothermal conversion of fly ash 165 samples (FA-A2, FA-A3, FA-A4 and FA-B1) at 180 OC in 0.5 M NaOH for 48 h. Note H-Cancrinite : hydroxylcancrinite. Figure 5.17a XRD patterns of WFA-A1 hydrothermally treated at various 167 temperatures in 0.5 M NaOH for 48 h. The peaks are labelled Ac (analcime), An (anhydrite), C (calcite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), Hcr (hydroxylcancrinite), He (hematite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite), Q (quartz), Sd (sodalite), Tb (tobermorite- 11Å) and Us (ulvöspinel). Figure 5.18a XRD patterns of washed fly ash samples (WFA-An and WFA-Bn) 169 hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h. The peaks are labelled Ac (analcime), Ab (albite), C (calcite), G (gehlenite), Gy (gypsum), Hcr (hydroxylcancrinite), Ilt (illite), Kt (katoite), Lp (lepidocrocite), M (magnetite), Po (portlandite), Tb (tobermorite-11Å) and Us (ulvöspinel). Figure 5.18b Minerals produced from hydrothermal conversion of washed fly 169 ash samples (WFA-An and WFA-Bn) at 180 OC in 0.5 M NaOH for 48 h. Note H-Cancrinite : hydroxylcancrinite. Figure 5.19 XRD patterns of (a) leached specimens of fly ash (FA-An and FA- 172 Bn), and (b) unleached and leached FA-A1 specimens. The peaks are labelled Al (alunite), An (anhydrite), Bo (boehmite), C (calcite), Ct (caracolite), Fe (iron), G (gehlenite), Go (goethite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), M (magnetite), Ms (monosulfate), Q(quartz), R (rutile), S (sylvite), Tb (tobermorite-14Å), Us (ulvöspinel) and Zc (ZnCl 2). Figure 5.20 XRD patterns of (a) solid residues leached from solidified 174 products (HFA-An and HFA-Bn) of fly ash samples at an age of 28 days, and (b) unleached and leached HFA-A1 specimens. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), Hz (hydrozincite), Hl (halite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite), Sd (sodalite), Tb (tobermorite-14Å), Th (thenardite) and Us (ulvöspinel). Figure 5.21 XRD patterns of (a) solid residues leached from the hydrothermal 176 products (HTFA-An and HTFA-Bn ) of fly ash samples treated at

xxi 180 OC in 0.5 M NaOH for 48 h; (b) unleached and leached HTFA-B1 samples. The peaks are labelled An (anhydrite), Ac (analcime), C (calcite), G (gehlenite), Gr (grossular), Gy (gypsum), Hcr (hydroxylcancrinite), Hl (halite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite) and Tb (tobermorite- 11Å). Figure 5.22 XRD patterns of (a) solid residues leached from washed fly ash 181 samples (WFA-An and WFA-Bn ), and (b) unleached and leached WFA-A1 samples. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), M (magnetite), Pa (potassium alum), Po (portlandite), Q(quartz), R (rutile), Tb (tobermorite-14 Å-CSH phase) and Us (ulvöspinel). Figure 5.23 XRD patterns of (a) solid residues leached from solidified 182 products (HWFA-An and HWFA-Bn ) of washed fly ash samples at an age of 28 days, and (b) unleached and leached HWFA-A1 samples. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), Pa (potassium alum), Po (portlandite), Q(quartz),R (rutile) and Us (ulvöspinel). Figure 5.24 XRD patterns of (a) solid residues leached from hydrothermal 183 products (HTWFA-An and HTWFA-Bn ) of washed fly ash samples treated at 180 OC in 0.5 M NaOH for 48 h, and (b) the unleached and leached HTWFA-A3 samples. The peaks are labelled Ab (albite), Ac (analcime), C (calcite), G (gehlenite), Gy (gypsum), Hcr (hydroxylcancrinite), He (hematite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite), Tb (tobermorite- 11Å) and Us (ulvöspinel).

xxii LIST OF ABBREVIATIONS AND ACRONYMS Abbreviations and acronyms ANC Acid neutralisation capacity APC Air pollution control BA-A1 Bottom ash collected from incinerator plant A (MHKW Iserlohn) BA-A2 Bottom ash collected from incinerator plant A (MHKW Iserlohn) BA-B1 Bottom ash collected from incinerator plant B (MHKW Essen) BNC Base neutralisation capacity BSE Backscattered electron CEC Cation exchange capacity C-S-H Calcium silicate hydrates EDX Energy dispersive spectroscopy EDTA Ethylenediaminetetraacetic disodium salt EPA Environmental Protection Agency EPMA Electron probe x-ray microanalysis ESP Electrostatic precipitator FA-A1 Fly ash collected from incinerator plant A (MHKW Iserlohn) FA-A2 Fly ash collected from incinerator plant A (MHKW Iserlohn) FA-A3 Fly ash collected from incinerator plant A (MHKW Iserlohn) FA-A4 Fly ash collected from incinerator plant A (MHKW Iserlohn) FA-B1 Fly ash collected from incinerator plant B (MHKW Essen) HFA-A1 Solidified product of FA-A1 HFA-A2 Solidified product of FA-A2 HFA-A3 Solidified product of FA-A3 HFA-A4 Solidified product of FA-A4 HFA-B1 Solidified product of FA-B1 HTFA-A1 Hydrothermal product of FA-A1 HTFA-A2 Hydrothermal product of FA-A2 HTFA-A3 Hydrothermal product of FA-A3 HTFA-A4 Hydrothermal product of FA-A4 HTFA-B1 Hydrothermal product of FA-B1 HWFA-A1 Solidified product of washed FA-A1

xxiii HWFA-A2 Solidified product of washed FA-A2 HWFA-A3 Solidified product of washed FA-A3 HWFA-A4 Solidified product of washed FA-A4 HWFA-B1 Solidified product of washed FA-B1 HTWFA-A1 Hydrothermal product of washed FA-A1 HTWFA-A2 Hydrothermal product of washed FA-A2 HTWFA-A3 Hydrothermal product of washed FA-A3 HTWFA-A4 Hydrothermal product of washed FA-A4 HTWFA-B1 Hydrothermal product of washed FA-B1 ICCD International Centre for Diffraction Data ICP Inductively coupled plasma spectroscopy ICSD Inorganic Crystal Structure Database LAGA Länderarbeitsgemeinschaft Abfall L/S Liquid to solid MHKW Müllheizkraftwerk MSW Municipal solid waste MSWI Municipal solid waste incinerator NIST National Institute of Standards and Technology p.a Pro analyse PCDD/F Polychlorinated dibenzo-p-dioxins and furans PDF Powder Diffraction File POP Persistent organic pollutants SEM Scanning electron microscopy SRM Standard Reference Material TCLP Toxicity characteristic leaching procedure TEQ Toxic equivalent concentration WDX Wavelength dispersive spectrometry WFA-A1 Washed FA-A1 WFA-A2 Washed FA-A2 WFA-A3 Washed FA-A3 WFA-A4 Washed FA-A4 WFA-B1 Washed FA-B1

xxiv w/c Water to cement XRD X-ray diffraction XRF X-ray fluorescence

xxv CHAPTER 1

CHAPTER 1 INTRODUCTION

1.1 Background to Study Municipal solid waste incineration (MSWI) is a commonly accepted solution for managing an increasing production of municipal solid waste (MSW). The incineration has become a more attractive technique for treating non-recycled wastes including plastics, compared to other waste treatments such as mechanical/biological processes and direct disposal option at landfills. The method becomes a common management strategy for treating MSW, because of providing a great reduction of volume, weight and hazards of the disposed waste and therefore contributes to prolonging lifetime of the disposal yard (Hjelmar, 1996; Li et al ., 2003; Sabbas et al ., 2003). Additionally, it presents the possible advantage to recover some of energy bound in the waste for generating electricity and steam. However, the incineration generates (i) a large quantity of bottom ash consisting of a slag-type material, and (ii) air pollution control (APC) residues including fly ash and acid gas scrubbing residues containing significant amounts of heavy metals (Cd, Zn, Pb, Hg, Cu, Cr and Ni) as well as trace amounts of organic pollutants (polychlorinated dibenso-p-dioxins and furans-PCDD/F) (Eighmy et al ., 1995; Kirby and Rimstidt, 1993). The quantity of fly ash and the potential leachability of heavy metals and toxic organic components have raised concern in many countries regarding to environmental acceptability as a landfill deposit or as a resource (Akiko et al ., 1996; Forestier, and Libourel, 1998; Ferreira et al ., 2003; Rincon, et al ., 1999). In Germany, MSWI fly ash is regarded as a hazardous waste which must be disposed of in a safe manner such as underground landfill (TA-Siedlungsabfall, 1993; LAGA, 1994). Hence the management of fly ash is one of the important environmental issues related to the incineration of MSW (Ferreira et al ., 2003). Many treatment technologies applicable to MSWI fly ash have been recently proposed (Erol et al. , 2001; Park and Heo, 2002a; Park and Heo, 2002b; S Ørensen et al. , 2000). Various options such as thermal treatment, physical/chemical separation, and stabilisation/solidification (S/S) techniques are now available for treatments of fly ash in view of their reuse or final disposal. The use of thermal treatment via vitrification has been tested and practised on a small scale. However, the method is considered more costly than other solutions because it requires a significant amount of energy and an

1 CHAPTER 1

expensive capital apparatus. It could also lead to a subsequent environmental destructive impact as a result of volatilisation of heavy metals from fly ash during vitrifying and melting processes. As an alternative to the thermal treatment, physical/chemical separation methods employed for extracting heavy metals (Cd, Pb and Zn) from fly ash have been attempted but with limited success (Hong et al. , 2000a; Hong et al. , 2000b; Katsuura et al., 1996). More recently, the stabilisation and solidification (S/S) methods have emerged as a viable alternative for MSWI fly ash treatment by solidification with Portland cement into stable complexes. In this context, toxic components are incorporated in a cement matrix through either physical or chemical immobilisation mechanisms, depending on the particular contaminant to be fixed and type of binder being used. This technique has long been known as an economical option for the waste management strategy (Glasser, 1997). However, the solidification with Portland cement presents some disadvantages specifically protection against humidity is required to prevent breakdown and leaching of heavy metals (Nzihou and Sharrock, 2002). In addition, salts present in fly ash will interfere with basic hydration reactions of cement, leading to an inadequate set and/or deterioration of the waste form over time.

Pozzolanic solidification of MSWI fly ash in a solution enriched with Ca(OH) 2 in order to take an advantage of its pozzolanic property has yet to be fully realised (Rémond et al ., 2002; Targan et al ., 2003; Ubbriaco and Calabrese, 2000; Ubbriaco et al ., 2001). An interesting potential for generating a pozzolanic property of MSWI fly ash is that ettringite and calcium-silicate-hydrate (C-S-H) may be formed using an activator of Ca(OH) 2 giving a good degree of stabilisation of the hardened material. This stabilised material becomes, in principle, more compatible for storage, landfill or reuse as a resource. A promising new approach to the utilisation of MSWI fly ash is to convert it into zeolite-like materials and neomorphic phases ( e.g ., tobermorite-11Å and katoite) (Yang and Yang, 1998; Miyake et al. , 2002). Conversion of fly ash into zeolite materials by a hydrothermal method may be used as a cost-effective option to immobilise heavy metals (Maenami et al. , 2000). The hydrothermal process has been employed extensively for treating coal fly ash (Querol et al ., 2002) and may be also beneficially applied for MSWI fly ash, in which SiO 2 and Al 2O3-bearing phases are abundant. For

2 CHAPTER 1

that, the efficiency of hydrothermal conversion depends upon the mass ratio of the concentrations of SiO 2 and Al 2O3. Coal fly ashes with similar bulk SiO 2/Al 2O3 ratios have been treated to produce different zeolites; this is attributed to differences in the composition of glassy matrix (Querol et al ., 2002). A significant amount of alumino- silicate glassy matrix in MSWI fly ash has made synthesis of zeolites attractive both technologically and commercially (Maenami et al. , 2000; Miyake et al. , 2002). Despite many recent investigations, technology and developing efforts on MSWI fly ash treatments, a viable ash management technology is in high demand by the industry and regulators to provide environmentally and economically sound solutions. Significant efforts have been devoted to the development of cost effective treatments of MSWI fly ash through pozzolanic solidification (Baur et al ., 2001; Ubbriaco et al ., 2001) and hydrothermal conversion (Yang and Yang, 1998; Miyake et al. , 2002). However, many fundamental aspects of conditional treatments and fly ash characteristics, which are essential to provide a scientific basis and establish confidence in the technology, remain poorly understood and lack quantification. The mineralogy of fly ashes and their final products have not been studied in-depth. A detailed knowledge of the minerals present, and their chemical compositions is necessary in order to be able to assess quantitatively the correlation between microstructure and leaching behaviour of the materials and thus, to optimise the stabilised products (Forestier and Libourel, 1998; Thipse et al ., 2002; Boccaccini et al ., 1997). The mineralogy of MSWI residues (bottom ash and fly ash) represents an assemblage of crystalline and amorphous phases. As a result of high temperature combustion, they are unstable under natural condition and susceptible to mineralogical changes in a disposal environment on contact with leachate and groundwater. However, such changes may also be beneficial, leading to significant reductions in trace elements for leaching (Meima and Comans, 1999; Zevenbergen and Comans, 1994; Zevenbergen et al ., 1996). Typically, aging and weathering processes may promote chemical reactions, which are responsible for immobilising heavy metals within the waste matrix. These chemical reactions may include hydration, carbonation, and oxidation/reduction processes, which in the landfill lead to more stable phase assemblages (Zevenbergen and Comans, 1994). These processes may cause a decrease in pH in addition to the sorption of contaminants onto the surface of the material and therefore lead to the

3 CHAPTER 1

reduction of contaminant leaching from the residues (Meima and Comans, 1997; Meima and Comans, 1999; Sabbas et al ., 2003). Accordingly, it is postulated that aging and weathering processes applied to promote a mineralogical change may cause MSWI residues to be less susceptible to leaching when disposed or reused as a resource. Research on the mineralogical alteration induced by aging and weathering processes has been hampered by the fact that the mineralogy of MSWI residues is very complex and thus cannot be easily evaluated quantitatively. Consequently, there are only relatively few studies that examined the mineralogical alteration using a mineralogical approach that relies on XRD methods to evaluate changes in the phase abundance. Use of a mineralogical approach may gain an insight into changes of amorphous and crystalline phases, which control to the alteration process. Based on the mineralogical analysis, development of stable mineral phases in order to provide effective and safe solutions for the management of MSWI fly ash could be established at laboratory level with the assurance of technical and environmental acceptability (Kirby and Rimstidt, 1993; Zevenbergen and Comans, 1994).

1.2 Characteristics of MSWI Residues The MSWI residues are a by-product generated during an incineration of MSW at an incinerator plant. The waste residues produced in most modern incinerators consist of: (i) bottom ash, (ii) grate siftings, (iii) boiler and economiser ash, (iv) fly ash, and (v) air pollution control (APC) residue (Hjelmar, 1996; Sabbas et al. , 2003). Bottom ash is generally collected at the base of the combustion chamber. Immediately after incineration, the hot bottom ash and grate siftings are quenched in a container of water. The generated bottom ash consists of relatively inhomogeneous large particles (<45mm), fused lumps of slag and pieces of scrap metal (Hjelmar, 1996; Sabbas et al. , 2003). Amounts of bottom ash, including the grate siftings, generated by waste incineration, are in the range of 30-35 % of burnt waste by weight. Prior to deposit or reuse, bottom ash is treated by removal of large pieces of unburned material and magnetic particles, and followed by crushing for particle size reduction (Meima and Comans, 1997). Bottom ash is not classified as a hazardous waste according to the European Waste Catalogue (E.U, 2000). However, amounts of bottom ash may require a high

4 CHAPTER 1

expenditure for their disposal, thereby being explored frequently for constructional applications. In Germany, about 50 % or more of bottom ash are currently employed as secondary building materials or other civil engineering applications (Reimann, 1994). Further, boiler and economiser ash are composed of the coarse particulate matter, which is present in the flue gas from the combustion chamber, cooled in a heat exchanger and finally collected at the heat recovery section. Moreover, fly ash represents the finer fraction of the particulate (< 200 µm) which is entrained in the flue gas and subsequently trapped in an electrostatic precipitator or fabric filter and later removed before any further treatment of the gaseous effluents. Specifically, the APC residue is obtained after reagent injection ( e.g. , lime and water) in the acid gas treatment facilities prior effluent gas discharge into the atmosphere. This resulting residue may be in the form of solid, liquid or sludge, depending upon whether dry, semi-dry or wet processes are employed for air pollution control. Depending upon types of solid waste, combustion unit and air pollution-control (APC) device, MSWI residues have a heterogeneous composition with varying degree physical nature and chemical compositions (Eighmy et al ., 1995; Kirby and Rimstidt, 1993; Meima and Comans, 1997; Song et al ., 2004). Miguel et al . (1992) classified chemical elements in the incineration residues as follows: (i) elements with high boiling point that are not volatilised in the combustion area and are main components of bottom ash, (ii) elements volatilised during combustion, which remain in small amounts in bottom ash, while the volatilised matter may condense on the surface of fly ash particles, as the combustion gas stream cools down, (iii) elements that undergo volatilisation but not condensation, thereby remaining in the gas phase throughout the entire process. The elements with high boiling point, including Si 4+ , Al 3+ , Fe 3+ and Ca 2+ may constitute for over 80% of the weight of bottom ash (Hjelmar, 1996; Sabbas et al. , 2003). However, low volatile inorganic contaminants in the waste feed are mainly found in bottom ash (Kirby and Rimstidt, 1993; Eighmy et al ., 1995). Bottom ash may be similar to igneous rocks, granite and basalt as well as natural glasses in chemical composition and geochemical properties (Belevi et al. , 1992; Kirby and Rimstidt, 1993; Zevenbergen and Comans, 1994). In contrast, fly ash comprises highly volatile elements - 6+ 4+ 3+ + 2+ such as Cl and S in addition to some constituents like Si , Al , K and Ca .

5 CHAPTER 1

Subsequently, fly ash and the APC residue contain high concentrations of heavy metals and organic micro-pollutants as a result of volatilisation and subsequent condensation of particles as well as concentration phenomena occurring during combustion. Fly ash is rich in some elements (Si 4+ , Al 3+ and Fe 2+, 3+ ) and compounds (metals and salts) and therefore has some potential to be used as a raw material. Hence, each potential application for fly ash presents advantages; (i) the use of a zero-cost raw material, (ii) the conservation of natural resources, and (iii) the reduction of likelihood of fly ash disposal (Ferreira et al. , 2003). Fly ash has a higher fraction of water-soluble compounds than bottom ash. This factor is considered an important aspect for treatment, disposal and possible utilisation of MSWI residues (Hjelmar, 1996; Sabbas et al. , 2003). As demonstrated previously, fly ash contains more leachable heavy metals than bottom ash (Hjelmar, 1996; Sabbas et al. , 2003). Leaching properties of heavy metals are internally affected by composition and chemical states of the metals themselves and external factors such as pH, ionic strength, and oxidation/reduction potential of the contact solvent or liquid/solid ratio (Akiko et al ., 1996; Johnson et al ., 1996).

1.3 Treatment Strategy of MSWI Residues Bottom ash generally satisfies all present requirements for utilisation or disposal (Reimann, 1994; Sabbas et al. , 2003). It has been thus possible to directly reuse bottom ash as a resource or deposit without significant treatment. However, bottom ash is often aged by storing for a period of several weeks up to a few months. This aging process is necessary to assure the specific volume to be constant and to reduce leachates. An additional treatment of bottom ash may include washing with water in order to remove salts and partially sulphates (Sabbas et al. , 2003). In contrast to bottom ash, fly ash contains basically two groups of hazardous compounds: (i) persistent organic pollutants (POP), and (ii) heavy metals. The POP may consist of polychlorinated dibenzo-p-dioxins and furans (PCDD/F), chlorobenzenes and chloronaphthalenes. These organic pollutants could be normally destroyed by a thermo- catalytic process with significant reduction of the toxic equivalent concentration (TEQ) by more than 99% (Stuetzle et al ., 1991). However, the heavy metals could not be destroyed and thus require a subsequent treatment process. In the present study, MSWI

6 CHAPTER 1

fly ash as a hazardous material is selected to be stabilised in order to meet the environmental regulation. Principally, three processes can be employed to treat fly ash (Sabbas et al ., 2003). These are (i) physical and chemical separation, (ii) thermal treatments, and (iii) solidification/stabilisation techniques.

Physical and chemical separation Physical and chemical separation methods are frequently employed for treating a heterogeneous material such as MSWI fly ash. The objectives of the methods are (i) to isolate the finer fraction in liquid suspensions, which are more concentrated in contaminants, and (ii) to improve engineering properties such as a particle gradation and hydraulic conductivity that are more suitable for a subsequent utilisation. The use of an electromagnet for separating magnetic particles of bottom ash and fly ash is also common practice (Kirby and Rimstidt, 1993; Sabbas et al ., 2003). Washing with water can be also applied as a pre-treatment stage of MSWI fly ash prior to further chemical stabilisation processes. The method can simply remove highly water-soluble salts from fly ash (Schneider et al. , 1994; Sabbas et al ., 2003). Such salts may be responsible for much of the negative effects on properties of fly ash such as high leachability, high water absorption and corrosiveness (Derie, 1996; Nzihou and Sharrock, 2002). Further treatment of fly ash may include a chemical extraction and chemical precipitation by an ion exchange process. Hazardous components of fly ash may be extracted by an addition of sodium carbonate or sodium bicarbonate for either mobilisation of sulphate through the formation of soluble Na 2SO 4 or precipitation of

CaCO 3. The use of inorganic acids such as hydrochloric, nitric or sulphuric acid and aqua regia for recovering heavy metals and detoxifying fly ash has been also reported (Hong et al. , 2000a; Hong et al. , 2000b; Katsuura et al., 1996). The important factors influencing the efficiency of heavy metal extraction are pH, liquid to solid (L/S) ratio, nature of the employed extracting agent and the particular metal concerned.

7 CHAPTER 1

Thermal treatment Thermal treatments of MSWI fly ash are widely employed to reduce leachability of hazardous components. This method involves basic processes such as vitrification, melting and sintering (Sabbas et al ., 2003). The benefits of thermal treatment are now being explored for stabilising complex materials such as MSWI fly ash and allowing a large volume reduction of the material for long-term storage or depositing at landfills (Barbieri et al. , 2000; Scarinci et al. , 2000). The thermal treatment of the incinerator ash in some cases yields dense, homogeneous products with less porosity as well as improved mechanical properties. The vitrification of fly ash results in a product, which has potential to be used as raw material for a glass-ceramic manufacture (Boccaccini et al. , 1995; Cheng et al ., 2002). However, as major drawback the method is energy- intensive. The vitrification and melting of fly ash with high salt contents may be accompanied by volatilisation of metals, thereby promoting air pollution.

Solidification/ Stabilisation Solidification/stabilisation (S/S) treatments of MSWI residues, in particular fly ash, have been extensively employed in some countries to obtain reduced heavy metal leaching from the residues (Sabbas et al ., 2003). These methods may yield the treated fly ash with improved physical, mechanical and chemical properties, and eventually lead to decrease the leachability of contaminants out of the waste matrix (Gilliam and Wiles, 1996). The common practice makes use of solidification of fly ash in cement, lime and/or pozzolanic materials. The purposes of the method are (i) to prevent dust- borne contamination, and (ii) to reduce the mobility of heavy metals at a relatively low cost, thereby facilitating landfilling. However, the method may contribute to increase amounts of waste to be landfilled. Additionally, the existence of soluble salts within fly ash makes the stabilisation process ineffective. Likewise, the stabilisation of zinc and lead by cement-and lime-based processes may cause serious problems of releasing compounds to the environment, due to the strong amphoteric behaviour of zinc and lead. Hence, an effective cement-based stabilisation could be achieved if heavy metal ions could be incorporated in the crystal lattice of the hydration products or an appropriate additive is employed (Sabbas et al ., 2003).

8 CHAPTER 1

The use of a mineral additive for exploring the pozzolanic behaviour of MSWI fly ash has been reported by a number of workers (Rémond et al ., 2002; Targan et al ., 2003; Ubbriaco and Calabrese, 2000; Ubbriaco et al ., 2001). Fly ash is pozzolan, which is a siliceous or siliceous-aluminous material that become cementitious if combined with an activator ( lime, or Portland cement ) in the presence of water. It has also been recycled as an engineering material for many years because of its pozzolanic characteristic (Hamernik and Frantz, 1991). The pozzolanic solidification of fly ash yields a cementitious material with improved mechanical strength and chemical stability (Mangialardi et al ., 1999; Ubbriaco et al ., 2001). Treatment of MSWI fly ash may be also performed by chemical stabilisation processes including evaporation, precipitation, ion exchange, or adsorption mechanisms on various mineral species. Common chemical agents employed for chemical stabilisation processes include sulphides, soluble phosphates, ferrous iron sulphate and carbonates (Derie, 1996; Katsuura et al. , 1996; Hjelmar et al. , 1999). Aging and weathering of MSWI residues can also promote chemical reactions responsible for the fixation of heavy metals within the waste matrix (Meima and Comans, 1997; Sabbas et al ., 2003). The chemical reactions may involve hydration, carbonation or oxidation/reduction. As demonstrated by previous experiments, the alteration of mineralogical phases in the residues during aging is favourable for reducing the leaching of trace elements such as Cd, Cu, Pb, Zn and Mo (Meima and Comans, 1999; Zevenbergen and Comans, 1994). One possible application for MSWI fly ash is to convert it into a zeolite-like material by hydrothermal treatment at low temperatures (below 200 OC). Yang and Yang (1998) and Miyake et al. (2002) demonstrated the synthesis of zeolite-like materials from MSWI fly ash by a hydrothermal alkaline processing, similar to that reported by Querol et al . (1997) for coal fly ash. MSWI fly ash has a high specific surface and a high content of aluminosilicate glass (Ferreira et al ., 2003), which are important factors in the synthesis of zeolites, although this glass may be less quantity in MSWI fly ash than in coal fly ash. In the present study, pozzolanic solidification and hydrothermal conversion were examined as environmentally compatible stabilisation processes of MSWI fly ash. Further details of the stabilisation treatments of MSWI fly ash by pozzolanic solidification and hydrothermal conversion are provided in chapter 2.

9 CHAPTER 1

1.4 Objectives The major objective of this research study was to propose a MSWI fly ash treatment strategy by converting the ash into stable mineral phases with particular reference to immobilising heavy metals. The research involved a systematic characterisation of chemistry and mineralogy of MSWI residues, treatment experiments and evaluation of the leaching behaviour. The objectives were as follows: (i) To perform a comprehensive study of MSWI residues (bottom ash and fly ash) collected from two different MSWI plants located in Germany with the intention of exploring the chemical and mineralogical properties. (ii) To investigate mineralogical alterations of bottom ash and fly ash associated with aging and weathering processes. (iii) To apply water-extraction methods as a pre-treatment stage for fly ash by removing soluble salts, and recovering some elements and compounds (salts). (iv) To apply pozzolanic solidification and hydrothermal method for converting raw and washed fly ashes into stable mineral phases with improved leaching resistance. (v) To investigate mineral stability and leaching behaviour for raw and washed fly ashes as well as the stabilised products.

1.5 Research Plan The proposed research plan was to characterise MSWI residues, to conduct pre- treatment of fly ash, synthesise stable mineral phases from fly ash and examine leaching behaviour. This would involve: (i) Studying the characteristics of MSWI residues (bottom ash and fly ash) locally sourced from two incinerator facilities in Ruhr-industrial area, Germany and to utilise this fly ash to develop stable mineral phases. It was of special interest to investigate the mineralogical alteration of the residues by XRD. XRD Rietveld analysis has been employed in this study to gain an insight into mineralogical phase compositions in the residues. (ii) Performing mineral fractionation procedures on MSWI residues using a magnetic separator, water-washing and Soxhlet water-extractor procedure as means of improving mineral identification. Specifically, the water washing and Soxhlet

10 CHAPTER 1

extractor have been used in this study to remove soluble salts from fly ash. It was envisaged that the Soxhlet water-extraction would help in the understanding of extraction of some elements and compounds from fly ash. (iii) Developing stable mineral phases from raw and washed fly ashes with regard to incorporated heavy metals in the waste matrix through controlled pozzolanic

reaction of the saturated Ca(OH)2 solution, and hydrothermal treatment in alkali (NaOH and KOH) solutions. (iv) Carrying out leaching tests for raw and washed fly ashes as well as the selected stabilised products according to the TCLP (toxicity characteristic leaching procedure) method (U.S. EPA, 1990).

1.6 Structure of Doctoral Thesis Chapter 1 outlines the necessity of studies on the characteristics of MSWI residues and the potential treatment methods for the immobilisation of hazardous components in MSWI fly ash with particular reference to pozzolanic solidification and hydrothermal conversion. Different treatment methods of MSWI residues available for immobilisation of toxic constituents are also reviewed. Chapter 2 reviews the literature on the immobilisation strategy of heavy metals in MSWI fly ash in the context of fly ash processing for the development of stable mineral phases by pozzolanic solidification and hydrothermal conversion. Chemical and mineralogical characteristics including leaching mechanism of MSWI residues are also discussed, followed by an evaluation of leaching properties according to the TCLP method. Chapter 3 is devoted to experimental design and to treatment strategies particularly the state of the art in the ash processing, characterisation and evaluation of leaching properties. Chapter 4 provides a summary of results on the characterisation of bulk chemical and mineralogical properties of MSWI residues, microanalysis of individual particles of the residues, mineralogical alteration, and water-extraction processes of fly ash. Chapter 5 presents the results of pozzolanic solidification and hydrothermal conversion of raw and washed fly ashes, evaluations of mineral stability and leaching properties using the TCLP method. It is shown that the stabilised materials produced by

11 CHAPTER 1

pozzolanic activation and hydrothermal methods exhibit improved leaching resistance compared to the parent materials. Chapter 6 provides concluding remarks on the analysis of bulk chemical and mineralogical compositions of MSWI residues, microanalysis of individual particles of the residues, examination of mineralogical phase evolution and water-extraction processes of fly ash. The main conclusions on the synthesis of stable mineral phases and evaluation of leaching behaviour of the raw and washed fly ashes as well as the stabilised products are also presented. Further work and suggestions of the research concerning characterising MSWI residues, pozzolanic and hydrothermal activation methods of fly ash, and leaching property evaluation are discussed.

12 CHAPTER 2

CHAPTER 2 STABILISATION TREATMENTS OF MSWI FLY ASH

This chapter addresses stabilisation treatments of MSWI Fly ash with particular reference to improved leaching resistance. The principal issues considered are characteristics of MSWI residues, synthesis strategies for cementitious and zeolite like- materials, immobilisation mechanisms of heavy metals and the evaluation of leaching properties.

2.1 Stabilisation of MSWI Fly Ash Cement-based technique The stabilisation process of MSWI fly ash by converting it into a relatively stable form of a cement-like material is one of the promising solutions for the fly ash management strategy. This method can promote the reusability of fly ash by providing a possible zero-cost raw material. It also offers the fly ash treatment at a relatively low cost, because of employing simple and relatively inexpensive equipment (Hjelmar et al ., 1999; Ferreira et al ., 2003; Sabbas et al ., 2003). The common strategy for the solidification and stabilisation of fly ash in a cement- based system has been performed by: (i) an addition of alkali matrices such as

Ca(OH) 2 or Portland cement in order to develop a pozzolanic material, and (ii) entrapping or encapsulating heavy metals substances (Pb 2+ , Zn 2+ , Cu 2+ and Cd 2+ ) within the solidified waste matrix and thereby reducing their mobility of heavy metals (Baur et al ., 2001; Sabbas et al ., 2003). The conversion of fly ash into the cementitious material is based on the fact that this material contains components such as SiO 2, Al 2O3, CaO, and Fe2O3, which are required for the synthesis of cement-like materials, although these appear in less quantity in fly ash than Portland cement. The fly ash reacted with an activator of

Ca(OH) 2 potentially exhibits cementitous characteristics (Glasser, 1997; Hamernik and Frantz, 1991). Here water is required to initiate the pozzolanic solidification reaction, which is mainly controlled by the mass ratio of water to fly ash referred to liquid to solid (L/S) ratio.

13 CHAPTER 2

Further pozzolanic reaction between the silica component of fly ash and Ca(OH) 2 in the solution produces a calcium silicate hydrate (C-S-H) phase. The composition of the C-S-H phase depends on the molar ratio of Ca/Si (from 1.8 to 2.0), while the structure of the C-S-H phase can be amorphous or fully crystalline (Taylor, 1990). In addition, ettringite and syngenite are frequently formed during the pozzolanic solidification process. The C-S-H phase has long been recognised for sorption and substitution of heavy metals from different types of waste materials including MSWI fly ash, because of its high specific surface area and variation in structures and compositions (Ma and Brown, 1997; Pomiès et al. , 2001a; Pomiès et al. , 2001b; Baur et al ., 2001). Consequently, heavy metal species may be dissolved, adsorbed into a glassy matrix, or be incorporated in the newly form of oxide, hydroxide and carbonate minerals. Further, important parameters for cement-like materials by ion-substituted mechanism are dissolution rates. The dissolution rates for several minerals including the calcium silicate hydrates (C-S-H) phase, feldspars, and clays have been reported in the magnitudes of 10 -10 to 10 - 13 mol m -2 s -1 at pH values of 11-12 (White and Brantley, 1995).

Hydrothermal processing An alternative approach for the stabilisation of MSWI fly ash is its conversion into zeolite-like materials through hydrothermal reaction very similar to that reported for coal fly ash (Yang and Yang, 1998; Miyake et al. , 2002; Querol et al ., 1997; Querol et al ., 2002). The use of fly ash for synthesis of zeolites is desirable in view of the possibilities of making beneficial products and immobilising of hazardous components. It has been demonstrated experimentally (Miyake et al ., 2002) that MSWI fly ash can be successfully converted into zeolite-like materials such as sodalite and zeolite A or P in addition to neomorphic phases such as tobermorite and katoite. Zeolites are microporous, crystalline aluminosilicates and consist of three components: tetrahedral framework, cations, and adsorbed molecules (Figure 2.1). Their structural models may be classified as: (i) a three-dimensional and regular framework formed by linked TO 4 tetrahedral (T= Si, Al), each oxygen being shared between two T elements, and (ii) channels and cavities with molecular sizes which can host the

14 CHAPTER 2

compensating cations, water or other molecules and salts (Vaughan, 1978). Here the structural formula of a zeolite is given by (Breck, 1984); ( )⋅ M n/x Al xSi y O2 ()x+y p H 2O (2.1) where M indicates the mono-valent cations (Na, K, Li) and/or the divalent cations (Ca, Mg, Ba, Sr), n is the cation charge: y/x is 1 to 6; p/x is 1 to 4 (Querol et al ., 1997; Querol et al ., 2002). Further, the zeolite structure may be broadly classified according to the type of cations. The exchangeable cations ( e g., Na +) would be easily substituted with other ions depending upon the presence of more electropositive cations in the solution for reaction of zeolite.

Figure 2.1 Idealised structure of zeolite framework (after Querol et al ., 2002).

The important feature of zeolites for the immobilisation of heavy metals is due to having a large internal pore volume where the cavities are interconnected by one or more channel systems. This open micro or nano porosity and the stability of the framework of zeolites allow potentially a transfer of matter between the interior and the exterior of crystals. Generally, the crystalline zeolites have unique absorbent properties with large void volumes in the range of 20-50 %. To date, a diversity of zeolite products has been developed with variations of either the starting raw material composition ( such as silica and alumina ) or reaction conditions (Barrer, 1982). It has been shown that kinetic mechanisms of hydrothermal

15 CHAPTER 2

reaction and zeolite compositions referred to the system of CaO-SiO 2-Al 2O3 are controlled by experimental parameters such as temperature, reaction time, liquid/solid mass ratio (L/S), hydroxide concentration, type of hydroxide solution and the intensity of the mixing procedure (Barrer, 1982). The zeolite product can be examined regarding to (i) its adsorption property, and (ii) its leaching behaviour (Steenbruggen and Hollman, 1998). The adsorption property may be explored by batch and column experiments for the selectivity series of elements such as Ba 2+ , Cd 2+ , Co 2+ , Cu 2+ , Ni 2+ , Pb 2+ and Zn 2+ , whilst the leaching behaviour could be investigated by pH titration experiments. The zeolites capacity of absorbing heavy metal elements can be commonly examined in term of a cation exchange capacity (CEC). This important property has provided a basis for the development of zeolitic materials as ion exchanger materials, gas sorption matrices and stereo selective catalysts (Ferreira et al ., 2003).

2.2 Characteristics of MSWI Residues Chemistry and mineralogy of MSWI residues have received much attention in assessing their potential toxicity, because toxicity is related to the polluting element concentration, the speciation of the pollutant element (s) and nature of the host phases (Chimenos et al ., 2000; Eighmy et al ., 1995; Forestier and Libourel, 1998; Miguel et al ., 1992; Sabbas et al ., 2003). For example, hexavalent chromium is more toxic than trivalent chromium, and zinc chloride (ZnCl 2) is very soluble in water, whereas ZnCO 3 is nearly insoluble. Therefore, a detailed knowledge of the chemistry and mineralogy of these residues is necessary for the development of any stabilisation-solidification processes (Hjelmar, 1996; Sabbas et al ., 2003).

Chemistry MSWI residues (bottom ash and fly ash) consist of particles with significant chemical variation. Table 2.1 presents chemical compositions of these residues (Hjelmar, 1996). The major elements of Si, Al, Fe, Ca, Mg, Na, and K are normally found in both bottom ash and fly ash as oxides, sulfates and chlorides. In particular, the elements S, Cl, Pb and Zn are more enriched in fly ash than that in bottom ash. Furthermore, some trace elements ( e.g., Ba and Cr) in both residues are present at same

16 CHAPTER 2

level. Chemical elements of notably Cu are commonly concentrated in bottom ash, while the volatile elements (Cd, Hg, As, Pb and Zn) are enriched in fly ash and APC residues. The chemistry of MSWI residues, as inferred above, could be also understood in view of their characteristics related to the melting and boiling points (Table 2.2). Elements or compounds with low melting and boiling temperatures (below 1000 OC) are volatile during incineration and are subsequently entrapped in the flue gas. Thus they are more concentrated in fly ash than bottom ash (Eighmy et al ., 1995; Song et al ., 2004). In contrast, elements or compounds with high melting and boiling temperatures (above 1000 OC) are less volatile and therefore are largely found in bottom ash. Furthermore, an important consideration in the disposal or reutilization of MSWI residues is water solubility of chemical compounds (Sabbas et al ., 2003; Song et al ., 2004). Only a small mass fraction of bottom ash dissolves in water. Fly ash is generally composed of a much higher proportion of soluble compounds than bottom ash. In fact, 15-25 % of the total mass of fly ash and 30-40 % of the total mass of the dry/semidry acid gas scrubbing APC residues containing salts are readily dissolved in water. The water solubility of an acid gas scrubbing APC residue produced by a wet scrubbing system (a mixture of sludge and fly ash) may achieve to 14 % of the total mass (Hjelmar, 1996; Sabbas et al ., 2003).

Mineralogy The mineralogy of MSWI residues represents an assemblage of crystalline and amorphous phases, while the leaching behaviour of the residues can be related to the mineralogical alterations which naturally occur as a consequence of contact with atmospheric agents. Hence, the extent of release of contaminants, namely heavy metals and salts, may vary significantly over time (Eighmy et al ., 1995; Song et al ., 2004). More attention should be paid to the major crystalline phases controlling leaching chemistry and the host phases of toxic elements. For example, chloride minerals such as halite and sylvite found in fly ash control the leaching of Cl, K and Na, while the Zn leaching may be controlled by the dissolution of K 2ZnCl 4 and ZnCl 2. Anglesite (PbSO 4) and cerussite (PbCO 3) have been also identified as controlling phases in the residues for Pb leaching at pH below 10 (Eighmy et al ., 1995).

17 CHAPTER 2

Table 2.1 Ranges of chemical elements in MSWI residues (Hjelmar, 1996). Element Bottom ash Fly ash Element Bottom ash Fly ash (wt.%) (wt.%) (ppm) (ppm) Si 21-29 9.5-19 Ag 4.1-14 31-95 Al 4.7-7.2 4.9-7.8 As 19- 80 49-320 Fe 2.7-15 1.8-3.5 Ba 900-2700 920-1800 Ca 6.5- 9.7 7.4-13 Be n. d n. d Mg 0.77-1.9 1.1-1.9 Cd 1.4- 40 250- 450 K 0.92- 2.2 2.3- 4.7 Co < 10- 40 29- 69 Na 2.2- 4.1 2.2- 5.7 Cr 230- 600 140-530 Ti 0.32- 0.72 0.75-1.2 Cu 900- 4800 860-1400 S 0.13- 0.8 0.11-0.32 Hg < 0.01-3 0.8-7 Cl 0.12- 0.32 4.5-10.1 Mo 2.5- 40 15- 49 P 0.29-1.3 0.48- 0.96 Ni 60-190 92- 240 Mn 0.07- 0.17 0.08- 0.17 Pb 1300-5400 7400-19000 Sc 0.6- 8 6.1-33 Sn < 100-1300 1400-1900 Sr 170-350 < 80 - 250 V 36- 90 32-150 W < 20-50 n. d Zn 1800- 6200 19000- 41000 n. d = not detected

18 CHAPTER 2

Table 2.2 Physical properties of metals and their compounds (Lide, 1997; Youcai et al. , 2002). Element Melting Boiling Oxides ( OC) Chlorides ( OC) Sulfates ( OC) point point (OC) (OC) Hg -39 357 Decomposable m.p 275, b.p 301 Decomposable at above 400 the m.p Zn 419 907 Volatilisation at m.p283, Decomposable 1800 sublimation under under calcinations calcinations Cu 1083 2595 m.p. 1026 m.p 620, Decomposable at decomposable at 560 993 Pb 327 1744 m.p. 886, m.p.501; b.p.950 m.p. 1170 b.p. 1516 Cd 321 767 Sublimation at m.p.570; b.p. 960 m.p. 1000 900 Ni 1555 2837 m.p1980 m.p. 1001 m.p. 99 Cr 1900 2480 m.p. 2435, b.p m.p.83 Decomposable at 3000 high temperature Al 660.37 2467 m.p.2072; m.p. 190 at 2.5 atm Decomposable at b.p.2980 sublimation at 177.8 the m.p.770 Fe 1535 3000 m.p. 1377, m.p.282; b.p.316 Decomposable at decomposable high temperature at 3410 Si 1410 2355 m.p.1610 m.p.-70; b.p.57.57 b.p.2230 Notes: m.p and b.p = melting and boiling points, respectively.

19 CHAPTER 2

Table 2.3 presents typical mineralogical phases of MSWI residues. Bottom ash is composed of glassy phase, magnetite, ferrite spinel and silicate minerals, in addition to metallic fragments such as silicon, aluminum and iron. These typical minerals are formed by a high temperature incineration process (above 1000 OC) (Kirby and Rimstidt, 1993; Speiser et al ., 2001). In contrast to bottom ash, the mineralogy of fly ash is influenced by several processes such as vaporisation, melting, crystallisation and precipitation during incineration and a subsequent treatment of flue gas (Forestier and Libourel, 1998; Song et al. , 2004). Such different processes generate preferentially fly ash composed of chloride, sulfate and aluminosilicate glass as well as silicate minerals (Eighmy et al ., 1995; Forestier and Libourel, 1998; Fermo et al ., 1999; Fermo et al ., 2000). Importantly, the toxic elements in MSWI fly ash are more concentrated in the glassy phases and magnetic fraction rather than in the crystalline phases (Zevenbergen and Commans, 1994). Therefore, both the amorphous phase and the magnetic fraction are the most important hosts for many toxic elements.

Table 2.3 Mineralogy of MSWI residues. Fly ash Bottom ash

Amorphous phase (CaO-Al 2O3-SiO 2) Amorphous phase (CaO-Al 2O3-SiO 2)

Chlorides (halite, sylvite, K 2ZnCl 4) Spinel (magnetite) Carbonates (calcite) Carbonates (calcite) Sulfates (anhydrite) Sulfates (anhydrite) Phosphates (apatite, whitlockite) Metallic fragments (Fe, Al, Si) Silicates (pyroxene, plagioclase, gehlenite) Silicates (pyroxene, plagioclase, gehlenite) Oxides (quartz , rutile) Oxides (quartz, rutile) Notes : Data sources for fly ash (Eighmy et al ., 1995; Forestier and Libourel, 1998; Piantone et al ., 2003) and for MSWI bottom ash (Kirby and Rimstidt, 1993; Speiser et al ., 2001).

Several studies (Speiser et al ., 2000; Sabbas et al ., 2003) demonstrated that a number of weathering reactions can promote mineralogical changes of MSWI residues, when they are exposed to atmospheric conditions. Weathering is a natural process as a consequence of several factors such as pH, redox potential, temperature and humidity conditions as well as the concentration of certain compounds ( e.g ., CO 2) in the environment. Such processes lead to a slow mineralogical change over time. The

20 CHAPTER 2

mineralogical alterations and formation of new phases eventually impacts on the leaching behaviour of the residues (Sabbas et al ., 2003). The modified leaching behaviour is the result of several interrelated processes, including hydrolysis, pozzolanic solidification, dissolution/precipitation, carbonation, oxidation/reduction, complexation, sorption, formation of solid solutions as well as mineral neo-formation (Meima and Comans, 1999; Zevenbergen and Comans, 1994; Sabbas et al ., 2003). However, there is disagreement with regard to forms and mechanisms in which the alterations occur in the residues. Further, pozzolanic solidification and hydrothermal conversion of MSWI fly ash can modify the mineralogical compositions which reflect to the relative proportion of the aluminosilicate glassy phase and crystalline phases (Forestier and Libourel, 1998). In the pozzolanic solidification of fly ash, the amorphous phase is the main contributor of silica to form the CSH phase, while anhydrite can be considered an important source of sulfate interacting with aluminates to produce ettringite (Ubbriaco et al ., 2001). Moreover, the important components of fly ash for zeolite synthesis are amorphous aluminosilicate glass, quartz, melilite, and metal oxides ( e.g ., Na 2O, K 2O, CaO). The highest yields of synthesis (> 80 wt.%) correspond to the highest glass content in fly ash. But the lower conversion (yields) of fly ash may be attributed to (i) the larger contents of non-reactive phases ( e.g ., hematite, magnetite), and (ii) the larger contents of resistant alumina-silicate phases ( e.g., melilite, quartz) (Querol et al ., 2002).

Heavy metals High concentrations of heavy metals ( i.e., As, Cd, Hg, Pb and Zn) are potentially found in fly ash and APC residues (Forestier and Libourel, 1998; Sabbas et al ., 2003). Amounts and distributions of metals in fly ash depend on the type of waste incinerated and the incineration process. Furthermore, the effect of incineration temperature and chemical environment in the combustion chamber may produce the partitioning of metals into solid and gas phases (Fermo et al ., 1999). In the combustion zone, metals or metal species behave differently related to their thermal behaviours. Consequently, the metal species entrapped in a glassy phase or adsorbed into fly ash particles depend upon a wide range of processes such as volatilisation, melting, subsequent condensation and

21 CHAPTER 2

precipitation (Eighmy et al ., 1995). Among these processes, the volatilisation and condensation mainly control both concentration and crystal chemistry of pollutants. The distribution of heavy metals in fly ash could be understood in view of their geochemical behaviour; lithophile, siderophile, chalcophile and atmophile (Forestier and Libourel, 1998; Sabbas et al ., 2003). The most lithophilic ( e.g ., Cr, Mn, Na, Ti and V) and siderophilic elements ( e.g ., Co, Mo, Ni, Fe and Sn) are more concentrated in the larger fly ash particles. A number of elements (W, Mo, Zr, Ti, Co, Al, Si, Ba, Sr), which are hardly vaporised during incineration, would preferentially melt and thereby tend to be tightly fixed in phases inherited from MSWI residues (Hjelmar, 1996). In contrast, the chalcophilic elements ( e.g ., As, Cd, Cu, Pb and Zn) are readily vaporised at an intermediate temperature and those become a major part of fly ash particles (Forestier and Libourel, 1998; Belevi and Moench, 2000). The highly volatile element Hg may be emitted almost totally in the vapour phase and therefore may be not found in the fly ash. Finally, the atmophilic elements (C, S, Cl and F) generally remain in the vapour or enter in the minute particles of condensed-phases ( e.g ., Cu, Fe, Sn) after electro filtration.

2.3 Stabilisation of MSWI Fly Ash in a Cement-Based System Cement matrices are widely used for immobilisation of heavy metals through a chemical sorption process. These matrices could be formed by mixing MSWI fly ash and alkaline matrices [Ca(OH) 2 or Portland cement] through pozzolanic reaction (Glasser, 1997; Macphee and Glasser, 1993). A gel-like calcium-silicate-hydrate phase (C-S-H) formed as a result of pozzolanic solidification has potential for sorption of heavy metals. Other solidification products may include calcium sulfoaluminate hydrates, ettringite and monosulfate (Glasser, 1993). Much of the current research in this area is focused on the elucidation of the microstructural evolution of cement pastes containing waste and the invention of a method for quantitative analysis of the hydrated cement pastes. Hence, considerable developments of the microstructural model, in particular for the C-S-H phase have taken place (Taylor, 1950; Powers, 1960; Feldman and Sereda, 1970; Jennings, 2000).

22 CHAPTER 2

2.3.1 Pozzolanic Solidification of MSWI Fly Ash

The value of inducing pozzolanic reaction by mixing fly ash and Ca(OH) 2 or Portland cement in the presence of water has been long recognised in waste management (Glasser, 1977). The pozzolanic solidification of fly ash is generally controlled by the mass ratio of water to solid material (w/c). The w/c ratio is usually adjusted in the range of 0.4 to 1.0 for a flowable mixture, whereas the w/c ratio of 0.25 is considered to be the minimum limit for fluidity. However, a lower w/c ratio is usually required for a low permeability barrier of the hardened product. The pozzolanic reaction principally begins with the formation of calcium hydroxide Ca(OH) 2 as a result of chemical reaction between CaO (lime) and water

(Baur et al ., 2001). Then the active compound of fly ash reacts with Ca(OH) 2 to form pozzolanic reaction products such as calcium silicate hydrate and calcium aluminate hydrate, thereby reducing the content of Ca(OH) 2 in the solution. Consequently, the pozzolanic reaction of fly ash continues until fly ash solidifies. The pozzolanic reactions for the formation of those products may be given as follows (Rémond et al ., 2002): χ ( ) + + χ ⋅ × ( + χ) Ca OH 2 SiO 2 n H 2O ⇒ CaO SiO 2 n H 2O (2.2) χ ( ) + + χ ⋅ ⋅ ( + χ) Ca OH 2 Al 2O3 m H 2O ⇒ CaO Al 2O3 m H2O (2.3) in the formula: χ≤3, ( ) + + + ⋅ ⋅ ⋅ ( + ) 3 Ca OH 2 Al 2O3 2 SiO 2 mH 2O ⇒ 3 CaO Al 2O3 2 SiO 2 m 3 H 2O (2.4) However, the presence of sulfate and chloride minerals in MSWI fly ash may significantly affect the pozzolanic reaction. For instance, gypsum formed from anhydrite can further react with the calcium aluminate hydrate to produce ettringite or monosulfate, respectively according to: + ( ) + + ⋅ ⋅ ⋅ Al 2O3 3 Ca OH 2 3 CaSO 4 29 H 2O ⇒ 3 CaO Al 2O3 3 CaSO 4 32 H 2O (2.5) + + + ⋅ ⋅ ⋅ Al 2O3 3 Ca (OH ) 2 3 CaSO 4 9 H 2O ⇒ 3 CaO Al 2O3 CaSO 4 12 H 2O (2.6)

Alternatively, the reaction between K 2SO 4 and CaSO 4 in the presence of water may produce syngenite [K 2Ca(SO 4)2⋅H2O)] (Taylor, 1990). The formation of Friedel’s salt

(3CaO ⋅Al 2O3⋅CaCl 2⋅10H 2O) due to a reaction between chlorides and aluminates released from MSWI fly ash has been also identified (Rémond et al ., 2002).

Several studies demonstrated that the excess of Ca(OH) 2 in the solution would significantly improve the solubility of SiO 2 from fly ash, thereby promoting a

23 CHAPTER 2

pozzolanic reaction for the development of C-S-H phase (Ma and Brown, 1997; Pomiès et al., 2001a; Pomiès et al., 2001b). The variability of composition and crystallinity of C-S-H phase may also be controlled by temperature. Hence, the C-S-H phase may be formed under hydrothermal conditions, which lead to the solidified materials having a good degree of stabilisation and improved mechanical properties (Jiang and Roy, 1992).

2.3.2 Microstructural Model of C-S-H Phase The microstructural model of the C-S-H phase is an important determinant for the study of pozzolanic solidification for MSWI fly ash. Since the 1950s several models for the microstructure of the C-S-H phase have been published (Taylor, 1950; Powers, 1960; Feldman and Sereda, 1970; Jennings, 2000). Here, the development of the microstructural model for the C-S-H phase in the cement solidification is still a matter of concern. Two distinct types of the C-S-H microstructure have been proposed according to the degree of atomic packing during solidification: (i) the low density (LD) of C-S-H, which is initially formed, and (ii) the high density (HD) of C-S-H, which is formed later (Jennings, 2000; Tennis and Jennings, 2000). Moreover, the C-S-H microstructure could be distinguished in terms of their chemical compositions (Ca/Si ratio), morphology and the XRD powder pattern (Taylor, 1990). Hence, the first group of C-S- H (I) may have a crumpled foil-like morphology for the ratio of Ca/Si (0.8-1.5), while the second group of C-S-H (II) with the fibrous aggregates has the Ca/Si ratio of 1.5. The typical broad basal reflexes of C-S-H (I) and C-S-H (II) are centred at about 12.5Å and 10.5Å, respectively. Two general approaches have been followed for the explanation of the microstructure of the C-S-H phase (Taylor, 1990). The first model involves nanometer sized areas of structurally imperfect tobermorite and jennite, which could be considered as silicate tetrahedral chains of wollastonite type (the so called “dreierkette”) (Taylor, 1990; Kirkpatrick et al ., 1997) (Figure 2.2). The imperfect tobermorite structure may be due to (i) missing bridging tetrahedra in the chains and (ii) missing chains in the structure (substituted by hydroxyl groups). These defects may cause the great variation of the Ca/Si ratio from 0.75 (tobermorite) to 1.5 (jennite).

24 CHAPTER 2

Figure 2.2 (A) and (B) structure of a single layer of 1.4-nm tobermorite in bc -and ac - projections, respectively. In (B), the chains are seen end on. (C) Suggested structure for a single layer of jennite, in ac projection; the chains are seen end on and the Ca-O sheets edge on, parallel to their corrugations, and circled ‘H’s represent hydroxyl groups. In (A), (B) and (C), full circles represent calcium atoms, P and B denote paired and bridging tetrahedra, respectively. Jennite axes relate to a monoclinic pseudocell with a=1.00nm, b=0.36nm, c=2.14nm, β=101.9 O (after Taylor, 1990).

25 CHAPTER 2

The second model for the microstructure of C-S-H phase has been proposed by Grutzeck et al . (1999) in terms of a sorosilicate structure (consisting of dimers) (Figure 2.3).

Figure 2.3 A sorosilicate model of C-S-H Phases: (a) fully stoichiometric sorosilicates

(Ca/Si=2.0), (b) sorosilicate with all Ca(OH) 2 removed (Ca/Si=1.5), (c) sorosilicate with half of its calcium removed (Ca/Si=1.0); (d) the sorosilicate has ”unzipped” to form dreierketten (Ca/Si =1.0) (after Grutzeck et al ., 1999).

2.3.3 Stabilisation Mechanism in a Cement-Based System The stabilisation of MSWI fly ash in the cement-based system is essentially controlled by a sorption process of heavy metals onto neo-formed cement minerals through either physical ( at micro scale ) or chemical ( at atomic scale ) immobilisation mechanisms, depending on the binder employed (Glasser, 1997). Based on this approach, the sorption of heavy metals into cement matrices may occur via a variety of mechanisms: (i) adsorption ( specific and non-specific ), (ii) precipitation, and (iii) absorption or incorporation ( with or without solid solution formation ) (Sposito, 1986).

Adsorption The adsorption is a process where a solute in the liquid phase becomes bonded to the surface of a solid. A model for the adsorption due to a diffusion mechanism of

26 CHAPTER 2

contaminant into a solid phase has been described by Sposito (1986) in terms of surface functional groups. From the Sposito’ analysis, the surface functional groups represent a chemically reactive molecular unit bound into the structure of a solid at its periphery. Here the most important inorganic surface functional groups in a cementitious system involve metal oxides and hydroxides, aluminates, and silicates. Furthermore, the surface functional groups are broadly classified as inner-(IS) and outer sphere (OS) complexes. The surface functional group can be directly related to the binding of the adsorbate in the IS complexes, whereas the adsorbate in the OS complexes is separated from the solid by at least one water molecule (Sparks, 1995). The IS complexes bind ions much more strongly than those held by the OS complexes, because these ions are bound by covalent or ionic bonds, while the OS complexes involve electrostatic coulombic interactions. The OS complexes can occur only on the surfaces that are of opposite charge to the adsorbate, while the IS complexes can occur on the surface without any surface charge to the adsorbate. Moreover, the OS complex is usually formed faster and reversible, but this is not always the case for IS complexes. Both complexation mechanisms can occur simultaneously. Theoretically, the adsorption process occurs due to the net accumulation of ions at the interface between a solid phase and an aqueous solution phase and may result from either physical or chemical interaction with the surface. The ions accumulating at the surface can have different molecular arrangements involving several kinds of weak forces. This mechanism can be understood with assumption that the adsorption sites (S) on the surface of an adsorbant become occupied by an adsorbate (A) from the solution. The adsorption process may be written as (Baur et al ., 2004): S + A ↔ SA (2.7) The Langmuir isotherm is often employed for describing the adsorption of a solute from a liquid solution and is expressed as K [A] [][]SA = S ads (2.8) max + [] 1 K ads A with the surface site density [S] max =[S] +[SA]. The surface concentration ([S], [SA],

[S] max ) are usually expressed as mol/g and K is the constant of adsorption. It can be seen from the equation 2.8 that the increased concentration in [A] causes the increased concentration in [SA] until [S] max is achieved. Hence, the

27 CHAPTER 2

adsorption capacity of a solid may be related to the number of surface sites available for the adsorption [S] max . Once all the sites are fully occupied by adsorbate, the surface is saturated. Therefore any additional adsorbate will not be adsorbed at the surface. The sorption dependence of ionic strength (I) is also an important parameter for distinguishing the IS and OS complexes: the higher I is, the weaker the formation of OS complexes are. However, this factor is not always consistent, and thus only spectroscopic methods, such as extended X-ray adsorption fine the structure spectroscopy (EXAFS) or infrared spectroscopy (IR), can be used to examine the real distinction between IS and OS complexion.

Precipitation Precipitation occurs when two or more solutes combine to form a solid phase. The solid phase and the mobile solute will have the same composition. Chemical substitution/co-precipitation is also a subset of precipitation when a separate trace element becomes part of the crystal structure of the precipitating solid. Further, the precipitation can occur if the solubility of a solid has been exceeded, i.e ., some degree of supersaturation is required. The solubility of the solid may be estimated by a solubility constant, K so which can be derived as follows (Baur et al ., 2004):

AaBb (s) ↔ a A (aq) + b B (aq) (2.9) a b Kso = {A} {B} (2.10) where A and B are the chemical species; a and b are the stoichiometric coefficients. Moreover, the parentheses refer to the chemical activity of the species such as solid (s) and aqueous (aq). As the supersaturation of a solution is achieved with respect to a solid phase leading to the formation of a critical cluster or nucleus, from which a crystal will grow (Stumm, 1992). Extreme supersaturation yields a high rate of formation of crystal nuclei and a large amount of very small crystals or even non-crystalline solids. In contrast, minimal supersaturation can provide extremely low rates of nucleation. Therefore the crystal growth takes place from a few nuclei only, resulting in a highly crystalline product with large crystals (Sparks, 1995). Generally, precipitation mechanisms are much slower than adsorption reactions.

28 CHAPTER 2

Absorption/incorporation and solid solution formation The absorption and incorporation are a process where a foreign constituent is taken up into solid particles through a diffusion mechanism (Sposito, 1986). Within this particle, the foreign constituent may be adsorbed into internal layers or substituted into a crystalline lattice as a result of an ion exchange. The ion exchange here represents any replacement of an ion of a solid phase in contact with a solution by another ion (Stumm, 1992). Principally, the concentration gradient of a solute at the surface of the adsorbent controls the rate of sorption. If there is no gradient, then a homogeneous solid solution can form. The formation of the homogeneous solid solution as a result of a solid (AB 1) in contact with a solution containing B 2 may be given by (Baur et al ., 2004):

AB 1(solid) + B 2 ↔ A B 2(solid) + B 1 (2.11)

Using relation of K so AB1 = {A}{B 1}, K so AB2 = {A} {B 2}, the equilibrium constant of D for this reaction may be written as: { }{ } B1 A B )s(2 K so AB D = = 1 (2.12) {}B {}A B K 2 )s(1 so AB 2 In the case of an ideal solid solution, the ratio of concentration in the solution and the mole fractions (X) of AB 1 and AB 2 may be defined as:

n A B = 1 XAB with n = number of moles, (2.13) 1 n + n A B1 A B2 Hence, the constant of D may be obtained by:

[B ] X AB K so AB 1 2 = 1 = D (2.14) []B X K 2 AB 1 so AB 2

Sorption Isotherms The sorption of a specific compound can be only characterised by an empirical equation of the sorption isotherm, because of difficulties in making a distinction of individual sorption mechanisms, acting in a highly complex system particularly in a cementitious material. This equation represents a relationship between the activity (concentration) of adsorbate in the aqueous phase and the amount sorbed to a solid phase at given temperature (Stumm, 1992).

29 CHAPTER 2

An empirical adsorption isotherm for non-ideal systems developed according to the Freundlich isotherm is given by: [SA] = m[A] n (2.15) where m and n are the positive empirical parameters. The Freundlich isotherm for adsorption can be subsequently determined from the linear regression of a plot of log [SA] versus log[A] yielding to the intercept m and the gradient n. Furthermore, the type of the predominant sorption mechanism may be estimated from the shape of plot of sorption data in a double logarithmic diagram (Figure 2.4).

Figure 2.4 Schematic sorption isotherms of metal ion (Me) at a mineral surface at constant pH for different cases (after Stumm, 1992). (a) adsorption only (b) adsorption and surface precipitation via ideal solid-solution (c) adsorption and heterogeneous precipitation for low activation energies (d) adsorption and precipitation of a metastable precursor (e) same as in (d) but with the transformation of the precursor into the stable phase The two vertical lines represent the solubility concentrations of Me for the stable Me oxide and a metastable precursor ( e.g ., a hydrated Me oxide phase).

2.4 Hydrothermal Treatment of MSWI fly ash The use of MSWI fly ash for synthesis of zeolites has an important economical and environmental implication, because of presenting possible advantages in waste minimisation and resource conservation. It has been demonstrated that there is a possibility of converting fly ash into zeolite-like materials through hydrothermal process (Yang and Yang, 1998; Maenami et al. , 2000; Miyake et al. , 2002).

30 CHAPTER 2

The hydrothermal method has been successfully applied for synthesis of different zeolites from coal fly ash (see review by Querol et al ., 2002), while properties of the coal fly ash are very different from those of the MSWI fly ash investigated in the present study, the synthesis method based on the hydrothermal alkaline conversion is of possible relevance. The present study considers the synthesis of zeolite-like materials from MSWI fly ash by hydrothermal alkaline processing, very similar to that reported for coal fly ash. Table 2.4 shows the zeolites types, which have been synthesised from coal fly ash for a wide range of industrial applications. These extensive applications of zeolites are mainly based on (Breck, 1984): • Ion exchange: Exchange inherent Na +/K +/Ca 2+ for other cations on the basis of ion selectivity. • Gas adsorption: Selectivity absorption of specific gas molecules. • Water adsorption: Reversible adsorption of water without any desorption chemical or physical change in the zeolite matrix. The use of zeolitic materials for water purification has been previously demonstrated by a number of workers (Moreno et al. , 2001; Querol et al ., 2002). For example, zeolite X having a large pore size (7.3Å) and a high CEC (5 meq g -1) is an important potential of molecular sieve and a high-cation exchange material.

Mechanism The mechanism for zeolite synthesis from MSWI fly ash in a batch hydrothermal synthesis process may follow three stages: (i) the dissolution of aluminium and silicon from fly ash, (ii) the deposition of aluminosilicate gel on the fly ash surface, and (iii) the crystallisation of zeolite from aluminosilicate gel (Querol et al ., 2002). Hence, amorphous aluminosilicate glass, quartz, and gehlenite are supposed to be the source for aluminium and silicon for zeolite synthesis (Maenami et al. , 2000; Miyake et al. , 2002). The largest and most unstable of these phases in the hydrothermal environment is the aluminosilicate glass phase, and therefore it has the highest rate of dissolution, and is the largest contributor to produce zeolites. In contrast, quartz and gehlenite are more stable than the glassy phase (Miyake et al. , 2002), and therefore only a little amount of these phases react during the hydrothermal synthesis.

31 CHAPTER 2

Table 2.4 Joint Committee of Powder Diffraction Society (JCPDS) card number for some zeolites (Querol et al ., 2002).

Zeolites JCPDS High industrial application

NaP1 zeolite (Na 6Al 6Si 10 O32 ·12H 2O) 39-0219

Phillipsite/KM-zeolite (K 2Al 2Si 3O10 ·H2O) 30-0902

K-Chabazite (K 2Al 2SiO 6·H2O) 12-0194

F linde zeolite (KAlSiO 4·1.5H 2O) 25-0619

Herschelite (Na 1.08 Al 2Si 1.68 O7.44 ·1.8H 2O) 31-1271

Faujasite (Na 2Al 2Si 3.3 O8.8 ·6.7H 2O) 12-0228

Zeolite A (NaAlSi 1.1 O4.2 ·2.25 H 2O) 43-0142 Low industrial application

Perlialite (K 9NaCaAl 12 Si 24 O72 ·15H 2O) 38-0395

Analcime (NaAlSi 2O6·H 2O) 19-1180

Hydroxy-sodalite (Na 1.08 Al 2Si 1.68 O7.44 ·1.8H 2O) 31-1271

Hydroxy-cancrinite (Na 14 Al 12 Si 13 O51 ·6H 2O) 28-1036

Kalsilite (KAlSiO 4) 33-0988

Tobermorite [Ca 5(OH) 2Si 6O16 ·4H 2O] 19-1364

Factors influencing the synthesis of zeolites The hydrothermal synthesis for different zeolites from coal fly ash has been reported by a number of workers (Moreno et al ., 2001; Querol et al ., 2002). All methodologies are developed on the basis of dissolution of Al-Si bearing phases in alkaline solutions (mainly NaOH and KOH solutions) and the subsequent precipitation of zeolitic materials. Following these methodologies, a variety of zeolites has been synthesised through variation of the molarity of alkaline reagents, activation- solution/material ratios, temperatures, and reaction times (Table 2.5). However, the hydrothermal processing route for synthesis of zeolites from combustion residues including MSWI fly ash may have some limitations regarding to: (i) the speed of reactions, and (ii) relatively high activation temperatures (125 - 200 OC) for dissolving Si-and Al-compounds from these materials.

32 CHAPTER 2

Table 2.5 Zeolites and other neomorphic phases synthesised from coal fly ash as a function of activation agent (NaOH or KOH), temperature and activation solution/fly ash ratio (ml/g) (Querol et al ., 2002)

Solution/fly ash ratio (10-18 ml/g) NaOH Temperature Product 0.5-3.0 M 90-175 OC NaP1 175-225 OC Analcime, hydroxy-sodalite, nepheline, tobermorite 3.0-5.0 M 150-200 O C Hydroxy-sodalite, hydroxy-cancrinite, tobermorite KOH 0.5-1.0 M 150-200 OC KM, tobermorite 3.0 M <150 OC Linde F, tobermorite 5.0 M <150 OC Linde F, kalsilite, tobermorite 3.0-5.0 M >150 OC Kalsilite, tobermorite

Solution/fly ash ratio (1-3 ml/g) NaOH 5.0 M 150-200 OC Low activation for all temperatures 1.0 M 150 OC Low activation, NaP1 (herschelite) 200 OC NaP1 and herschelite for 8 h activation 2.0-3.0 M 90 OC A zeolite 150 OC NaP1 (herschelite traces), faujasite (if aging) 200 OC NaP1, herschelite, 5.0 M 150-200 OC Herschelite, analcime, hydroxy-sodalite, hydroxy-cancrinite KOH 2.0 M 150-200 O C KM zeolite 5.0 M 150 O C KM, chabazite and linde F traces 200 O C Kalsilite and KM, perlialite and tobermorite traces

33 CHAPTER 2

Due to the limitation of parameter conditions, many of high-CEC and large pore zeolites ( e.g ., zeolites A and X) could not be synthesised. If the temperature is reduced considerably, then a very long activation time is required. Hence, to optimise the zeolite production at a low cost requires a degree of compromise between reaction time and temperature, since improvements in one may adversely influence the other. In view of limitations of the zeolite synthesis of MSWI fly ash, the following parameters are to be considered in the study (Querol et al ., 2002): (i) the activation time, (ii) the concentration of the activation agents, and (iii) liquid/solid (L/S) ratio. The activation time for synthesis is inversely proportional to the glass content of materials. MSWI fly ash with the high glass content is readily converted to zeolites within the short period (6-8h), while the longer reaction times (24- 48 h) are required for obtaining a similar zeolite from fly ash with the high content of Si and Al. Other factors controlling the dissolution of the Al-and Si-bearing phases are synthesis time and the variation of activation solution/fly ash (L/S) ratio. With increasing L/S ratio and time (12-24 h), the high productivity of zeolites is achieved because of the increased dissolution rate of these phases. Further, temperature and concentration of an activation agent become important factors for the zeolite conversion. Hence, a compromise required for optimising the efficiency of zeolite formation is evident with activation time, temperature and concentration of the activation agent being major determinants for the zeolite preparation strategy.

Cation exchange capacity (CEC) The cation exchange capacity can be used for assessment of the ability of a zeolite to attract and exchange cations. The CEC of zeolites is the measure of site quantities on its surface for retaining cations as a result of electrostatic forces. The CEC of zeolites can be determined using an ammonium solution according to the methodology of international soil reference and information centre (ISRIC) (1995). Similar tests can be conducted using distilled water to determine the leachable Na + levels from the zeolitic products, which could be erroneously attributed to ionic exchange processes.

34 CHAPTER 2

2.5 Leaching Behaviour of MSWI Residues The leaching behaviour has received much attention in the assessment of the environmental acceptability whether the MSWI residues to be deposited at landfills or reused as a resource (van der Sloot et al ., 1997). Strategies for disposal or recycling of MSWI residues depend on the understanding of both short-and long term leaching behaviour of these residues. A large number of published works are currently available for the short-term leaching behaviour of MSWI residues, but only limited data exist regarding to the long-term leaching behaviour, because of the complexities of a synthesis information on leaching principles, leaching tests results, field experiments and hydrogeochemical modeling of mineral changes and speciation ( Hjelmar, 1996; Sabbas et al ., 2003). Table 2.6 provides typical ranges of concentrations of inorganic salts, trace elements and non-volatile organic carbon (NVOC) obtained at the early stage of leachates of the MSWI residues from laboratory leaching test and field investigations (Hjelmar, 1996). These maximum levels of concentrations were basically obtained from the availability for leaching as a function of L/S (liquid/solid ratio). However the leaching behaviour of contaminants may be also affected by other factors as pH, redox potential, ionic strength, complexing inorganic ions and organics, the presence of various minerals. At low L/S values, the leaching of several contaminants, particularly trace elements, is solubility controlled and strongly influenced by the pH of the leachates (Tan et al ., 1997). Hence, to predict the leaching quality as a function of time requires an additional information related to other factors such as pH and redox conditions. Once the relationship between the controlled contaminants and the controlling factor has been established, it could facilitate the identification of the process controlling leachate composition.

35 CHAPTER 2

Table 2.6 Maximum concentration levels of contaminants in leachates from various MSWI residues (Hjelmar, 1996). Typical maximum MSWI fly ash and Mixture of MSWI levels of MSWI bottom ash residues from dry and fly ash and sludge concentration in semidry APC from wet scrubbing leachate processes process >100 g/l Cl -, Ca 10-100 g/l Na, K, Pb Cl -, Na, K 1-10 g/l 2− - Zn 2− SO 4 , Cl , Na, K, Ca SO 4 , Ca 100-1000 mg/l − 2− NVOC, NH 4 N NVOC, SO 4 10-100 mg/l 1-10 mg/l Cu, Mo, Pb Cu, Cd, Cr, Mo NVOC, Mo 100-1000 µg/l Mn, Zn As 10-100 µg/l As, Cd, Ni, Se As, Cr, Zn 1-10 µg/l Cr, Hg, Sn Pb <1 µg/l Hg Cd, Cu, Hg

2.5.1 Leaching Mechanism Leaching characteristics of MSWI residues in aqueous and acidic solutions are considered to be due to chemical interactions of individual waste particles and contaminant transport processes via the fluid moving through solid particles (Sabbas et al ., 2003). Hence, the transport medium of pollutants in a natural disposal environment is generally represented by water. The mechanism governing the leaching process at the interface of a material may be illustrated using figure 2.5. It is assumed here that the chemical interaction between the solid waste and the aqueous solution predominantly occurs at the interface. If a porous matrix containing heavy metals is exposed to an aqueous solution, the pore water will be contaminated as a consequence of desorption of metals or/and dissolution of metal compounds.

36 CHAPTER 2

Figure 2.5 Schematic of leaching mechanisms (after Conner, 1990).

The process of desorption of metals or dissolution of metal compounds in the pore water is mainly controlled by the surface reaction (Conner, 1990). Due to the difference in chemical potential between the pore fluid and the fluid surrounding the porous matrix, a diffusion of metals through the pore fluid can occur and finally result in the leaching process. However, when the aqueous solution or water passes through the porous matrix, the contaminant transport occurs as a result of an advection process along with the diffusion mechanism of the contaminants through water.

2.5.2 Factors Influencing Leaching Metals solubility The solubility of metals in water may be due to the surface hydrolysis, the presence of other organic and inorganic ligands, their coordination chemistry, and the pH of solution. An assessment of metal leaching from MSWI residues can be made using a concept of availability for leaching 1 (van der Sloot et al ., 1997). Hence, the potential leachability of a given element as a result of contact with fluid may be classified as: (i) a solubility-controlled leaching indicating the amount of contaminants

37 CHAPTER 2

released from a loosely packed granular waste material, (ii) an availability-controlled leaching which is related to the leaching resistance of very soluble mineral phases, and (iii) a diffusion-controlled leaching which corresponds to the leaching performance of compacted granular materials or treated ( e.g. , solidified ) wastes. In the case of the diffusion-controlled release, the leachate percolation takes place at the surface of a solid material, thereby causing a molecular diffusion to be the dominant process, which controls the release of contaminant from the waste. However, a chemical equilibrium between solid and liquid phases has been not achieved, because there is a reduction in the surface area of material, which contacts with the leachate. Accordingly, the release of contaminants from the residues occurs at extremely slow rate during this process. Further, an important parameter that controls the metal release during leaching is pH of pore water (Sabbas et al ., 2003). The presence of the stabilised fly ash in water may modify the pH of the pore water and thereby resulting in the leachate composition (Theis and Wirth, 1977). The leachability for strongly soluble species (e.g., alkali salts ) is almost independent on the pH. However, most the leaching of contaminants are largely dependent on the pH. The pH-dependence on the leaching of inorganic species in the alkaline solution is shown in figure 2.6. The leaching behaviour of these elements may be defined as: (i) cation-forming species and non-amphoteric metal ions ( e.g., Cd), (ii) amphoteric metals (including Al, Pb, Zn), and (iii) oxyanion-forming elements ( e.g ., As,Cr, Mo, V, B, Sb). Principally, the concentrations of metal and non-amphoteric metals usually remain constant at pH value less than 4, but decrease significantly at the pH ranges (8 - 9), then remain steady or slightly increase at the higher pH values. In contrast, the solubility of amphoteric metals increases in both strongly acidic and alkali conditions, resulting in a V-shaped solubility curve. For oxyanion-forming elements in alkaline solution, the solubility usually decreases in the pH value (above 10) (Hjelmar, 1996; Sabbas et al. , 2003).

1 The availability for leaching of a constituent indicates the leaching ability with respect to a specific component or element of the solid waste in water.

38 CHAPTER 2

Figure 2.6 Solubility of (a) Cd, (b) Al, and (c) B concentration in leaching solution from samples of fresh and aged MSWI fly ash (after Sabbas et al ., 2003).

39 CHAPTER 2

The dependence of leaching resistance of MWSI residues on the pH can be also examined in terms of acid or base neutralisation capacity (ANC/BNC). The ANC/BNC can be used to measure the ability of a system to neutralise the effect of acids or bases. Because MSWI residues are alkaline or basic, the ANC may be more an appropriate measure for the neutralisation capacity. The presence of complexing agents including dissolved ligands may also change significantly the extent of contaminant leaching from MSWI residues. Ligands are formed when a metal is bound by anions or molecules in a coordination compound. The formation of complexes with the dissolved ligands will improve the solubility of the metal. The presence of soluble metal complexes often reduces metal adsorption (Reed and Nonavinakere, 1992), particularly when the metal complexes have less affinity for the sorption sites than the free metal ion. Therefore the high concentration of the dissolved ligands can cause the high metal concentration in leachates.

Liquid-to-solid ratio (L/S) dependence of leachability The leaching behaviour of contaminants in water may be related to the liquid to solid (L/S) ratio, particularly in the case of solubility-controlled leaching. In practice, the L/S ratio may be the result of climatic conditions, hydrology and hydrogeology of application site as well as the physical characteristics of the waste material. Time dependence of release rate from a residue in the field can be theoretically estimated using the cumulative release rate at different L/S ratios given by (Kosson et al ., 1996): L (inf ) (t ) = year (2.16) ρ S Hfill  L  M ∑ =   x ()S (2.17)  S  x

3 2 where inf is the annual infiltration at a given location (m /m year), t year is the exposure ρ time in the field (year), Hfill is the height or depth of residue fill material (m), and is

3 the bulk density of the material concerned (kg/m ). M∑ is the estimated cumulative release (mg/kg), and Sx is determined by solubility or availability (mg/l). The linear dependence of the cumulative release rate on the L/S ratio may be found in the solubility-controlled leaching. However, this linear trend may be altered due to the presence of other species. This makes a delayed release when the presence of

40 CHAPTER 2

a sparingly soluble phase for controlling solubility is depleted after a relatively short period. Hence, the cumulative release versus L/S plots exhibits the lower slope with the low value of L/S ratio, but the rising slope with the higher L/S ratios may be achieved by the depleted sparingly soluble spesies. Conversely, an enhanced initial release may occur due to the presence of other complexing agents, which would be expected to improve the solubility of particular compounds. Accordingly, a transition release rate from a higher to lower value on the curve slope corresponds with the rising L/S ratio (Sabbas et al ., 2003). For availability-controlled leaching, the release of contaminants into the solution can occur at its maximum level, because of the high contaminant solubility and the independence of pH solution. At a given L/S ratio, there is a transition from solubility- controlled to availability-controlled leaching when the concentration of solution remains constant with decreasing pH. Finally at low L/S ratios, a rapid washout of the soluble constituents can occur in the availability-controlled leaching. The available amount in this case can be achieved at L/S ratio of 1 to 2. Conversely, for the higher L/S ratio, the cumulative release remains at this maximum value. The availability-controlled leaching pattern for Na, Cl and K on weathered bottom ash as function L/S ratio can be illustrated in figure 2.7.

Figure 2.7 Leaching behaviour of Na, K, Cl from weathered MSWI bottom ash as function of the L/S ratio (after Sabbas et al ., 2003).

41 CHAPTER 2

In view of leaching from compacted granular residues or monolithic forms, neither solubility-controlled nor availability-controlled leaching is present, but leaching could rather be attributed to the molecular diffusion and surface dissolution mechanism. Hence, the rate of contaminant release via diffusion kinetically controls the leaching, which may be measured through the effective diffusion coefficient. Such diffusion process may be influenced by release mechanisms, physical ( e.g., porosity, pore structure, degree of compaction and tortuosity ) and chemical retardation.

Adsorption of metals The leaching behaviour of MSWI residues may be influenced by different mechanisms of adsorption. The adsorption of metals occurring on the surface of amorphous phases or secondary precipitates may be related to: (i) the properties of the solid ( particles size, nature of inorganic oxide coating, organic carbon content, and zero point charge of the solid ), and (ii) the properties of the liquid, including pH of pore water and total dissolved metal concentrations ( i.e ., the sum of the free metal pool, inorganic ion pairs as well as the concentration of the complexing ligands). The total dissolved concentrations can be related to the sum of many different components, thereby being affected by any factor that would impact one of the individual components. Here the dominant factor of the adsorption and solubility of most chemical compounds is the pH of pore water (Sauve et al ., 2000). The relative preference of particles to adsorb heavy metal ions can be related to the selectivity coefficient of the metal and mineral surface, where a mixture of metals and other ions may be found in a pozzolanic solidification or hydrothermal process. A method of depicting an affinity of ions at the surface, for example, the most clay minerals, is by means of the Hofmeister series (Stumm and Morgan, 1996). The Hofmeister series is given by: Ba 2+ >Sr 2+ >Ca 2+ >Mg 2+ >Cs 1+ >K 1+ >Na 1+ >Li 1+ >H 1+ The higher the valence of ions is, the greater affinity to adsorb metals, but is the smaller hydrated radius at equal concentrations. Ions with higher valence tend to be adsorbed first, similar to ions with smaller hydrated radius. Moreover, retardation of a metal in a cementitious matrix is strongly affected by the composition of the solution.

42 CHAPTER 2

Chemistry of the solid phase The factor controlling leaching of MSWI residues is mainly based on the distribution of metals on the solid phase. For example, the metal leaching from fly ash is affected by the total amount of metals in this material and the relative amounts of metals in glassy and crystalline phases. The glassy phase has the larger available surface areas and therefore has many reactive sites for adsorption of trace metals ions. Generally, the heavy metals fixed within the glassy phase are stable and could be only released through long-term weathering processes, while metals or metal compounds accumulated at the surface of fly ash particles during combustion are more chemically reactive (Theis and Wirth, 1977; Belevi and Moench, 2000). Iron and aluminium oxides or hydroxides present in MSWI residues are also considered as major sinks for trace elements (Theis and Wirth, 1977). They have certain cation exchange capacities (Kirby and Rimstidt, 1993). The cation exchange capacity increases with the formation of clay-like minerals during chemical weathering (Zevenbergen and Comans, 1994). The effect of surface charge on the adsorption of metals regarding to the influence of leaching resistance can be examined in the context of pH and a point of zero charge (pzc). Typically, the point of zero charge for some minerals as a function of pH is indicated by the net surface charge of a solid at zero value as illustrated in figure 2.8. The pH at the intersection of the dashed line is denoted as net zero of the surface charge for a particular mineral (pH zpc ). For example, the pzc value of 3 is for mineral of montmorillonite. The nature of a more negative surface charge may be also observed due to the increase in the pH value above of the pH zpc , and thereby increasing the adsorption of cations on the solid surface. Here, the pzc value of MSWI fly ash may vary between 2 and 4 (Ricou et al ., 1999).

Figure 2.8 Point of zero charge and effect of pH on the surface charge for some common minerals (after Stumm and Morgan, 1996).

43 CHAPTER 2

2.5.3 Evaluation of Leaching Resistance The leaching or extraction test is often used as a practical method to evaluate the potential of MSWI fly ash to release contaminants to the environment. The leaching test is principally designed according to the Environmental Protection Agency (EPA)’s TCLP (toxicity characteristic leaching procedure) to determine the maximum concentration levels of contaminants in water, which can be released from the tested sample. The contaminant concentrations obtained are then compared with the toxicity characteristics set by the EPA (U.S. EPA,1990). Various techniques have been proposed for evaluating leaching resistance of the material (van der Sloot et al. , 1997). These are: (i) a water leach test (ASTM D 3987- 85); (ii) a toxicity characteristic leaching procedure (TCLP); (iii) a extraction procedure toxicity (EP Tox) test; (iv) a multiple extraction procedure (MEP); and (v) a synthetic acid precipitation leaching test. The present study considers the TCLP test.

Toxicity characteristic leaching procedure (TCLP) The TCLP test according to EPA Method 1311 (U.S. EPA,1990) is used extensively to determine the mobility of both organic and inorganic compound present in liquid, solid, and multiphase wastes. In this method, the crushed sample with particles size less than 9.5 mm are subsequently extracted by using an acetate buffer solution of a pH 5 or an acetic acid solution with a pH of 3, depending on the alkalinity of the waste. Hence, the acetate buffer is added only once, at the start of the extraction. A liquid-to- solid ratio of 20:1 is normally selected and the time of extraction is 18 hr. The leachate is filtered through a 0.6 µm to 0.8µm filter paper prior to conducting chemical analysis.

2.6 Summary The principles developing the treatment strategy of MSWI residues, particularly MSWI fly ash have been discussed. Bottom ash contains only limited amounts of trace contaminants, therefore the disposal or reuse of bottom ash is considered less problematic than fly ash because of the release of contaminants at an environmentally acceptable rate. In contrast, fly ash is composed of significant amounts of readily soluble compounds and leachable trace heavy metals, and therefore it requires a subsequent treatment for improving the mineral stability and the leaching resistance.

44 CHAPTER 2

Characteristics of MSWI residues clearly vary in chemical and mineralogical composition. The subsequent mineralogical changes that undergo as a consequence of weathering and aging over time become a major concern of investigation, which should be based on the chemistry and mineralogical characterisations. By understanding on the chemistry and mineralogy, a proper treatment strategy for MSWI fly ash can be determined. A literature review of the treatment strategy on MSWI fly ash has shown that pozzolanic solidification and hydrothermal methods appear to be the economical and technical sound solutions for reducing release of contaminants. The key variables controlling developments of the cement matrix are components of SiO 2, Al 2O3, CaO, and Fe 2O3. These components, in conjunction with an activator of Ca(OH) 2 and water may generate a cement like-material through a pozzolanic reaction that is potential for immobilisation of heavy metals. In addition, a large abundance of SiO 2 and Al 2O3 components in an aluminosilicate glassy phase of fly ash is an important factor for the formation of zeolites by the hydrothermal method. The pozzolanic solidification and hydrothermal experiments of fly ash have been investigated in this work to study the systematic influence of pozzolanic and hydrothermal reactions on the associated mineral stability with the improved leaching resistance. The study was designed to provide a detailed knowledge of the characteristics of MSWI residues and the development of pozzolanic and zeolitic materials from fly ash for an environmental protection and technical applications.

45 CHAPTER 3

CHAPTER 3 METHODOLOGY OF RESEARCH

This chapter describes methods employed in the study. Experimental design (involving state of the art in the powder preparation; powder processing strategy; characterisation measurements and evaluation of leaching properties) is first outlined, followed by descriptions of the experimental methods used.

3.1 Experimental Design State of the art in the powder preparation The MSWI residues selected for the study consisted of (i) bottom ash and (ii) fly ash collected from two different MSWI plants located in the town of Iserlohn and Essen, Germany, hereafter are designated as incinerator plants A and B, respectively. In particular, batches of fly ash from the incinerator plant of Iserlohn (incinerator A) were collected from three different locations of electrostatic precipitator (ESP) so that the chemical and mineralogical characteristics, basicity and heavy metal contents with the different particle morphology and size could be investigated. This would be one of the important factors for selection of further treatment and disposal. Bottom ash and fly ash with particle sizes less than 2 mm and 200 µm respectively were sampled and processed into powder. Due to a large of particle size especially for bottom ash, a substantial grinding was required prior to subsequent processing and characterisation. In addition to significant reduction ash particles that might be achieved by grinding, an agate vibratory disc mill was employed for homogenising the mineral particles. Subsequently, a mineral fractionation procedure was employed in view of the fact that separation of the particular particles may provide the possibility of detection on the mineral fraction occurring in the lower concentration within residues. A hand magnet was utilised for separating magnetic particles of bottom ash, while magnetic fraction of fly ash was extracted by conducting a water-washing treatment using a magnetic stirrer. The water-washing treatment was also considered as a simple technique for extracting soluble salts and minimising heavy metal leachability (Wang et al ., 2001). The treatment stage benefit of fly ash by the washing procedure was somewhat offset by generating a new waste residue containing chlorides, water-soluble

46 CHAPTER 3

sulfate and alkali ions (Wang et al ., 2001; Sabbas et al. , 2003). Hence, a degree of compromise was involved in the selection of the pre-washing treatment due to the competition between quality benefits of the fly ash for subsequent treatment and producing other pollutants. However, it was envisaged that the pozzolanic and hydrothermal activity of fly ash might be improved by such pre-treatment. A further investigation was conducted by a Soxhlet of water-extractor on the water-extractability of minerals and heavy metals. It was regarded important to elucidate on the solubility of minerals and trace metals in water so that their effects on the leaching performance could be assessed. This approach was considered to be of particular significance in order to enhance chemical properties of fly ash by removal of salts, and partly sulfate from the materials. Here, efficiency of extractions was related to the selection of liquid to solid (L/S) mass ratio and time. In this study, the L/S ratio of 10 was selected because it allows extracting almost 90% of soluble salts (Nzihou and Sharrock, 2002). The solution residues were then evaporated at low temperature to produce crystalline minerals for subsequent x-ray fluorescence (XRF) and x-ray diffraction (XRD) characterisations.

Powder processing strategy The powder processing procedure was a key aspect of the study for pozzolanic solidification and hydrothermal conversion of fly ash. The success level to produce stable mineral phases depended on the powder processing route, synthesis parameters and drying method. The principal considerations for the powder processing stage were particle size distribution and the specific surface area. Mechanical activation using a magnetic stirrer was chosen in view of the result reported by Blanco et al . (2005) that the technique is simple and efficient at achieving fine particle size and also yields less contaminant than a dry mixing. Use of fine particle size powders was regarded important to enhance the pozzolanic activity and hydrothermal reaction rate. Longer stirring times at particular rotating speed of the stirrer were selected to produce more effective particle size distribution as means of achieving the homogeneity of the powder precursor and increasing specific surface area as a requirement for synthesis of cementious and zeolite like-materials.

47 CHAPTER 3

The synthesis of cementitious material was examined in terms of pozzolanic behaviours and the degree of solidification due to the presence of reactive SiO 2 from the fly ash and the addition of Ca(OH) 2. Hence, water was essential to initiate a pozzolanic reaction, which was mainly controlled by the ratio of liquid to solid (L/S) material. Selection of L/S ratio was based on achieving pozzolanic solidification and porosity, ultimately flow ability of solidified product subject to the powder processing sequence employed.

It was regarded important to prepare a saturated solution of Ca(OH) 2 with distilled water so that a cement-like material could be produced by mixing it with fly ash at L/S ratios of 3 and 10. This approach was considered to be of particular significance in order to promote a pozzolanic reaction between fly ash components of SiO 2, Al 2O3 and

Fe 2O3 and Ca(OH) 2 available in the solution. The products of this reaction would be calcium silicates and aluminates, which can be considered as the equivalence of Portland cement in initiating the cementation process (Glasser, 1997; Hamernik and Frantz, 1991). Solidification time was explored as a key step in examining evolution of hydrate phases ( e.g ., CSH phase and ettringite) through a systematic study of pozzolanic reaction. The time chosen was based on achieving high yields of these hydrate phases at room temperature. The direct conversion of fly ash into zeolite like-materials and other neomorphic phases by hydrothermal treatment was also investigated. The methodology was based on a direct alkaline hydrothermal activation with the different molarity of the alkaline reagents (0.5-2.5 M), solution to fly ash (L/S) ratio, temperature and time. Alkali metal hydroxide solutions of either NaOH or KOH were employed to activate the reactions because it has been shown to be effective in the conversion of coal fly ash (Moreno et al ., 2001). The ratio of solution to fly ash (L/S) was explored as a key step in evaluating zeolite evolution through a study of reactivity of alkali solution. The selection of hydrothermal temperatures and times was based on achieving zeolite formation with the variation of crystalline and amorphous phase composition subject to the ranges of reaction parameters employed. Hence, the choice of parameter ranges was guided by the outcome of the experiments and by what was considered still realistic for an economically feasible process.

48 CHAPTER 3

Characterisation The characterisation of chemistry and mineralogy played an important role for gaining insight in the characteristics of the MSWI residues, particularly fly ash to be stabilised for immobilisation of heavy metals. Two evaluation methods were mainly used for investigating: (i) mineralogical phase alteration of MSWI residues and (ii) formation of stable mineral phases generated from the treatments of fly ash. Mineralogical phase changes were to be examined in terms of weathering and aging behaviour over time and the degree of crystallisation that may result due to the presence of an amorphous phase in MSWI residues. It was regarded important to mix the fresh MSWI residues with water so that the effect of water reactivity on promoting mineralogical alteration could be assessed. This approach was also considered to be particularly important to understand the mechanism of pozzolanic activity and hydrothermal conversion of fly ash over a period of time. Crystalline and amorphous phase compositions were determined using the Rietveld method and also observed by an optical microscopy. Further formation of stable mineral phases as a result of pozzolanic solidification and hydrothermal treatments was characterised by x-ray diffraction (XRD) and scanning electron microscopy (SEM). Assessment of characteristics of MSWI residues was also conducted by various analytical techniques. Optical microscopy and scanning electron microscopy (SEM) with energy dispersion spectrometry (EDX) were employed to characterise morphology and chemical compositions being the major aspects of interest. Bulk chemical elements were also determined by x-ray fluorescence (XRF) analysis. Electron probe x-ray microanalysis (EPMA) was employed to determine the chemical composition of individual particles of MSWI residues. Such analysis was important to present both qualitative and quantitative chemical data for major and trace elements in a wide range of particle sizes, which could be obtained by a large number of spot WDX (wavelength dispersion spectrometry) analyses (Piantone et al. , 2003). Importantly, this method can provide invaluable information about the chemical composition of amorphous phase in the ash particles. It is through by this approach that the XRD phase analysis coupled with the EPMA technique were employed for understanding distribution of mineralogical composition within particles of ashes.

49 CHAPTER 3

Leaching properties Assessment of leaching properties was undertaken to ascertain the immobilisation of heavy metals. It was emphasised previously that the relationship between processing and mineralogical evolutions clearly determines the resultant properties of pollutants in the form of less toxic compounds which were of paramount importance in the fly ash treatment strategy. Leachability was examined by using pH dependence test, because of having been standardised for characterisation of MSWI residues (Sabbas et al. , 2003). Here, the samples to be tested were firstly necessary to be ground for obtaining the small particle size, so that it could ensure a good contact between water and the sample, and may also minimise the diffusion barrier of the particles. The water-to-sample ratio was subsequently adjusted in terms of a liquid/solid (L/S) ratio as recommended by Environmental Protection Agency (EPA) (US EPA, 1990; EPA-SW 846 Method 1311). The leached solution was analysed by inductively coupled plasma mass spectroscopy (ICP-MS) for concentrations of selected toxic elements.

3.2 Powder Processing Sampling The MSWI residues examined were sampled from two German MSWI plants. The incinerator plant A opened in 1970 is located in Iserlohn. It has three combustion lines with a nominal capacity rating of 8 tons/h (for the two combustion lines), whereas the third combustion line has capacity of 16 tons/h. On the other hand, the incinerator plant B is located in Essen and began operation in 1987. This plant has four combustion lines, which have a nominal capacity rating of each 26 tons/h. The incinerated material may consist of household waste (50-70%), bulky refuse (8-10 %), industrial waste including colours and lacquer, filter material with oil and organic solvents, chemical waste, and shredder waste. Both plants consist of a primary combustor with movable grates and the incinerator normally operates at temperatures between 850 and 1000 OC. The residence time of the waste in the incinerator varies from 40 to 60 min. The flue gases are thus generated containing different gaseous species such as H, C, S, N, Cl and O, where HCl,

CO 2, SO x and NO x are dominant. Gaseous forms of metals, organic species and dust particles are also produced. The flue gases are then cooled through heat exchangers (boiler and economiser), electrostatic precipitators (ESP) and scrubbers. Consequently,

50 CHAPTER 3

three solid residues are produced: (i) boiler and economiser ash collected at the heat recovery section, (ii) fly ash collected from electrofiltration of flue gas having temperature between 250 and 400 OC, and (iii) APC (air pollution control) residues, which are generated by treatment of the downstream flue gas in a scrubber at temperatures below 80 OC. In this scrubber, lime and water are used to neutralise acid gases. Further, the hot bottom ash is quenched in a wet slag extractor together with the grate siftings. Similar to the incineration process, the quenching is done in a continuous process. The ratio of water to ash varies between 2 to 12 percent according to the heterogeneous composition of feed waste. The residence time of the ash in the water varies between 4 and 8 minutes, depending on the throughput of waste and on its calorific value. The ash is transported by belt conveyors to an ash hall for stocking. The transport for final treatment is done by an external company. The final treatment is performed by the removal of large pieces of unburned material, the removal of magnetic particles and size reduction by crushing. Immediately after quenching, the quenched bottom ash has been treated by the removal of large pieces of unburned material, the removal of magnetic particles, followed by size reduction by crushing. In both facilities, there is no fly ash mixed with the bottom ash In this study, three batches of fresh bottom ash and fly ash, each being approximately 15 kg, were sampled during two distinct periods between 2003 and 2004. The first batches of bottom ash and fly ash were firstly sampled from incinerator facility A in March 2003, hereafter referred to the samples of BA-A1 and FA-A1 respectively. Furthermore, the second batches of bottom ash (BA-A2) and fly ash (FA-A2, FA-A3 and FA-A4) were collected from incinerator A in August 2004. Here, the FA-A2, FA- A3 and Fa-A4 samples from facility A were collected from the three different locations of electrostatic precipitator (ESP) devices. Finally, the third batches of bottom ash (BA- B1) and fly ash (FA-B1) from the ESP device were sampled from incinerator facility B in August 2004. In all cases, it was assumed that the samples collected represent MSWI residues as produced rather than time averaged compositions.

51 CHAPTER 3

Ash processing A conventional powder processing was selected to prepare bottom ash and fly ash powder batches. The wet bottom ash was initially crushed with a crusher mill for 30 minutes and dried at room temperature for 24 hours. The resulting dried powder was again ground in an agate bowl for 30 minutes and sieved through 100 µm to break up the hard agglomerates for subsequent characterisation. Similarly, the fly ash was homogenised by the agate bowl for 30 minutes and passed through 100 µm sieving. As a part of examination for mineralogical alteration under influence of water interaction over time, sample powders of the first batches of fresh bottom and fly ashes (BA-A1 and FA-A1) were prepared. Hence, each sample of either bottom ash or fly ash was mixed with three parts of distilled water and left to dry at room temperature for 1 month. The relative humidity of the environment was typically 60-70 %. The alkaline bottom ash had pH ranging from 10 to 11 in distilled water, while the pH of fly ash varied from 9 to 10. The materials solidified by reaction with water and they were subsequently examined by XRD analysis (see section 3.5.4). A further investigation of mineralogical alterations was performed on the second and third batches of fresh bottom and fly ashes (BA-A2, BA-B1, FA-A2, FA-A3, FA- A4, FA-B1). Hence powder of each fresh ash was prepared and front loaded into an aluminium XRD sample holder, and finally compacted with a glass slide. Three suites of specimens prepared in this way were subsequently examined by XRD analysis (see section 3.5.4). These specimens were then aged in atmosphere condition for a period of 6 months.

Fractionation A fractionation scheme was developed to increase the opportunity for detection and identification of phases occurring in lower concentration in the MSWI residues. Particle sizes of bottom ash with less than 100 µm were separated with a hand magnet. For extracting magnetic particles of fly ash, a procedure was developed by conducting a parallel experiment of water-washing treatment in a glass beaker equipped with a magnetic stirrer for mixing at 300 rpm. Here, 70 grams of the fly ash powder batches were prepared. They were initially placed in the glass beaker containing distilled water. The water to solid (L/S) mass ratio (ml/g) was adjusted to 10. Subsequently, this

52 CHAPTER 3

mixture was stirred to stand for 24 hours at room temperature. At the end of stirring, the resulting slurries was filtered with a paper filter (Schleicher&Schuell no. 604), washed repeatedly with distilled water, and dried at room temperature for subsequent stabilisation and hydrothermal experiments. While the magnetic particles that adhered to the stirrer blade were washed with distilled water and dried for XRD examination. Soluble salts were extracted from the fly ash by a Soxhlet type water-extractor (Figure 3.1). The general procedure is described as follows; 70 grams of the fly ash powder batches were initially placed in the paper filter capsule and subsequently put in a glass extractor. The glass flask was filled with distilled water, which has mass ratio of water to fly ash of 10. The water was then boiled for 7 days. The water steam rising from the boiling water of the flask was condensed in a Liebig cooler. The condensing H 2O continuously dripped down into the filter capsule and filled the cylindrical glass container around the filter until a certain fluid level was reached, and the fluid drained into the boiling flask. Soluble components were thus quickly transported from the sample to the solution in the boiling flask. This process repeated itself periodically. At the end of extraction process, the flask device was cooled to room temperature and left for a further 24 hours. The resulting precipitated particles were filtered through a paper filter (Schleicher&Schuell no. 604) and centrifuged and dried at room temperature, while the water residues were evaporated at 50 OC for subsequent XRD and XRF analyses.

Figure 3.1 A Soxhlet of water-extraction apparatus. A: Liebig cooler, B: paper filter capsule, C: glass flask, and D: glass extractor.

53 CHAPTER 3

3.3 Pozzolanic Solidification Experiment

The reactant selected as pozzolanic activator was Ca(OH) 2 powder with p.a. grade supplied by Merck, Germany. Saturated solution of Ca(OH) 2 was prepared by dissolving Ca(OH) 2 powder in a glass beaker containing distilled water. A magnetic stirrer mixed the slurry at 300 rpm for 24 hours. The solutions were then filtered through a paper filter (Schleicher&Schuell no. 604) and centrifuged. Here, the saturated solution was to be used for subsequent solidification and stabilisation of fly ash specimens. Paste samples of cementitous materials were prepared by mixing fly ash with the saturated solution. For this purpose, two different kinds of fly ash were employed: (i) the raw fly ash, and (ii) the water-washed fly ash. The mass ratio of fly ash to the solution was adjusted to be 1: 3 and 1:10. For obtaining the homogenous slurry, a mixture of fly ash-saturated solution was stirred at a constant speed of 300 rpm for 5 hours. At the allotted time periods, the resulting slurry was placed in an open plastic container and then dried at room temperature holding for a period of time (from 7 to 28 days, and from 1 to 3 months). At the end of the pozzolanic solidification experiment, the dried cake products were crushed and stored within a plastic container until analysed.

3.4 Hydrothermal Processing The experimental synthesis was conducted using two different alkaline solutions of KOH and NaOH with p.a grade supplied by Merck, Germany. For this purpose, two different kinds of fly ash were examined: (i) the raw fly ash, and (ii) the water-washed fly ash. All of the hydrothermal experiments were conducted using the non-agitated batch hydrothermal process as follows; 15 grams of the fly ash powder batches were added to alkali solution (either KOH or NaOH) in a glass beaker, of which the liquid/solid (L/S) ratio was adjusted to be 10 L/kg. The mixture was placed in the glass beaker and further stirred using the magnetic stirrer at 300 rpm for 24 h. The hydrothermal reaction was conducted in a Teflon-line autoclave under an autogenous pressure. The reaction temperatures (90 O-180 OC), reaction times (2-7 days), and 0.5 M-2.5 M of alkali metal hydroxide solutions were varied as control parameters. After the hydrothermal treatments, the reaction mixtures

54 CHAPTER 3

were filtrated through a paper filter (Schleicher&Schuell no. 604) and centrifuged. The resulting slurries were dried at room temperature. The dried cake products were then ground and stored in airtight polypropylene containers until analysed.

3.5 Characterisation

3.5.1 Chemical Analysis Chemical composition analysis of the fresh MSWI residues, the washed fly ash and the solid residues from the Soxhlet extraction process was conducted by wavelength dispersive x-ray fluorescence (XRF) for all size fractions of the materials. Batches of 8 grams of the materials were mixed with 2 g of acetone as a binder. The mixture was finely homogenised in an agate bowl. The powder pellets were subsequently pressed at 20 Kg/cm 2 for 2 minutes and finally analysed by XRF for the all elements present.

3.5.2 Optical and SEM Microscopy The fresh bottom ash and fly ash particles were investigated by optical polarisation microscope (LEICA) and by (SEM) scanning electron microscope (A LEO DSM microscope) with EDX system. For this observation, the selected ash particles were placed onto aluminium stubs using double-sided adhesive carbon discs and then sputter-coated with gold. Powder fractions were also made into petrographic thin sections. In this case, the bottom ash and fly ash were initially sieved through 100 µm to separate out the coarse-grained intact particles from the fine-grained ash. Some thin sections contained large (<100 m), medium (<50 m) and fine (<20 m) particles. Subsequently, the particles were embedded into an inert epoxy resin, cut and mounted onto a glass slide with size of 48 x 28 x 1.0 mm, and polished with a final diamond powder grain size 25-30 µm. The polished sample was then carbon coated for subsequent optical and SEM examination. SEM analysis was also carried out on specimens obtained from the pozzolanic solidification and hydrothermal conversion of fly ash. In this case, the specimens were mounted onto aluminium stubs using double-sided adhesive carbon discs and sputter coated with gold.

55 CHAPTER 3

3.5.3 Electron Probe X-Ray Microanalysis (EPMA) Polished thin samples in the glass slide were also examined using the electron probe x-ray microanalysis (EPMA). EPMA was employed to determine the mineralogical phase compositions (solid solution, alloys, glasses etc) and the nature of minor phases. The chemical composition of individual particles on the polished section was determined with a Cameca SX-50 electron microprobe automated by a Sun workstation, operated in wavelength-dispersive (WDX) mode. Quantitative chemical analysis was performed at 15 kV and 2-10 nA using either a defocused electron beam of 5 or 10 µm spot size or a rastered beam at 20000 - 40000 times magnification. The conventional ZAF correction method (Z = atomic number factor, A=absorption factor, F= characteristic fluorescence correction) was employed and based on the peak-to- background method (P/U-ZAF correction). Elements with Z < 11 were not analysed, but oxygen content was determined by stoichiometry.

3.5.4 Analytical Method Qualitative evaluation All samples including products obtained by the pozzolanic solidification and hydrothermal conversion of fly ash were subjected to XRD analysis. Each sample was then ground for 10 min with a mortar and pestle. The powder of each specimen was loaded into an aluminium/plastic well mount and flattened and compacted with a glass slide. Data collection for phase identification and subsequent quantitative analysis was performed using CuK α monochromated radiation in a conventional Bragg-Brentano (BB) parafocusing geometry (a Philips MPD and Siemens D500 Diffractometer) using the following conditions: Cu-anode tube operating at 45 kV and 30 mA; unfiltered radiation; 2 O divergence slit for incident beam and 0.05 O receiving slit; diffracted beam graphite monochromator set to eliminate CuK β; NaI scintillation detector with PHA; goniometer range 2 θ = 5 – 85 O 2 θ; counting time of 1-10 s/step and step size of 0.020- 0.022. The optics of the Bragg-Brentano diffractometer is shown in figure 3.2.

56 CHAPTER 3

Figure 3.2 Geometry of the Bragg-Brentano diffractometer (after Siemens, 1986).

Identification of the diffraction peaks was conducted either using the peak finder feature or on screen in the Philips X’Pert Software (Philips Electronics N.V). The Philips software involves a set of program for use in evaluating the background, finding the diffraction pattern for significant peaks. The system involves also a search match program that can be used to help identify the possible crystalline phases in the ash samples. In the search procedure, peak positions were located automatically using minima in the second derivative of the diffraction traces. The procedure also involved the chemical elements exclusion based on the XRF data for a reference pattern. The search was initially based on the n strongest peaks of each phase and applied in a reiterative way in many automatic and manual cycles, assigning identified peaks from major phases first, then concentrating on remaining ones, reducing from 5 in steps down to 1, and increasing tolerances for peak positions and heights from step to step. Phase identification analysis was conducted by comparing all diffraction peaks obtained from

57 CHAPTER 3

line pattern identifications with the reference pattern from the ICDD (2002) database on the powder diffraction file (PDF) sets 1-41. The peak positions and peak heights were checked against the entries in the ICDD. The final step was to select from the “candidates ” lists of minerals those which had a high score and match the peaks/features of the measurement. However, these conventional search match procedures reach their limitations if peak overlaps occur and when the lattice parameters shift in a phase due to solid solution members not present in the ICDD database.

Rietveld analysis of mineralogical phase composition Conventional x-ray diffraction (XRD) analysis provides only qualitative mineralogical composition, whereas quantitative data are essential for clarifying the environmental significance of the MSWI residue. This can be achieved by the Rietveld pattern-fitting (Rietveld, 1969). The calculated intensity for point ‘i’ of the diffraction pattern ( yc ) for a thick specimen, with Bragg-Brentano diffractometer optics is described by the following formula (Warren, 1969):

 κ  2  C   y =  ∑ L  F ϕ ()2θ − 2θ  k P  + y (3.1) ci  µ  jk  jk  ijk ik jk  2  jk bi   jk  Vk   where j and k are indices representing the Bragg peak j of the phase k, respectively; κ is µ an instrumental constant; is the linear attenuation coefficient of the mixture; L jk is the θ Lorentz-polarization-multiplicity factor (a function of Bragg angle ); Fjk is the structure factor-a function of atomic scattering factors, positions, site occupancies and vibrational parameter; ϕijk is the profile spread function for Bragg peak ‘jk’ and function of optical design, crystallite size and residual strain; 2 θik -2θjk is the angular separation θ of point ‘i’ from the Bragg peak (angle 2 k); Ck is the volume fraction for phase k; Vk is the corresponding unit cell volume of phase k; Pjk is the preferred orientation (PO) factor for peak ‘jk’ and ybi is the intensity of background at point i. For computational convenience, the calculated intensity at each step ‘i’ can be expressed as:

= 2 ϕ + yci ∑Sk L jk Fjk ijk Pjk ybi (3.2) jk

58 CHAPTER 3

where Sk is the scale factor for phase k according to the equation: κ C S = k (3.3) k µ 2 Vk Equation 3.3 is employed in the Rietveld program used in the study-see below. The Rietveld refinement procedure is based on least squares minimisation of the residual (R) defined by:

n = ()− 2 R ∑ w i yi yci (3.4) i=1 where w i represents the weighting factor, yi is the measured intensity and yci is the calculated intensity for measurement i.

In general, the calculation of yci involves two groups of parameters. The first group comprises the structural parameters: the overall isotropic temperature parameter,

B; the fractional co-ordinates of the i the atom in the asymmetric unit, Xi Y i Z i ; the atomic temperature parameters, Bi , assuming isotropic vibrations; and the occupation number, ni . The second parameter group consists of instrument/materials parameters. These θ are (i) the zero position of the detector, 2 i ; and (ii) for each phase k, the scale factor,

Sk ; the half-width parameters, Uk ,V k ,W k ; the peak asymmetry, As; the preferred orientation parameter, Pjk ; and (iii) for each phase, the coefficients in the d-spacing expression:

1 = Ah 2 + Bk 2 + Cl 2 + Dkl + Ehl + Fhk (3.5) d 2 where h k and l are the Miller indices. During the course of the refinement, the six coefficients are converted to reciprocal cell parameters, and the real-space parameters, according to A=a*.a*, B =b*.b*, C=c*.c*, D=2b*.c*.sin α*, E= 2c* a*sin β* and F=2 a* b* sin γ* where a*, b*, c*, α*, β*, and γ* are the reciprocal cell parameters. The figures-of-merit used to evaluate the agreement between the observed and the calculated pattern during the course of Rietveld refinements are (Young and Wiles, 1982):

59 CHAPTER 3

  − 1 () ∑ yi (obs )  yi cal S  The profile R-factor, R = k (3.6) p () ∑ yi obs

2 2/1     () −  1  () ∑ yi obs  yi cal    Sk   The weighted profile R-factor R = (3.7) wp {}() 2 ∑ w i yi obs

2/1 ()N − P The expected R-factor, R = (3.8) exp 2 ∑ w i yi

∑ I − I = ko kc The Bragg R-factor for each phase R B (3.9) Iko where Iko is the observed integrated intensity; Ikc is the calculated intensity; N is the number of profile measurements; and P is the number of parameters refined.

The goodness-of-fit (GOF), an additional figure-of-merit, is defined as:

2  R  GOF =  wp  (3.10) R exp  The success and completeness of the refinement is adjudged in part by the value of GOF which ideally is unity (Prince, 1993). However, in practice, GOF may vary over a rather wide range for a variety of reasons, for example, due to an inadequate peak shape function, the complex character of the texture, or finally, the incorrect structure solution. Hence, the quality of each refinement may be further assessed by the agreement between measured and calculated data on difference plots whereby the difference between the two patterns should essentially be flat for a well-refined model (see figure 3.3 for a representative Rietveld refinement plot).

60 CHAPTER 3

Figure 3.3 Quality of the Rietveld pattern-fitting results for the aged BA-A1 bottom ash where ( ······ ) and ( −−−−) denote the observed and calculated patterns, respectively, along with the associated difference plot.

In this study, the Rietveld refinements were performed using the SIROQUANT Quantitative XRD software program as developed by Taylor and Clapp (1992). The SIROQUANT contains an extensive data base of crystal structure models from which the full diffraction profiles are calculated, additional structure models were taken from the inorganic crystal structure database (ICSD, 1999). The parameters refined were: (i) the 2 θ scale zero position, (ii) the phase scale factors, (iii) the lattice parameters, (iv) the peak asymmetry and peak shape functions. SIROQUANT fits the diffraction line widths (FWHM) as a function of tan( θ) using the u-v-w formula of Caglioti et al . (1958), for phases with less than 1 weight-% only the w parameter was fitted, while u and v were fixed to the values of the measured quartz, unless more than three significant isolated peaks were available to constrain the u, v, w values. Preferred orientation was assumed to be absent in the powder sample and in all cases, atomic positional parameters were kept fixed to the literature values. To ensure correct reference data, Rietveld refinements were also conducted on various pure phases and synthetic phase mixtures.

61 CHAPTER 3

Two profile functions in the SIROQUANT software package are available; the Pearson-m line shape and pseudo-Voigt shape parameter. The pseudo-Voigt profile function was employed in this study as it facilitates relatively rapid computations and is generally regarded as providing acceptable fits to measured line profile shapes (Young and Wiles, 1982). The levels of mineralogical phases including the glass content in the sample were determined by XRD Rietveld phase analysis using an internal standard method (Hill and Howard, 1987) and the SIROQUANT quantitative XRD analysis software. The strategy adopted for obtaining weight fractions of the crystalline phases and the amorphous fraction makes use of diffractograms collected from a mixture of the ash with 10 wt. % of CeO 2 (NIST SRM 674) added as internal standard. The weight fraction wk of a crystalline phase k in the system is derived from the known weight fraction w s of the phase added as internal standard by the relation: φ S w = k k w (3.11) k φ s s Ss where S k and S s are the phase scale factors matching the calculated to the observed intensity, and φk and φs are the product of the number of formula units per unit cell, the mass of the formula unit, and the unit cell volume for each phase. The CeO 2 powder was assumed to be fully crystalline. The amorphous level or glassy phase content was then determined by the difference between the sum of the weight fractions wk for all phases and 100 %.

3.5 Leaching Experiment Leaching tests were performed according to the standard of the toxicity characteristic leaching procedure (TCLP) of the Environmental Protection Agency (U.S. EPA, 1990) (EPA-SW 846 Method 1311). 5 grams of the powder batches were initially ground to particles less than 100 µm in size. In this method, the ground sample was subsequently extracted by using an acetic acid solution (CH 3COOH supplied by J.T.Beaker, Holland) mixed with distilled water. Here, the acetate buffer was added only once in order to have a solution with pH of 3, at the start of the extraction. The tested sample was first placed in a glass beaker containing the acetic acid solution and stirred in the magnetic stirrer at 300 rpm. The solution with a liquid-to-solid mass (L/S)

62 CHAPTER 3

ratio of 20:1 (ml/g) was selected and the time of extraction was 18 ± 3 h. After the leaching process, the leachants were filtered through a paper filter (Schleicher&Schuell no. 604) and pH was again measured prior to conducting chemical analysis. The concentrations of heavy metals in the leachants were determined by inductively coupled plasma mass spectroscopy (ICP-MS), while the dried filter cakes of the leached ash were then examined by XRD.

63 CHAPTER 4

CHAPTER 4 CHARACTERISTICS OF MSWI RESIDUES

This chapter presents the chemical and mineralogical characteristics of the MSWI residues (bottom ash and fly ash) with particular reference to bulk and individual ash particle compositions, mineralogical reaction of the aging process and pre-treatment processing of the fly ash. Furthermore, the pre-treatment procedure of the fly ash is examined in the context of enhanced chemical properties and extraction of heavy metals for subsequent stabilisation processes.

4.1 Characterisation of MSWI Residues In this section, the characteristics of the MSWI solid residues investigated in this study and the chemical and physical behaviour during incineration are described along with the formation of mineral assemblages.

4.1.1 Bulk-Chemistry Chemical compositions of the MSWI residues were analysed for major and trace elements. The bulk chemical compositions of the raw bottom ash and fly ash powders were determined by XRF according to the procedure described in section 3.5.1. The compositions are presented in the form of element weight percents, independent of the actual form of chemical binding in the residues. The light elements oxygen and carbon were not analysed. However, the cations are mainly bound in oxidic compounds (oxides, silicates and sulfates). Table 4.1 presents the chemical compositions of the bottom ash and fly ash collected in the incinerator plants A and B. The sum of the major components and trace element concentrations for these materials is notably less than 100 %, because the remainder of the materials may comprise largely water and carbon dioxide (which were not analysed for). Bottom ash samples are similar in compositions between two different incinerators. Additionally, fly ash samples from two different incinerators showed a relatively small variation in chemical compositions, except for the concentrations of the chlorine and certain hazardous heavy metals ( e.g., Pb and Zn).

64

CHAPTER 4

Table 4.1 Chemical composition of the MSWI solid residues. Element Bottom ash Fly ash wt.% BA-A1 BA-A2 BA-B1 FA-A1 FA-A2 FA-A3 FA-A4 FA-B1 Si 22.38 18.20 19.98 5.63 9.30 12.09 5.52 12.39 Ca 9.73 10.60 10.06 12.41 11.42 12.76 6.10 16.45 Fe 8.92 5.37 3.94 3.71 2.41 3.18 1.97 1.94 Al 6.36 4.19 5.70 2.43 3.55 4.36 2.07 3.68 Na 3.35 2.44 2.92 10.38 6.40 5.20 10.12 4.44 K 0.85 1.20 0.86 6.11 5.57 4.49 7.73 4.96 Mg 1.47 1.16 1.04 1.08 1.05 1.19 0.80 1.42 Ti 0.81 0.98 0.44 1.02 1.24 1.34 0.60 1.05 Mn 0.14 0.10 0.07 0.12 0.09 0.11 0.06 0.10 P 0.39 0.46 0.30 0.40 0.58 0.62 0.48 0.55 Cl 0.32 0.54 0.34 8.32 7.03 5.86 7.50 6.75 Pb 0.10 0.43 0.17 1.36 1.39 1.05 2.36 0.71 S 0.49 1.30 1.13 4.11 5.58 4.61 7.95 4.27 Zn 0.77 0.72 0.80 4.91 5.29 3.92 8.23 2.26 ppm As 21 101 52 307 477 389 744 252 Ba 3926 3754 1731 3470 2183 2455 1847 1404 Bi 204 355 306 1473 Cd 14 17 6 456 506 358 881 354 Ce 51 75 47 6 Co 67 63 62 262 37 42 51 31 Cr 1158 1602 605 2026 1308 1606 1486 895 Cs 13 15 37 114 271 185 370 97 Cu 7743 3118 2300 3513 3882 3940 6224 1071 F 439 540 530 3042 3351 2637 1928 2992 Ga 5 3 30 27 20 43 29 Gd 2 8 6 9 13 11 20 9 Hf 42 49 21 34 70 66 54 23 Mo 99 76 12 489 108 107 259 19 Nd 16 28 24 24 32 32 34 30 Ni 356 223 143 614 223 232 380 98 Pr 5 9 3 6 15 20 3 6 Rb 32 45 36 265 317 226 343 276 Sb 70 104 42 1630 899 602 2401 723 Sc 22 33 30 57 47 45 67 51 Sm 5 7 10 Sn 42 47 38 333 328 251 936 192 Sr 403 307 378 542 281 311 182 408 V 76 82 50 86 47 62 43 51 Y 79 319 120 885 914 703 1544 467 Yb 5 8 8 15 13 13 21 8 Zr 310 1100 262 232 1599 1440 137 239 *BA-An; FA-An and BA-Bn; FA-Bn stand for the samples of bottom ash and fly ash collected from the incinerator plants A and B, respectively (see section 3.2).

65

CHAPTER 4

For the bottom ash, it can be seen that the major constituents (>1 wt. %) in the samples collected from incinerator A and B are Si 4+ , Ca 2+ , Fe 2+, 3+ , Al 3+ , Na +, and Mg 2+ , while minor concentrations of Cl -, S n+ , K +, Ti 4+ , Zn 2+ , Cu 2+ , Ba 2+ , P 5+ , and Pb 2+ are observed. There is a clear chemical distinction between the bottom ash samples from incinerator A and those from incinerator B. The bottom ash from incinerator B contained less Fe 2+,3+ and Ti 4+ compared with bottom ash from incinerator A. Furthermore, the major chemical constituents of the bottom ash (with large Fe-metal scraps removed) are considered to be similar to those of the earth’s crust, i.e ., Si 4+ , Al 3+ , 2+,3+ 2+ 2+ + + 4+ n+ 5+ Fe , Mg ,Ca , Na , K and three minor elements (Ti , Mn , P ). The significant enrichment of the major elements Fe and Ca compared to those of the earth’s crust is evident for example in the bottom ash of sample BA-A1 (Figure 4.1).

30

25 Bottom ash Earth's crust

20

15

10 Weight (%) Weight

5

0 Si Al Fe Mg Na Ca Chemical Element

Figure 4.1 Major chemical elements of bottom ash (BA-A1) in comparison to the element abundance in the Earth’s crust (Rudnick and Fountain, 1995).

66

CHAPTER 4

Chemical differences between bottom ash and fly ash were noticeable in all samples collected from the incinerators A and B. The concentrations of Cl, Na, K and S, which belong to the volatile elements were increased in the fly ash while the less volatile elements such as Si, Al and Fe were enriched in the bottom ash. Obviously, the elements S and Cl accounted for a large percentage in the fly ash samples. Differences in the chemical compositions from bottom ash to fly ash are clearly related to the factors described in section 2.2 (refer to chemistry ). Concentration of Si (about 20 wt.%) in the bottom ash of incinerators A and B was much higher than Si in the fly ash (about 10 wt.%). In particular, the content of Fe (5-8 wt.%) in the bottom ash of incinerator A was higher than the bottom ash of incinerator B. Moreover, the fly ash samples collected from incinerators A and B have different Ca contents. The difference in Ca content may be attributed to the treatment process for the flue gas, where lime-water is injected into a unit of the acid gas treatment to reduce air pollution (Forestier and Libourel, 1998; Li

- 2− et al ., 2004). The contents of Cl and SO 4 in the fly ash may be also linked to the efficiency of the flue gas treatments to neutralise acid gases such as HCl and SO 2 (Forestier and Libourel, 1998; Li et al ., 2004). In the present study, the chemical compositions of German MSWI residues are comparable with the data presented by Hjelmar (1996) (see table 2.1 in section 2.2). It is reported in table 4.1 that the Cl - content ranges from 5 to 8 wt. % in the fly ash, whereas less than 1 wt.% of Cl - is contained in the bottom ash. Data of Hjelmar (1996) revealed that the Cl - contents vary from 4.5 to 10.1 wt. % for the fly ash and from 0.12 to 0.32 wt. % for the bottom ash, respectively. The high chlorine content is considered as the typical characteristic of MSWI fly ash, in which 40-50 wt. % of Cl - originate from the incineration of plastic materials, mainly PVC (Forestier and Libourel, 1998). Concentrations of heavy metals in the MSWI solid residues are shown in table 4.1. The elements Cd, Pb, and As are more likely to be volatile at the incinerator temperature and therefore are enriched in the fly ash (Eighmy et al. , 1995). However, Hg was not found in the present study suggesting that it may have not been caught in the particles precipitated in the electric filter, and will reach the subsequent stages of gas cleaning. The concentration of Pb obtained in the present study ranges from 0.7 to 2.3 wt.%, which is somewhat higher than data of Hjelmar (1996) (see table 2.1 in section 2.2, 0.74 and 1.9 wt. %). Furthermore, the concentrations of some toxic heavy metals in

67

CHAPTER 4

the fly ash are higher than in the study of Hjelmar (1996), such as Cd (0.04-0.09 wt.%) and As (0.03-0.07wt.%). The concentrations of trace elements Sr, As, Co, V and Mo are below 1 wt.%, which is in good agreement with the work of Forestier and libourel (1998) and Hjelmar (1996). In general, concentration levels of heavy metals found in the present study are similar to the results provided by other studies (Forestier and Libourel, 1998; Hjelmar, 1996), but there are also differences in the concentrations of Cr and Cu, which may be due to the different waste compositions. Further, Ni is a non-volatile heavy metal, which was detected both in fly ash and bottom ash. Elements Cu and Cr, according to their lithophile behaviour, remained at values of about 0.06-0.77 wt.% both in bottom ash and fly ash. The greatest fractionation into the fly ash was found for Zn which belongs to the moderately volatile elements (2-8 wt.%).

4.1.2 Particle Chemistry and Morphology The incineration process has strong effects on particle chemistry, and morphology of the MSWI residues. In this regard, particle morphology and size are important aspects that control the leaching behaviour of the residues. A variety of particle shapes and chemical compositions of the MSWI residues was observed by SEM equipped with EDX according to the procedure described in section 3.5.2. Comparison of the different bottom ash samples from two incinerators A and B revealed that most bottom ash particles were polycrystalline with large grains set in a very fine-grained matrix. This may result from the different nature of MSW burn in the incineration system (Chang et al ., 2001; Song et al ., 2004). Figures 4.2a and b display typical SEM photographs for the bottom ash and fly ash from incinerator A. The majority of the bottom ash particles were angular-shaped, porous aggregates < 2 mm in diameter, with a flaky morphology. Spherical particles were not observed in the bottom ash. The angular and/or porous fragments in the bottom ash may represent inorganic components that were partially or completely melted and then become porous due to the release of trapped gases. The most common elements identified by EDX in the particles are Ca, Si, Al, Fe and Mg, in agreement with the bulk compositions determined by XRF. An oxygen K α-peak was frequently present in EDX spectra of particles, whereas Cl and C K α-peaks were uncommon. Furthermore, Mg and Ti were commonly and

68

CHAPTER 4

small levels of Mn or Cr were occasionally detected in association with Fe-rich oxide particles. Fly ash particles collected from two incinerators A and B have various forms. A wide range of particle shapes, including spheres, prisms and needles was observed in the fly ash. However, most fly ash particle shapes were spherical. A porous texture was also observed in some these spheres. Figure 4.2b shows fly ash particles of the FA-A1 sample collected from incinerator A. Particles in the size range of 1-20 µm were frequently observed in the sample and their surfaces were coated or agglomerated with very fine-grained particles less than 0.1 µm in diameter. These spherical glassy particles represent melt droplets that have formed during incineration of the municipal solid waste (Eighmy et al ., 1995; Kirby and Rimstidt, 1993). Differences in both particle size and morphology exist in the fly ash collected from three different electrostatic precipitator units in the incinerator A (FA-A2, FA-A3 and FA-A4). The average particle size decreased from 20 µm in FA-A2 to approximately 1 µm in FA-A4. The FA-A4 sample lacked spherical glassy particles. It will be shown in section 4.1.3 that evidence for the presence of an amorphous phase in the FA-A4 sample was not found by the quantitative XRD method. The most common elements in the fly ash particles i.e ., Si, Al, Fe, K, O, Zn, Cl, Ca and S were frequently observed in SEM/EDX spectra, in agreement with the XRF analyses, verifying that salts are largely present in the fly ash. The most common compounds were Fe and Ca silicates and aluminates. Larger spheres (>10 µm) were commonly composed of two or more phases (glass and crystalline phase). The sphere shown in figure 4.3a consists of a Al 2O3/SiO 2 matrix with varying proportions of either Fe and/or Ca. There are also K and Zn present in this sphere.

69

CHAPTER 4

(a)

(b)

Figure 4.2 SEM micrographs of particles for (a) bottom ash (BA-A1) and (b) fly ash (FA-A1) collected from incinerator A. Note the difference in scale for the two micrographs.

70

CHAPTER 4

(a)

(b)

Figure 4.3 (a) SEM micrograph of the FA-B1 sample collected from incinerator B: a spherical particle is visible; on its surface some small aggregated particles are present, and (b) EDX spectrum of the spherical particle shown in figure 4.3a.

71

CHAPTER 4

4.1.3 Mineralogy of MSWI Residues The mineralogical characteristics of the residues were examined to identify: (i) the major phases controlling the leaching chemistry, and (ii) the host phases of toxic elements which directly associate with the release of the elements. Initially, the classical (albeit computerised ) XRD search-match method described in section 3.5.4 was employed to explore the presence of crystalline phases by the diffraction lines which (i) should be clearly resolved from diffraction peaks of other phases and (ii) well defined in terms of counting statistics. Figure 4.4 represents the XRD patterns of the samples of bottom ash and fly ash collected from two incinerator plants A and B. Each peak has been identified by the search-match program and labelled with the mineralogical phase name according to the ICDD-PDF (2002). In the freshly quenched bottom ash, the main crystalline phases are anhydrite (CaSO 4) (PDF#86-2270), calcite (CaCO 3) (PDF#86-2334), corundum ( α-

Al 2O3) (PDF#83-2080), ettringite (3CaO ·Al 2O3·3CaSO 4·32H 2O) (PDF#72-0646), gehlenite (melilite group) [(Ca,Na) 2(Mg,Fe,Si,Al) 3O7](PDF#89-5917), hydrocalumite

[Ca 8Al 4(OH) 24 (CO 3)Cl 2 (H 2O) 1.6 (H 2O) 8] (PDF#78-2051), magnetite (Fe 3O4) (PDF#89-

3854), quartz ( α-SiO2) (PDF#88-2487), and rutile (TiO 2) (PDF#89-4920).

The principal minerals in the fresh fly ash are alunite [NaAl 3(SO 4)2(OH) 6]

(PDF#89-3952), anhydride (CaSO 4) (PDF#86-2270), boehmite [Al(OH) 3] (PDF#83-

2384), calcite (CaCO 3), caracolite [Na 3Pb 2(SO 4)3Cl] (PDF#73-1935), cristobalite (SiO 2)

(PDF#89-3606), gehlenite (melilite group) [(Ca,Na) 2(Mg,Fe,Si,Al) 3O7)], halite (NaCl)

(PDF#88-2300), hematite (Fe 2O3) (PDF#89-2810), hydrocalumite [ Ca 8Al 4(OH) 24 (CO 3)Cl 2

(H 2O) 1.6 (H 2O) 8], lazurite (sodalite) [Na 5Ca 2Al 6Si 6O24 (SO 4)2] (PDF#77-1703), lime (CaO)

(PDF#82-1690), minium (Pb 3O4) (PDF#89-1947), (K 2ZnCl 4) (PDF#86-0166), quartz 2+ (α-SiO 2), rutile (TiO 2), sylvite (KCl) (PDF#89-3619) and ulvöspinel (Fe 2 TiO 4) (PDF#71-1141). However, many mineral phases could not be reliably identified due to the large number of overlapping peaks and due to shifts of peak positions in solid solution members which are not contained in the ICDD-PDF (2002). The remaining peaks could be assigned to some phases only with a high level uncertainty. The computerised search-match routine was therefore employed as a first step in the mineral identification.

72

CHAPTER 4

(a)

(b)

Figure 4.4 XRD patterns of the fresh (a) bottom ash and (b) fly ash. BA-An, FA-An and BA-Bn, FA-Bn refer to the bottom ash and fly ash collected from the incinerator plants A and B respectively (see section 3.2). The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Bo (boehmite), C (calcite), Co (corundum), Cr (cristobalite), Ct (caracolite), E (ettringite), Fe (iron), G (gehlenite), He (hematite), Hc (hydrocalumite), Hl (halite), L(lime), Lz (lazurite), M (magnetite), Mi (minium), Pz (K 2ZnCl 4), Q (quartz), R (rutile), S (sylvite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

73

CHAPTER 4

The ‘ candidate and/or probable ’ minerals found in the search-match program were subsequently judged by the full profile Rietveld refinement, as peaks of phases which have been missed in the search match or mistakenly assigned phases clearly stand out in the difference curve of the calculated and the measured diffraction profile (Prince, 1993; Rietveld, 1969; Winburn et al ., 2000). In the present study, the adopted Rietveld procedure was limited to fixed structural and thus compositional models of the minerals in calculations, because it could not correct easily for the variation of structure factors as a function of composition in the solid solutions. Even if the elemental compositions were exactly known, many degrees of freedom would remain for the variation of positional co- ordinates and occupation factors. Any "exact" approach requires the synthesis of the phases with the chemical composition and particle size distribution and morphology encountered in the material AND a reliable single crystal structure analysis. However, the changes of the results expected in doing so would be too small to affect our conclusions in this study. Further, the inhomogeneous composition of the complex multiphase of the MSWI residues (Figure 4.4) made phase identification more difficult. Therefore, a fractionation scheme was developed for the bottom ash and fly ash, where magnetic ash particles were separated by an electromagnet and water soluble salts were removed by washing treatment as described in section 3.2. The fractionation method was essential for identifying the mineral phases in the MSWI residues by XRD, because of providing the reduction in the number of overlapping peaks. XRD patterns of synthetic pure phases or minerals from the mineral collections, as well as synthetic mixtures were also used for comparison.

Mineralogy of bottom ash XRD phase analysis was performed on powdered BA-A1 samples to obtain an estimate of the changes in phase content. Four specimens, namely freshly sampled, aged (6 months-old), water-treated and magnetically separated bottom ash had been prepared according to the procedure described in section 3.2 and were subsequently measured by XRD. XRD patterns of these four samples are presented in figure 4.5a. Phase abundances wt.% (crystalline and amorphous phases) were then estimated from powder

74

CHAPTER 4

XRD traces using the Rietveld method (see section 3.5.4). The strategy adopted made use of refinements of the diffraction data collected from samples (i) without internal standard and (ii) with internal standard for amorphous phase determination. For instance, the quality of refinements of the aged bottom ash may be gauged from the diffraction plot in figure 4.5b. The abundance of each phase in the BA-A1 sample calculated from the SIROQUANT program, along with amorphous phases are reported in table 4.2. The quantitative results are given in weight percent normalised to 100% including the amorphous fraction estimated with the aid of an internal standard. In addition to the amorphous constituent, the major mineralogical phases in the fresh sample, arbitrarily defined as those present with more than 3 wt.%, are quartz, magnetite, corundum, and sheet silicates close to illite and/or muscovite. The significant amount of amorphous phase may include small amounts of unidentified crystalline minerals. XRD analysis indicated that weathering reactions lead to alterations in the original mineralogy of the bottom ash. In the aged bottom ash, the major new phases were ettringite, hydrocalumite and calcite, with a significant amount of coquimbite (1.2 wt.%), while gypsum was found as a major phase in the water-treated bottom ash.

Reasonably, the formation of calcite may be related to the CO 2 uptake by the initially alkaline bottom ash. A small amount of crystalline portlandite [Ca(OH) 2] could be identified in both aged and water-treated samples. This phase is likely to form in both bottom ash samples because of a result of hydrolysis of calcium oxide (CaO) (Belevi et al ., 1992; Comans and Meima, 1994). Table 4.2 presents further additional phases which are present with less than 3 wt. % each. Some minerals, e.g., periclase (MgO) and marcasite (FeS 2), and metals such as Si are not detected in all samples; this may be due to the heterogeneity of the bottom ash particles. For the heavy magnetic fraction, XRD analysis provided a clear identification of magnetite, wüstite, and the weakly magnetic hematite. Surprisingly, no metallic iron was found. The spinel mineral may also contain elements such as Ca, Al, Ti, according to EDX analysis. Non-magnetic crystalline phases which were attached to the magnetic phases are anhydrite, augite , calcite, corundum, diopside, melilite, marcasite, plagioclase, quartz, silicon carbide in addition to amorphous phase.

75

CHAPTER 4

(a)

Figure 4.5a Diffractograms of (a) fresh, (b) 6-months storage, (c) water-treated and (d) magnetically separated bottom ash of BA-A1. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Co (corundum), Cq (coquimbite), E (ettringite), G (gehlenite), Gy (gypsum), He (hematite), Hc (hydrocalumite), Lp (lepidocrocite), M (magnetite), Q (quartz), R (rutile) and W (wüstite).

(b)

Figure 4.5b Quality of the Rietveld pattern-fitting results for the 6-months storage of the BA-A1 sample measured without internal standard. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern.

76

CHAPTER 4

Table 4.2 XRD-based mineralogy of the BA-A1 sample. The minerals are arranged in the following sequence (1) major minerals in the fresh bottom ash, (2) major new mineral after aging or water treatment, (3) minor phases (<3 wt.% each) in alphabetical order. Mineral Fresh Aged Treated Magnetic Crystal § (wt.%) Formula (simplified) Water Fraction Structure Reference Amorphous *42.2(7) 35.3(6) 30.1(9) 69.7(4) Quartz α-SiO 2 25.5(3) 9.6(0) 25.0(1) 1.6(0) S-1 Corundum α-Al 2O3 4.1(2) 4.4(1) 3.0(2) 0.7(1) S-40 Muscovite/illite KAl 3Si 3O10 (OH) 2 3.6(3) 2.1(1) 6.4(3) S-116 Magnetite Fe 3O4 3.4(1) 1.2(1) 3.5(2) 8.5(1) S-50

Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)Cl 2 0.6(1) 15.4(2) 0.6(1) S-778 (H 2O) 1.6 (H 2O) 8 Ettringite 3CaO ·Al 2O3·3CaSO 4·32H 2O 0.4(1) 6.6(1) 0.3(0) S-195 Calcite CaCO 3 0.6(2) 4.4(1) 4.1(1) 0.3(0) S-11 Gypsum CaSO 4·2H 2O 4.3(2) S-26

Alunite NaKAl 3(OH) 6(SO 4)2 0.6(2) 0.5(2) 1.9(2) S-459 Albite NaAlSi 3O8 1.8(2) 2.1(1) 2.5(2) 3.4(2) S-155 (plagioclase) Anhydrite CaSO 4 2.1(1) 1.4(1) 1.4(1) 0.6(1) S-80 Augite Ca 3Na 3Mg 3FeAl 1.6 Si 7O24 2.7(2) 1.5(1) 1.8(3) 3.2(2) S-622 Barite BaSO 4 0.4(0) 0.5(0) 0.7(1) S-297 Bicchulite Ca 2Al 2SiO 6·(OH) 2 0.2(0) 0.4(1) S-358 Chalcocite Cu 2S 2.6(1) S-278 3+ Coquimbite Fe 2(SO 4)3·9H 2O 1.2(1) 1.0(1) S-439 Covellite CuS 0.9(0) 0.9(1) S-455 Cristobalite SiO 2 0.3(0) 0.3(0) 0.5(1) S-37 Diopside CaMgSi 2O6 2.0(2) 0.4(2) 4.2(2) 3.2(2) S-133 Gehlenite (Ca,Na) 2(Mg,Fe,Si,Al) 3O7 2.4(1) 2.5(1) 1.8(2) 3.0(1) S-165 (melilite) Gibsite AlOOH 0.4(1) S-39 Hematite Fe 2O3 1.2(1) 1.6(1) 1.0(1) 0.7(1) S-41 Hercynite FeAl 2O4 0.4(1) 0.7(1) 0.5(1) 0.2(0) S-426 Lepidocrocite FeOOH 0.5(0) 0.5(1) S-45 Marcasite FeS 2 1.7(0) S-907 Olivine group (Mg,Fe) 2SiO 4 1.5(1) 1.2(1) 0.4(1) S-53 Periclase MgO 0.6(1) S-96 Portlandite Ca(OH) 2 0.2(0) 0.5(1) S-124 Rutile TiO 2 0.8(1) 0.9(2) 1.2(2) S-15 Silicon Si 0.3(0) S-256 Silicon carbide SiC 0.5(1) S-244 Sodalite Na 4Al 3Si 3O12 Cl 0.5(1) S-1031 Tenorite CuO 0.4(0) S-13 Tobermorite Ca 5(OH) 2SiO 16 ·4H 2O 2.5(2) 0.5(1) S-309 Wüstite Fe 0.83 O 0.8(1) 0.2(0) 0.5(0) 2.7(1) S-142 Total 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter S for the SIROQUANT database (2002) followed by the entry number of the respective database.

77

CHAPTER 4

When considering the acid neutralisation capacity (ANC) of bottom ash, a first comparison was made between the own pH of the bottom ash, i.e ., the pH values measured in a suspension of distilled water with no acid added. In the case of the fresh BA-A1 sample, the own pH ranged from 10.5 to 11.5. This indicates that, as suggested by previous studies (Comans and Meima, 1994; Meima and Comans, 1997), the suspension was strongly undersaturated in portlandite, which would control the equilibrium suspension pH at as a high value as 12.5. This result has been confirmed by XRD analysis indicating that no significant amount of the crystalline portlandite

[Ca(OH) 2] was present in bottom ash (Table 4.2). Moreover, the pH value of the suspension may be influenced by the formation of ettringite and an amorphous Al- hydroxide (including AlOOH or Al(OH) 3), which may lead to the pH of suspension in the range from 10 to 10.5 (Chimenos et al ., 2000; Polettini and Pomi, 2004). In naturally aged and water-treated bottom ash samples, the own pH was in the range of from 8 to 9. This can be reasonably well explained by the presence of a significant amount of calcite (CaCO 3) in equilibrium with atmospheric CO 2 which would theoretically control the suspension pH at a value of 8.25 (Polettini and Pomi, 2004). Additionally, neo-formation of, for example, gypsum (in the water-treated bottom ash) and gibbsite (in the aged bottom ash) in addition to ettringite contributed to the reduction of the own pH of the bottom ash (Chimenos et al ., 2000). Two more batches of the fresh bottom ash collected from incinerator plants A and B (BA-A2 and BA-B1) were measured by XRD in an attempt to document the variability in mineralogical composition. The results of mineralogical phase compositions by the Rietveld analysis are reported in table 4.3. Both bottom ash samples have equal amounts of amorphous and crystalline components. However, the amorphous contents shown in table 4.3 may also include small quantities of unidentified crystalline minerals. High glass contents in the bottom ash varying from 40 to 75 wt.% have also been reported by Zevenbergen et al . (1994) and Pfrang-Stotz and Schneider (1995). Both samples were mainly comprised of quartz, plagioclase, calcite, and muscovite/illite. Significant amounts of corundum and ettringite were detected in sample BA-B1; but only very minor ettringite was present in the BA-A2 sample. Halite was detected in both samples. Most crystalline phases are high-temperature constituents

78

CHAPTER 4

of the fresh bottom ash except for calcite, goethite and portlandite, which formed during aging. These secondary phases are likely to have formed immediately after quenching of the hot bottom ash and may further react under atmospheric conditions (Speiser et al ., 2000). Ettringite and hydrocalumite were also observed in the fresh bottom ash. However, they may take some time to form at room temperature rather than upon quenching and those phases may be further altered due to aging (Polettini and Pomi, 2004; Sabbas et al ., 2003). It will be shown in section 4.3.1 that mineralogical changes and formation of new phases occurred in the bottom ash samples (BA-A2 and BA-B1), when exposed for several months to atmospheric conditions.

Mineralogy of fly ash The mineralogy of fly ash was investigated by XRD analysis according to the procedure outlined in section 3.5.4. XRD patterns of (i) fresh fly ash FA-A1, (ii) aged (6 months-old) fly ash, (iii) water-treated fly ash, and (iv) magnetically separated fly ash (see section 3.2 for description of the samples) are shown in figure 4.6a. The quantitative phase analysis was subsequently carried out with the XRD Rietveld method. The quality of refinements of the water-treated FA-A1 sample, taken as an example, may be gauged from the diffraction plot in figure 4.6b. The intensity of most peaks is well represented in the calculated diffractogram. For the fresh FA-A1 sample, the main crystalline phases identified in figure 4.7a are anhydrite (CaSO 4), calcite (CaCO 3), a melilite group mineral close to gehlenite

[(Ca,Na) 2(Mg,Fe,Si,Al) 3O7], halite (NaCl), hematite (Fe 2O3), lime (CaO), potassium tetrachlorozincate (K 2ZnCl 4), portlandite [Ca(OH) 2], quartz ( α-SiO 2), rutile (TiO 2), and 2+ ulvöspinel (Fe 2 TiO 4). K 2ZnCl 4 is not known as a mineral and its identification in fly ash appears to be unusual. However, if K 2ZnCl 4 was included in the set of model phases, an excellent agreement between the observed and calculated pattern can be observed in the diffraction plot (see figure 4.7b), thus the identification is reliable. Some of the crystalline phases ( e.g ., anhydrite, halite, K 2ZnCl 4, melilite and quartz) identified here have also been reported as major constituents of fly ash by Eighmy et al . (1995) and Kirby and Rimstidt (1993).

79

CHAPTER 4

Table 4.3 XRD-based mineralogy of bottom ash samples. §Crystal Mineral (wt.%) Formula (simplified) BA-A2 BA-B1 Structure Reference Amorphous *50.9(9) 50.0(9) Anhydrite CaSO 4 0.7(1) 0.9(1) S-80 Albite NaAlSi 3O8 5.2(3) 5.2(2) S-155 (plagioclase) Augite Ca 3Na 3Mg 3FeAl 1.6 Si 7O24 0.6(3) S-622 Barite BaSO 4 0.7(1) 0.6(1) S-297 Calcite CaCO 3 5.2(1) 3.8(1) S-11 C3S Ca 3SiO 5 1.9(3) 0.9(3) S-128 C2S Ca 2SiO 4 1.0(3) 1.8(2) S-69 Corundum α-Al 2O3 4.3(2) S-40 Cristobalite SiO 2 0.4(1) 0.2(0) S-37 Diopside CaMgSi 2O6 2.7(3) 1.3(3) S-133 Ettringite 3CaO ·Al 2O3·3CaSO 4·32H 2O 0.7(2) 3.3(2) S-195 Gehlenite (Ca,Na) 2(Mg,Fe,Si,Al) 3O7 1.6(2) 1.6(1) S-165 (melilite) Goethite FeOOH 0.8(1) S-42 Halite NaCl 0.5(1) 0.5(1) S-109 Hauyne (sodalite) Na 6Ca 2 Al 6Si 6O24 (SO 4)2 0.4(1) I-024441 Hematite Fe 2O3 0.3(1) 1.1(1) S-41 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)Cl 2(H 2O) 1.6 (H 2O) 8 0.2(0) 0.2(0) S-778 Hercynite FeAl 2O4 0.8(1) 0.5(1) S-426 Kalsilite KAlSiO 4 2.3(2) 0.5(1) S-961 Marcasite FeS 2 1.8(1) 0.7(1) S-907 Magnetite Fe 3O4 1.3(1) 0.3(1) S-50 Muscovite/illite KAl 3Si 3O10 (OH) 2 4.1(4) 3.2(3) S-116 Nepheline KNa 3Al 4Si 4O16 1.1(2) S-89 Olivine group (Mg,Fe) 2SiO 4 0.9(2) S-53 Portlandite Ca(OH) 2 0.4(1) S-124 Quartz α-SiO 2 11.4(2) 12.4(0) S-1 Rutile TiO 2 0.4(1) S-15 Sanidine KAlSi 3O8 2.9(3) 2.4(2) S-21 Silicon Si 0.3(0) 0.3(0) S-256 2+ Ulvöspinel Fe 2 TiO 4 0.3(0) S-362 Wollastonite (Ca,Fe)SiO 3 2.2(2) S-262 Total 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter (S for the SIROQUANT database, I for the ICSD database) followed by the entry number of the respective database.

80

CHAPTER 4

(a)

Figure 4.6a XRD patterns of (a) fresh, (b) aged (c) water- treated, and (d) magnetically separated fly ash. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite) An (anhydrite), Ba (bassanite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), L (lime), M (magnetite), Po (portlandite), Pz (K 2ZnCl 4), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Sy (syngenite) and Us (ulvöspinel).

(a)

Figure 4.6b Quality of the Rietveld pattern-fitting results for the water-treated FA-A1 sample without internal standard. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern.

81

CHAPTER 4

(a)

Figure 4.7a XRD pattern for the fresh FA-A1 sample. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), G (gehlenite), Hl (halite), He (hematite), L (lime), Pz (potassium tetrachlorozincate), Q (quartz), Po (portlandite), R (rutile) and Us (ulvöspinel).

(b)

Figure 4.7b Quality of the Rietveld pattern-fitting results for the fresh FA-A1 sample without internal standard. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern.

82

CHAPTER 4

Mineralogical compositions of the FA-A1 sample were subsequently estimated by XRD Rietveld phase analysis using the internal standard method as described in section 3.5.4 (Table 4.4). The fresh FA-A1 sample contains 40 wt.% glass and 60 wt.% crystalline phases. Principal Ca-rich phases detected are anhydrite (7.7 wt.%), calcite

(1.5 wt.%), melilite close to gehlenite (3.3 wt.%), C 2S (2.3 wt.%) and other calcium- bearing minerals. Furthermore, corresponding to the large fractions of Na, K, Zn and Cl, significant amounts of K 2ZnCl 4 (9.3 wt.% ) and NaCl (4.5 wt.%) and a minor amount of caracolite [Na 3Pb 2(SO 4)3Cl] are present. Enrichment of chloride minerals in the fly ash has been shown to result from condensation of the hot flue gas by heterogeneous nucleation on particles (Eighmy et al ., 1995). Significant amounts of quartz and sanidine feldspar (> 1wt.%) were also detected. Lead minerals were minium (1 wt.%), caracolite (0.2 wt.%) and massicotite (0.2 wt.%). ZnS (0.5 wt.%) was observed as a second Zn-rich phase. The presence of Zn may be related to particles containing Zn 2+ -ions, which was derived from the vaporisation of a waste material ( e.g ., galvanised zinc) at the combustion zone.

Various oxides such as lime (CaO), cristobalite (SiO 2) and rutile (TiO 2) were also found in minor concentrations. The magnetic fly ash particles extracted by the magnetic stirrer (see section 3.5) (wt.% abundance in table 4.4) contain magnetite (21.0 %) and the weakly magnetic hematite (4.3 %). In addition, non-magnetic phases attached to the magnetic particles are glass constituent (56.7 %), melilite (4.9%), diopside-pyroxene (3.3 %), calcite

(2.3 %), C 3S (2.6 %) and quartz (1.4 %). Other minor minerals (<1 wt.%) detected in this fraction include cristobalite, olivine (fayalite), halite, lepidocrocite, minium

(Pb 3O4), rutile, and ulvöspinel.

The K 2ZnCl 4 detected in the fresh fly ash (Figure 4.6a) disappears in the aged and water-treated fly ash, which indicates that this phase was converted into more stable K-and Cl-bearing minerals. Aging also resulted in the formation of new minerals namely bassanite, hydrocalumite, syngenite, sodalite, and muscovite/illite. The

2− formation of Ca-Al-SO 4 -Cl-hydrate phases ( e.g., hydrocalumite) is likely initiated by dissolution of Ca-Cl-and Al-rich minerals. Gordaite and gypsum are detected only in the

2− water-treated material. Relatively high concentrations of Na, Zn, SO 4 and Cl in sample FA-A1 (Table 4.1) may result in gordaite formation. In particular, aging resulted in the

83

CHAPTER 4

formation of high numbers of crystalline phases in the naturally aged and water-treated samples, relative to those in the fresh one. A variety of minerals could be identified in the bulk FA-A1 sample, implying that the fly ash material exhibited a significant mineralogical variation between ash particles. Hence, the variability in mineralogical compositions and glass content may be attributed to the heterogeneity of fly ash particles. The heterogeneity can be shown by typical examples of minerals such as cristobalite and iron (Fe) that were not found in all fly ash samples and it was confirmed by SEM/EDX analysis. As far as the acid neutralising capacity (ANC) of the FA-A1 sample was concerned, a first comparison was made between the own pH of the material suspended in distilled water with no acid added. The fresh FA-A1 had a pH range of 10.5-11.5.

This pH value may be controlled by the presence of alunite [NaAl 3(SO 4)2(OH) 6], portlandite [Ca(OH) 2], and probably amorphous aluminium hydroxide [Al(OH) 3]. For the aged and water-treated samples, a number of reactions decreased the pH to 9.5-10.5.

In this state, buffering equilibria between calcite (CaCO 3) and atmospheric CO 2 under the influence of the Ca- and SO 4-bearing minerals ( e.g. , gypsum) are governing the chemical system (Sabbas et al ., 2003). However, interactions between the different solid phases present in the aged FA-A1 sample may also contribute to the strong variability in the pH value. Variability of mineralogical compositions in different batches of fresh fly ash samples collected at the incinerator plants A and B (FA-A2, FA-A3, FA-A4 and FA- B1) was also examined by XRD. The XRD patterns of these samples and the representative Rietveld refinement plot of the FA-A3 sample are shown in figure 4.8. Mineral abundance in the fly ash samples is reported in table 4.5. Most of the fly ash materials have a significant content of amorphous phase (40-60 wt.%) with exception of the FA-A4 specimen. This result was confirmed also by SEM analysis. Despite their different bulk chemistry and geographical origin, the mineralogical composition of all samples was similar. The main difference lied in the proportions of Ca-, Na- and K–rich chlorides and sulfate compounds. Fly ash samples from the incinerators A and B typically had large fractions of Na, K and Cl, corresponding to significant amounts of halite (NaCl) and sylvite (KCl), except for FA-A4 in which only halite was present.

84

CHAPTER 4

Table 4.4 XRD-based mineralogy of the FA-A1 Sample. Water Magnetic Crystal Mineral Formula Fresh Aged Treated Fraction Structure (wt.%) (simplified) Model§ Amorphous 44.2(6)* 45.3(6) 35.2(9) 56.7(6) Alunite-natro NaKAl 3(OH) 6(SO 4)2 1.8(1) 0.9(1) 1.2(1) S-459 Anhydrite CaSO4 8.5(1) 5.6(1) 8.3(1) 0.7(1) S-80 Albite (plagioclase) NaAlSi 3O8 1.4(1) 1.6(2) 1.3(1) S-155 Apatite Ca 5(PO 4)3OH 2.3(2) 1.3(1) 1.1(2) S-215 Bassanite CaSO 4·0.5H 2O 1.3(1) S-292 C2S Ca 2SiO 4 2.2(2) 1.4(1) S-69 C3S Ca 3SiO 5 2.6(2) 2.1(2) 1.8(2) 2.6(2) S-128 C3A Ca 3Al 2O6 0.9(1) 1.3(1) 0.7(1) S-71 Calcite CaCO 3 1.5(1) 1.4(1) 2.2(1) 2.3(2) S-11 Caracolite Na 3Pb 2(SO 4)3Cl 0.2(0) 0.2(0) 0.2(0) I-024459 Cristobalite SiO 2 0.4(0) 0.4(0) 0.4(1) 0.2(1) S-37 Diopside CaMgSi 2O6 2.3(2) 1.5(2) 1.6(2) 3.3(2) S-133 Digenite CuS 0.2(0) 0.2(0) S-562 Dolomite CaMg(CO 3)2 1.1(1) 0.6(1) 1.2(2) S-31 Enstatite (Mg,Fe) SiO 3 2.3(1) 2.5(2) 2.1(2) S-151 Garnet Ca 3(Al,Fe) 2 (Si,P) 3O12 0.4(1) 0.6(1) 0.5(1) S-552 Gehlenite (melilite) (Ca ,Na) 2(Mg,Fe,Si,Al) 3O7 3.3(1) 4.4(2) 8.1(4) 4.9(2) S-165 Gypsum CaSO 4·2H 2O 1.3(1) S-26 Gordaite NaZn 4(SO 4)(OH) 6Cl(H 2O) 6 1.2(1) I-406090 Halite NaCl 4.6(1) 5.3(0) 6.2(1) 0.3(1) S-109 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)Cl 2 0.6(1) 1.8(2) S-778 (H 2O) 1.6 (H 2O) 8 Hematite Fe 2O3 0.5(1) 0.6(0) 0.8(1) 4.3(1) S-41 Hercynite FeAl 2O4 0.6(1) 0.4(1) S-426 Iron Fe 0.4(0) S-140 Kalsilite KAlSiO 4 0.4(1) 0.4(1) 0.4(0) S-961 Lepidocrocite FeOOH 0.3(0) 0.5(1) S-45 Lime CaO 0.3(0) 0.3(0) S-150 Magnetite Fe 3O4 0.5(1) 0.9(1) 1.0(1) 21.0(2) S-50 Massicotite PbO 0.2(0) 0.3(1) S-809 Muscovite/Illite KAl 3Si 3O10 (OH) 2 0.6(1) S-116 Minium Pb 3O4 1.0(0) 0.2(0) 0.4(0) 0.3(0) S-713 Nepheline KNa 3Al 4Si 4O16 1.4(1) 0.8(1) S-89 2+ Olivine group (Mg,Fe )2SiO 4 0.5(2) S-53 Portlandite Ca(OH) 2 0.2(0) S-124 Potassium – K2ZnCl 4 9.3(2) I-080861 tetrachlorozincate

Quartz α-SiO 2 2.1(0) 2.9(1) 2.3(0) 1.4(1) S-1 Rutile TiO 2 0.6(1) 0.6(1) 0.6(1) 0.6(1) S-15 Sanidine KAlSi 3O8 1.3(0) 2.1(1) 1.6(3) S-21 Sodalite Na 4Al 3Si 3O12 Cl 0.3(1) S-1031 Sylvite KCl 0.6(0) S-164

Syngenite K2Ca(SO 4)2·H 2O 9.6(2) 14.0(3) S-325 Tobermorite-14Å Ca 5(OH) 2Si 6O16 ·4H 2O 0.3(1) 0.6(1) S-309 2+ Ulvöspinel Fe 2 TiO 4 0.7(0) 0.8(0) 0.9(1) 0.4(1) S-362 Wurtzite ZnS 0.5(0) 0.3(0) 0.3(0) S-703 Total 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter (S for the SIROQUANT database, I for the ICSD database) followed by the entry number of the respective database.

85

CHAPTER 4

Calcite, quartz, rutile and ulvöspinel were present in all ash samples from incinerator A. Likewise, the higher concentration of Si in the FA-B1 sample resulted in a higher quartz content (4 wt.%). The aluminosilicate minerals plagioclase (albite-anorthite) and melilite (gehlenite) were observed in all fly ash samples. The only sulfide mineral found in the FA-A4 and FA-B1 samples is marcasite (FeS 2). Sample FA-A4 had a noticeable content of metallic iron (1.5 wt.%), whereas in all other fly ash samples, metallic iron was close to or below the detection limit ( ≤0.2 wt.%) .

The principal sulfate phase in the fresh fly ash samples was anhydrite (CaSO 4).

Syngenite [K 2Ca(SO 4)2·H2O] was present in all fly ash samples from incinerator A, while monosulfate [3CaO ·Al 2O3·CaSO 4·12H 2O] and thenardite (Na 2SO 4) were additionally detected in the fresh FA-A4 sample only. Syngenite and monosulfate formed in some of the fresh fly ash samples may be attributed to the dissolution of K- Al-and Ca-minerals ( e.g ., gypsum, aluminates and glass) and subsequent reactions during aging.

A small quantity of portlandite [Ca(OH) 2], likely the result of hydrolysis of calcium oxide (CaO) in the presence of water (moisture), was found in all fly ash samples. A minor amount of lime could be also seen in the FA-B1 sample. The Al- fraction available in all fly ash samples may have also undergone a redox reaction, leading to the formation of boehmite (AlOOH) (see section 4.3). In XRD analysis, mineral compounds of the heavy metals Pb and Zn were detected. Three lead phases found in fly ash samples collected from incinerator A are

Pb 3O4 (FA-A4), PbCl 2 (FA-A2) and PbCO 3 (FA-A4). They were not observed in the fly ash from incinerator B which had a lower Pb bulk concentration (see table 4.1).

Furthermore, caracolite [Na 3Pb 2(SO 4)3Cl] was present in samples FA-A4 and FA-B1. Zinc, the most abundant polluting element in fly ash samples from incinerator A, was present in gordaite and hydrozincite. The Cu-sulfide digenite was detected in sample FA-A4. Cd-,Ni-and Cr-rich compounds were not detected by XRD, suggesting that these elements were incorporated as lower concentration in amorphous and/or other complex crystalline phases (see also in EPMA analysis section 4.2).

86

CHAPTER 4

(a)

Figure 4.8a XRD patterns of the fresh fly ash samples from incinerator A and B. FA- An and FA-Bn stand for samples of fly ash from the incinerator facilities A and B respectively. The peaks are labelled An (anhydrite), Bo (boehmite), C (calcite), Cr (cristobalite), Ct (caracolite), Fe (iron), G (gehlenite), Gd (gordaite), Hl (halite), Hc (hydrocalumite), He (hematite), L (lime), Lz (lazurite), Po (portlandite), Q (quartz), R (rutile) and S (sylvite).

(b)

Figure 4.8b Quality of the Rietveld pattern-fitting results for the fresh sample of FA- A3. The curves are observed ( ······ ) and calculated ( ) patterns respectively; the bottom curve shows the difference between observed and calculated pattern.

87

CHAPTER 4

Table 4.5 XRD-based mineralogy of fly ash samples. Crystal Mineral (wt.%) Formula (simplified) *FA-A2 FA-A3 FA-A4 FA-B1 Structure Model§ Amorphous 43.3(9) 61.6(5) n.d 58.2(7) Alunite K Al 3(OH) 6(SO 4)2 1.0(2) 0.9(1) 0.6(1) S-459 Anhydrite CaSO 4 4.9(1) 4.1(1) 3.1(1) 4.0(1) S-80 Albite (plagioclase) NaAlSi 3O8 2.5(3) 2.2(2) 2.4(4) 0.9(2) S-155 Apatite Ca 5(PO 4)3OH 2.3(2) S-215 Bassanite CaSO 4·0.5H 2O 1.5(1) 1.8(1) S-292 Boehmite AlOOH 0.8(1) 0.3(0) 0.6(2) 0.4(1) S-43 3+ Butlerite Fe (OH) SO 4·2H 2O 1.0(2) 0.4(1) S-377 Calcite CaCO 3 2.9(1) 3.1(1) 5.4(2) 3.0(1) S-11 Calcium titanite CaTiO 3 1.5(2) S-757 Caracolite Na 3Pb 2(SO 4)3Cl 0.3(0) 0.3(0) I-024459 Cerussite PbCO 3 0.3(0) S-676 C2S Ca 2SiO 4 3.1(3) 2.8(2) 4.8(3) 3.1(2) S-69 C3A Ca 3Al 2O6 1.0(1) 1.0(1) S-71 Contunnite PbCl 2 0.3(0) S-883 Cristobalite SiO 2 1.5(1) S-37 Diopside CaMgSi 2O6 2.0(3) 1.4(2) 10.3(4) 1.9(2) S-133 Digenite CuS 0.9(1) S-562 Dolomite CaMg(CO 3)2 1.8(2) 1.0(2) S-31 Enstatite (Mg,Fe) SiO 3 5.4(4) S-151 Garnet Ca 3(Al,Fe ) 2(Si,P) 3O12 0.6(1) S-552 Gehlenite (melilite) (Ca,Na) 2(Mg,Fe,Si,Al) 3O7 2.6(2) 3.2(1) 8.9(3) 2.3(2) S-165 Gordaite NaZn 4(SO 4)(OH) 6Cl(H 2O) 6 1.0(1) 0.2(1) 2.1(1) I-406090 Halite NaCl 9.3(1) 5.6(1) 9.4(2) 4.4(1) S-109 Hematite Fe 2O3 2.0(1) 1.6(1) 0.5(2) 0.6(1) S-41 Hercynite FeAl 2O4 0.4(1) S-426 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)Cl 2 (H 2O) 1.6 (H 2O) 8 0.5(1) S-778 Hydrophilite CaCl 2 0.3(0) I-026686 Hydrozincite Zn 5(OH) 6(CO 3)2 1.3(1) S-389 Iron Fe 1.5(1) 0.2(0) S-140 Lazurite (sodalite) (Na 6Ca 2Al 6Si 6O24 ) (SO 4)2 2.1(1) 0.5(1) 2.8(2) 0.8(1) S-550 Lime CaO 0.6(1) S-150 Kalsilite KAlSiO 4 1.5(2) 1.2(2) S-961 Kornelite Fe 4(SO 4)6·15H 2O 0.7(2) S-695 Lepidocrocite FeOOH 0.3(1) 0.2(0) 0.8(1) 0.2(0) S-45 Magnetite Fe 3O4 0.8(1) S-50 Marcasite FeS 2 1.2(2) 0.7(1) S-362 Minium Pb 3O4 0.6(0) 1.0(1) S-713 Monosulfate 3CaO ·Al 2O3·CaSO 4·12H 2O 8.9(9) I-100138 Olivine group (Mg,Fe) 2 SiO 3 1.0(1) S-53 Periclase MgO 0.7(1) S-96 Portlandite Ca(OH) 2 0.4(1) 0.5(1) 1.0(1) 0.4(1) S-124 Quartz α-SiO 2 2.5(1) 2.0(1) 1.7(1) 4.0(1) S-1 Rutile TiO 2 0.6(1) 0.4(1) 0.5(1) S-15 Sanidine KAlSi 3O8 4.0(3) 2.2(2) S-21 Sphalerite FeS 0.9(2) S-65 Sylvite KCl 3.3(1) 2.0(1) 3.0(1) S-164 Syngenite K2Ca(SO 4)2·H 2O 7.0(3) 3.1(2) 10.2(4) S-325 Thenardite Na 2SO 4 2.0(2) S-115 2+ Ulvöspinel Fe 2 TiO 4 2.1(1) 0.4(1) 2.4(1) S-362 Wurtzite ZnS 0.3(1) S-703 Total 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter (S for the SIROQUANT database, I for the ICSD database) followed by the entry number of the respective database. n.d = not detected.

88

CHAPTER 4

4.1.4 Summary Characteristics of the MSWI residues from two waste incinerator plants located in the Ruhr industrial area of Germany have been examined in terms of grain morphology, grain size, chemical and mineralogical phase compositions. SEM/EDX analysis indicates that all bottom ash samples contain vitreous, and micro-crystalline components. The average particle size for a particular incinerator plant decreases from the samples of bottom ash (< 2 mm) to fly ash (< 200 µm). The proportion of angular particles is highest in the bottom ash samples and lowest in the fly ash samples. Two general components found in the fly ash: (i) an extremely fine-grained, polycrystalline (< 1µm), platy material containing many volatile elements i.e., Cl, K, Zn, Na, S and Pb, and (ii) globules of aluminosilicates glass. Concentrations of the toxic heavy metals (Zn, Pb and Cd) in the fly ash are rather high, thereby causing potential hazards, when fly ash material is exposed to the environment. The main phases in the bottom ash are glass, melilite, quartz and magnetite. These are commonly present in metallurgical slags and bottom ashes formed during waste incineration. The principal minerals in fly ash are highly soluble salts ( e.g., NaCl,

K2ZnCl 4 and KCl) that are present in the extremely fine-grained aggregate particles. Larger spherical particles are mainly comprised of calcium, sodium and potassium aluminosilicates and a glass phase. The mineral sub-assemblage containing the heavy metals Pb and Zn is detected. The mineral assemblage of the MSWI residues appears to be thermodynamically unstable and subject to aging and weathering processes, leading to the formation of secondary minerals ( e.g., ettringite and hydrocalumite). The alteration of the mineral phases occurring in the selected bottom ash and fly ash samples is examined in detail in section. 4.3.

89

CHAPTER 4

4.2 Microscopic Characteristics of MSWI Residues In this section, microscopic characterisation of bottom ash and fly ash particles is presented. The exercise involved observation of individual particles of the BA-A1 and FA-A1 specimens. The characteristics of the individual bottom ash and fly ash particles are discussed separately.

4.2.1 MSWI Bottom Ash For microscopic examinations, individual particles of the selected BA-A1 sample were prepared in polished thin-sections as described in section 3.5.2. Polarisation microscopy, SEM and electron probe x-ray microanalysis (EPMA) were employed to evaluate characteristics of individual particles according to the procedure outlined in section 3.5.3. Under the optical microscope, bottom ash particles revealed a large diversity of minerals embedded in ash fragments. Within the fragments, there were particles of charred wood, metallic fragments and vitreous phases with embedded crystalline grain. Most of these grains were optically identified as quartz. Moreover, a glass phase was frequently observed in bottom ash particles. For relatively large bottom ash particles (< 200 µm), additional observations with regard to shape and texture could be made in BSE (back scattered electron) images and/or element distribution maps. In all BSE images, the vitreous type of ash particle had a bright surface and a very dense structure. The vitreous bottom ash particles have been produced by melting in the combustion chamber. Porous ash particles having large and small degassing voids were observed quite often, also a flow texture was frequently found. In the present study, eighty five spots were analysed to examine elemental compositions of the phases encountered. Representative spot analyses of selected particles figured in the BSE images of figures 4.9 and 4.10, are given in table 4.6. In general, the ash particles consisted of a large variety of minerals that were mostly silicates and oxides. Among these minerals, quartz, magnetite and corundum were readily distinguished, as shown by the k α-emission lines Si, Fe or Al and O in energy- and wavelength-dispersive spectra. Small fragments of Al, Cu, and Si metal, as well as alloys were identified. The silicon and copper alloys may have been produced from the incinerated wastes of

90

CHAPTER 4

electronics and electric appliances. Large fragments of iron implements could be seen macroscopically in the bottom ash particles; however, in the < 100 µm grain size fractions investigated by EPMA, iron metal particles were not observed. In the mineral assemblage, sulfide and sulfate particles ( i.e., Cu 2S or BaSO 4) were frequently encountered (see analysis in table 4.7). The presence of BaSO 4 in the bottom ash may originate from incinerated paint or medical waste. From the WDX analytical data (EPMA), mineral formulae of clay group mineral, melilite and wollastonite were calculated according to the procedure described by Deer et al. (1992). In particular, some particles were characterised by a rather low analytical total (93-99 wt.%), suggesting that some undetected light elements, or perhaps water, may be present. The structural formulae frequently corresponded to a melilite group mineral, plagioclase, wollastonite, clay group minerals and oxides such as magnetite and spinel. The calculated chemical formula of melilite group may be 2+ represented as (Ca 3.5 Na 0.65 K 0.01 ) (Mg 0.4 Fe 0.6 Al 1.4 Si 3.6 ) O 14 . It represents a solid solution in the 5-component system spanned by the end-members akermanite

(Ca 2MgSi 2O7), gehlenite (Ca 2Al 2SiO 7), soda melilite (CaNaAlSi 2O7), iron-akermanite 2+ 3+ (Ca 2Fe Si 2O7) and iron-gehlenite (Ca 2Fe 2 SiO 7). Melilites of similar composition are commonly found in blast furnace slags. They also crystallise from alkaline magmas rich in calcium or form in thermally metamorphosed silicate-bearing carbonate rocks (Yoder, 1979). Analyses of glasses revealed mainly aluminosilicate compositions. Some of the glasses came from original glass fragments in the waste. These particles were optically clear and have SiO 2 contents exceeding 70 weight percent (Table 4.6; spot 6). Other glass particles were in optically inhomogeneous, contained inclusions, and displayed dark or brownish striations suggesting that they were melted during incineration; their

SiO 2 contents are significantly lower than those of technical glasses (Table 4.6; spots 1 and 8). These particles were melt products formed during incineration. These types of glasses are identical with one of the two groups of incineration-generated glasses described in detail by Eusden et al . (1999), which these authors termed "(transparent) isotropic melt products" (table 4.6; spot 1) and "opaque glassy melt product".

91

CHAPTER 4

(a)

(b)

Figure 4.9 BSE images of bottom ash polished thin sections illustrating the location of EPMA spot analyses. The numbers refer to analyses in table 4.6.

92

CHAPTER 4

(a)

(b)

Figure 4.10 BSE images of bottom ash polished thin sections illustrating the location of EPMA spot analyses. The numbers refer to analyses in table 4.6.

93

CHAPTER 4

Table 4.6 Chemical composition of phases in bottom ash particles analysed by EPMA as illustrated in figures 4.9 and 4.10. Composition is given in wt. % of oxides except for the alloy (spot 2), where the numbers refer to element wt.%.

Spot Phase Na 2O MgO Al 2O3 SiO 2 SO 3 K2O CaO TiO 2 Cr 2O3 MnO FeO CuO ZnO BaO

1 Glass 5.45 0.01 27.83 55.31 0.03 0.50 9.79 0.05 0.02 0.01 0.50 0.00 0.00 0.00 3 Hematite 0.01 0.00 0.00 0.05 0.04 0.00 0.13 0.00 0.00 0.68 91.99 0.03 0.17 0.00 4 Magnetite 0.01 0.04 0.00 0.06 0.02 0.00 0.07 0.01 0.02 0.69 96.39 0.01 0.16 0.00 5 Corundum 0.01 0.01 101.90 0.01 0.03 0.00 0.08 0.01 0.03 0.01 0.34 0.08 0.11 0.11 6 Glass 2.71 0.11 1.06 79.81 0.66 0.50 9.61 0.00 0.03 0.02 0.08 0.03 0.00 2.05 7 aClay group 0.29 3.81 16.26 57.22 0.01 2.70 5.22 0.22 0.00 0.08 8.06 0.07 0.15 0.02 8 Glass 5.16 0.02 40.77 44.71 0.06 1.47 0.68 0.07 0.00 0.01 1.42 0.01 0.16 0.21 9 Clay group 0.42 0.07 32.96 55.14 0.00 1.46 0.35 0.14 0.07 0.04 1.02 7.65 0.23 0.45 10 cWollastonite 1.27 2.42 5.54 31.96 0.43 0.29 21.56 0.22 0.05 0.35 33.38 0.10 0.06 0.92 11 bMelilite 3.64 2.86 12.36 38.21 0.01 0.08 34.26 0.12 0.03 0.05 7.98 0.09 0.30 0.00

Spot Na Mg Al Si S K Ca Ti Cr Mn Fe Cu Zn Ba

2 Alloy 0.24 0.17 7.49 28.82 0.00 0.57 0.36 0.00 0.00 0.06 1.39 17.10 0.00 0.56 The mineralogical formula is presented as follows: a total Clay group (Ca 0.78 Ba <0.01 Na 0.10 K 0.48 ) (Mg 0.79 Mn <0.01 Fe 0.011 Ti 0.01 Zn 0.02 Cu <0.01 ) (Al 2.69 Si 8.03 ) (O,OH) 24 b 2+ Melilite (Ca 3.5 Na 0.65 K 0.01 ) (Mg 0.4 Fe 0.6 Al 1.4 Si 3.6 ) O 14 c 2+ 3+ Wollastonite(Ca 3.15 Mn 0.04 Fe 2.68 Na 0.34 K 0.05 ) (Al 0.89 Fe 0.74 Si 4.36 ) O 18

94

CHAPTER 4

Table 4.7 Mineralogy of bottom ash particles from EPMA analysis (composition given in wt. % of oxides, except for the metallic phases in wt.% element)

Phase Na 2O MgO Al 2O3 SiO 2 SO 3 K2O CaO TiO 2 Cr 2O3 MnO FeO CuO ZnO BaO (Cu,Fe) 2S 0.00 0.01 0.07 0.08 44.16 0.01 0.07 0.00 0.02 0.04 1.63 96.64 0.00 0.00 CuS 0.00 0.03 0.06 0.05 60.94 0.00 0.10 0.00 0.00 0.03 0.00 89.09 0.00 0.13 aCa-Al-Sulfate 0.16 0.22 6.79 1.00 12.06 0.09 19.81 0.38 0.00 0.00 0.41 0.00 1.21 0.08 Ca-Al-Sulfate 0.19 0.00 22.08 0.10 3.01 0.00 43.89 0.04 0.05 0.00 0.26 0.00 0.00 0.00 bFeldspar 2.21 0.37 29.36 46.44 0.00 0.23 15.91 0.32 0.00 0.05 2.65 0.02 0.00 0.24 Fe-Ti-Zn-O 0.00 0.03 0.28 0.17 0.00 0.03 0.08 10.39 0.08 0.00 64.55 7.40 6.65 0.00 cMuscovite 5.16 0.02 40.77 44.71 0.06 1.47 0.68 0.07 0.00 0.01 1.42 0.01 0.16 0.21 dOlivine 0.02 3.34 0.09 29.67 0.00 0.00 2.12 0.07 0.04 0.55 64.20 0.07 0.00 0.00 ePlagioclase 5.25 2.58 2.38 49.87 0.34 0.76 4.34 0.02 0.00 0.06 1.66 0.15 0.04 0.09 fPyroxene 4.64 0.18 12.48 37.26 0.56 1.75 13.76 0.14 0.00 0.21 28.37 0.08 0.00 0.33 gSpinel 0.02 0.96 9.94 0.24 0.00 0.00 0.18 1.66 0.09 0.19 80.77 0.00 0.03 0.00

Metallic Na Mg Al Si S K Ca Ti Cr Mn Fe Cu Zn Ba Si 0.01 0.01 0.14 70.20 0.00 0.02 0.12 0.03 0.00 0.04 0.05 0.10 0.06 0.00 Ca-Cu-Fe 0.05 0.04 0.12 0.19 0.35 0.02 1.09 0.08 0.00 0.08 0.69 1.68 0.00 0.00 Alloy 0.74 0.21 0.48 14.09 0.36 0.18 0.91 0.02 0.04 0.00 0.42 0.14 0.31 0.16 Fe-Si 0.02 0.01 1.15 45.45 0.00 0.02 0.06 0.00 0.04 0.35 48.32 0.22 0.00 0.05 hGraphite 0.01 0.11 0.17 0.54 0.03 0.15 0.14 0.02 0.03 0.01 0.09 0.18 0.02 0.14 The calculated mineralogical formula, except for ettringite and graphite may be presented as follows: a Ettringite (idealized) (3CaO·Al 2O3·3CaSO 4·32H 2O) b 3+ Feldspar (Na 0.8 Ca 3.26 K0.1 Ba 0.01 )(Fe 0.28 Al 6.6 Si 0.9 )Si 8O32 c total Muscovite (Ca 0.09 Ba 0.01 Na 1.35 K 0.25 ) (Fe 0.16 Ti <0.01 Al 4.56 ) (Al 1.95 Si 6.5 ) (O,OH) 24 d total Olivine (Mg 0.16 Mn 0.02 Ca 0.07 Fe 1.77 ) Si O 4 e Plagioclase (Ca 1.9 Na 1.9 K 0.1 ) (Al 5.9 Si 10.0 ) O 32 f total Pyroxene (Ca 0.62 Na 0.37 K0.04 Mg 0.13 Fe 0.99 Mn <0.01 Ti <0.01 Al 0.43 ) (Al 0.43 Si 1.59 ) O 6 g 2+ 3+ Spinel (Ca 0.07 Mn 0.06 Zn <0.01 Fe 7.11 )(Fe 11.27 Cr 0.04 Al 4.16 Si 0.08 )O 32 hGraphite Although the light element C could not be analysed, graphite is recognised in the BSE image

95

CHAPTER 4

The plagioclase feldspar found in the bottom ash particles is close to a member of the albite-anorthite series, pyroxene is a member of the augite-diopside series, and olivine group is compositionally close to fayalite. Spinels have compositions close to 2+ 3+ 2+ magnetite [Fe (Fe )2O4] and hercynite (Fe Al 2O4). Additionally, a Ca-Al-sulfate possibly close to ettringite was found. This phase is commonly considered as product of mineralogical alteration of bottom ash induced by aging (Eusden et al ., 1999; Speiser et al ., 2000). Complex oxides of extremely Fe-rich composition with varying proportions of Ti and Zn, and in some cases Cr have been found. Graphite was also observed in the bottom ash particles. Some minerals were not encountered in polished thin sections, although they were identified by XRD, ( e.g., anhydrite). Probably, anhydrite had been removed during sample preparation. Overall, a reasonably good agreement is seen for the minerals or solid phases identified by EPMA and XRD.

4.2.2 MSWI Fly Ash Polished thin-sections of FA-A1 were examined with the optical microscope, SEM and EPMA to explore phase identity and composition within composite and single-phase ash particles (see section 3.5.2 and 3.5.3). In transmitted light under the polarising microscope, most fly ash particles appeared as either translucent or opaque spherical grains. Among the translucent particles, there were both isotropic and anisotropic grains. Most of the latter could be optically identified as quartz. SEM analysis showed that the majority of fly ash particles have a spherical in shape and exhibit a relatively smooth surface. Their diameters frequently ranged from 20 to 50 µm. Loose agglomerates of particles, a few them to hundreds of nanometer in diameter were usually attached to the spheres. These characteristics of polished thin sections were obvious also in the back-scattered electron (BSE) images in figures 4.11 and 4.12. The BSE images further revealed that the spherical particles were either hollow or filled, and that they exhibited ranges in diameter of sphere (1– 50 µm). The spherical particles are mainly composed of calcium aluminosilicate glass as indicated by the WDX spectra and the EPMA analyses (Table 4.8; spots 9, 10, 15 and 17). Similar aluminosilicate glass spheres have been found in the MSWI fly ash by Eighmy et al . (1995).

96

CHAPTER 4

(a)

(b)

Figure 4.11 BSE images of fly ash polished thin sections illustrating the location of EPMA spot analysis. The numbers refer to analyses in table 4.8.

97

CHAPTER 4

(a)

(b)

Figure 4.12 BSE images of fly ash polished thin sections illustrating the location of EPMA spot analysis. The numbers refer to analyses in table 4.8.

98

CHAPTER 4

BSE images of quantitatively analysed grains in the polished thin-sections obtained by EPMA are used to correlate the different phases with different particle morphologies (Figures 4.11 and 4.12). A selection of 181 spots analysed is given in table 4.8. Mineral formulae were calculated according to the procedure described by Deer et al. (1992). There is significant chemical and mineralogical variation between different ash particles and even within single particles. Quartz and magnetite grains were most readily distinguished during this work. Quartz typically occurs as spherical grains and is sometime enclosed by a glass rim (Figure 4.12). Metallic aluminum and silicon particles in silicate agglomerates were also identified. Importantly, the spherical particles are glasses which have a Ca-rich aluminosilicate composition (Table 4.8; spots 9, 10, 15 and 17). Similar compositions of glass particles have been previously observed by Forestier and Libourel (1998). A large number of irregularly shaped particles were analysed by EPMA. Most of them displayed a high reflectance and had an angular shape. The prominent characteristic of these particles is their high Fe-content (Table 4.8; spots 6 and 14), but some of them may contain calcium and silicon as well. Other particles are chemically close to Ca-Fe-rich garnet (Table 4.8; spots 4, 12 and 13). Mg-Fe-rich pyroxene with

total the formula (Ca 0.13 Na 0.21 K0.15 Fe 0.85 Mg 1.2 )(Al 0.2 Si 1.8 )O 6 has been tentatively identified. In particles rich in Ca, Na, K, Fe, Al, Si and O, the following crystalline phases have been tentatively identified: sodian sanidine, melilite to akermanite-gehlenite, fayalite-rich olivine, and anorthite-rich plagioclase. In addition, the oxides rutile, lime (portlandite), escolaite (Cr 2O3), and the calcium-rich phases CaAl 2O4, hatrurite (Ca 3SiO 5) and wollastonite (CaSiO 3) are identified. The magnetic grain consists essentially of magnetite (spinel) in addition to hematite, which has similar FeO content with magnetite, however hematite has less Al 2O3 content than magnetite (Table 4.10). Moreover, alloys consisting of zinc, iron, aluminium, nickel and chromium have been found (Table 4.9). These alloys could not be detected by XRD. Anhydrite/bassanite, Ca- Al-sulfates, possibly ettringite, chlorides, apatite and complex silicates could be also analysed (Tables 4.8, 4.9 and 4.10). Similar observations were reported for comparable fly ash particles by Forestier and Libourel (1998). Some of these phases are confirmed by the XRD analysis presented in table 4.4 (refer to section 4.1.3), but the mineralogical identification of many particles or compound particles remains uncertain.

99

CHAPTER 4

Table 4.8 Chemical composition (wt.%) of phases in fly ash particles analysed by EPMA spot analysis as illustrated in figures 4.11 and 4.12. Spot Phase Na 2O MgO Al 2O3 SiO 2 P2O3 SO 3 Cl° K2O CaO TiO 2 Cr 2O3 FeO NiO CuO ZnO CdO SnO BaO PbO 1 Pyroxene 2.23 4.34 12.31 33.48 0.35 0.06 0.09 0.08 37.72 2.99 0.08 3.37 0.13 0.05 2.01 0.05 0.03 0.06 0.00 2 Quartz 0.00 0.02 0.00 99.90 0.03 0.01 0.04 0.03 0.08 0.00 0.04 0.01 0.00 0.00 0.01 0.00 0.00 0.19 0.00 3 Ca-Zn-Si-SO 4 0.25 1.09 3.17 11.65 0.28 20.16 3.16 1.51 28.95 0.42 0.06 0.50 0.02 0.00 13.19 0.21 0.01 0.09 0.00 4 aGarnet 1.21 1.72 9.60 21.80 7.81 0.31 0.08 0.60 25.17 2.65 0.08 20.39 0.18 0.15 0.14 0.08 0.02 1.45 0.00 5 Ca-Ba-Fe-Si-O 1.06 1.34 4.29 26.18 1.66 0.10 0.06 0.20 20.49 8.30 0.03 10.93 0.05 1.54 0.65 0.00 0.04 18.26 0.40 6 Ca-Fe-Si-O 0.02 5.96 2.27 2.44 0.15 0.16 0.01 0.05 2.35 2.16 0.49 63.98 0.66 4.39 3.60 0.07 0.13 2.01 0.00 7 Ca-Al-SO 4-Cl-O 1.65 0.19 72.54 0.13 0.07 8.03 1.54 1.88 7.16 0.00 0.00 0.16 0.00 0.43 4.08 0.00 0.00 1.09 0.00 8 Ca-Al-SO 4-Cl-O 0.24 0.54 30.41 0.64 0.05 13.55 5.86 1.74 19.32 0.00 0.05 0.14 0.08 1.10 3.25 0.11 0.01 4.00 1.34 9 Glass 4.64 4.90 15.20 36.83 3.33 0.06 0.04 0.91 18.78 1.71 0.02 9.27 0.06 0.21 0.67 0.06 0.07 0.33 0.14 10 Glass 0.64 2.71 8.07 27.21 2.53 4.08 0.08 0.20 45.71 0.34 0.09 4.11 0.01 0.25 1.20 0.00 0.07 0.00 0.00 11 Quartz 0.19 0.01 0.13 94.59 0.03 0.10 0.12 0.16 0.31 0.07 0.00 0.64 0.03 0.08 0.41 0.07 0.02 0.00 1.39 12 Garnet 1.09 0.80 7.85 24.59 0.75 0.04 0.03 0.25 15.12 1.95 0.13 42.63 0.04 0.16 0.48 0.08 0.07 0.32 0.01 13 Garnet 2.13 0.93 8.34 30.57 0.71 0.04 0.04 0.43 18.25 1.77 0.08 32.50 0.01 0.32 1.52 0.06 0.13 0.30 0.24 14 Ca-Fe-Si-Al-O 0.55 0.67 3.59 8.82 0.31 0.03 0.00 0.15 5.84 3.29 0.14 67.47 0.08 0.29 0.89 0.03 0.16 0.17 0.18 15 Glass 3.33 1.44 3.65 48.33 0.83 0.05 0.07 0.76 30.22 1.15 0.10 5.52 0.01 0.20 2.00 0.03 0.04 0.39 0.52 16 bPyroxene 2.67 19.30 2.16 41.97 0.12 0.09 0.02 2.79 2.97 0.10 0.22 24.57 0.05 0.26 0.27 0.01 0.01 0.07 0.00 17 Glass 0.41 5.46 15.45 29.40 1.04 0.11 0.07 0.10 38.45 0.89 0.02 3.04 0.00 0.11 2.20 0.00 0.07 0.79 0.00

The compositions of the observed minerals may be presented as follows: a 2+ 3+ Garnet: (Ca 2.5 Mg 0.25 Fe 0.2 Zn 0.01 )(Al 0.9 Fe 0.9 Ti 0.16 Sn <0.01 ) (Al 0.2 P0.64 Ti 0.04 Si 2.1 )O 12 b total Pyroxene: (Ca 0.13 Na 0.21 K0.15 Fe 0.85 Mg 1.2 )(Al 0.2 Si 1.8 ) O 6

100

CHAPTER 4

Table 4.9 Chemical composition of phases in fly ash particles analysed by EPMA (composition given in wt. % of oxides, except for the alloy given in wt.% elements). Phase Na 2O MgO Al 2O3 SiO 2 P2O3 SO 3 Cl° K2O CaO TiO 2 Cr 2O3 FeO NiO CuO ZnO CdO SnO BaO PbO Silicates aAlkali Feldspar 1.72 0.06 19.23 65.34 0.00 0.12 0.08 10.71 0.92 0.00 0.05 0.12 0.00 0.05 0.00 0.00 0.04 0.19 0.02 Mg-Al-Si-O 2.13 11.56 18.94 45.00 0.03 0.55 1.43 3.46 0.16 0.02 0.00 0.56 0.00 0.13 0.66 0.02 0.03 0.06 0.94 C2S 0.05 4.02 2.24 27.00 1.67 0.38 0.21 0.41 59.61 0.33 0.04 1.55 0.00 0.02 1.13 0.00 0.11 0.10 0.19 bClay group 0.89 0.54 37.76 49.50 0.08 0.16 0.35 3.93 0.17 0.00 0.03 0.83 0.00 0.07 0.31 0.02 0.04 0.00 0.24 cOlivine 2.89 0.35 2.22 29.87 0.87 0.05 0.00 0.49 6.47 0.46 0.06 46.72 0.28 1.94 0.12 0.00 0.03 0.40 0.25 Hatrurite (Ca 3SiO 5) 0.04 0.78 3.17 26.93 1.38 0.17 0.16 0.27 62.47 0.39 0.12 2.19 0.00 0.13 0.40 0.06 0.20 0.00 0.16 Al-Si-O 2.78 3.17 20.89 30.04 0.08 3.69 4.40 5.03 5.31 0.04 0.02 0.98 0.00 0.05 0.87 0.02 0.09 0.00 0.77 dMica group 6.17 18.41 11.75 57.74 0.15 0.05 0.02 2.08 0.92 0.68 0.02 1.35 0.01 0.00 1.34 0.04 0.02 0.07 0.20 eMelilite 1.69 8.16 6.43 34.76 3.11 0.03 0.05 0.20 33.55 0.71 0.03 7.93 0.02 0.10 0.29 0.09 0.00 0.32 0.00 fSpinel 2.18 1.43 9.39 13.92 0.80 0.07 0.09 0.79 10.14 1.02 8.32 43.00 3.18 0.31 1.62 0.00 0.07 0.13 0.03 gWollastonite 1.17 0.69 1.76 39.91 0.26 1.79 0.85 0.05 46.03 4.95 0.04 1.96 0.01 0.05 0.23 0.15 0.00 0.17 0.02 Alloys Na Mg Al Si P S Cl K Ca Ti Cr Fe Ni Cu Zn Cd Sn Ba Pb Ni-Zn-Fe 0.60 0.17 0.02 0.12 0.01 0.41 0.52 0.27 0.13 0.00 0.00 0.47 7.56 0.05 0.97 0.02 0.01 0.00 0.04 Fe-Si 0.37 0.59 0.81 2.61 0.26 0.05 0.02 0.13 2.60 0.62 0.02 6.65 0.01 0.05 0.08 0.00 0.01 0.02 0.00 Fe-Zn 0.00 0.19 0.02 0.14 0.01 0.07 0.10 0.06 0.16 0.02 0.00 10.15 0.00 0.06 1.24 0.00 0.02 0.01 0.01 The calculated formula of some mineral compositions may be presented as follows: a 2+ Alkali feldspar (K 2.8 Na 0.99 Ca 0.2 Ba 0.01 ) (Si 3.9 Al 4.03 Fe 0.04 ) Si 8O32 b Clay group (K 0.6 Na 0.23 Ca 0.02 )(Al 4.5 Fe 0.1 Mg 0.1 Zn 0.03 )(Al 1.4 Si 6.6 )O 20 (OH,Cl) 4 c 2+ Olivine (Ca 0.23 Mg 0.2 Fe 1.34 )SiO 4 d 2+ Mica group (Na 1.6 K0.34 Ca 0.12 )(Zn 0.12 Mg 3.6 Fe 0.14 Al 1.4 )(Al 0.43 Si 7.6 )(OH,Cl) 4O20 e 2+ Melilite (Ca 3.5 Na 0.32 K0.03 )(Mg 1.2 Fe 0.65 Al0.75 Si 3.4 )O 14 fSpinel Fe-Cr-Al-Ni-rich phase g 2+ 3+ Wollastonite (Ca 5.9 Mg 0.13 Fe 0.02 Na 0.3 )(Al 0.25 Fe 0.12 Si 4.8 )O 18

101

CHAPTER 4

Table 4.10 Chemical composition of phases in fly ash particles analysed by EPMA (composition given in wt. % of oxides).

Phase Na 2O MgO Al 2O3 SiO 2 P2O3 SO 2 Cl° K2O CaO TiO 2 Cr 2O3 FeO NiO CuO ZnO CdO SnO BaO PbO Chlorides Sulfates and Phosphate KCl-NaCl 4.02 0.10 0.13 3.10 0.13 6.87 5.40 10.39 0.64 1.44 0.04 0.52 0.04 0.49 4.18 0.00 0.56 0.13 2.60 aAnhydrite/ bassanite 0.18 0.24 0.12 0.87 0.23 42.84 0.20 1.07 43.22 0.01 0.00 0.61 0.12 0.02 1.72 0.00 0.21 0.23 0.00 bApatite 1.23 9.92 2.96 6.53 28.13 0.01 0.01 0.13 37.31 0.33 0.04 1.59 0.00 0.02 0.40 0.00 0.06 0.33 0.08 Ca-Si-Al-S-O 0.54 0.66 13.45 13.71 0.50 19.96 2.81 1.28 34.97 0.08 0.00 0.64 0.00 0.32 5.32 0.09 0.17 0.03 0.00 K-Zn-S-Cl-O 0.00 1.05 0.22 1.64 0.21 6.02 14.96 16.42 7.07 0.05 0.08 1.35 0.02 0.46 41.57 0.14 0.31 0.00 1.40 Ca-Zn-S-Cl-O 0.00 0.35 0.12 0.39 0.21 4.11 5.44 0.87 45.24 0.04 0.07 0.06 0.01 0.09 8.81 0.16 0.08 0.00 0.20 Zn-S-Cl-O 0.45 3.02 0.08 2.42 0.52 6.47 10.83 4.08 0.67 0.12 0.07 0.81 0.05 0.78 57.03 0.23 0.37 0.00 5.06 Oxides Rutile 0.49 0.08 2.76 1.05 0.05 1.10 0.74 0.53 3.87 86.03 0.00 0.36 0.00 0.02 0.69 0.08 0.00 0.19 0.24 Magnetite 0.27 0.03 4.21 0.76 0.22 0.03 0.04 0.14 0.24 0.01 0.23 86.44 0.13 0.31 0.30 0.01 0.04 0.14 0.05 Hematite 0.06 0.05 0.02 0.17 0.00 0.47 0.05 0.03 3.25 0.00 0.05 85.15 0.07 0.50 0.76 0.06 0.05 0.25 0.19 cMassicotite 0.97 0.75 1.88 16.11 0.12 1.02 1.74 0.60 2.19 0.17 0.02 0.44 0.01 0.14 0.43 0.21 0.13 0.06 34.98 dMinium 0.12 0.01 0.31 4.60 0.06 1.06 3.38 1.45 1.71 0.03 0.08 1.37 0.03 0.28 2.18 0.11 0.04 0.00 31.25 eEscolaite 0.10 0.06 0.53 0.38 0.12 0.15 1.21 0.26 0.28 0.14 78.38 0.48 0.01 0.07 1.23 0.04 0.03 0.00 0.39 fLime/ portlandite 0.21 0.14 0.15 0.28 0.32 1.20 0.64 0.45 60.94 0.03 0.01 0.18 0.04 0.08 0.51 0.06 0.03 0.13 0.06

CaAl 2O4 0.46 2.52 62.70 4.74 0.11 1.66 0.16 0.82 23.28 0.62 0.07 0.68 0.03 0.03 0.73 0.03 0.01 0.11 0.04 The compositions of the observed minerals may be presented as follows: a Anhydrite/bassanite CaSO 4/ CaSO 4·0.5H 2O b Apatite (Ca 9Na 0.5 )[(P 5.2 Si 1.5 ) O 4]6(OH,Cl) 2 cMassicotite PbO d Minium Pb 3O4 e Escolaite Cr 2O3 f Lime/portlandite CaO/Ca(OH) 2

102

CHAPTER 4

Further WDX data revealed detectable concentrations of several trace metals in the fly ash particles. A complex distribution of pollutant elements (Cd, Cr, Ni, Pb, Zn) is observed (Tables 4.8, 4.9 and 4.10). A Cd-rich mineral could not be detected, as Cd is chemically related to Zn, the highest Cd-content can be observed in the Zn-rich particle (Table 4.8; spot 3). Cr and Ni are mainly identified in alloys and in complex silicates and/or oxides. Lead (Pb) is found in metal, chlorides and sulfates and silicates ( e.g.,

Mg-Al-Si-O) and/or oxides ( i.e., Pb 3O4 and PbO) (Tables 4.8, 4.9 and 4.10). Perhaps, the PbO contents obtained in some analyses are due to “impurities” (compound particles). For example, Pb is unlikely to be present in escolite (Cr 2O3). The most abundant pollutant element Zn was found in alloys, chlorides, sulfates, spinels and

2− calcium-bearing aluminosilicate glass. It is supposed that Ca-rich and SO 4 -rich phases are important hosts for the trace elements Pb and Zn, as proposed by the previous researchers (Eighmy et al ., 1995; Forestier and Libourel, 1998).

4.2.3 Summary Microanalysis of MSWI residues has been performed with particular reference to microchemistry and mineralogy, which can be inferred from the chemical composition of individual particles. The individual bottom ash particles observed by optical and EPMA analysis confirmed the presence of Si-rich aluminosilicate glass. Among the bottom ash particles Al-, Fe- or Si-rich grains occur. Some of these grains chemically correspond to melilite, plagioclase, quartz and spinel, which have also been identified by XRD. These phases are constituents of the original waste and/or have been formed during combustion at high temperature. Early alteration product formed by quenching of the bottom ash are ettringite and portlandite, which are found in the bottom ash particles. Likewise, there is a significant chemical and mineralogical heterogeneity between different fly ash particles and even within single particles. Microanalysis suggests that fly ash particles have high contents of volatile elements, notably K, Cl, Zn and S, which are present as extremely fine-grained mixtures of chloride and sulfate minerals. Among the crystalline phases are oxides, silicates, alloys and metallic compounds. Spherical particles consist of calcium-rich aluminosilicate glass. Some mineral phases enriched in heavy metals (Pb 3O4 and PbO) are identified, while the

103

CHAPTER 4

heavy metals are mostly incorporated in the glass phase, alloys and complex crystalline phases. These results have practical implications for the understanding of the leaching behaviour of fly ash, where identification of the host phases of toxic elements is of major importance. Detailed investigations on the heavy metal release by leaching tests are presented in chapter 5.

4.3 Mineralogical Alteration of the MSWI Residues In this section, the effect of weathering and aging reactions on the mineralogical alteration of bottom ash and fly ash examined by XRD is presented.

4.3.1 Mineralogical Changes of Bottom Ash Aging experiments were conducted on the fresh BA-A2 and BA-B1 samples with the aim of promoting the natural weathering reactions occurring over time (from fresh to the 6-months). The mineralogical composition of a fresh bottom ash samples is presented in table 4.3 (refers to section 4.1.3). Alteration of crystalline phases was then analysed by the XRD method according to the procedure outlined in section 3.2. The XRD patterns of the BA-A2 specimen against aging time are shown in figure 4.13. Each peak of the principal minerals was identified and judged by the Rietveld method. The integral intensity of these peaks is proportional to the amount of the phase present in the bottom ash. For the freshly quenched bottom ash of BA-A2, the main crystalline phases were anhydrite, ettringite, calcite, gehlenite, halite, hematite and magnetite. The presence of ettringite in the quenched bottom ash results from the dissolution of Ca-and Al-bearing minerals and subsequent reaction between the dissolved species according to the equation (Meima and Comans, 1997; Sabbas et al ., 2003): 2+ + 3+ + 2− + ⇔ + + ⋅ ⋅ ⋅ 6Ca 2Al 3SO 4 38 H2O 12 H 3 CaO Al 2O3 3 CaSO 4 32 H 2O (4.1) Hence, ettringite formed in the fresh sample follows the pattern normally encountered in alteration products of bottom ash, whilst the alteration process starts immediately after the quenching of the hot bottom ash in the reservoir (Speiser et al ., 2000).

104

CHAPTER 4

(a)

(b)

Figure 4.13 Diffractograms of BA-A2 aged at room temperature over times; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Cr (cristobalite), D (diopside), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), M (magnetite), Q (quartz) and R (rutile). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

105

CHAPTER 4

Diffraction peaks of portlandite [Ca(OH) 2] could not be identified, although hydrolysis of calcium oxide (CaO) is likely to start immediately after bottom ash quenching (Belevi et al ., 1992; Comans and Meima, 1994). However, portlandite may be present as nanocrystalline particles, which resulted in low-intensity, broad diffraction peaks. Hydrocalumite [Ca 8Al 4(OH) 24(CO 3)Cl 2(H 2O) 1.6 (H 2O) 8], which was formed due to quenching is recognised at 11.20° 2 θ (PDF#78-2051) in the fresh bottom ash of BA- A2 (Figure 4.13b) . As aging progressed, the peaks of ettringite progressively decreased in intensity, while peaks of gypsum became sharper and more intense, for example, the peak at 20.74 O 2 θ (PDF#74-1905) (Figure 4.13b). Anhydrite and calcite were still detectable in the aged materials, and their peak intensities remained constant relative to fresh bottom ash. Previous work, however, has indicated that anhydrite and calcite may vary in mineralogical compositions of bottom ash during short-term (< 1 month) natural weathering. Both phases have been shown to be subject to the mineralogical alteration induced by aging (Chimenos et al ., 2000; Polettini and Pomi, 2004). Gehlenite, quartz, cristobalite, magnetite, hematite and rutile remained almost unchanged during the aging process, indicating that silicates and the oxides were stable mineral phases in the BA- A2 sample. Halite detected in the fresh sample was still present in the aged BA-A2 materials. Figure 4.14a presents XRD patterns of the sample of BA-B1 as a result of mineralogical alteration experiments dependent on aging time. The fresh BA-B1 sample mainly consists of anhydrite, ettringite, calcite, corundum, gehlenite and quartz (see also table 4.3). The formation of ettringite is due to reaction (equation 4.1) as discussed above in the content of sample BA-A2. Additionally, portlandite [Ca(OH) 2] was formed in the fresh BA-B1 sample, which is recognised by a low-intensity peak at 34.11 O 2θ (PDF#87-0673). Reasonably, portlandite is formed as a result of hydrolysis of CaO

[i.e., CaO →Ca(OH) 2] during bottom ash quenching (Belevi et al ., 1992; Comans and Meima, 1994). Alteration reactions obviously leaded to the oxidation of metallic iron in the BA-B1 sample toward the formation of goethite (FeOOH).

106

CHAPTER 4

(a)

(b)

Figure 4.14 Diffractograms of BA-B1 aged at room temperature over times; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Co (corundum) Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Po (portlandite), Q (quartz) and R (rutile). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

107

CHAPTER 4

With increasing aging time, it could be seen that the ettringite peaks disappeared progressively, while the gypsum peaks appeared and became more intense very similar to the sample of BA-A2 (Figure 4.14b). Peaks of anhydrite were still observed in the aged materials, although with decreased peak intensities. The decrease of peak intensities of anhydrite suggests that during aging anhydrite contributes also to the formation of gypsum.

4.3.2 Mineralogical Changes of Fly Ash A mineralogical alteration-time study on the fly ash samples was performed by XRD according to the procedure outlined in section 3.5.1. The evolution of XRD patterns of the selected samples of FA-A2, FA-A3, FA-A4 and FA-B1 as a result of the aging process (from fresh to 6-months) is presented in figures 4.15- 4.18, respectively. The mineralogical phase compositions of all fresh fly ash samples are presented in table 4.5 (refer to section 4.1). Figure 4.15 shows the mineralogical alterations of the FA-A2 sample as a function of time. In the fresh FA-A2 sample, the XRD patterns display a series of sharp diffraction peaks identified as those of anhydrite (CaSO 4), halite (NaCl), quartz ( α-

SiO 2) and sylvite (KCl). The remainder of peaks include alunite [KAl 3(OH) 6(SO 4)2], boehmite (AlOOH), calcite (CaCO 3), cristobalite (SiO 2), gehlenite (Ca 2Al 2SiO 7), hematite (Fe 2O3), iron (Fe), lazurite close to sodalite group [Na 6Ca 2Al 6Si 6O24 (SO 4)2], magnetite (Fe 3O4) and rutile (TiO 2). The presence of boehmite (AlOOH) in the sample may support the hypothesis that a significant waste composition of Al has undergone the following redox reaction (Sabbas et al ., 2003): o + → + Al 2H 2O AlOOH 5.1 H 2 (4.2)

In the aged FA-A2 sample, new peaks of bassanite (CaSO 4·0.5H 2O), gordaite

[NaZn 4(SO 4)(OH) 6Cl(H 2O) 6], hydrocalumite [Ca 8Al 4(OH) 24 (CO 3)Cl 2(H 2O) 1.6 (H 2O) 8] and syngenite [K 2Ca(SO 4)2.H 2O] developed by weathering reactions (see figure 4.15b). It can be seen that aging of the FA-A2 sample has resulted in a marked increase of peak intensity of syngenite and bassanite, together with a disappearance progressive of the alunite and lazurite peaks. The dissolution of alunite and lazurite with aging may contribute to precipitation of syngenite and bassanite.

108

CHAPTER 4

(a)

(b)

Figure 4.15 Diffractograms of FA-A2 aged at room temperature as a function of time; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Ba (bassanite), Bo (boehmite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gd (gordaite), Hc (hydrocalumite), He (hematite), Hl (halite), Lz (lazurite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite) and Sy (syngenite). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

109

CHAPTER 4

Moreover, peak intensities of anhydrite do not show such a pronounced change, whilst the sulfate mineral syngenite is developed, probably because the sulfate is provided by other crystalline compounds such as alunite or amorphous phases rather than by anhydrite. It can be further seen in figure 4.15b that gordaite and hydrocalumite are formed, while peak intensities of halite and sylvite are deceased during aging. The main peak of sylvite at 28.49 O 2 θ (PDF# 89-3619) is overlapped with the growing peak of syngenite at 28.39 O 2 θ (PDF# 74-2159). Sabbas et al . (2003) suggested that Ca-, Na-, Zn-, Cl-bearing phases like gordaite and hydrocalumite, formed by reactions from halite, sylvite and other Ca- phases containing varying amounts of Zn. There is a slight increase in peak intensities of calcite and gehlenite after 1 week of aging and they are likely to increase progressively at later ages. The increased peak intensity of gehlenite (and thus of its weight fraction) may be attributed to the dissolution of other phases induced by aging. Very similar XRD patterns were observed for the fresh samples FA-A3 and FA- A2 (compare figure 4.15b with figure 4.16b). For the fresh sample FA-A3, the main crystalline phases are anhydrite, calcite, gehlenite, halite and quartz. A small amount of boehmite (AlOOH) is indicated by a low-intensity peak at 14.49 O 2 θ (PDF# 83-2384). In the aged FA-A3 material, the newly formed minerals are very similar to those of the aged sample FA-A2 (Figure 4.16b). Hydrocalumite and syngenite were formed during aging, accompanied by a decrease in the content of halite and sylvite. However, gordaite was not formed during aging. This can be probably attributed to the chloride content of the raw sample FA-A3, in which the high chloride content is an important factor to form gordaite (see table 4.1). Likewise, the increased peak intensities of calcite and gehlenite were observed and this is probably influenced by the dissolution of other phases. Figure 4.17 shows diffractograms obtained from sample FA-A4. A series of diffraction peaks of the typical mineral assemblage such as anhydrite, halite and quartz are shown. The remaining peaks identified are cristobalite (SiO 2), caracolite 2+ [Na 3Pb 2(SO 4)3Cl], rutile (TiO 2), ulvöspinel (Fe 2 TiO 4), iron (Fe) and lazurite (sodalite)

[Na 6Ca 2Al 6Si 6O24 (SO 4)2]. An important feature of aging of this sample is the formation of gordaite.

110

CHAPTER 4

(a)

(b)

Figure 4.16 Diffractograms of FA-A3 aged at room temperature as a function of time; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), Ba (bassanite), Bo (boehmite), C (calcite), Cr (cristobalite), G (gehlenite), Hc (hydrocalumite), He (hematite), Hl (halite), Lz (lazurite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite) and Sy (syngenite). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

111

CHAPTER 4

Peaks of gordaite were not identified in the fresh FA-A4 sample. After 7 days, sharp diffraction peaks of gordaite appeared (strongest peak at 6.78 O 2 θ, PDF# 88-1359) and then increased progressively (Figure 4.17b), accompanied by the reduction of halite peak intensity. Furthermore, anhydrite, caracolite and lazurite occurring in the fresh FA- A4 sample disappeared as a result of aging. They may contribute to the formation of the

Na, K, Ca, Cl and SO 4-bearing minerals gordaite and syngenite. Diffractograms of sample FA-B1 as a function of aging time are shown in figure 4.18a. For the fresh FA-B1, the main crystalline phases are anhydrite, calcite, gehlenite, sylvite and quartz. Several minor phases ( e.g., lime CaO and rutile) have low-intensity peaks. Aging of this sample induced halite growth. The peak intensity of sylvite was fluctuating during the six-months aging. In contrast to other fly ash samples, gypsum is the major product, but neither gordaite nor syngenite were formed in aging experiments of sample FA-B1. Gypsum is indicated by its strongest diffraction peak occurring at 20.74 O 2 θ (PDF#74-1905) (Figure 4.18b). This peak became sharper and more intense with increasing aging time, while peak intensities of anhydrite and lime were reduced. This supports the hypothesis that anhydrite is converted into gypsum (Sabbas et al ., 2003). An important feature derived from XRD patterns of sample FA-B1 is the increasing peak (101) intensity of quartz during aging. The reason for the increased quartz content may be related to (i) the depletion of the SiO 2 in the amorphous phase of the fresh material, and (ii) the dissolution effect of other crystalline phases (Polettini and Pomi et al. , 2004). The peak intensities of gehlenite at all stages remained constant. Likewise, hydrocalumite was initially observed after one week, and then it remained unchanged further aging.

112

CHAPTER 4

(a)

(b)

Figure 4.17 Diffractograms of FA-A4 aged at room temperature as a function of time; (a) from fresh to 6 months, and (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), Ct (caracolite), Fe (iron), G (gehlenite), Gd (gordaite), Hl (halite), Lz (lazurite), M (magnetite), Q (quartz), R (rutile), Sy (syngenite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

113

CHAPTER 4

(a)

(b)

Figure 4.18 Diffractograms of FA-B1 aged at room temperature as a function of time; (a) from fresh to 6 months, (b) fresh and 6 months. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Ba (bassanite), Bo (boehmite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), L (lime), Lp (Lepidocrocite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

114

CHAPTER 4

4.3.3 Summary The mineralogical alteration of bottom and fly ash subjected to aging at room temperature and natural humidity has been demonstrated. The aging effects on the MSWI residues collected at two different incinerator plants A and B resulted in the formation of new minerals. Alteration products of bottom ash during or shortly effect quenching are anhydrite, calcite, ettringite, and iron oxides. Gypsum and hydrocalumite formed later from ettringite. Early stages alteration of fresh fly ash generated anhydrite, calcite, ettringite, and iron oxides. Subsequent aging leaded to the growth of gordaite, gypsum and syngenite. The experimental results indicate that natural aging of a few weeks to months may be employed as a pre-treatment strategy to decrease chemical reactivity, and enhance subsequent leaching characteristics of the residues, prior to their utilisation or final disposal. Investigations on mineral stability and leaching properties of all aged fly ash samples are presented in chapter 5.

4.4 Water-Extraction Processes of Fly Ash In this section, the water-washing and Soxhlet water-extraction processes are examined as a means of chemical and mineralogical stabilisation prior to disposal or reuse. Bulk chemical and mineralogical compositions of residues produced by these two extraction processes are discussed separately.

4.4.1 Water-Washing Process Five different types of MSWI fly ash (FA-A1, FA-A2, FA-A3, FA-A4 and FA- B1) were selected for the water-washing experiments. The batch washing experiments were conducted with a fly ash to water mass ratio of 1:10 (g/ml) according to the procedure outlined in section 3.5. The chemical characteristics of raw fly ash powder employed in the experiments are provided in table 4.1 (refer to section 4.1). The alkaline fly ash samples had a pH of 10.5-11.5 in a suspension of distilled water, except for the FA-A4 sample which had a lower pH value of 6.5-7.5. The bulk chemical compositions of the dried, insoluble residue of the washed fly ash were analysed by the XRF method (see section 3.5).

115

CHAPTER 4

Table 4.11 lists the chemical composition of the washed fly ash residues along with their raw fly ash compositions. In the experiments, approximately 30 wt.% of the sample weight was lost because of dissolution of chloride and sulfate compounds. In particular, nearly 99 % of Cl - was washed out from the fly ash. Consequently, a number of elements (Ca, Si, Al, P, Mg and Fe) tend to increased weight percentage due to the diluting effect of water-soluble compounds (Cl -, Na and K). It is noted here that the increased weight percentages of some elements in the washed fly ash residues are not generally indicating “insolubility” of these elements. Hence, elements of Al and Si in the forms of Al 2O3 and SiO 2 may be partly extracted during washing. As a consequence of the high release of water-soluble compounds and the low release of heavy metals, all washed fly ash residues tend to be enriched in heavy metals (Cd, Cr, Cu, Ni, Pb and Zn), thereby yielding a potentially hazardous material. However, the removal of chlorides from the fly ash may be beneficial for the fixation of heavy metals through immobilisation processes in the cementitious material, because salts (mainly chlorides) in the fly ash will negatively interfere with the solidification of fly ash material

2− (Mangialardi, 2003; Sabbas et al ., 2003). Additionally, the removal of SO 4 -phases may improve the characteristics of fly ash-cement mixtures in terms of setting time, early hardening and mechanical strength (Mangialardi, 2003) . The mineralogy of the washed fly ash samples was subsequently investigated by XRD (see section 3.4). XRD patterns of the washed fly ash samples hereafter designated as WFA-A1, WFA-A2, WFA-A3, WFA-A4 and WFA-B1, respectively are shown in figure 4.19a. Peaks of halite, sylvite and K 2ZnCl 4 are absent in all washed fly ash samples, suggesting that the chlorides were readily removed by washing. The washing process also resulted in the occurrence of the new phases ettringite (WFA-A1, WFA-A3 and WFA-B1), gypsum (in all washed fly ashes), and gordaite (WFA-A1, WFA-A2 and WFA-A4). Closer examination of the XRD patterns of the raw and washed FA-A3 samples indicates that syngenite and hydrocalumite detected in the unwashed sample were dissolved in water during the washing process (Figure 4.19b). Correspondingly, gypsum and ettringite precipitated and can be observed in the XRD pattern.

116

CHAPTER 4

Table 4.11 The quantity of chemical elements in the raw and washed fly ash samples. Element Sample (wt.%) FA-A1 WFA-A1 FA-A2 WFA-A2 FA-A3 WFA-A3 FA-A4 WFA-A4 FA-B1 WFA-B1 Si 5.63 6.95 9.30 10.38 12.09 9.46 5.52 8.65 12.39 11.30 Ti 1.02 1.53 1.24 1.76 1.34 1.68 0.60 1.09 1.05 1.29 Al 2.43 3.73 3.55 3.80 4.36 4.42 2.07 2.52 3.68 4.34 Fe 3.71 3.99 2.41 3.61 3.18 3.59 1.97 3.73 1.94 2.53 Ca 12.41 18.97 11.42 17.02 12.76 18.29 6.10 11.51 16.45 21.23 Mg 1.08 1.57 1.05 1.33 1.19 1.44 0.80 0.84 1.42 1.82 Mn 0.12 0.18 0.09 0.14 0.11 0.15 0.06 0.12 0.10 0.12 K 6.11 0.99 5.57 1.20 4.49 1.20 7.73 1.98 4.96 1.34 Na 10.38 3.09 6.40 3.17 5.20 2.65 10.12 4.47 4.44 2.00 P 0.40 0.55 0.58 0.56 0.62 0.55 0.48 0.60 0.55 0.52 Cl 8.32 0.56 7.03 0.39 5.86 0.28 7.50 0.86 6.75 0.55 Pb 1.36 2.17 1.39 2.18 1.05 1.58 2.36 4.53 0.71 0.95 S 4.11 5.49 5.58 4.75 4.61 3.84 7.95 4.87 4.27 3.30 Zn 4.91 8.09 5.29 8.22 3.92 6.02 8.23 11.17 2.26 3.09 ppm As 307 1029 477 928 389 778 744 1781 252 425 Ba 3470 5365 2183 3855 2455 3671 1847 3870 1404 2095 Bi 204 364 355 583 306 479 1473 2813 Cd 456 751 506 740 358 541 881 343 354 486 Co 262 265 37 235 42 65 51 190 31 364 Cr 2026 2924 1308 2181 1606 2407 1486 2959 895 1282 Cs 114 184 271 295 185 210 370 236 97 92 Cu 3513 5660 3882 6143 3940 5746 6224 12327 1071 1533 F 3042 5131 3351 5485 2637 5305 1928 5272 2992 5542 Ga 30 43 27 41 20 27 43 73 29 37 Mo 489 609 108 170 107 116 259 503 19 29 Ni 614 799 223 496 232 350 380 815 98 330 Rb 265 27 317 12 226 11 343 276 58 Sb 1630 2678 899 1442 602 925 2401 4408 723 1002 Sn 333 557 328 521 251 369 936 1740 192 266 Sr 542 744 281 397 311 398 182 315 408 489 V 86 122 47 100 62 105 43 94 51 83 Y 885 1436 914 1457 703 1072 1544 3002 467 636 Zr 232 336 1599 2499 1440 1937 137 248 239 317 *WFA-An and WFA-Bn stand for the water-washed MSWI fly ash along with the raw fly ash (FA-An and FA-Bn) extracted in water using liquid-to-solid ratio (L/S) = 10 (ml/gr).

117

CHAPTER 4

Specifically, gordaite was found in some of the washed fly ash samples (WFA-A2 and WFA-A4), where samples FA-A4 and WFA-A4 have highest Zn content of all fly ash samples (Table 4.11). The formation of gordaite in these materials may be attributed to soluble salts (mainly sulfate and chloride), which are still present in minor concentrations, and subsequently reacted with Na + and Zn 2+ to form gordaite. Table 4.12 summarises mineral abundance of the washed fly ash samples determined by Rietveld analysis (see section 3.4). A high amount of gypsum (>2 wt.%) was found in all washed fly ash samples, while a significant amount of ettringite (>1 wt.%) was only formed in the samples WFA-A1, WFA-A3 and WFA-B1. Findings of the hydrate phases gypsum and ettringite have been reported for a chemically similar washed MSWI fly ash by Mangialardi (2003). A significant amount of hydrozincite

[Zn 5(OH) 6(CO 3)2] was observed in the WFA-A4 sample. Less soluble minerals ( e.g., quartz and calcite) appear to increase in concentration because the diluting effect of the chloride minerals washed out from the samples is dropped. It has been previously reported that typical water-soluble compounds comprise CaClOH, CaCl 2·

Ca(OH) 2·H2O, NaCl, KCl, CaCl 2, while the less soluble compounds are reported to be

Ca(OH) 2, CaSO 4, CaCO 3, 7CaO ·2Al 2O3·CaCO 3·24H 2O, and SiO 2- polymorphs (Alba et al ., 1997; Eighmy et al ., 1995; Kirby and Rimstidt, 1993).

Finally, the formation of potassium alum [KAl(SO 4)2·12H 2O] was observed in washed samples FA-A4 and FA-B1. This mineral is similar to the phase encountered by Eighmy et al . (1995) in the acid-leached fly ash. According to the quantitative XRD analysis, the content of the amorphous phase remained unchanged during the washing process. A slight increase in the amorphous contents of the washed fly ash samples can be attributed to the removal of soluble chlorides and sulfates. In XRD analysis, the amorphous phase of sample WFA-A4 was not detected. This agrees well with the results of SEM/EDX and XRD analyses conducted on the raw sample FA-A4 (see section 4.1.2 and 4.1.3).

118

CHAPTER 4

(a)

(b)

Figure 4.19 Diffractograms of (a) washed (WFA-An and WFA-Bn) fly ashes and (b) the raw and washed FA-A3 samples. The principal diffraction peaks of the predominant minerals are shown for reference purposes. The peaks are labelled An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), M (magnetite), Po (portlandite) Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel).

119

CHAPTER 4

Table 4.12 XRD-based mineralogy of water-washed fly ash samples. Formula WFA- WFA- WFA- WFA- WFA- §Crystal Mineral (wt.%) (simplified) A1 A2 A3 A4 B1 structure reference Amorphous *43.2(6) 39.6(7) 41.0(9) n.d 52.7(7) Gypsum CaSO 4·2H 2O 16.4(2) 14.8(2) 6.5(1) 21.0(2) 2.0(1) S-26 Albite (plagioclase) NaAlSi 3O8 6.3(3) 5.0(2) 2.9(2) 4.6(3) 2.5(2) S-211 Anhydrite CaSO 4 5.2(1) 5.0(1) 4.5(1) 0.4(1) 4.6(1) S-80 C2S Ca 2SiO 4 3.1(2) 4.3(2) 2.5(2) 5.9(3) 4.4(2) S-69 Calcite CaCO 3 3.3(1) 3.0(1) 5.7(1) 1.9(1) 4.7(1) S-11 Gehlenite (melilite) (Ca,Na) 2(Mg,Fe,Si,Al) 3O7 2.9(1) 4.0(2) 2.9(1) 3.9(2) 3.6(1) S-165 C3S Ca 3SiO 5 2.8(2) 2.5(3) 1.9(2) S-128 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)Cl 2 S-778 (H 2O) 1.6 (H 2O) 8 2.1(1) 1.3(1) 0.3(1) Quartz α-SiO 2 2.0(1) 5.1(1) 4.1(1) 2.4(1) 4.4(1) S-1 Magnetite Fe 3O4 1.5(1) 1.8(1) 0.4(1) 5.2(1) S-50 Kalsilite KAlSiO 4 1.9(0) 1.9(1) 2.0(1) 3.9(1) 1.9(1) S-961 Augite (pyroxene) Ca 3Na 3Mg 3FeAl 1.6 Si 7O24 1.7(2) 7.2(2) 2.2(2) S-622 2+ Ulvöspinel Fe 2 TiO 4 1.2(1) 0.7(1) 0.3(0) 4.3(1) S-362 Ettringite 3CaO·Al 2O3·3CaSO 4·32H 2O 1.0(1) 3.4(2) 2.5(1) S-195 Nepheline KNa 3Al 4Si 4O16 1.0(1) 1.9(2) S-89 Alunite KAl 3(OH) 6(SO 4)2 2.0(2) S-996 Bassanite CaSO 4·0.5H 2O 2.9(2) 1.2(1) S-292 Boehmite AlOOH 0.4(1) S-43 Burkeite Na 6(CO 3) (SO 4)2 1.5(1) 0.9(1) 1.7(1) S-681 Butlerite FeSO 4(OH)·2H 2O 2.0(1) S-343 C3A Ca 3Al 2O6 0.6(1) 1.4(1) 2.0(1) 1.0(1) S-71 Corundum α-Al 2O3 0.8(2) S-40 Cristobalite SiO 2 0.4(1) S-37 Diopside CaMgSi 2O6 2.5(2) 2.3(2) S-971 Dolomite CaMgCO 3 0.5(1) S-31 Enstatite (Mg,Fe)SiO 3 3.2(3) 3.5(2) S-151 Fayalite (olivine) (Mg,Fe) 2 SiO 4 0.7(2) 1.3(2) 1.9(1) S-53 Galena PbS 0.2(0) S-66 Gordaite NaZn 4(SO 4)(OH) 6Cl(H 2O) 6 0.5(1) 1.0(1) 5.3(1) I-406090 Hematite Fe 2O3 0.8(1) 1.9(1) 2.9(1) 0.3(1) S-41 Hydrozincite Zn 5(OH) 6(CO 3)2 1.6(1) S-389 Iron Fe 1.1(0) S-140 Lepidocrocite FeOOH 0.6(1) 0.9(1) 0.4(1) 0.9(0) 0.6(1) I-31136 Marcasite FeS 2 0.4(1) 1.5(1) 1.6(1) 2.5(1) 0.3(1) S-362 Minium Pb 3O4 0.3(0) 0.5(0) 0.5(0) 3.1(0) S-713 Muscovite/illite KAl 3Si 3O10 (OH) 2 1.8(2) S-116 Periclase MgO 0.4(1) 0.6(1) S-96 Potassium Alum KAl(SO 4)2·12H 2O 2.7(2) 1.6(2) S-448 Portlandite Ca(OH) 2 0.6(1) 0.5(1) S-124 Rutile TiO 2 0.7(1) 1.0(1) 0.7(1) S-15 Sanidine KAlSi 3O8 2.4(2) 2.7(2) 3.3(2) S-21 Sodalite group Na 4Al 3Si 3O12 Cl 0.5(1) 0.6(1) 1.1(1) 1.2(1) 0.8(1) S-1031 Wüstite FeO 1.3(1) S-142 Total 100 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter (S for the SIROQUANT database, I for the ICSD database) followed by the entry number of the respective database. n.d= not detected WFA-An and WFA-Bn stand for water-washed MSWI fly ash of FA-An and FA-Bn.

120

CHAPTER 4

4.4.2 Soxhlet Water Extraction The solubility of salts in the fly ash and the extractable heavy metal concentrations in water were investigated with the Soxhlet apparatus according to the procedure outlined in section 3.2. The extraction experiments of the FA-A1 sample resulted in three fractions; (i) the extracted fly ash, (ii) particles precipitated from the solution, and (iii) the soluble components dissolved in the solution. The solution residue was then dried by evaporating at low temperature. The bulk chemical compositions of the three fractions were analysed by the XRF technique (see section 3.4) (Table 4.13). The repeated extraction processes with water at a liquid/solid mass (L/S) ratio of 10 (ml/g) resulted in a constant weight loss of the sample of approximately 35 %. This weight loss is mainly due to the solubility of sulfate and chloride minerals in water. As most of the highly soluble salts were extracted, only a small fraction corresponding to 0.07 wt.% Cl remained in the residual fly ash. Consequently, Cl was almost totally transferred to the evaporated solution. The compounds containing elements Fe, Mn and Ti were clearly insoluble. The data in table 4.13 indicate that the Soxhlet extraction is not effective for the removal of heavy metals such as As, Cu, Cr and Ni from the fly ash. Very low concentrations of Cr, Cu and Ni are found in either precipitated residue or in evaporated solution (close to detection limit). Only a low percentage of Pb could be extracted, because lead is incorporated in the almost water-insoluble phases (Pb 3O4 and PbO see table 4.4). The heavy metals with highest release are Zn and Cd, which occur with concentrations of 3.76 wt.% and 0.08 wt.% (765 ppm) respectively in the precipitated residue. The dissolution of some solid phases ( e.g ., hydrocalumite and caracolite) from the raw fly ash may be responsible for the release of Zn and Pb. Similarly, the release of

Cd may be attributed to the dissolution of minerals [ e.g ., Cd 5(AsO 4)3Cl and CdCO 3] as proposed by Eighmy et al . (1995). As most heavy metals are present as low-soluble compounds, they are strongly concentrated in the residues. However, the complexity of the dissolution/precipitation processes and the unidentified phases to which the heavy metals are bound may also contribute to the difficulty in determining the solid phase controlling solubility.

121

CHAPTER 4

Table 4.13 Bulk chemical composition of solid residues obtained from the Soxhlet extraction process of sample FA-A1. Element Raw Extracted Precipitated Evaporated Fly ash Fly ash Residue Solution wt.% Si 5.98 10.91 2.03 n.d Ti 1.09 1.82 0.01 0.01 Al 2.52 5.65 2.24 n.d Fe 4.02 6.62 <0.01 n.d Ca 12.49 14.09 32.61 6.64 Mg 1.15 2.15 0.33 0.17 Mn 0.13 0.21 <0.01 <0.01 K 5.70 0.73 0.65 18.74 Na 8.46 3.12 1.19 15.70 P 0.39 0.65 n.d <0.01 Cl 4.21 0.07 0.04 10.12 Pb 1.26 1.94 0.34 n.d S 6.75 1.99 18.68 13.66 Zn 4.57 7.40 3.76 <0.01 ppm As 486 713 118 19 Ba 3915 6280 2161 27 Bi 195 310 n.d n.d Cd 426 604 765 n.d Co 306 493 5 1 Cr 2282 3671 10 12 Cs 103 82 216 28 Cu 3613 5885 49 16 F 2841 4714 2919 443 Mo 486 480 262 293 Nd 24 31 17 6 Ni 735 1183 31 16 Rb 255 3 7 477 Sb 1535 2397 210 6 Sc 65 95 71 11 Sn 309 488 1 2 Sr 546 612 1467 233 V 96 154 16 2 Y 824 1281 218 n.d Zr 235 349 151 23 n.d = not detected

122

CHAPTER 4

Mineralogical phase compositions of the extracted residual fly ash, the precipitated residue, and the evaporated solution were also investigated by XRD (Figure

4.20a). The chloride minerals ( i.e. , halite and sylvite) detected in the raw fly ash disappeared on extraction, correspondingly these minerals reappeared in the evaporated solution. Anhydrite was only partly dissolved. Additionally, the Soxhlet extraction process resulted in the formation of gypsum, bassanite, ettringite and monosulfate in the precipitated residue. Figure 4.20b shows that gypsum and perhaps a minor amount of ettringite are also formed in the extracted fly ash sample, but aluminosilicates, carbonates and oxide minerals remain stable. The abundance of each phase was subsequently determined by XRD Rietveld analysis and the results are presented in table 4.14. The results show that in the extracted fly ash sample, the amorphous phase still dominates the bulk together with gehlenite, C 3S, feldspar, quartz and oxide minerals. This indicates that the amorphous phase remains nearly unchanged during the extraction process. Quantitative XRD analysis provided also a direct evidence of the dissolution and/or precipitation of solid phases, which control the chemical element solubility.

Solubility of K, Na and Cl in water was controlled by concentration products of halite and sylvite. Evidently, the evaporation of the solution yielded a white crystalline mass composed of halite (60.6 wt.%) and sylvite (19.6 wt.%). Also, anhydrite, syngenite, arcanite (K 2SO 4), hydrophilite and calcite were found in small concentrations in the evaporated solution. Anhydrite was converted to gypsum in hot water. Correspondingly, relatively high concentrations of bassanite (38.8 wt.%) and gypsum (12.9 wt.%) were found in the precipitated material. Bassanite was formed here by dehydration of gypsum due to the repetitive heating process of the precipitate material at the bottom of flask for longer time (7 days). On the other hand, the hydration of bassanite after cooling may not be able to occur, due to the limited access to water as a result of the thicker crust at the precipitated residue. Further, gypsum may react with calcium aluminate hydrate to produce ettringite and/or monosulfate according to (Taylor, 1990): + + + ⇒ ⋅ ⋅ ⋅ Al 2O3 3Ca (OH ) 2 3CaSO 4 29 H 2O 3CaO Al 2O3 3CaSO 4 32 H 2O (4.3) + + + ⇒ ⋅ ⋅ ⋅ Al 2O3 3 Ca (OH ) 2 3CaSO 4 9H 2O 3CaO Al 2O3 CaSO 4 12 H 2O (4.4)

Additionally, perhaps the [Al(OH) 3] concentration or other soluble ionic species controlled the precipitation of katoite [Ca 3Al 2(SiO 4)(OH) 8] and a minor amount of illite.

123

CHAPTER 4

(a)

(b)

Figure 4.20 XRD patterns of (a) solid residues extracted from sample FA-A1 by the Soxhlet device and (b) raw and extracted fly ash. The principal diffraction peaks for some of the more predominant minerals are shown for reference purposes. The peaks are labelled Al (alunite), An (anhydrite), Ac (arcanite), Ba (bassanite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), Kt (katoite), M (magnetite), Ms (monosulfate), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel). Note: SQRT refers to the square root of the intensity scale for the XRD patterns.

124

CHAPTER 4

Table 4.14 XRD based mineralogy of residues from the Soxhlet experiment of FA-A1. Mineral Formula Raw Extracted Precipitated Solute §Crystal (wt.%) fly ash fly ash residue evaporated Structure Reference Amorphous 45.3(6)* 50.2(8) n.d n.d Alunite-natro NaKAl 3(OH) 6(SO 4)2 0.9(1) S-459 Anhydrite CaSO4 5.6(1) 1.5 (1) 1.0(2) S-80 Albite 1.6(1) 1.2(2) S-155 (plagioclase) NaAlSi 3O8 Apatite Ca 5(PO 4)3OH 1.3(2) S-215 Arcanite K2SO 4 7.9(3) S-288 Bassanite CaSO 4·0.5H 2O 1.3(1) 38.8(5) S-292 Burkeite Na 6(CO 3) (SO 4)2 0.9(1) S-681 C2S Ca 2SiO 4 1.4(1) 0.5(2) S-69 C3S Ca 3SiO 5 2.1(2) 2.5(3) S-128 C3A Ca 3Al 2O6 1.3(1) 1.2(1) S-71 Calcite CaCO 3 1.4(1) 3.8(1) 3.6(1) 0.8(2) S-11 Caracolite Na 3Pb 2(SO 4)3Cl 0.2(0) I-024459 Cristobalite SiO 2 0.4(0) 0.6(1) S-37 Diopside CaMgSi 2O6 1.5(2) 1.5(1) S-133 Digenite CuS 0.2(0) S-562 Dolomite CaMg (CO 3)2 0.6(1) 1.5(2) S-31 Enstatite (Mg,Fe) SiO 3 2.5(1) 2.7(2) S-151 Ettringite 3CaO ·Al 2O3·3CaSO 4·32H 2O 0.3(0) 4.3(1) S-195 Garnet Ca 3(Al,Fe) 2(Si,P) 3O12 0.6(1) 0.7(1) S-552 Gehlenite (Ca ,Na) 2(Mg,Fe,Si,Al) 3O7 4.4(2) 7.3(2) S-165 (melilite) Gypsum CaSO 4·2H 2O 0.4(1) 14.9(2) S-26 Halite NaCl 5.7(1) 61.8(4) S-109

Hematite Fe 2O3 0.6(1) 2.0(1) S-41 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)Cl 2 0.6(1) S-778 (H 2O) 1.6 (H 2O) 8 Hydrophilite CaCl 2 0.3(0) I-026686 Hercynite FeAl 2O4 0.6(1) 1.4(1) S-426 Kalsilite KAlSiO 4 0.4(1) 1.5(2) S-961 Katoite Ca 3Al 2(SiO 4)(OH) 8 4.3(2) S-307 Lime CaO 0.3(0) S-150

Magnetite Fe 3O4 0.9(1) 2.8(1) S-50 Massicotite PbO 0.3(1) 0.3(1) S-809

Monosulfate 3CaO ·Al 2O3·CaSO 4·12H 2O 31.1(8) I-100138 Muscovite/Illite KAl 3Si 3O10 (OH) 2 0.6(1) 1.2(1) S-116 Minium Pb 3O4 0.2(0) 0.6(0) S-713 Nepheline KNa 3Al 4Si 4O16 1.4(1) 2.0(2) S-89 Portlandite Ca(OH) 2 0.7(1) S-124 Quartz α-SiO 2 2.9(1) 2.6(1) S-1 Rutile TiO 2 0.6(1) 0.5(0) S-15 Sanidine KAlSi 3O8 2.1(1) 4.1(2) S-21 Sodalite Na 4Al 3Si 3O12 Cl 0.3(0) S-1031 Sylvite KCl 0.6(0) 19.6(2) S-164 Syngenite K2Ca(SO 4)2·H 2O 9.6(2) 8.8(4) S-325 2+ Ulvöspinel Fe 2 TiO 4 0.8(0) 1.7(1) S-362 Wairakite Ca(Al 2Si 4O12 )(H 2O) 2 1.2(2) I-54152 Wurtzite ZnS 0.5(0) S-703 Zincite ZnO 0.8(0) 1.8(1) S-5 Total 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. n.d= not detected. §The reference is given as a letter (S for the SIROQUANT database, I for the ICSD database) followed by the entry number of the respective database .

125

CHAPTER 4

4.4.3 Summary Use of simple water-washing and Soxhlet water-extraction processes for improving chemical properties of fly ash for subsequent treatments has been demonstrated. Particularly, the water-washing strategy for fly ash can: (i) remove significant amounts of water-soluble salts, notably sodium and potassium chlorides, and (ii) promote the formation of hydrate phases such as gypsum and ettringite. However this process resulted in the heavy metals to be enriched in the washed fly ash due to the high release of water-soluble minerals and the low release of heavy metals. Further, the Soxhlet water-extraction process offers an alternative strategy to recover significant amounts of alkali chlorides and sulfates from fly ash. The extraction of sulfates from fly ash by water produced significant amounts of bassanite and gypsum with small amounts of monosulfate and ettringite, while the recovery of chloride minerals could be obtained by evaporating the solvent. However, similar to a water- washing process, the Soxhlet extraction is not effective in extracting the heavy metals, thus the extracted fly ash residue is a hazardous waste and needs to be properly treated before reuse or disposal. These results have practical implications in developing the fly ash powder precursor for a subsequent stabilisation process, where the low chloride content is of importance. The effect of water-washing as a pre-treatment of fly ash on the improvement of immobilisation process of heavy metals is considered in chapter 5.

4.5 Discussion Most previous studies of MSWI residues (bottom ash and fly ash) have focussed on the bulk chemical and mineralogical phase compositions to determine the overall properties of materials either as a waste product or a secondary raw material. In the first case, studies addressed primarily on the potential environmental problem that results from the disposal of residues in landfills, and on the possible release of toxic components to the groundwater (Forestier, and Libourel, 1998; Hjelmar, 1996; Meima and Comans, 1999; Sabbas et al ., 2003). In the second case, MSWI residues were investigated as a possible resource for the production of so-called ecomaterials (Ferreira et al ., 2003) such as an additive to concrete formulation (Dyer and Dhir, 2004; Ghgafoori and Cai, 1998) or as a raw material for ceramics, glass ceramics, and glasses (Andreola et al ., 2001; Boccaccini et al ., 1997).

126

CHAPTER 4

The results of the bulk chemical characterisation of the German MSWI residues presented here demonstrate that the non-volatile elements ( e.g., Al, Fe, and Si) are enriched in the bottom ash, while the volatile ones ( e.g., Cl, Na, K and S) are concentrated in the fly ash in addition to the heavy metal elements Pb, Zn, and Cd (Table 4.1). Moreover, the element contents of the MSWI residues are closely related to melting and boiling temperatures of the elements and/or chemical compounds (Table 2.2 in section 2.2). In the incinerator plants A and B, the temperature of incineration is set around 1000 OC. Hence, metals such as Ni, Cr and Fe, and oxides of Zn, Ni, Cr, Fe, Al and Si would be relatively less volatilised, because the melting points of oxides and metals are above the incinerator temperature. However, the melting points of all the chlorides, most sulfates and metallic Zn, Cu, Pb and Cd are below this temperature, which imply that these substances would be more volatile under the incineration conditions. Thus the heavy metals in fly ash may result from the volatilisation of chloride and sulfate compounds of Zn, Ni, Cu, Pb, Cd, Cr and Fe. However, the metallic forms of Zn, Cu, Pb and Cd may be oxidised into oxides in the incineration process and then volatilised in the flue gas. Although finding of the high volatile metal Hg in the fly ash has been previously reported (Hjelmar, 1996; Sabbas et al ., 2003), it could not be observed in the present study. The absence of Hg suggested that this element may not be caught in particles precipitating in the electrofilter, because it may be difficult to capture elemental mercury by adsorption or absorption. Mercury in municipal solid waste (MSW) is easily converted into mercury vapour in the combustor chamber, subsequently the vapour reacted with HCl in off-gas to produce mostly HgCl 2. Therefore Hg may be trapped in wet scrubbers and found in the APC (air pollution control) residue (see section 1.1) (Song et al ., 2004). Further significant amounts of volatile elements found in the fly ash suggested that the electrostatic precipitator captured a significant portion of metal and chloride compounds with deposition into solid particles. SEM observations show the presence of compounds containing the elements Cl, Na, K and S at the surface of spherical fly ash particles (see figure 4.3). This is probably due to volatilisation-condensation processes occurring in a region between the combustion area and the electrofilter (Eighmy et al ., 1995; Forestier and Libourel, 1998). In particular, the high content of the Cl in the fly

127

CHAPTER 4

ash becomes a more serious issue for possible reuse as construction materials, where Cl content in the fly ash should be less than 0.2 wt.% if it is mixed with the ordinary Portland cement (OPC) (Wang et al ., 2001). This implies that the high Cl content in the fly ash has to be removed prior to the resource recovery and reutilisation of the fly ash. A water-washing treatment concurrently with a stabilisation technique is very considerably promising to reduce the hazardous nature of fly ash. The macroscopic examination of the bottom ash revealed two major components: (i) refractory waste products, and (ii) melt products. The refractory waste products are generated by an incineration process whose melt temperatures have not been reached, while melt products or glasses are produced by melting of the bottom ash during incineration. XRD mineralogical analysis indicates that glasses, magnetite, quartz, and melilite are most abundant in the bottom ash (Tables 4.2 and 4.3). The results presented here prove that the studied MSWI fly ash contains significant proportions of crystalline and glass phases that could be determined by XRD Rietveld analysis (Tables 4.4 and 4.5). The major glassy materials are particularly present in the fly ash as a consequence of melt droplets formed through condensation of the flue gases (Eighmy et al ., 1995; Forestier and Libourel, 1998; Kirby and Rimstidt, 1993). In addition to glass phase, magnetite and ferrite spinel, as well as some silicates are identified. Also the formation of chloride minerals such as NaCl, KCl, and K 2ZnCl 4, which have been previously reported by Eighmy et al . (1995), is evident. Several processes occurred during incineration such as vaporisation, melting, crystallisation, vitrification, condensation, precipitation, and the flue gas clean up are supposed to influence the formation of complex mineralogy in the fly ash (Eighmy et al ., 1995; Forestier and Libourel, 1998). Lead and zinc are volatile at the incinerator temperature and could be identified in the crystalline form of minium (Pb 3O4), caracolite

[Na 3Pb 2(SO 4)3Cl], massicotite (PbO), K 2ZnCl 4 and wurtzite (ZnS). Specifically, the different contents of Pb in the FA-A1 sample that are obtained by the XRD quantitative phase analysis and by the XRF method, respectively indicate that almost 80 % of Pb is present in the crystalline phase. Similarly, it is estimated that almost 50 % of Zn is fixed in crystalline rather than in amorphous phases. The characterisation of individual bottom ash particles by optical microscopy and EPMA demonstrates the presence of Si-rich aluminosilicate glass (Figure 4.9 and

128

CHAPTER 4

table 4.6). This typical glass phase seems to be developed from a variety of mineral and glassy constituents already present in the parent waste material. The mineralogical composition of bottom ash particles in which Al-, Fe- or Si-rich grains could be clearly distinguished, is very heterogeneous. Some of these grains correspond chemically to melilite, plagioclase, quartz and spinel, which have also been identified by XRD. The feldspar is mainly plagioclase. The crystalline phases are likely preserved from the fresh bottom ash rather than have formed during aging. The temperature in the combustion chamber is considered to play important role in the development of mineral phases existing in the bottom ash. Pfrang-Stotz and Schneider (1995) have reported that the formation of new glasses indicates a waste bed temperature of 1000 OC or more. Magnetite and aluminous spinel are commonly generated in the temperature range of 600-1000 OC, while melilite is formed at 700-800 OC. Hematite is likely formed from magnetite by an oxidation process in the combustion chamber (Eusden et al ., 1999). The EPMA results presented here demonstrate that the studied fly ash particles contain a large number of phases that may be amorphous or crystalline (see figures 4.11 and 4.12; table 4.8). Moreover, the identified amorphous and crystalline phases considerably differ in chemical compositions within and between fly ash particles. The identified glass phase has a Ca-rich aluminosilicate composition. Similar observations have been reported for a comparable fly ash by Forestier and Libourel (1998). Observations on the polished thin-sections show that most particles contain Ca-, Na-, K- ,S-or Cl-rich grains (Tables 4.8, 4.9 and 4.10). Some of these grains chemically correspond to chloride and sulfate minerals which have been also identified by XRD. These phases, or grains that typically exist at the surface of fly ash particles may have crystallised from melt droplets formed through condensation (Eighmy et al ., 1995).

2+ 2− Based on the very limited number of fly ash particles examined, Ca and SO 4 may be concentrated in the crystals attached to the surface of glass spheres and not within the glass itself. A significant result of the present study is that some key trace elements of environmental or health concerns appear to be preferentially associated with a specific mineralogy. Tables 4.8 and 4.9 reveal, for example, that both Pb and Zn are present either in the form of amorphous or crystalline phases. These relations are also characteristic of the other heavy metals in the studied sample, as they were observed in

129

CHAPTER 4

many different areas observed by spot analyses. Specifically, the data document that in the studied fly ash, the Fe-rich particles are significantly enriched in Cr and Ni and to a lesser extent in Zn, relative to the Fe-poor particles (Table 4.10). Further EPMA investigations do not exhibit a distribution of trace elements in fly ash particles. Other studies, however, suggested that many trace elements are concentrated at or near the surface of fly ash particles (Forestier and Libourel, 1998).

2− Surface enrichment of several potentially toxic elements (Pb, Cr) as well as of SO 4 has been documented by various surface analytical techniques (Eighmy et al ., 1995). Such a nanometer-scale enrichment of chemical elements at the surface of fly ash particles is attributed to the condensation of volatilised chemical species from the flue gas (Eighmy et al ., 1995; Forestier and Libourel, 1998). Consequently, these zones of surface enrichment with heavy metals have a high potential leachability (Eighmy et al ., 1995).

Anhydrite (CaSO 4), calcite (CaCO 3), ettringite (3CaO ·Al 2O3·3CaSO 4·32H 2O) are initially formed after quenching of the hot bottom ash (see figures 4.13 and 4.14). In addition, the formation of portlandite [Ca(OH) 2] resulting from the hydrolysis of calcium oxide (CaO) is evident in the BA-B1 sample. Furthermore, gypsum

(CaSO 4·2H 2O) is formed with increasing aging time, correspondingly the gradual disappearance of ettringite phase is observed. The mineralogical alteration processes of the studied bottom ash involve the formation of Ca-hydrate phases (gypsum and ettringite) and portlandite (Speiser et al ., 2000). The alteration processes of the fly ash result in newly formed secondary minerals such as ettringite, gordaite and syngenite. It is supposed that calcite, Ca- sulfates, Na-and K-chlorides have taken part in the reactions of the weathering process (White and Brandley, 1995). Moreover, the mineralogical changes in the fly ash induced by aging may be employed for a simple pre-treatment of soluble salts including compounds of Pb and Zn leading to a stabilised mineral with improved leaching resistance (Lamers and Born, 1994; Sabbas et al ., 2003). The results of the water-washing process demonstrate that significant amounts of alkali chloride and sulfates could be removed from the fly ash. However, as can be deduced from the data in table 4.11, the washing process always produces a significant heavy metal enrichment in the fly ash, due to a significant removal of chlorides (85- 95 %) and to a lesser solubility of sulfates. Additionally, the washing step promotes the

130

CHAPTER 4

formation of hydrate phases such as ettringite and gypsum. The presence of ettringite is advantageous for incorporating and/or converting heavy metal compounds into less reactive forms (Mangialardi, 2003). Furthermore, the washing process of fly ash would yield a powder precursor with improved chemical characteristic required for the subsequent stabilisation and utilisation (Mangialardi, 2003). The water-extraction experiments with the Soxhlet device for the FA-A1 sample demonstrate that the solubility of phases containing elements Cl, Na, K and heavy metals can be assessed (Mangialardi, 2003; Speiser et al ., 2000). The principal phases of Ca, Cl, K, Na, and Zn were readily dissolved in water (Table 4.13). Correspondingly, the water-soluble minerals halite and sylvite are absent in the extracted sample and are largely found in the evaporated solution (Table 4.14). The precipitation of the Ca 2+ ,

3+ 2− Al - and SO 4 -bearing hydrate phases bassanite, ettringite, gypsum, monosulfate may be attributed to the dissolution of anhydrite and alunite. Ettringite appears to react to

2+ + 2− monosulfate at the boiling temperature of water. The dissolution of Ca , K and SO 4 in water corresponds to the formation of syngenite, as shown in the evaporated solute (Table 4.14). The XRD quantitative phase analysis of the extracted FA-A1 sample confirms that the content of amorphous phase remains stable and is not altered by the extraction process. Similar to the water-washing process, the Soxhlet extraction process is not capable of recovering significant amounts of heavy metals ( e.g., Zn, Pb and Cd) from the fly ash. The highest release of heavy metals is particularly observed for zinc, which may be related to the dissolution of the compounds K 2ZnCl 4 and gordaite

[NaZn 4(SO 4)(OH) 6Cl(H 2O) 6] (see table 4.4). Dissolution of caracolite and chloride minerals may be also responsible for the release of Pb and Cd. The remaining fractions of Pb and Cd and almost total Ni, Cr and Cu bound to glass phases, alloys, and oxides, which have very low solubility in hot water. The ineffective of Soxhlet extraction process to recover heavy metals from fly ash may be possibly due to the extraction agent employed. Kersch et al . (2002) reported that the complexing agent i.e., supercritical-fluid extraction (SFE) with CO 2 was effective in removing lead and zinc from municipal waste incinerator ash. Thus, attention to the use of different complexing agents for extracting heavy metals is an area of concern.

131

CHAPTER 5

CHAPTER 5

SYNTHESIS AND LEACHING ANALYSIS OF STABILISED MATERIALS FROM RAW AND WASHED MSWI FLY ASHES

The first and second parts of the chapter present results for the synthesis experiments of the stabilised materials from the raw and washed fly ashes through pozzolanic solidification and hydrothermal treatments. The chapter also presents results of the TCLP (toxicity characteristic leaching procedure) analysis for an assessment of the mineral stability and leachability of the stabilised materials in the acidic environment. Observations are presented on the heavy metal leaching from the tested materials compared with the respective regulatory limits.

5.1 Pozzolanic Solidification of MSWI Fly Ash The pozzolanic solidification of the raw and washed fly ashes for developing cement-like materials is investigated in this section. Results are presented for the solidification response of the materials in contact with a saturated solution of Ca(OH) 2. The pozzolanic behaviour of the raw and washed fly ashes is discussed separately.

5.1.1 Raw Fly Ash Pozzolanic solidification experiments for the FA-A1 material composition were initially performed at variations of the saturated solution/fly ash ratio, hereafter refers to the liquid/solid (L/S) ratio, and solidification time according to the procedure outlined in section 3.3. The phases formed after the solidification process were then characterised by means of XRD analysis and SEM/EDX (see section 3.5). The raw FA-A1 sample employed for the solidification experiments was the aged fly ash (one-year old). The chemical compositions of this sample are provided in table 4.1 (refer to section 4.1.1). Significant concentrations of heavy metals are present in this sample, where most abundant heavy metals are zinc and lead. Major phases identified in the raw FA-A1 sample are anhydrite, gehlenite, halite and sylvite in addition to minor minerals such as alunite, bassanite, calcite, hydrocalumite, portlandite and ulvöspinel (Figure 5.1a and table 4.4 in section 4.1.3). A significant amount of

132 CHAPTER 5

amorphous material (> 40 wt. %), probably Ca-rich aluminosilicate glasses, was recognised. Figure 5.1a presents XRD patterns of sample FA-A1 as a result of pozzolanic solidification at L/S ratio of 3 and various times. Peaks corresponding to syngenite

[K 2Ca(SO 4)2·H2O] (PDF#74-2159), gypsum(CaSO 4·2H 2O)(PDF#70-0982) and ettringite (3CaO ·Al 2O3·3CaSO 4·32H 2O) (PDF#72-0646) were observed at the first 7 days. All those hydrate phases are commonly found in the solidified products of Portland cement materials (Taylor, 1990). By increasing time (beyond 7 days), a slow increase of the production of hydrate phases was recognised, probably because of the solidification of pastes. The formation of a calcium silicate hydrate (CSH) phase (modelled by tobermorite-14Å with strongest peak at 7.76 O 2 θ, PDF#83-1520) was identified. Further, bassanite peaks disappeared, while a small reduction in peak intensities of anhydrite and sylvite was observed. It appears that those minerals may have undergone a base-induced dissolution process, which contributes compounds of

Ca, K and SO 4 to the formation of hydrate phases. However, many phases such as calcite, quartz, and gehlenite exhibited a weak reactivity with the saturated Ca(OH) 2 solution. Typically, hydrocalumite present in the raw FA-A1 sample was also resistant to the solidification process and therefore the peak intensity of hydrocalumite persisted in all the XRD patterns. Likewise, the pozzolanic solidification of sample FA-A1 at fixed L/S ratio of 10 for various times (1-3 months) resulted in hydrate phases (syngenite and gypsum) (Figure 5.1b). The XRD pattern of this solidified sample after 3 months showed a O strongest peak of bernalite [Fe(OH) 3(H 2O) 0.25 ] at 23.57 2 θ (PDF#81-2022). Principally, the overall pozzolanic reactions of FA-A1 greatly slowed down, in comparison with the experiment with L/S ratio of 3. Anhydrite slowly converted to gypsum. The disappearance of bassanite peaks and a significant reduction of peak intensities for halite in the solidified sample could be also observed in XRD patterns. The formation of the CSH phase close to tobermorite-14Å (PDF#83-1520) was also observed in all ages of the solidified sample. In the experiment with L/S ratio of 10, ettringite was deserved only after 3 months, probably because of the lower concentration of aluminate in the solution (Ubbriaco et al ., 2001).

133 CHAPTER 5

(a)

(b)

Figure 5.1 XRD patterns of FA-A1 solidified with the saturated solution of Ca(OH) 2 using liquid/solid ratios of (a) L/S = 3, and (b) L/S =10 at various times. The peaks are labelled Al (alunite), An (anhydrite), Ba (bassanite), Be (bernalite), C (calcite), Cr (cristobalite), E (ettringite), Fe (iron), Gy (gypsum), G (gehlenite), Hc (hydrocalumite), He (hematite), Hl (halite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Sy (syngenite), Us (ulvöspinel), Tb (tobermorite 14Å-CSH phase) and Th (thenardite).

134 CHAPTER 5

SEM observation (Figure 5.2a) on the solidified FA-A1 sample after 1 month at L/S ratio of 10 indicated syngenite of prismatic crystals with 2-5 m in diameter with main elements of S, K, Ca, O and Al. This indicates syngenite, which has a prismatic morphology according to Taylor, (1990). The small amounts of Al and Si in the EDX spectrum may be due to the small crystal on the surface of the prismatic crystal (Figure 5.2b). Phase compositions of solidified products of the FA-A1 sample were then determined using XRD Rietveld analysis as described in section 3.5.4. The XRD pattern of the FA-A1 sample solidified with the saturated solution at fixed L/S ratio of 3 for 14 days are shown in figure 5.3a. The quality of the Rietveld pattern-fitting may be gauged from the diffraction pattern plot in figure 5.3b. The intensity of most peaks is well represented in the calculated diffractogram. The phase compositions of the FA-A1 sample as a function of solidification time are shown in figure 5.4a. Hydrocalumite was initially formed; afterward the content remained roughly constant after course of the pozzolanic reaction. As the solidification proceeded, significant amounts of syngenite and gypsum (> 10 wt. %) were formed at times between 7- 28 days. Syngenite is salt that may develop rapidly in the alkaline environment (Taylor, 1990). The high yields of syngenite and gypsum appear to be associated with the high concentrations of Ca, K- and SO 4 in the parent material that were largely provided by the glassy phase. Only little part of the chloride compounds interacted with the hydrate phases. Further, the highest quantities of syngenite (> 20 wt. %) and gypsum (>15 wt. %) were achieved at 28 days of solidification. In contrast, a low abundance of ettringite (< 2 wt.%) was produced, probably because of the low quantity of aluminate phase available in the sample and the low amount of lime coming from hydrolysis. The pozzolanic solidification of sample FA-A1 resulted in the small quantity of the CSH phase. Interestingly, the pozzolanic solidification of the FA-A1 sample at L/S ratio of 10 for longer times (1-3 months) clearly resulted in the reduction of syngenite content opposed by the increased content of gypsum (Figure 5.4b). It appears that a significant quantity of sulfate from syngenite promoted the increased content of gypsum. Similarly, the contents of hydrocalumite and CSH phase are more likely to increase, but the abundance of glass remained constant.

135 CHAPTER 5

(a)

(b)

Figure 5.2 (a) SEM image of prismatic crystals of syngenite, and (b) EDX spectrum obtained from FA-A1 solidified with the saturated solution of Ca(OH) 2 at the L/S ratio of 10 for 1 month.

136 CHAPTER 5

(a)

Figure 5.3a Diffractogram of FA-A1 solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 14 days. The peaks are labelled An (anhydrite), C (calcite), E (ettringite), Gy (gypsum), G (gehlenite), Hc (hydrocalumite), Hl (halite), M (magnetite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel).

(b)

Figure 5.3b Quality of the Rietveld pattern-fitting results for FA-A1 solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 14 days where (·········) and ( ) denote the observed and calculated patterns, respectively.

137 CHAPTER 5

(a)

(b) .

Figure 5.4 Main hydrate phases as a result of FA-A1 solidified with the saturated solution of Ca(OH) 2 using ratios of: (a) L/S = 3, and (b) L/S = 10 at various times. Note that the amorphous phase in the solidified FA-A1 sample at L/S ratio of 3 and various times was not detected by the quantitative XRD analysis.

138 CHAPTER 5

The absence of any dramatic reduction of the glass content for this solidified

FA-A1 sample at L/S ratio of 10 may support the hypothesis that an excess of Ca(OH) 2 may form a protective surface layer around glass particles, thereby significantly reducing the dissolution of glassy phase during the pozzolanic reaction. Additionally, a new stabilised glass may be formed in the fly ash. Unexpectedly, the formation of ettringite was no longer found in pastes with L/S ratio of 10. The L/S ratio is a key variable for the development of ettringite and CSH phase for reducing the hazardous potential of fly ash (Mangialardi et al ., 1999; Ubbriaco et al ., 2001). Accordingly, smaller L/S ratios (< 3) are more promising result for the development of suitable solidification and stabilisation processes of the fly ash. To examine the pozzolanic behaviour of different fly ash samples, four other batches of fly ash (FA-A2, FA-A3, FA-A4 and FA-B1) were solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 3 according to the procedure outlined in section 3.3. The solidified products were then characterised by means of XRD analysis (see section 3.5). The chemical compositions of all raw fly ash samples employed in the synthesis experiments are provided in table 4.1 (refer to section 4.1.1). The mineralogical compositions of all raw samples are given in table 4.5 (section 4.1.3). More than 40 wt. % of all raw fly ash samples employed, except for the FA-A4 sample, are amorphous materials, which are probably Ca-rich aluminosilicate glasses. Specifically, major phases found in all samples are chloride and sulfate compounds. A number of hydrate phases have been developed in some samples during aging such as syngenite (FA-A2, FA-A3 and FA-A4), gordaite and monosulfate (FA-A4). All remaining phases such as quartz, magnetite and hematite remained inert during treatments with Ca(OH) 2 solution. The evolution of solidified products of four different fly ash samples at L/S ratio of 3 for various times (7-28 days) revealed a similar trend in the formation of hydrate phases (syngenite, gypsum and ettringite), whereas the solidification of pastes completed at around 7 days. Syngenite was readily formed in some fly ash samples (FA-A2, FA-A3 and FA-A4) at about 7 days, while anhydrite was found to disappear progressively. The pozzolanic solidification of the FA-A4 sample has resulted in

139 CHAPTER 5

marked increase of peak intensity of gordaite with a progressively disappearance of the halite and monosulfate peaks (Figure 4.8a and table 4.5 for comparison). XRD patterns of all solidified fly ash samples at an age of 28 days are presented in figure 5.5a. Peaks corresponding to gypsum (PDF#70-0982) and syngenite (PDF#74- 2159) were observed in all solidified samples, while the main peak of ettringite (PDF#72-0646) developed only in the solidified FA-B1. The formation of hydrate phases appears to follow the patterns normally encountered in the pozzolanic solidification of Portland cement (PC) (Taylor, 1990). Hydroxyzincite [Ca(Zn(OH) 3)2] (PDF#72-1100) was also detectable in the solidified FA-A4 sample. The formation of hydroxyzincite is probably due to the presence of Zn(OH) 2 in the solution that reacted with Ca(OH) 2 (Ubbriaco et al ., 2001). Furthermore, potassium alum [KAl

(SO 4)2·12H 2O] (PDF#75-2080) was formed in the solidified FA-A3 sample, whilst gordaite (main peak at 6.78 O 2θ, PDF#88-1359) formed in the solidified FA-A2 and FA-A4 samples. The major hydrate phases formed in all fly ash samples at an age of 28 days were determined by XRD Rietveld analysis (see section 3.5.4). A remarkable amount of syngenite was found in the solidified FA-A4 sample (>30 wt. %) (Figure 5.5b), which contained typically the highest content of K and SO 4. In contrast, the quantity of gypsum formed in the solidified FA-A4 sample is the lowest, probably due to the lowest content of the Ca-and SO 4 present in the raw material (Table 4.1 in section 4.1.1). The highest abundance of gypsum (>20 wt. %) was found in the solidified FA-A2 sample having the highest concentrations of Ca- and SO 4. The pozzolanic solidification of the FA-B1 sample at an age of 28 days resulted in a significant amount of ettringite (> 5 wt. %) (Figure 5.5b). This may be related to the significant amount of aluminates available in the solution at L/S ratio of 3. Moreover, a small amount of hydrocalumite was formed. There was probably a competition between the formation of ettringite (reaction between aluminates and sulfates) and the development of hydrocalumite (reaction between aluminates and chlorides) (Ubbriaco et al ., 2001). Further, gypsum was still found to be a major phase. However, syngenite was not recognised in the solidified FA-B1 sample, according to the smaller concentrations of Ca, K- and SO 4 in this sample (Table 4.1 in section 4.1.1).

140 CHAPTER 5

(a)

Figure 5.5a Diffractograms of FA-An and FA-Bn samples solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days. The peaks are labelled An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), Hl (halite), Hz (hydrozincite), Pa (potassium alum), Q (quartz), R (rutile), S (sylvite), Sy (syngenite) and Us (ulvöspinel).

(b)

Figure 5.5b Main hydrate phases as a result of pozzolanic solidification of FA-An and

FA-Bn samples with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days.

141 CHAPTER 5

5.1.2 Washed Fly Ash Pozzolanic solidification experiments were initially conducted on the WFA-A1 sample with the saturated Ca(OH) 2 solution of at L/S ratios of 3 and 10 with various times (see section 3.3). The solidified products were then characterised by XRD and SEM/EDX according the procedure described in section 3.5. The chemical composition of the WFA-A1 sample is provided in table 4.11 (refer to section 4.4.1). Approximately 40 wt. % of the WFA-A1 sample is amorphous phase; probably silicate glass (Table 4.12 in section 4.4.1). A number of phases such as gypsum, ettringite, hydrocalumite, anhydrite and Ca-bearing materials are present in the sample. A minor of gordaite was also found in the WFA-A1 sample. Quartz appeared to be poorly reactive in the saturated Ca(OH) 2 solution. Figure 5.6a reveals XRD patterns of the WFA-A1 sample solidified at the L/S ratio of 3 and various times. The dissolution of Ca-and SO 4-ions from the parent material yielded to a marked increase of gypsum (by about a factor of two). After 7 days, ettringite and hydrocalumite were still detectable, but the disappearance of gordaite was observed. Additionally, no increase of the hydrate phases could be observed, likely because the paste had completely solidified at around 7 days. Clearly, a number of the remaining parent phases such as gehlenite, calcite and portlandite remained stable during the pozzolanic reaction. Further abundance of gypsum, ettringite and hydrocalumite in the solidified WFA-A1 sample as determined by XRD Rietveld method (see section 3.5.4) is presented in figure 5.6b. Initially, the content of gypsum increases at the first 7 days of solidification, after that its content remains constant with solidification time. Similarly, the contents of ettringite, hydrocalumite and anhydrite remain constant for beyond 7 days. A dramatic decrease in the content of amorphous phase in the initial 7 days was observed just like in the unwashed samples at L/S = 3. The association of the disappearing amorphous phase and the increase of the gypsum content suggested that sulfate compounds reside in amorphous materials, or unstable sulfates are encapsulated in glasses. This is partly evidenced by EPMA, indicating that most sulfates appeared to be concentrated on the surface of glass particles (Table 4.12; spot 10 in section 4.4).

142 CHAPTER 5

(a)

Figure 5.6a Diffractograms of WFA-A1 solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 3 and various times. The peaks are labelled An (anhydrite), C (calcite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), M (magnetite), Po (portlandite), Q (quartz), R (rutile) and Sp (sphalerite).

Figure 5.6b Main hydrate phases as a result of pozzolanic solidification of WFA-A1 with the saturated solution of Ca(OH) 2 at L/S ratio of 3 and various times. Note that the amorphous phase in the solidified WFA-A1 sample at L/S ratio of 3 and various times was not detected by the quantitative XRD analysis.

143 CHAPTER 5

The pozzolanic behaviour of the WFA-A1 sample as a function of solidification time at fixed L/S ratio of 10 was also examined by XRD. The XRD analysis confirmed the presence of gypsum and ettringite in all solidified WFA-A1 samples (Figure 5.7a). The formation of ettringite involved a consumption of sulfate at the initial solidification of the paste, and then in the subsequent period, the pozzolanic solidification of the WFA-A1 sample produced the calcium silicate hydrate (CSH) phase. The formation of the CSH phase modelled by tobermorite-14Å became evident with a low-intensity peak at 7.76 O 2 θ (PDF#83-1520), while hydrocalumite detected in the parent sample was found to disappear. The phase abundance of each phase in the solidified WFA-A1 sample was then determined by XRD Rietveld analysis (see section 3.5.4). All solidified WFA-A1 samples contained gypsum as the dominant phase in addition to ettringite, hydrocalumite and calcium silicate hydrate (CSH) phase (Figure 5.7b). With increasing time, the production of hydrate phases (ettringite and CSH phase) only showed a limited increase due to the insufficient availability of aluminates in the sample and likely because of the solidification of the paste. Consequently, a significant quantity of sulfates remained in the solidified samples. This can be seen by no marked reduction in quantity of anhydrite with solidification time. Further pozzolanic solidification of WFA-A1 at L/S ratio of 10 did not significantly modify the glass content of the parent material. The reason for the absence of any marked reduction in the glass content is not obvious, but may be associated with the formation of protective surface layers over the glass particles due to an excess of Ca(OH) 2, and therefore the glass is converted to be less reactive form. SEM observation (Figure 5.8a) on the solidified WFA-A1 after 2 months at L/S of 10 indicates gypsum of prismatic crystals (tablets) with 10-20 m in diameter with main elements of S, Ca, and O. This indicates gypsum, which has a prismatic morphology according to Taylor, (1990). The small amounts of K, Fe, Al and Si in the EDX spectrum (Figure 5.8b) may be due to the small crystal on the surface of the prismatic crystal. Pozzolanic behaviour of different washed fly ash samples (WFA-A2, WFA-A3, WFA-A4 and WFA-B1) at fixed L/S ratio of 3 and various times was investigated by XRD. The chemical compositions of these samples are provided in table 4.11 (refer to section 4.4.1).

144 CHAPTER 5

(a)

Figure 5.7a Diffractograms of WFA-A1 solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 10 and various times. The peaks are labelled An (anhydrite), C (calcite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), He (hematite), M (magnetite), Q (quartz), R (rutile) and Tb (tobermorite-14Å-CSH phase).

(b)

Figure 5.7b Main hydrate phases as a result of pozzolanic solidification of sample

WFA-A1 with the saturated solution of Ca(OH) 2 at L/S ratio of 10 and various times.

145 CHAPTER 5

(a)

(b)

Figure 5.8 (a) SEM image of a prismatic morphology of gypsum, and (b) EDX spectrum obtained from WFA-A1 solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 10 for 2 months.

146 CHAPTER 5

The mineralogical compositions of all washed fly ash samples are given in table 4.12 (section 4.4.1). More than 40 wt. % of all washed fly ash samples employed, except for the WFA-A4 sample, are amorphous phases; probably Si-rich aluminosilicate glasses. A number of hydrate phases having developed during water-washing are gypsum (for all washed fly ash samples), gordaite (WFA-A1, WFA-A2 and WFA-A4), ettringite (WFA-A1, WFA-A2, WFA-A3 and WFA-B1) and hydrocalumite (WFA-A1, WFA-A4 and WFA-B1) (Figure 4.19 in section 4.4.1). All remaining phases such as magnetite, hematite, quartz and other minerals do not appear to contribute to the formation of hydrate phases. The XRD analysis indicated that a rapid formation of gypsum and ettringite in all washed samples was noted during the first 7 days of solidification, but no further change was observed for the longer solidification time (7-28 days). This is probably due to the solidification of pastes, which has occurred at the first 7 days. Figure 5.9a presents XRD patterns of all washed fly ash samples solidified for 28 days. Gypsum (PDF#70-0982) was a dominant phase in all solidified samples, while ettringite (PDF#72-0646) was only formed in some solidified samples (WFA-A2, WFA-A3 and

WFA-B1). Furthermore, potassium alum [KAl(SO 4)2·12H 2O] (PDF#75-2080) was observed in the solidified WFA-B1 sample. Gordaite (PDF#88-1359) and cerussite (PDF#85-1088) were detectable in the solidified WFA-A4 sample. The abundance of each main crystalline phase for all washed fly ash samples solidified for 28 days as determined by XRD Rietveld analysis (see section 3.5.4) is presented in figure 5.9b. An increase quantity of ettringite (>10 wt. %) was recognised in the solidified WFA-B1 sample. This may be ascribed to the availability of reactive aluminate phases supplied by the glassy phase. Interestingly, the content of gypsum formed in the solidified WFA-B1 sample was the lowest, probably due to the significant amount of gypsum being consumed for the production of ettringite (Ubbriaco et al ., 2001). A significant amount of potassium alum was also formed in the WFA-B1 sample. Furthermore, the pozzolanic solidification of WFA-A4 sample for 28 days resulted in the reduction of gordaite content as compared to that in the parent WFA-A4 material (see Table 4.12).

147 CHAPTER 5

(a)

Figure 5.9a Diffractograms of WFA-An and WFA-Bn samples solidified with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cs (cerussite-PbCO 3), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), M (magnetite), Pa (potassium alum), Q (quartz), R (rutile), Sp (sphalerite) and Tb (tobermorite 14Å-CSH phase) .

(b)

Figure 5.9b Main hydrate phases as a result of pozzolanic solidification of WFA-An and WFA-Bn samples with the saturated solution of Ca(OH) 2 at L/S ratio of 3 for 28 days.

148 CHAPTER 5

5.1.3 Summary The raw and washed fly ash samples solidified with a saturated solution of

Ca(OH) 2 resulted in the desired evolution of the hydrate phases. The main outcomes for the pozzolanic solidification of these samples are that: (i) A pozzolanic reaction and solidification occurred within 7 days, and no alteration reaction proceeded after 7 days. (ii) Formation of significant amounts of syngenite and gypsum, but small amount of ettringite and CSH could be achieved by controlling the liquid/solid (L/S) ratios. (iii) Amorphous fraction disappears at L/S ratio of 3. However, the amorphous fraction appears constant at L/S ratio of 10. This may be associated to the formation of a protective layer on the surface of glass particles as a consequence

of an excess of Ca(OH) 2 in the solution. Such a glass interaction may form a new stabilised amorphous phase. (iv) The pozzolanic behaviour of raw (unwashed) and washed fly ashes is similar, apart from a lower amount of ettringite formed in the washed fly ash. The mineral stability and leachability of heavy metals for the pozzolanic products examined by a leaching test are considered in section 5.3.

5.2 Hydrothermal Treatments of MSWI fly ash The benefits of using MSWI fly ash for synthesis zeolites and other neomorphic phases are examined in this section. Results are presented for the hydrothermal treatments of the raw and washed fly ashes investigated by a classical alkaline conversion method. The hydrothermal products obtained from the raw and washed fly ashes are discussed separately.

5.2.1 Raw Fly Ash Hydrothermal experiments were initially conducted on the FA-A1 material composition using an activation agent of NaOH and KOH at various temperatures and times as described in section 3.3. The chemical compositions of the raw FA-A1 sample employed in the experiments are provided in table 4.1 (refer to section 4.1.1). The hydrothermal products after the experiments were then characterised by XRD and SEM/EDX according the procedure outlined in section 3.5.

149 CHAPTER 5

The raw FA-A1 sample contained an amorphous phase as the major phase plus various SiO 2 and Al 2O3 bearing phases such as quartz and gehlenite (Table 4.4 in section 4.1.3). Approximately 55 wt. % of the FA-A1 sample consisted of probably a Ca-rich aluminosilicate glass. Other minerals present in the sample such as magnetite, hematite, chloride minerals and Ca-sulfates are considered as impurities for the present study as they do not contribute to the formation of zeolites or neomorphic phases. Figure 5.10a displays XRD patterns of the FA-A1 sample hydrothermally treated at temperatures of 90 to 180 OC in 0.5 M NaOH for 48 h. Halite, anhydrite and sylvite detected in the raw sample disappear at 90 OC. Correspondingly, hydrocalumite

[Ca 8Al 4(OH) 24 (CO 3)(Cl) (H 2O) 9.6 ] and monosulfate (3CaO ·Al 2O3·CaSO 4·12H 2O) were formed (strongest peaks at 11.19 O 2 θ, PDF#78-2051 and 9.89 O 2 θ, PDF#83-1289, respectively). However, gehlenite, which is expected to play an important role in providing the Ca 2+ -ion, remained unchanged after the treatments. In contrast, the peak intensity of quartz seems to increase after the treatments. This finding suggests that silicate from the amorphous phase may have been extracted by the NaOH solution to contribute an increased quartz concentration. Further, hydrothermal treatments of the FA-A1 sample at 140 OC yielded the hydrated compounds close to the synthetic minerals of tobermorite-11Å

(Ca 5Si 6O18 ·5H 2O) and illite [KAl 3Si 3O10 (OH) 2], as can be estimated from the peaks at 7.82 O 2 θ (PDF#45-1480) and 10.27 O 2 θ (PDF#26-091), respectively. Taking the positions of other peaks into account, Al-substituted tobermorite-11Å seems to be the most likely candidate with structural formula [Ca 5Na xAl xSi 6-x(OH) 2O16 ·4H 2O] (Taylor, 1990; Yao et al ., 1999). In addition, a hydrogarnet mineral close to katoite O [Ca 3Al 2(SiO 4)(OH) 8] could be observed in the XRD trace at 32.31 2 θ (PDF# 84- 0917). The reason for the formation of katoite may be estimated from the molar component ratios of CaO/[SiO 2+Al 2O3] and Al/[Al+Si], respectively. The ratios calculated from the XRF data of the FA-A1 sample provided values of 1.53 and 0.37 respectively, which are beyond the values required for the optimum yields of Al- substituted tobermorite- 11Å (Coleman and Brassington, 2003) (see section 5.4 for the discussion). Furthermore, analcime (PDF#89-6324), hydroxylcancrinite (PDF#88-1931) and sodalite (PDF#88-2088) developed up to the temperature of 140 OC, in addition to gypsum (CaSO 4·2H 2O) (PDF# 76-1746).

150 CHAPTER 5

(a)

(b)

Figure 5.10 XRD patterns of FA-A1 hydrothermally treated at various temperatures for 48 h in (a) 0.5 M NaOH, and (b) 0.5 M KOH. The peaks are labelled Ac (analcime), Al (alunite), An (anhydrite), Ba (bassanite), C (calcite), Cr (cristobalite), Fe (iron), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), Hcr (Hydroxylcancrinite), He (hematite), Hl (halite), Ilt (illite), Kt (katoite), Lp (lepidocrocite) M (magnetite), Ms (monosulfate), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Us (ulvöspinel) and Tb (tobermorite- 11Å).

151 CHAPTER 5

The formation of Al-substituted tobermorite-11Å was also confirmed by SEM/EDX analysis. Figure 5.11 shows a platy or lath-like morphology developed in the FA-A1 sample treated at 180 OC, which is a typical synthetic tobermorite-11Å structure (Yao et al ., 1999; Coleman and Brassington, 2003). Additionally, the resulting tobermorite-11Å structure may accommodate cation of Zn 2+ as shown by the EDX spectrum. Figure 5.10b reveals XRD patterns of the FA-A1 sample treated at various temperatures in 0.5 M KOH for fixed time of 48 h. At 90 OC, peaks of halite, anhydrite and sylvite clearly disappear; correspondingly new peaks of hydrocalumite and monosulfate appear. Up to a temperature of 140 OC, a mixed product of Al-substituted tobermorite- 11Å and katoite was readily formed in the sample. Treated at 160 OC, gypsum was formed and an increase in peak intensity with rising temperature is recognised. This is probably related to the observed depletion in hydrocalumite and monosulfate. SEM/EDX analysis also confirmed a typical lathlike structure of synthetic tobermorite-11Å developed at 180 OC, in that Ca and Si with varying small proportions of Al, Fe and Zn are shown in the EDX spectrum (Figure 5.12). Phase abundance for the hydrothermal products of the FA-A1 sample treated in either 0.5 M NaOH or 0.5 M KOH solution under various synthesis conditions was determined by XRD Rietveld analysis. The Rietveld analysis made use of refinements of the XRD data for the mixture of the treated FA-A1 sample with the internal standard O of CeO 2 (see section 3.5.4). The diffractogram of the FA-A1 sample treated at 180 C in 0.5 M NaOH for 48 h and mixed with internal standard is presented in figure 5.13a. The quality of refinements for this sample may be gauged from the diffraction pattern plot given in figure 5.13b. The intensity of most peaks is well represented in the calculated diffractogram. The phase abundance of the FA-A1 sample treated in the temperature range of 90-180 OC with 0.5 M NaOH for 48 h along with the abundance of the starting material are presented in table 5.1.

152 CHAPTER 5

(a)

(b)

Figure 5.11 (a) SEM image of a platy network Al-substituted tobermorite-11Å crystal morphology, and (b) EDX spectrum obtained from FA-A1 hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h.

153 CHAPTER 5

(a)

(b)

Figure 5.12 (a) SEM image of a platy network Al-substituted tobermorite-11Å crystal morphology, and (b) EDX spectrum obtained from FA-A1 hydrothermally treated at 180 OC in 0.5 M KOH for 48 h.

154 CHAPTER 5

(a)

Figure 5.13a Diffractogram of FA-A1 hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h. The peaks are labelled C (calcite), Ce (internal standard of CeO 2), G (gehlenite), Gy (gypsum), Ilt (illite), Kt (katoite), M (magnetite) and Tb (tobermorite- 11Å).

(b)

Figure 5.13b Quality of the Rietveld pattern-fitting results for FA-A1 hydrothermally treated with 0.5 M NaOH at 180 OC for 48 h where (······) and ( ) denote the observed and calculated patterns, respectively.

155 CHAPTER 5

Table 5.1 XRD-based mineralogy of the FA-A1 sample treated with 0.5 M NaOH. Mineral Formula Starting 90 OC 140 OC 160 OC 180 OC Structure (wt. %) Material 48 h 48 h 48 h 48 h Model§ Amorphous 57.4(9)* 45.0(9) 43.5(8) 34.8(9) 31.4(8) Albite (plagioclase) NaAlSi 3O8 1.1(2) 1.0(4) 0.4(2) 0.4(2) S-155 Alunite NaKAl 3(OH) 6(SO 4)2 1.4(1) 0.6(1) S-459 Analcime NaAlSi 2O6·H 2O 1.1(1) 0.9(1) 0.5(1) S-370 Anhydrite CaSO 4 4.5(3) 1.2(1) 0.7(1) 0.4(1) S-80 Apatite Ca 5(PO 4)3OH 2.0(2) 0.5(1) 0.6(1) 0.5(1) 0.6(1) S-215 Bassanite CaSO 4·0.5H 2O 0.7(1) S-292 Calcite CaCO 3 1.7(2) 3.3(1) 2.2(1) 2.0(2) 2.2(1) S-11 Caracolite Na 3Pb 2(SO 4)3Cl 0.4(1) 0.2(0) I-024459 Calcium titanite CaTiO 3 0.4(1) 1.4(1) 1.7(1) 1.3(1) 1.7(0) S-757 C3A Ca 3Al 2O6 0.9(1) 0.6(1) 0.6(1) 0.6(2) 0.6(1) S-71 C2S Ca 2SiO 4 0.8(1) 2.8(2) 2.2(2) 2.7(2) 2.2(2) S-69 C3S Ca 3SiO 5 2.5(2) 2.3(2) 2.0(2) 1.7(2) 2.3(2) S-128 Diopside CaMgSi 2O6 1.8(2) 2.3(2) 2.2(2) 2.2(2) 1.2(2) S-133 Enstatite (Mg,Fe)SiO 3 1.2(2) 3.0(2) 1.3(2) 1.3(2) 1.2(2) S-151 Forsterite (olivine) (Mg,Fe) 2SiO 4 2.0(2) 2.1(2) 1.6(2) 2.0(2) 1.7(2) S-53 Garnet Ca 3(Al,Fe) 2(Si,P) 3O12 0.4(1) S-552 Gehlenite (melilite) (Ca,Na) 2(Mg,Fe,Si,Al) 3O7 1.3(2) 2.4 (1) 3.1(1) 3.4(1) 2.9(1) S-165 Grossular Ca 3Al 2Si 3O12 1.1(1) 1.4(1) 1.5(1) 1.3(1) S-59 Gypsum CaSO 4⋅2H 2O 0.8(1) 1.2(1) S-355 Halite NaCl 5.3(1) 0.8(1) 1.2(1) 0.7(1) 0.8(1) S-105 Hematite Fe 2O3 1.8(1) 1.3(1) 0.8(1) 1.2(1) 0.7(1) S-41 Hercynite FeAl 2O4 0.7(1) 0.9(1) 1.0(1) 0.7(1) S-426 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)(Cl) 0.7(1) 12.1(2) S-778 (H 2O) 9.6 Hydroxylcancrinite Na 8(Al 6Si 6O24 )(OH) 1.4 (CO 3)0.3 0.8(2) 1.3(2 1.8(2) 1.5(2) S-383 (H 2O) 6..35 Illite KAl 3Si 3O10 (OH) 2 0.4(0) 0.7(1) 3.5(4) 8.9(2) S-116 Iron Fe 0.5(1) 0.2(0) S-140 Kalsilite KAlSiO 4 0.3(0) 1.1(1) 0.2(0) 0.6(2) 0.6(1) S-961 Katoite Ca 3Al 2(SiO 4)(OH) 8 0.4(2) 5.2(1) 5.8(1) 4.7(1) S-308 Lepidocrocite FeOOH 0.3(1) S-45 Minium Pb 3O4 0.8(1) 0.3(0) 0.3(0) 0.3(0) 0.3(0) S-713 Magnetite Fe 3O4 0.9(1) 1.4(1) 2.0(1) 2.3(1) 2.3(1) S-50 Monosulfate 3CaO ⋅Al 2O3⋅CaSO 4⋅12H 2O 0.3(0) I-100138 Nepheline KNa 3Al 4Si 4O16 1.2(2) 1.7(2) 1.6(2) 1.8(2) 1.1(2) S-89 Portlandite Ca(OH) 2 0.4(1) 0.7(1) 0.2(0) S-124 Pyrite FeS 2 0.5(1) 0.4(1) 0.4(1) 0.7(1) 0.8(2) S-29 Quartz α-SiO 2 1.5(1) 2.6(1) 1.7(1) 0.7(1) 0.5(1) S-1 Rutile TiO 2 0.4(1) 0.7(1) 0.4(1) 0.8(1) 0.9(1) S-15 Sanidine KAlSi 3O8 2.0(3) 2.2(2) 1.2(2) 1.8(2) 0.9(2) S-21 Sodalite Na 4Al 3Si 3O12 Cl 0.4(1) 0.6(1) 0.3(1) 0.3(1) 0.3(0) S-1031 Sylvite KCl 1.2(1) S-164 Thenardite Na 2SO 4 0.2(0) 0.5(1) 0.3(0) S-115 Tobermorite- 11Å Ca 5Na xAl xSi 6-x(OH) 2O16 15.4(2) 18.5(2) 21.8(2) S-309 ·4H 2O 2+ Ulvöspinel Fe 2 TiO 4 0.9(1) 1.2(1) 1.2(1) 1.3(1) 1.2(1) S-362 Zincite ZnO 0.3(0) 0.2(0) 0.2(0) 0.3(0) 0.3(0) S-5 Total 100 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter S for the SIROQUANT database (2002) and I for the ICDD followed by the entry number of the respective database.

156 CHAPTER 5

Quantitatively, the mineralogical phase compositions of the raw FA-A1 material changed after the hydrothermal treatment at 90 OC. There is an evidence for the reduction of glass content and an increased amount of quartz, compared to those in the starting material. The quartz content then reduced gradually with rising temperature. Synthesis at 90 OC also yielded a relatively high abundance of hydrocalumite (12.1 wt. %) and a small amount of monosulfate, together with a dramatic reduction of anhydrite, bassanite, halite and sylvite. It seems that a small amount of monosulfate was formed, likely related to the decomposition of ettringite, as this phase is unstable above temperature of 90 OC (Taylor, 1990). Evidently, contents of hydrocalumite and monosulfate decrease with increasing temperature, whereas the gypsum content increases. Other minor phases such as portlandite and iron hydroxide (lepidocrocite, FeOOH) were identified. The abundances of some minerals such as Ca-bearing minerals (e.g ., gypsum and anhydrite), magnetite and hematite remained constant with rising temperature and they would be still present in the hydrothermal product. The major phases arising from the treatment at 140 OC were found to be tobermorite- 11Å and katoite, but only minor amounts of zeolites i.e., analcime and hydroxylcancrinite were produced. Starting at a temperature of 160 OC, minor phases of illite and gypsum developed. Moreover, the hydrothermal treatment at 180 OC yielded a relatively high abundance of tobermorite-11Å (21.8 wt. %), illite (8.9 wt. %) and katoite (4.7 wt. %). Further, the mineral abundance produced in the FA-A1 sample treated at 90- 180 OC in 0.5 M KOH for 48 h is given in table 5.2. Treated at 90 OC, the amount of glassy phase reduced, while the quartz content increased. With increasing temperature, the mineralogical phase compositions of the raw FA-A1 sample were altered as a result of the dissolution of some minerals ( e.g ., NaCl and KCl) and the formation of new secondary precipitates ( i.e ., hydrocalumite and monosulfate). Hence, hydrocalumite (12.1 wt. %) was formed predominantly plus a small amount of monosulfate (0.3 wt. %) after the hydrothermal treatment at 90 OC. The reduction in hydrocalumite and monosulfate contents with rising temperature could be observed in association with increased gypsum content. Hydrothermal treatments of the FA-A1 sample at 140 OC produced significant amounts of tobermorite-11Å (11.0 wt.%), illite (8.2 wt.%) and katoite (4.9 wt.%). Up to 160 OC, tobermorite- 11Å and katoite became the major

157 CHAPTER 5

products. By treating at 180 OC in 0.5 M KOH for 48 h, only minor amounts of zeolitic materials ( e.g ., analcime and hydroxylcancrinite) could be generated, suggesting that the synthesis process was relatively insensitive to the raw material compositions, more so a function of temperature and time (Yao et al ., 1999; Querol et al ., 2002). The formation of tobermorite-11Å and zeolitic materials in 0.5 M NaOH was investigated at different reaction temperatures. Figure 5.14a shows the change in the main crystalline and glassy phase contents as a function of temperature and fixed time of 48 h. It can be seen that the content of glassy phase gradually decreases with rising reaction temperature to 180 OC. By contrast, the fraction of tobermorite-11Å, illite and katoite gradually increases with rising temperature. The reduction in glass content with reaction temperature appears to be associated with alumina and silica dissolution into the alkaline solution toward formation of tobermorite-11Å and katoite (Rimstidt and Barnes, 1980). On the other hand, the gehlenite content remains constant with increasing reaction temperature, suggesting that the amorphous phase is the only main contributor of Ca, Si-and Al-compounds for the formation of tobermorite-11Å and katoite. Moreover, the reduction of glassy phase with reaction temperature may result in a lower content of analcime and hydroxylcancrinite. Similar results have been reported by Yao et al . (1999). The formation of analcime and hydroxylcancrinite requires longer reaction time and higher concentration of NaOH than required for the development of tobermorite-11Å. It can be also seen that the quartz fraction was dramatically decreased in association with the increase of tobermorite-11Å content. Therefore, the SiO 2 and Al 2O3-bearing phases contributed to the tobermorite-11Å formation may be classified according to their solubility potential in the alkaline environment as follows: glassy phase > quartz, as proposed by Rimstidt and Barnes (1980). The effect of reaction temperatures on the formation of tobermorite-11Å and other zeolitic materials using the 0.5 M KOH solution for fixed time of 48 h was also examined. Alteration of the mineralogical phase compositions for the FA-A1 sample as a function of reaction temperature is shown in figure 5.14b.

158 CHAPTER 5

Table 5.2 XRD-based mineralogy of the FA-A1 sample treated with 0.5 M KOH. Mineral Formula Starting 90 OC 140 OC 160 OC 180 OC Structure (wt. %) Material 48 h 48 h 48 h 48 h Model§ Amorphous 57.4(9)* 51.8(7) 42.2(8) 26.0(9) 43.3(6) Albite (plagioclase) NaAlSi 3O8 1.1(2) 1.1(2) 1.0(2) 0.4(0) 0.4(1) S-155 Alunite NaKAl 3(OH) 6(SO 4)2 1.4(1) 0.3(1) 0.6(1) S-459 Analcime NaAlSi 2O6·H 2O 0.7(1) 1.4(2) 0.5(1) S-370 Anhydrite CaSO 4 4.5(3) 0.8(1) 0.6(1) 1.4(2) 0.8(1) S-80 Apatite Ca 5(PO 4)3OH 2.0(2) 0.6(1) 0.8(2) 0.7(2) 0.8(1) S-215 Bassanite CaSO 4⋅0.5H 2O 0.7(1) S-292 Calcite CaCO 3 1.7(2) 4.5(1) 3.5(1) 4.5(1) 2.2(1) S-11 Caracolite Na 3Pb 2(SO 4)3Cl 0.4(1) 0.2(0) I-024459 Calcium titanite CaTiO 3 0.4(1) 0.5(1) 0.6(1) 0.5(1) 0.4(1) S-757 C3A Ca 3Al 2O6 0.9(1) 0.8(2) 0.9(2) 1.0(2) 0.6(2) S-71 C2S Ca 2SiO 4 0.8(1) 0.3(1) 1.6(2) 3.4(2) 2.2(2) S-69 C3S Ca 3SiO 5 2.5(2) 2.1(2) 2.2(2) 3.3(3) 2.4(2) S-128 Diopside CaMgSi 2O6 1.8(2) 1.0(2) 2.0(2) 2.3(2) 1.9(2) S-133 Dolomite CaMg (CO 3)2 0.4(1) 0.2(0) 0.9(1) 0.2(0) S-31 Enstatite (Mg,Fe)SiO 3 1.2(2) 1.3(2) 1.1(2) 2.0(2) 0.7(2) S-105 Forsterite (olivine) (Mg,Fe) 2SiO 4 2.0(2) 1.4(1) 1.3(1) 2.6(1) 1.1(1) S-41 Garnet Ca 3(Al,Fe) 2(Si,P) 3O12 0.4(1) S-426 Gehlenite (melilite) (Ca,Na) 2(Mg,Fe,Si,Al) 3O7 1.3(2) 2.4(1) 3.0(1) 4.4(1) 2.7(1) S-778 Grossular Ca 3Al 2Si 3O12 0.5(1) 1.8(1) 2.1(1) 1.1(1) S-59 Gypsum CaSO 4·2H 2O 1.1(2) 0.8(1) S-355 Halite NaCl 5.3(1) 0.6(1) 0.6(1) 0.3(0) 0.3(0) S-105 Hematite Fe 2O3 1.8(1) 1.5(1) 1.1(1) 1.3(1) 0.9(1) S-41 Hercynite FeAl 2O4 0.7(1) 0.9(1) 1.0(1) 0.7(1) S-426 Hydrocalumite Ca 8Al 4(OH) 24 (CO 3)(Cl) 0.7(1) 12.4(2) 0.4(1) S-778 (H 2O) 9.6 Hydroxylcancrinite Na 8(Al 6Si 6O24 )(OH) 1.4 (CO 3) 0.4(1) 0.4(1) 0.5(1) 0.7(2) S-383 0.3 (H 2O) 6..35 Illite KAl 3Si 3O10 (OH) 2 0.4(0) 8.2(3) 9.7(4) 10.5(2) S-116 Iron Fe 0.5(1) 0.6(0) S-140 Kalsilite K AlSiO 4 0.3(0) 0.7(1) S-116 Katoite Ca 3Al 2(SiO 4)(OH) 8 0.6(1) 4.9(1) 6.3(2) 3.5(1) S-140 Lepidocrocite FeOOH 0.3(0) S-961 Minium Pb 3O4 0.8(1) 0.5(0) 0.3(0) 0.6(0) 0.2(0) S-713 Magnetite Fe 3O4 0.9(1) 0.7(1) 1.3(1) 1.3(1) 1.4(1) S-50 Monosulfate 3CaO ⋅Al 2O3⋅CaSO 4⋅12H 2O 1.7(4) I-100138 Nepheline KNa 3Al 4SiO 4O16 1.2(2) 1.0(2) 1.1(1) 1.2(2) 0.9(2) S-89 Portlandite Ca(OH) 2 0.4(1) 0.6(1) S-124 Pyrite FeS 2 0.5(1) 0.7(1) 0.5(1) 0.8(1) 0.4(0) S-29 Quartz α-SiO 2 1.5(1) 3.3(1) 0.9(1) 0.3(0) 0.3(0) S-1 Rutile TiO 2 0.4(1) 0.6(1) 0.7(1) 0.8(1) 0.7(1) S-15 Sanidine KAlSi 3O8 2.0(3) 1.3(2) 1.2(2) 1.4(2) 0.6(2) S-21 Sodalite Na 4Al 3Si 3O12 Cl 0.4(1) 0.6(1) S-1031 Sylvite KCl 1.2(1) 0.5(0) 0.9(1) S-164 Thenardite Na 2SO 4 0.4(1) 0.6(1) 0.4(1) S-115 Tobermorite- 11Å Ca 5Na xAl xSi 6-x(OH) 2 11.00(2) 14.0(2) 14.3(2) S-309 O16 ·4H 2O 2+ Ulvöspinel Fe 2 TiO 4 0.9(1) 0.9(1) 1.1(1) 1.3(1) 1.0(1) S-362 Zincite ZnO 0.3(0) 0.3(0) 0.4(0) 0.6(1) 0.4(0) S-5 Total 100 100 100 100 100 *Figures in parentheses indicate the least-squares estimated standard deviation (esd) referring to the least significant figure to left, a zero indicates an esd < 0.05%. §The reference is given as a letter S for the SIROQUANT database (2002) and I for the ICDD followed by the entry number of the respective database.

159 CHAPTER 5

(a)

(b)

Figure 5.14 Minerals produced from hydrothermal conversion of sample FA-A1 in (a) 0.5 M NaOH, and (b) 0.5 M KOH at different reaction temperatures and fixed time of 48 h. Note H-Cancrinite : hydroxylcancrinite.

160 CHAPTER 5

In the sample treated from 90 to 160 OC, a significant decrease in the glass content was observed together with an increase in the contents of tobermorite-11Å and katoite. However, the extent of reaction for the FA-A1 sample in the presence of KOH was considerably lower than that in the presence of NaOH, especially for the formation of tobermorite-11Å and katoite. Upon treatment at 180 OC, no remarkable reduction of the glass content was noted, while the tobermorite-11Å content remained constant. Insignificant changes in the amounts of analcime, hydroxylcancrinite, and illite with rising reaction temperature were observed. Further, XRD patterns of the FA-A1 sample treated in various NaOH concentrations at 180 OC for 48 h are presented in figure 5.15a. In 0.5 M NaOH the highest yield of tobermorite-11Å was obtained as evidenced by a high-intensity peak at 7.82 O 2 θ (PDF#45-1480). However, a decreased peak intensity of tobermorite-11Å is observed. It appears that the content of tobermorite-11Å was lower, when the sample was treated with higher concentrations of NaOH (1- 2.5 M). This unexpected departure from increasing NaOH concentration was probably due to ineffective extraction of Si and Al from the glassy phase. Upon treating in 1 M NaOH, however, katoite with the best crystallinity may be obtained, together with the formation of gypsum. No significant formation of analcime, hydroxylcancrinite, illite and sodalite was observed by increasing the concentration of NaOH. Moreover, sylvite appeared again upon treating in 2.5 M NaOH, possibly due to precipitation of chloride. Figure 5.15b reveals XRD patterns of the FA-A1 sample treated at 180 OC in 0.5 M NaOH solution for 36, 48 and 168 h, respectively. The strongest peaks of tobermorite-11Å and katoite are evident at 7.82 O 2 θ (PDF#45-1480) and 32.31 O 2 θ (PDF# 84-0917), respectively. With increasing time to 168 h, peak intensities of tobermorite-11Å and katoite remain unchanged. Similarly, peak intensities of analcime, hydroxylcancrinite and sodalite are stable. It appears that the best crystallinity of tobermorite-11Å, katoite and illite was formed at 180 OC in 0.5 M NaOH for 48 h.

161 CHAPTER 5

(a)

(b)

Figure 5.15 XRD patterns of FA-A1 hydrothermally treated (a) at 180 OC in 0.5, 1 and 2.5 M NaOH for 48 h, and (b) at 180 OC in 0.5 M NaOH for 36, 48 and 168 h. The peaks are labelled Ac (analcime), Al (alunite), An (anhydrite), Ba (bassanite), C (calcite), Cr (cristobalite), Fe (iron), Gy (gypsum), G (gehlenite), Hc (hydrocalumite), Hcr (hydroxylcancrinite), He (hematite), Hl (halite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sd (sodalite), Us (ulvöspinel) and Tb (tobermorite- 11Å).

162 CHAPTER 5

From the results of the hydrothermal treatments of fly ash in the respective range of 90 O-180 OC in 0.5 M NaOH for 48 h, the following scenario may be scheduled: (i) The amorphous content is dissolved during hydrothermal treatment at 90 OC and the major phase of hydrocalumite is formed whilst the quartz content increases. (ii) At intermediate temperature (140 O-160 O C), tobermorite-11Å and katoite as well as a few of zeolitic materials are initially formed. (iii) At high temperature (180 OC), Ca, Al- and Si-ions, which are available in the amorphous phase, contribute to the formation of clay “illite”, katoite and tobermorite- 11Å.

Hydrothermal treatments of other batches of fly ash The use of 0.5 M NaOH as a mineralising agent for generating tobermorite-11Å and other zeolitic materials from different batches of raw fly ash samples was examined. The selected synthesis conditions were reproduced in the four different fly ash samples of FA-A2, FA-A3, FA-A4 and FA-B1 (see section 3.2). The chemical compositions of the raw fly ash samples used in the synthesis experiments are provided in table 4.1 (see section 4.1.1). The mineralogical compositions of the raw fly ash identified by XRD analysis are given in table 4.5 (section 4.1.3). Quantitative XRD analysis indicated that more than 40 wt. % of the raw fly ash samples, except for the FA-A4 sample, were non- crystalline materials, which are probably an aluminosilicate glassy phase. In addition, various SiO 2 and Al 2O3-bearing phases such as quartz and gehlenite were significantly abundant in the raw bulk samples. The resulting minerals of the synthesis were then characterised by XRD according to the procedure described in section 3.5. Figure 5.16a shows XRD patterns of all fly ash samples treated at 180 OC in 0.5 M NaOH for 48 h. The XRD patterns of the treated FA-A2, FA-A3, FA-A4 and FA-B1 samples are principally almost identical, where the main phase of tobermorite-11Å is clearly resolved corresponding to peak at 7.81 O 2 θ (PDF#45-1480). An additional peak as a result of splitting of the peak of illite [KAl 3Si 3O10 (OH) 2] (PDF#26-091) is observed at 10.27 O 2 θ. However, tobermorite-11Å developed in the FA-A4 sample appears to be a poorly ordered crystalline phase, probably because of the insufficient amorphous phase available in this sample (see section 4.1.3). A number of other phases, which exist as secondary minerals, include katoite, gypsum and lepidocrocite. Additionally,

163 CHAPTER 5

hydrothermal treatments of the FA-A4 sample yielded vuagnatite [CaAl(OH)SiO 4]

(PDF#71-1795) and metahalloysite [Si 2Al 2O5(OH) 4] (PDF#74-1023) as can be estimated from the peaks in the vicinity of 29.79 O and 20.14 O 2 θ, respectively. Generally, hydrothermal treatments of all fly ash samples produced small amounts of zeolite-like materials such as analcime and hydroxylcancrinite. The main products in all four treated samples at 180 OC in 0.5 M NaOH for 48 h determined by XRD Rietveld analysis (see section 3.5.4) are presented in figure 5.16b. The dominant phase in all hydrothermally treated fly ash samples is tobermorite-11Å. The highest yields (> 40 wt. %) could be obtained in the treated FA-B1 sample. Katoite and illite accompanying the production of tobermorite-11Å were significantly produced in the treated FA-A2, FA-A3 and FA-B1 samples. On the other hand, hydroxylcancrinite was the only zeolitic mineral, which could be significantly produced in the FA-B1 sample. Moreover, only small amounts of analcime and sodalite were found in some of the treated samples, indicating that the amorphous phase of most fly ash starting materials is inappropriate to develop zeolitic materials under the experimental conditions examined in the study.

5.2.2 Washed Fly Ash Synthesis experiments were conducted on all washed fly ash samples using 0.5 M NaOH activation agent under various temperatures and times according the procedure described in section 3.3. The resulting minerals after synthesis experiments were characterised by means of XRD analysis and SEM/EDX (see section 3.5). The chemical compositions of all washed fly ash samples employed are provided in table 4.11 (refer to section 4.4.1). The mineralogical compositions of the washed fly ash samples are composed of a significant amount of amorphous phase and various SiO 2 and Al 2O3 bearing phases such as quartz and gehlenite (Table 4.12 in section 4.4.1). Approximately 40 wt. % of all washed fly ash samples, except for the WFA-A4 sample, consisted of a silica-rich glasslike phase. A number of phases such as gypsum, anhydrite, magnetite and hematite are regarded as impurities for the present study as they do not contribute to formation of zeolites or neomorphic phases.

164 CHAPTER 5

(a)

Figure 5.16a XRD patterns of fly ash samples (FA-A2, FA-A3, FA-A4 and FA-B1) hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h. The peaks are labelled Ac (analcime), C (calcite), Fe (iron), G (gehlenite), Gy (gypsum), Hcr (hydroxylcancrinite), He (hematite), Hl (halite), Hs (metahalloysite), Ilt (illite), Kt (katoite), Lp (lepidocrocite), M (magnetite), Po (portlandite), Tb (tobermorite- 11Å) and V (vuagnatite).

Figure 5.16b Minerals produced from hydrothermal conversion of fly ash samples (FA- A2, FA-A3, FA-A4 and FA-B1) at 180 OC in 0.5 M NaOH for 48 h. Note H-Cancrinite: hydroxylcancrinite.

165 CHAPTER 5

The effect of reaction temperatures on the synthesis of tobermorite-11Å and other zeolitic materials using the WFA-A1 sample under a fixed condition of 0.5 M NaOH for 48 h was examined by XRD (Figure 5.17a). At 90 OC, the disappearance of phases such as gordaite, ettringite and hydrocalumite is evident, while gypsum is still present. The reflection of quartz shows an increase in intensity. However, a number of the remaining parent phases such as gehlenite, calcite and portlandite remain stable and therefore they did not react in the alkali solution. Further, formation of tobermorite- 11Å and katoite was initially observed after treatment at 140 OC of WFA-A1. Moreover, the peak intensity of quartz was greatly lowered. This finding indicated that the silicate from quartz may have dissolved in the NaOH solution to form hydrogels, which is responsible for the formation of tobermorite-11Å, katoite and a few of zeolites such as analcime and hydroxylcancrinite. On the other hand, gehlenite remained unchanged after the reactions. The WFA-A1 sample treated at 180 OC resulted in the highest production of tobermorite-11Å, being evidenced by the highest intensity peak at 7.81 O 2 θ (PDF#45-1480) together with the formation of illite. The formation of Al-substituted tobermorite-11Å has been confirmed by SEM analysis, showing the existence of the typical platy morphology that is commonly found in the synthetic one (Figure 5.17b). The mixed product of Al-substituted tobermorite- 11Å and katoite formed in the treated WFA-A1 sample could be estimated over a typical range of the Ca/(Si+Al) and the Al/(Al+Si) ratios. From the XRF data, these ratios were computed to be 1.77 and 0.27, respectively. The ratio estimates are beyond of the range values (0 - 0.17) and (0.80 - 0.85), respectively required for the synthesis of optimum tobermorite-11Å (Taylor, 1990; Coleman and Brassington, 2003). Evidently, katoite was the additional product of tobermorite-11Å which might also accommodate Zn 2+ metal cations as revealed by EDX. Another synthetic route to the formation of tobermorite-11Å and zeolitic materials by hydrothermal treatments for the different washed fly ash samples (WFA- A2, WFA-A3, WFA-A4 and WFA-B1) was investigated by XRD according to the procedure outlined in section 3.5. The treatments of those samples were performed under a fixed condition at 180 OC in 0.5 M NaOH, for 48 h (see section 3.3).

166 CHAPTER 5

(a)

Figure 5.17a XRD patterns of WFA-A1 hydrothermally treated at various temperatures in 0.5 M NaOH for 48 h. The peaks are labelled Ac (analcime), An (anhydrite), C (calcite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), Hcr (hydroxylcancrinite), He (hematite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite), Q (quartz), Sd (sodalite), Tb (tobermorite- 11Å) and Us (ulvöspinel).

(b)

Figure 5.17b SEM image of a platy network Al-substituted tobermorite-11 Å crystal morphology obtained from WFA-A1 hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h.

167 CHAPTER 5

XRD analysis indicated that tobermorite-11Å and katoite could be reproduced successfully in all washed samples treated at 180 OC in 0.5 M NaOH, for 48 h (Figure 5.18a). In contrast, a poorly ordered phase of tobermorite-11Å could be seen in the treated WFA-A4 sample, probably due to relatively low aluminosilicate glass content. Moreover, hydrothermal treatments of the WFA-A3 sample yielded the highest yield analcime as can be estimated from the peak in the vicinity of 25.91 2θO (PDF#89-6324) in addition to the presence of albite at peak of 27.99 2 θ O(PDF#89-6423). A high- intensity peak of hydroxylcancrinite (PDF#88-1931) could be identified in the treated WFA-B1 sample. The main minerals of the synthesis from all washed fly ash samples determined by XRD Rietveld analysis (see section 3.5.4) are presented in figure 5.18b. In this case, the phase abundance of the treated WFA-A4 sample could not be determined by the XRD Rietveld analysis due to its poor profile shape of the diffraction pattern. Tobermorite-11Å and katoite were found to be the major phases in those four samples. The most abundance of tobermorite-11Å was observed in the hydrothermal product of the WA-A2 sample, while the largest amount of katoite could be produced in the WFA- B1 sample. Interestingly, most washed fly ash samples exhibited analcime and hydroxylcancrinite growth after treatment at 180 OC for 48 h. Hence, the WFA-A3 fly ash is probably the most suitable material for producing analcime. On the other hand, the highest abundance of hydroxylcancrinite could be gained in the treated WFA-B1 sample. The reason for easy formation of analcime and hydroxylcancrinite may be discussed in terms of the chemical compositions as presented in table 4.11. One of the important synthesis parameters considered is the Si/Al molar ratio. The Si/Al molar ratio in the powder precursor will mainly control the degree of achievement for the formation of analcime and hydroxylcancrinite. For example, analcime can be synthesised in the alkali solution containing the powder precursor with the Si/Al molar ratio ranging from 1.94 to 2.26 (Neuhoff et al ., 2004), while the synthesis of hydroxylcancrinite requires the powder precursor with the Si/Al molar ratio at 1:1 (Yao et al ., 1999).

168 CHAPTER 5

(a)

Figure 5.18a XRD patterns of washed fly ash samples (WFA-An and WFA-Bn) hydrothermally treated at 180 OC in 0.5 M NaOH for 48 h. The peaks are labelled Ac (analcime), Ab (albite), C (calcite), G (gehlenite), Gy (gypsum), Hcr (hydroxylcancrinite), Ilt (illite), Kt (katoite), Lp (lepidocrocite), M (magnetite), Po (portlandite), Tb (tobermorite-11Å) and Us (ulvöspinel).

(b)

Figure 5.18b Minerals produced from hydrothermal conversion of washed fly ash samples (WFA-An and WFA-Bn) at 180 OC in 0.5 M NaOH for 48 h. Note H- Cancrinite : hydroxylcancrinite.

169 CHAPTER 5

5.2.3 Summary A synthetic hydrothermal route to the formation of a stabilised material from the raw and washed fly ashes has been demonstrated. Hydrothermal treatments of these starting materials using mineralising agents (NaOH and KOH) under various reaction temperatures and times resulted in the desired evolution of the stabilised materials containing Al-substituted tobermorite-11Å, katoite and zeolitic materials, whilst a number of heavy metals may have been incorporated in those phases. The main outcome for the synthesis of the stabilised materials is that the high abundance of tobermorite-11Å can be developed by controlling reaction temperatures and times. NaOH was found to promote the rapid decomposition of the raw material and subsequent formation of the mixed product of tobermorite-11Å and katoite. An equivalent concentration of KOH resulted in a comparatively slower rate of decomposition of the parent phases and promoted the formation of the mixed minerals of tobermorite-11Å and katoite. A greater efficiency for dissolving quartz was achieved by NaOH rather than KOH. The use of washed fly ash is necessary for a significant development of analcime and hydroxylcancrinite. The XRD quantitative analysis indicated that the amorphous alumina-silica gel precursor did not completely crystallise to form tobermorite-11Å, katoite and zeolitic materials, and therefore it would be still present in the hydrothermal products. The stability and toxic potential of hydrothermal products are considered in section 5.3.

5.3 Stability and Toxic Potential of Stabilised Materials Stability and toxicity of the stabilised materials from the raw and washed fly ashes are examined in this section. Results are presented for the leaching resistance and heavy metal leaching of the materials investigated by the TCLP tests. The leaching property of the raw and washed fly ashes including the stabilised materials is discussed separately.

5.3.1 Stability of Raw Fly Ash and Stabilised Materials under Acidic Conditions Single-step TCLP tests were conducted on the raw fly ash samples, which had been naturally aged (6-12 months), and on the stabilised materials according to the procedure outlined in section 3.5. The stability analysis presented in this study may be

170 CHAPTER 5

used to provide a first ( qualitative ) explanation for the toxicity of the materials with particular reference to the heavy metal leaching-controlling phases. The major soluble crystalline phases, which control the heavy metal leaching under the acidic conditions, were then investigated by XRD, whereas the heavy metal concentrations in the leaching solution were determined by inductively couple plasma (ICP) spectrometry. The chemical compositions of the raw fly ash samples employed in the leaching experiments are provided in table 4.1 (refer to section 4.1.1). The significant amounts of heavy metals found in the raw fly ash samples are Zn, Pb and Cd. Moreover, the mineralogical compositions of the raw samples are summarised in tables 4.4 and 4.5, respectively (refer to section 4.1.3). The main crystalline phases were anhydrite, halite, sylvite, together with a lesser quantity of many other minerals, quartz, calcite, gehlenite and various oxides rutile and magnetite. Importantly, Zn-bearing phases were found to be gordaite [NaZn 4(SO 4)(OH) 6Cl(H 2O) 6] (in samples FA-A1, FA-A2, FA-A3 and FA-

A4), hydrozincite [Zn 5(OH) 6(CO 3)2] (FA-A2) and ZnCl 2 (FA-A1). Lead-bearing minerals found in the samples are minium (Pb 3O4) (FA-A1, FA-A3, and FA-A4), caracolite [Na 3Pb 2(SO 4)3Cl] (FA-A4 and FA-B1) and cerussite (PbCO 3) (FA-A4). Figure 5.19a presents XRD patterns of the leached raw samples obtained from the TCLP tests with the pH values between 5 and 7. A rapid formation of gypsum

(CaSO 4·2H 2O) in the leached samples is observed. Evidently, chloride minerals (NaCl, and KCl) detected in all raw fly ash samples disappear. By comparison of the phases identified in the FA-A1 sample before and after leaching, removal of the alkali chlorides might contribute to an increase of the content of mineral fractions in the leached sample (Figure 5.19b). Clearly, anhydrite is the less soluble phase (Wang et al ., 2001) and increases in its peak intensity as evidenced by an additional peak at 25.45 O 2 θ (PDF#86-2270). An increase in the gypsum content was observed, likely due to the dissolution of hydrocalumite. Moreover, no peaks of caracolite and ZnCl 2 are evident in the leached FA-A1 sample, suggesting that they were readily removed. Similarly, Zn- bearing minerals ( e.g ., gordaite and hydrozincite) disappear in the leached samples, so that those phases were dissolved in the leaching solution over the range of pH values (5- 7) investigated. Apparently, the crystalline phases, which accommodate most of the Pb and Zn, have a lower leaching resistance than the amorphous glass matrix, thereby leading to the high potential toxicity of the material.

171 CHAPTER 5

(a)

(b)

Figure 5.19 XRD patterns of (a) leached specimens of fly ash (FA-An and FA-Bn), and (b) unleached and leached FA-A1 specimens. The peaks are labelled Al (alunite), An (anhydrite), Bo (boehmite), C (calcite), Ct (caracolite), Fe (iron), G (gehlenite), Go (goethite), Gy (gypsum), Hc (hydrocalumite), He (hematite), Hl (halite), M (magnetite), Ms (monosulfate), Q (quartz), R (rutile), S (sylvite), Tb (tobermorite-14Å), Us (ulvöspinel) and Zc (ZnCl 2).

172 CHAPTER 5

The leaching-controlling phases for solidified products of the raw fly ash samples at an age of 28 days, hereafter denoted as HFA-A1, HFA-A2, HFA-A3, HFA- A4 and HFA-B1, were investigated by XRD. The main hydrate phases identified (see section 5.1) are syngenite, gypsum, hydrocalumite and ettringite. Furthermore, all solidified products contained mainly anhydrite, halite, sylvite, together with a lesser quantity of many other minerals, quartz, calcite, gehlenite and various oxides rutile and magnetite. Lead (Pb) present is related to minerals of minium (Pb 3O4) in the solidified samples (HFA-A1, HF-A3, HFA-A4); and cerussite (PbCO 3) in HFA-A4. Gordaite

[NaZn 4(SO 4)(OH) 6Cl(H 2O) 6] was found in the solidified samples HFA-A2 and HFA-

A4. Additionally hydrozincite [Zn 5(OH) 6(CO 3)2] found in HFA-A4 and ZnCl 2 in HFA- A1 may control the Zn-leaching. Figure 5.20a shows major minerals of the solidified products identified by XRD after leaching with pH values between 5 and 7. Gypsum, portlandite and anhydrite remain stable after leaching. In addition, hydrozincite and thenardite existing in the HFA-A4 sample remain stable. However, syngenite, hydrocalumite, ettringite, and chloride minerals (NaCl and KCl) are absent in all leached specimens. Apparently, the disappearance of many hydrate phases from all solidified products during leaching suggested that the pozzolanic solidification of fly ash do not guarantee to yield a low leachability of material. Closer examination of the HFA-A1 sample (figure 5.20b) reveals that the dissolution of some of hydrate phases and chloride minerals resulted in the increased peak intensities of anhydrite and gypsum. The peak intensity of calcite detected in some specimens is reduced after leaching. The depletion in calcite may possibly be related to the precipitation of gypsum, which required the dissolution of other less stable Ca- bearing minerals. The stability of materials obtained from the hydrothermal conversion hereafter designated as HTFA-A1, HTFA-A2, HTFA-A3, HTFA-A4 and HTFA-B1 was examined by the TCLP tests. The hydrothermal products of all fly ash samples after treatments with 0.5, 1 and 2.5 M NaOH at 180 OC for 48 h were selected for the leaching examinations. The pH values measured in the suspension at the end period of leaching processes ranged from 7 to 8. The solid residues obtained from the leaching tests of the hydrothermal products were then investigated by XRD.

173 CHAPTER 5

(a)

(b)

Figure 5.20 XRD patterns of (a) solid residues leached from solidified products (HFA- An and HFA-Bn) of fly ash samples at an age of 28 days, and (b) unleached and leached HFA-A1 specimens. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), Hz (hydrozincite), Hl (halite), Po (portlandite), Q (quartz), R (rutile), S (sylvite), Sy (syngenite), Sd (sodalite), Tb (tobermorite-14Å), Th (thenardite) and Us (ulvöspinel).

174 CHAPTER 5

Figure 5.21a shows the leached hydrothermal products presenting strong XRD peaks with some modifications of the patterns with respect to the unleached minerals. Sulfates, aluminosilicates, carbonates and oxides are still present, but sylvite and halite are absent. It appears that the release of K +, Na + and Cl - may be solely controlled by the solubility of halite and sylvite. The XRD analysis of the leached samples proved that tobermorite-11Å, and katoite together with illite, analcime and hydroxylcancrinite have a high leaching resistance in the acidic solution. Closer examination of XRD patterns of the HTFA-B1 sample recorded before and after the leaching tests (Figure 5.21b) provides evidence that no significant alteration of the main mineral compositions is observed, except for the absence of chloride minerals ( e.g ., halite and sylvite). Expectedly, the main phases ( i.e ., tobermorite-11Å and katoite) have high chemical resistance in the acidic solution. Importantly, the present results show that the hydrothermal treatments of fly ash would generate a low leachability of the minerals. Table 5.3 summarises the TCLP results, along with the limit concentrations of the heavy metals established by the US law for solid waste disposal. For all raw fly ash samples, the leaching property of most elements differs significantly between the samples examined. Cd, Pb and Zn concentrations in the leaching solution exceeded the US limits. Reasonably, the high leaching characteristics of Zn and Pb are probably due to most Zn and Pb incorporated in high soluble minerals ( e.g ., gordaite and caracolite) (Tables 4.4 and 4.5), which may be expected to have a lower chemical resistance than the amorphous phases (Rincon et al ., 1999). Hence, the Zn leaching from the FA-A4 is strongly influenced by the solubility of gordaite, which is the main phase of the aged FA-A4 sample (see section 4.2.2). Therefore, the toxic potential of the FA-A4 material correlates directly with the chemical durability of the gordaite mineral. Further Zn-leaching from the FA-A1 sample may be related to the dissolution of

ZnCl 2 (Figure 5.19b). Additionally, the high leaching of Cd may be influenced by the solubility-controlling minerals [ e.g ., Cd 5(AsO 4)3Cl and CdCO 3] as proposed by Eighmy et al . (1995), although specific Cd-bearing minerals were not detected by XRD, likely they are present below the detection limit. However, concentrations of Cu, Cr, Ni and As in the leaching solution from most raw fly ash samples have passed the TCLP tests, except for the FA-A4 sample.

175 CHAPTER 5

(a)

(b)

Figure 5.21 XRD patterns of (a) solid residues leached from the hydrothermal products (HTFA-An and HTFA-Bn ) of fly ash samples treated at 180 OC in 0.5 M NaOH for 48 h; (b) unleached and leached HTFA-B1 samples. The peaks are labelled An (anhydrite), Ac (analcime), C (calcite), G (gehlenite), Gr (grossular), Gy (gypsum), Hcr (hydroxylcancrinite), Hl (halite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite) and Tb (tobermorite- 11Å).

176 CHAPTER 5

Table 5.3 Concentrations of heavy metals in the leaching solution after 18-h reaction of leaching process for the raw fly ash samples and their stabilised materials. Element ppm ppb pH Sample Cu Cd Cr Ni Pb Zn As FA-A1 14 12 1 3 17 397 212 5 HFA-A1, 28 days 8 7 <1 <1 11 80 45 5 HTFA-A1 <1 <1 <1 <1 <1 27 19 7 0.5 M NaOH, 180 OC, 48 h HTFA-A1 2 2 <1 <1 1 15 5 8 1 M NaOH, 180 OC, 48 h HTFA-A1 <1 <1 <1 <1 <1 2 2 9 2 M NaOH, 180 OC, 48 h Hydrothermal solution 1 n.d <1 n.d n.d <1 3 n.d

FA-A2 27 14 2 <1 29 243 20 6 HFA-A2, 28 days 15 6 2 <1 7 50 5 6 HTFA-A2 <1 3 <1 <1 2 5 2 7 0.5 M NaOH, 180 OC,48 h

FA-A3 28 9 1 <1 34 342 14 7 HFA-A3, 28 days 21 5 <1 <1 4 261 7 7 HTFA-A3 <1 <1 <1 <1 2 8 9 8 0.5 M NaOH,180 OC, 48 h

FA-A4 166 38 4 2 35 806 137 6 HFA-A4, 28 days 133 29 4 1 34 391 66 5 HTFA-A4 2 8 <1 <1 12 150 30 7 0.5 M NaOH, 180 OC, 48 h

FA-B1 5 10 <1 <1 20 265 149 6 HFA-B1, 28 days <1 4 <1 <1 4 45 46 8 HTFA-B1 <1 <1 <1 <1 <1 3 19 8 0.5 M NaOH,180 OC, 48 h ppm *U.S. TCLP limits 100 1 5 100 5 5 5 *United State for the toxicity characteristic leaching procedure (TCLP) limits -FA-An and FA-Bn denote the raw MSWI fly ash -HFA-An and HFA-Bn represent solidified products of the FA-An and FA-Bn samples followed by the letters indicating the time of solidification. -HTFA-An and HTFA-Bn represent hydrothermal products of the FA-An and FA-Bn samples followed by the letters indicating the reaction conditions ( i.e., concentrations of NaOH, temperature and time) employed. -n.d = not detected.

1 It is noted here that no significant amounts of heavy metals leached in the solution during hydrothermal process, indicating that those metals are fixed in the minerals of hydrothermal product.

177 CHAPTER 5

The reasons of the low leachability of Cu, Cr, Ni and As are not obvious, but they might be mostly incorporated in the amorphous phase, alloys or oxides, which are insoluble in the acidic solution (Youcai et al ., 2002). The effect of pozzolanic solidification of fly ash samples on the immobilisation of heavy metals was also examined by leaching tests. Results for the leaching tests of solidified products of fly ash samples (HFA-A1, HFA-A2, HFA-A3, HFA-A4 and HFA-B1) at an age of 28 days are presented in table 5.3. The heavy metals leaching concentrations decrease dramatically compared to those from the raw fly ash samples. However, the leaching concentrations of Zn, Pb and Cd are still higher than the TCLP limits. The high leachability of these metals may be attributed to the presence of chloride minerals and hydrate phases, which were obviously leached out according to the qualitative XRD analysis. The toxicity potential for hydrothermal products (HTFA-A1, HTFA-A2, HTFA- A3, HTFA-A4 and HTFA-B1) acquired at 180 OC in 0.5, 1 and 2.5 M NaOH for 48 h was also assessed by the leaching tests. The results of those toxicity measurements are listed in table 5.3. In general, the hydrothermal products have the decreased toxicity, indicating that the concentrations of some heavy metals in the leaching solution are lower than the TCLP limit. Therefore, it can be inferred that the hydrothermal process enhanced the heavy metal’s fixation/stabilisation. According to the qualitative XRD analysis, all main phases of the hydrothermal products remained stable after leaching; indicating their relative stabilities to the aggressive leaching environment. It seems that Zn, Pb and Cd may replace parent ions (Al 3+ and Ca 2+ ) enclosed in the framework of silicates, which have a low leachability (Coleman and Brassington, 2003; Yao et al ., 1999). However, the Zn concentration found in the leaching solution from the HTFA-A4 sample has exceeded the TCLP limit. This may be due to the poor ordered crystalline structure of tobermorite-11Å and no glassy phase present in this sample for trapping metal Zn (see table 4.5). Interestingly, no marked differences of heavy metal leaching from the HTFA-A1 sample, which has been treated with the different molarity (0.5, 1 and 2.5 M) of NaOH at 180 OC for 48 h were observed. The leaching characteristics of heavy metals seem to be independent of the material structures in the HTFA-A1 sample (crystalline versus amorphous glassy matrix) (see figure 5.15a).

178 CHAPTER 5

5.3.2 Stability of Washed Fly Ash and Stabilised Materials under Acidic Conditions The TCLP tests were conducted on the washed and stabilised materials according to the procedure described in section 3.5. The major soluble phases of the washed fly ash and the stabilised materials, which control the heavy metals leaching, were then identified by XRD. The pH suspension after 18-h reaction of the leaching process was also measured. The chemical compositions of the washed fly ash samples are summarised in table 4.11. The concentrations of trace metals in the washed fly ash samples are higher than in the raw fly ash samples. Moreover, gypsum is a major phase in all washed fly ash samples (Table 4.12 in section 4.3.1). A large number of hydrate phases were also observed in the washed fly ash samples such as hydrocalumite (WFA-A1, WFA-A4 and WFA-B1), ettringite (WFA-A1 and WFA-A3), gordaite (WFA-A1, WFA-A2 and WFA-A4) and hydrozincite (WFA-A4) (see-figure 4.19). However, these hydrate phases, except for gypsum, disappear after leaching (Figure 5.22a). In one instance, for the leached WFA-B1 sample, the CSH phase peak close to tobermorite-14Å appears at 7.76 O 2 θ (PDF#83-1520) and was tentatively identified. Apparently, leaching has allowed the identification of the CSH phase, which may be originally present at low concentration. In addition, the potassium alum remained in this sample after leaching. A comparison of XRD patterns between the unleached and leached WFA-A1 samples shows that ettringite and hydrocalumite peaks completely disappear after leaching, correspondingly the peak intensity of gypsum increases (Figure 5.22b). This agrees with previous studies showing that hydrocalumite and ettringite are unstable in the solution with the pH ranges (5-7) (Rémond et al ., 2002; Taylor, 1990). Further, there is an increase of peak intensities of some minerals such as anhydrite, quartz and gehlenite, relative to those in the unleached WFA-A1 sample. Figure 5.23a shows XRD patterns of the leached solidified products of the washed fly ash samples at an age of 28 days (HWFA-A1, HWFA-A2, HWFA-A3, HWFA-A4 and HWFA-B1). Peaks of ettringite and hydrocalumite are clearly absent, while peaks of gypsum are still present. Similarly, other phases such as anhydrite and quartz were insoluble during leaching. But gordaite was dissolved in the acidic solution. Closer examination of the XRD patterns (Figure 5.23b) shows that ettringite and hydrocalumite are absent in the leached sample of HWFA-A1, suggesting that they are

179 CHAPTER 5

not resistant to the acidic solution at pH less than 7. Correspondingly, peaks of gypsum increase in intensity. In general, the stability of minerals controlling heavy metals leaching in the solidified products of the washed fly ash samples are similar in principal to those present in the washed fly ash samples. The present results also suggest that the solidified products of these samples may have the high leachability. The stability of hydrothermal products of the washed fly ash samples in the acidic environment was also investigated by XRD. The samples hereafter labelled as HTWFA-A1, HTWFA-A2, HTWFA-A3, HTWFA-A4 and HTWFA-B1 obtained from the hydrothermal treatment with 0.5 M NaOH at 180 OC for 48 h were selected for the leaching examinations. XRD analyses conducted before leaching indicated that all hydrothermal products contained tobermorite-11Å and katoite as the dominant phases plus minor zeolitic materials. In particular, the HTWFA-A2 sample contained a significant amount of analcime. After leaching, the XRD spectra depicted that no alteration of major phases occurred in the leached specimens (Figure 5.24a). Expectedly, the hydrothermal products of the washed fly ash contained the stable minerals including quartz, anhydrite, gehlenite and gypsum. This evidence agrees very well with the previous findings for the leached hydrothermal products of the raw fly ash samples. The comparison between the minerals identified in the unleached and leached HTWFA-A2 samples ( e.g ., tobermorite-11Å and katoite) provides the supporting evidence that these minerals are stable in the acidic environment (Figure 5.24b). Table 5.4 presents the TCLP results for the heavy metal concentrations in the leaching solution from the washed fly ash samples and their stabilised materials. It is noted that As was not detected in the leaching solution. The results reveal that, for all washed fly ash samples investigated, the concentrations of some heavy metals (Zn, Pb, Cu and Cd) in the solution exceed the U.S regulatory limits. The only exception is presented by the WFA-B1 sample, which has lower concentrations of Pb and Cu leached in the TCLP tests than the TCLP limits. The high heavy metals leaching of the washed fly ash samples in the solution may be also influenced by the high initial contents of the heavy metals (Table 4.11 in section 4.4.1).

180 CHAPTER 5

(a)

(b)

Figure 5.22 XRD patterns of (a) solid residues leached from washed fly ash samples (WFA-An and WFA-Bn ), and (b) unleached and leached WFA-A1 samples. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gd (gordaite), Gy (gypsum), Hc (hydrocalumite), M (magnetite), Pa (potassium alum), Po (portlandite), Q (quartz), R (rutile), Tb (tobermorite 14 Å-CSH phase) and Us (ulvöspinel).

181 CHAPTER 5

(a)

(b)

Figure 5.23 XRD patterns of (a) solid residues leached from solidified products (HWFA-An and HWFA-Bn ) of washed fly ash samples at an age of 28 days, and (b) unleached and leached HWFA-A1 samples. The peaks are labelled Al (alunite), An (anhydrite), C (calcite), Cr (cristobalite), E (ettringite), G (gehlenite), Gy (gypsum), Hc (hydrocalumite), Pa (potassium alum), Po (portlandite), Q (quartz), R (rutile) and Us (ulvöspinel).

182 CHAPTER 5

(a)

(b)

Figure 5.24 XRD patterns of (a) solid residues leached from hydrothermal products (HTWFA-An and HTWFA-Bn ) of washed fly ash samples treated at 180 OC in 0.5 M NaOH for 48 h, and (b) the unleached and leached HTWFA-A3 samples. The peaks are labelled Ab (albite), Ac (analcime), C (calcite), G (gehlenite), Gy (gypsum), Hcr (hydroxylcancrinite), He (hematite), Ilt (illite), Kt (katoite), M (magnetite), Po (portlandite), Tb (tobermorite- 11Å) and Us (ulvöspinel).

183 CHAPTER 5

Table 5.4 Concentrations of heavy metals in the leaching solution after 18-h reaction of the leaching process for the washed fly ash samples and their stabilised materials. Element ppm Sample Cu Cd Cr Ni Pb Zn As pH WFA-A1 138 35 3 9 393 1504 n.d 5 HWFA-A1, 28 days 8 22 <1 1 48 386 n.d 6 HTWFA-A1 2 5 <1 <1 2 <1 n.d 6 0.5 M NaOH,180 OC, 48 h

WFA-A2 103 42 <1 13 270 1636 n.d 6 HWFA-A2, 28 days 57 6 <1 <1 194 112 n.d 6 HTWFA-A2 <1 3 <1 <1 3 <1 n.d 7 0.5 M NaOH,180 OC, 48 h

WFA-A3 46 29 <1 1 237 956 n.d 6 HWFA-A3, 28 days 43 26 <1 1 143 387 n.d 6 HTWFA-A3 1 2 <1 1 2 12 n.d 7 0.5 M NaOH,180 OC, 48 h

WFA-A4 167 29 3 3 401 1200 n.d 5 HWFA-A4, 28 days 44 <1 <1 1 172 185 n.d 5 HTWFA-A4 7 <1 <1 <1 46 <1 n.d 7 0.5 M NaOH,180 OC, 48 h

WFA-B1 <1 17 <1 <1 3 114 n.d 8 HWFA-B1, 28 days <1 <1 <1 <1 <1 69 n.d 7 HTWFA-B1 <1 <1 <1 <1 <1 9 n.d 8 0.5 M NaOH,180 OC, 48 h ppm *US TCLP limits 100 1 5 100 5 5 5 * United State for the toxicity characteristic leaching procedure (TCLP) limits -n.d = not detected -WFA-An and WFA-Bn denote the washed fly ash samples (FA-An and FA-Bn) -HWFA-An and HWFA-Bn represent solidified products of the WFA-An and WFA-Bn samples followed by the letters indicating the time of solidification. -HTWFA-An and HTWFA-Bn represent hydrothermal products of the WFA-An and WFA-Bn samples followed by the letters indicating the reaction conditions ( i.e., concentration of NaOH, temperature and time) employed.

184 CHAPTER 5

The leaching behaviour of the washed fly ash samples could be explained by recalling that the mobilisation of the particular heavy metal (Zn, Pb and Cd) is influenced by the stability of hydrate phases ( i.e., ettringite and hydrocalumite). Ettringite and hydrocalumite are probably soluble phases controlling leaching of particularly Cd 2+ , Pb 2+ , and Zn 2+ -ions at the pH values examined. However, the concentrations of Cr and Ni in the leachate are below the limits, suggesting that these elements may be distributed in the amorphous phase, alloys (Fe-Cr-Ni) or spinel. The effectiveness of pozzolanic solidification of the washed fly ash samples in reducing the aqueous heavy metal concentrations was also examined (Table 5.4). The solidified products have lower concentrations of heavy metals leached in the TCLP tests than the washed fly ash samples. Thus the mobility of heavy metals could be considerably reduced by pozzolanic solidification. The concentrations of Pb and Cd in the leaching solution from the HWFA-B1 are very low and much lower than the limits. On the other hand, the leaching concentration of Zn from all solidified samples could not pass the limits. It is presumably that a significant amount of Zn was concentrated in the unstable phase of ettringite, as confirmed by SEM/EDX analysis. Hence, the high leaching property of Zn can be simply explained by the poorer leaching resistance of ettringite, or only small quantities of Zn and Pb could be fixed in the amorphous CSH phase (Taylor, 1990). However, the concentrations of Ni and Cr in the leaching solution are well below the limits. These elements should be largely present in the solidified samples as likely amorphous CSH phase, spinel or alloys with low leachability. Nevertheless, the pozzolanic solidification of some washed fly ash samples are ineffective for immobilising all heavy metals examined. Leached amounts of heavy metals from all hydrothermal products of the washed fly ash samples are given in table 5.4, where the leaching solution had pH values ranging from 6 to 8. Results show that the heavy metal TCLP leaching decreases considerably. The hydrothermal products have lower concentrations of Zn, Pb and Cd leached in the TCLP tests than the washed fly ash samples. For the HTWFA-A1 and HTWFA-A2 specimens, Zn and Pb concentrations in the leaching solution are below the TCLP limits. It suggests that a significant part of Zn and Pb could be incorporated into tobermorite-11Å and katoite phases as evidenced by SEM/EDX analysis. By contrast, the Pb concentration in the leaching solution from the HTWFA-A4 sample

185 CHAPTER 5

exceeds allowable limits and thus this hydrothermal product is still considered as hazardous waste. Some possible explanations for the high solubility of Pb in the HTWFA-A4 material are related to the lack of a well-crystallised structure of tobermorite-11Å and katoite or no amorphous phase, which can fix the Pb-ion as confirmed by SEM/EDX analysis. Nevertheless, low leachability of other elements (Cu, Cd, Ni, Cr and Zn) could be observed in the HTWFA-A4 sample. They are probably due to be present in the alloys ( e.g ., Fe-Cr-Ni, Cu-Zn and ZnCrO 4) or spinel, which have a low leachability (Sabbas et al ., 2003). Further, HTWFA-A3 and HTWFA-B1 are characterised as a toxic material because they exhibited Zn solubility slightly above the TCLP limit (Table 5.4), but they contained elements (Cu, Cd, Ni, Cr and Zn) leached in the solution well below the TCLP limits. As regards to these samples, most of the Zn release may be attributed to the dissolution of neo-formed unstable phases (amphoteric metal, hydroxides and carbonates). The solubility of amphoteric metals (Al, Pb and Zn) has been shown to increase under strongly acidic conditions (Sabbas et al ., 2003). Moreover, the solubility of Cd in the HTWFA-A1, HTWFA-A2 and HTWFA-A3 exceeded the TCLP regulatory limits, suggesting that a significant amount of this element could not be fixed in tobermorite-11Å and katoite.

5.3.3 Summary It was found that the leaching characteristics for the raw and washed fly ashes as well as their stabilised materials exhibited some interesting features based on the treatment methods employed (pozzolanic solidification and hydrothermal processes). The raw fly ash samples are hazardous because they presented TCLP leachabilities of Zn, Pb and Cd above the respective regulatory limits. A significant reduction in the TCLP leachabilities of the heavy metals to below the respective regulatory limits could be achieved by pozzolanic solidification and hydrothermal treatments of fly ash samples. Likewise, the washed fly ash samples presented high concentrations of heavy metals and high leachability of Zn, Pb and Cd exceeding the respective TCLP regulatory limits. However, the pozzolanic solidification and hydrothermal treatments of the washed fly ash samples resulted in stabilised products with reduced leaching of heavy metal ions to below the respective regulatory limits. The leaching of Zn, Pb and

186 CHAPTER 5

Cd exceeding the TCLP limits were also observed in some stabilised materials of the raw and washed fly ash samples. But no clear evidence of the effect of washing on the improved leaching resistance of fly ash materials by pozzolanic solidification and hydrothermal treatments could be established.

5.4 Discussion The presented results of pozzolanic solidification experiments of the raw and washed fly ashes in the saturated solution of Ca(OH) 2 showed that the materials are potentially cement-like materials. The overall pozzolanic processes are greatly slowed by solidification. The principal phases of solidified products were gypsum, syngenite, in addition to small amounts of ettringite, CSH phase and portlandite. The pozzolanic behaviour of the fly ash samples as a function of time agrees in general with the expected pozzolanic solidification found in the mixtures of Portland cement and the MSWI fly ash materials (Ubbriaco and Calabrese, 2000; Ubbriaco et al. , 2001). The results obtained for the pozzolanic solidification of the raw and washed fly ashes provide the mineralogical information to ascertain the fixation of particular heavy metals in the hydrate phase. The compositions and glass nature of the raw fly ash samples were considered an important source of sulfate interacting with the saturated solution of Ca(OH) 2 to produce gypsum and syngenite. Initially, the large quantity of gypsum was formed, but once the soluble sulfates within the pastes were exhausted, further reaction of gypsum and amorphous K 2SO 4 resulted in the formation of syngenite. In this context, the crystalline form of arcanite (K 2SO 4) could not be detected by XRD in any fly ash sample; therefore only the amorphous K 2SO 4 phase may have contributed to the formation of syngenite. Moreover, the quantity of anhydrite was slightly reduced, corresponding to the formation of syngenite in one part, while the remaining part was completely hydrated to gypsum. With the advancement of pozzolanic solidification, the content of alkali chlorides such as halite and sylvite remained unchanged. Similarly, hydrocalumite appears to be not reactive and remained stable during the pozzolanic solidification. The presence of chloride minerals and monosulfate in some fly ash samples may have also contributed to the production of syngenite and gordaite.

187 CHAPTER 5

Moreover, these hydrate phases appear to be much more abundant after 28 days of solidification. Unexpectedly, the formation of ettringite was not favoured under low alkalinity condition of the fly ash samples examined (pH values ranging from 10 to 11). This feature may be also attributed to the insufficient content of aluminate phases in some fly ash samples to react with sulfate. The presence of heavy metals (Pb and Zn) from the fly ash, in solution, could be also considered to interfere with the normal equilibrium of calcium for the ettringite formation (Glasser, 1993). Moreover, the remaining phases of solidified products identified are amorphous CSH gel or another amorphous phase which is probably due to a pozzolanic or latent hydraulic reaction in combination with

Ca(OH) 2. Both ettringite and the CSH phase are considered important phases for the immobilisation of the heavy metals such as Zn, Pb and Cd (Ubbriaco and Calabrese, 2000; Ubbriaco et al. , 2001). As regards to the washed fly ash samples, which are characterised by the low content of KCl and NaCl, the formation of ettringite during treatments with saturated solution of Ca(OH) 2 began after some days and with slower development. In contrast, a rapid development of gypsum was observed, while anhydrite was not altered by the pozzolanic reaction. The quick development of ettringite was observed in the solidified samples of WFA-A3 and WFA-B1. Apparently, the presence of the saturated Ca(OH) 2 solution favoured the attack of the glass particles, and then provided the greater availability of the aluminate phase to react with the whole sulfate for ettringite formation. However, a moderate quantity of gordaite was also noticed in some solidified products because a small quantity of chloride still remained in the materials. The limited quantity of ettringite formed in the particular types of samples (the raw and the washed fly ash samples) may possibly be due to the inappropriate quantity of Ca(OH) 2 and a less favourable amount of aluminate reactive phase available in the materials (Ubbriaco and Calabrese, 2000; Ubbriaco et al. , 2001). It has been reported by

Ubbriaco et al. , 2001 that the excess of Ca(OH) 2 has a significant influence on the formation of ettringite including calcium silicate hydrate (CSH). Thus the mass ratio of fly ash/lime-water is an area of concern with regard to the paste preparation for the development of ettringite and CSH phase.

188 CHAPTER 5

A simple synthetic route to the formation of zeolites and other neomorphic phases by hydrothermal treatments of MSWI fly ash has been demonstrated. The main results of Al-substituted tobermorite-11Å and katoite have been shown to be consistent and in agreement with previous experimental research of using waste materials (Coleman and Brassington, 2003; Miyake et al ., 2000; Querol et al ., 2002; Yao et al ., 1999). In addition, small quantities of illite, analcime and hydroxylcancrinite were produced from the raw fly ash by changing the synthesis parameters such as temperature, time, and concentrations of the activation solution. Although the fly ash samples employed have significant chemical and mineralogical variations, the composition variations of the hydrothermal products under same synthesis conditions are low. The low variations in the hydrothermal products may possibly be attributed to the insignificant variation of composition of the glassy matrix (Querol et al ., 2002; Yao et al ., 1999). Obviously, the FA-A1 and FA-B1 materials showed the highest potential for the syntheses of Al-substituted tobermorite-11Å and katoite. Mineralising agents can be used to control the relative abundance of potentially competing hydrothermal reaction products. In this study, NaOH was found to promote the rapid decomposition of the parent material of MSWI fly ash and subsequent formation of the mixed minerals of Al-substituted tobermorite-11Å and katoite. An equivalent concentration of KOH resulted in a comparatively slower rate decomposition of the parent phases and promoted the formation of Al-substituted tobermorite-11Å with smaller proportions of katoite. All of the hydrothermal experiments presented here were performed using the non-agitated batch hydrothermal treatments, which were expected to follow three stages mechanisms for the zeolite development through; (i) the dissolution of aluminium and silicon from fly ash, (ii) the deposition of aluminosilicate gel on the fly ash surface, (iii) the crystallisation of zeolites from aluminosilicate gel (Murayama et al ., 2002a; Murayama et al ., 2002b). Figure 5.11 confirms this in part, however, the zeolitic and other neomorphic phases with respect to the tobermorite-11Å have not grown to surround ash particles, encapsulating them and isolating them from the dissolution. Alternatively they have grown into the spaces between particles, and joined fly ash particles (present in close proximity to each other as sediment at the bottom of the non-

189 CHAPTER 5

agitated autoclave together like cement). This feature may also be attributed to the absence of viscous effects in the non-agitated system. An examination of the current literature indicates that there are three phases in the fly ash being a very important for contributing the aluminium and silicon to hydrothermal syntheses; (i) amorphous aluminosilicate glass, (ii) quartz (iii) gehlenite (Coleman and Brassington, 2003; Querol et al ., 2002; Yao et al ., 1999). All of the fly ash samples employed in the study revealed that the aluminosilicate glassy phase is the most abundant and most unstable phase in the hydrothermal environment. Accordingly it has the highest rate of dissolution and it is the largest contributor of aluminium and silicon to the hydrothermal products. There is also a general reduction in the quartz content during hydrothermal treatments. However, gehlenite did not react with alkali solution to form zeolites and other neomorphic phases. The residual non-reacted components of the fly ash sample corresponding to Ca-sulfate bearing minerals, carbonate, magnetite, ulvöspinel and hematite remained unchanged after hydrothermal treatments. The synthesis strategy employing the washed fly ash yielded a mixed product of tobermorite-11Å and katoite with minor contents of analcime and hydroxylcancrinite. The result also demonstrated that analcime could be produced in the WFA-A3 material with higher yield. The use of the washed fly ash appears to be promising for the development of zeolitic materials such as analcime and hydroxylcancrinite. The ratio of Si/Al has been considered to be the most important factor for the development of zeolitic materials, while synthesis parameters ( e.g., time and reaction temperature) played smaller role. According to Coleman and Brassington (2003) the optimum result of Al- substituted tobermorite-11Å could be generally achieved from the reaction compositions falling within the following molar component ratios: CaO 80.0 < < 85.0 (5.1) + SiO 2 Al 2O3 Al 00.0 < < 17.0 (5.2) Al + Si It has been reported by Coleman and Brassington (2003) that reaction times and proportions of aluminium and calcium have a significant influence on the production of

190 CHAPTER 5

Al-substituted tobermorite-11Å. However, the long reaction times and an increased proportion of aluminium and calcium would promote the formation of hydrogarnet, such as katoite and other calcium silicate hydrate (CSH) as additional phases to Al- substituted tobermorite -11Å. Hence, the flexibility of hydrothermal conversion affords a possibility for controlling composition and morphology of the product phases. Accordingly, the yield and extent of Al-substituted tobermorite-11Å product phase derived from MSWI fly ash could be modified by an appropriate manipulation of reagent compositions and reaction conditions in order to maximise its ion exchange performance. As calculated from the XRF data (Table 4.1), the molar component ratios of

CaO/[SiO 2+Al 2O3] and Al/[Al+Si] for all MSWI fly ash samples employed in the study are beyond the above values. Accordingly, an improved yield of tobermorite could be expected from the stoichiometric additions of a reactive source of silica and a calcium oxide to adjust the composition to within the optimal ranges quoted above. However, the development of tobermorite-11Å from a waste material like MSWI fly ash for marketable sorbents requires a degree of compromise between the effectiveness of the product in a given application and the cost and the complexity of processing. The use of an additional feedstock material and processing steps generally implies an additional cost, which requires compensating by the improved service performance. The XRD technique has been employed for the identification of solubility- controlling phases during leaching tests in order to understand heavy metal leachability. Leaching behaviour of some heavy metals from the raw fly ash samples appears to be governed by the dissolution of certain soluble phases such as K 2ZnCl 4, ZnCl 2 and hydrocalumite. The dissolution of hydrocalumite from the raw fly ash resulted in the precipitation of a large amount of gypsum. After leaching, a large change in the mineralogical phase compositions of the raw fly ash could be observed including re- precipitation of alunite. Syngenite, ettringite and hydrocalumite developed in the solidified products of the raw fly ash samples. These product phases were unstable in the acidic solution, but gypsum was quite stable. The decompositions of syngenite, hydrocalumite, and ettringite in acidic conditions have also been reported in experimental research (Taylor, 1990; Rémond et al ., 2002). Qualitative estimates of the solubility minerals of ettringite and hydrocalumite in the leaching solution at pH below

191 CHAPTER 5

7 are considerably lower than the maximum values of the solubility constants (log K) of these minerals reported in the literature (Geelhoed et al. , 2002). Therefore the solution was undersaturated for ettringite and hydrocalumite. However, all hydrothermal products behaved quite differently in that they were quite resistant in the acidic solution. As all raw fly ash samples were enriched with chlorides, the residual hydrothermal product contents increased by the dissolution of halite and/or sylvite after leaching. Furthermore, the leaching characteristic of the washed fly ash is mainly influenced by the presence of gypsum, hydrocalumite and ettringite. The XRD analysis of the solids after leaching provided similar evidence of the dissolution of hydrocalumite and ettringite in the acidic environment. But after leaching, the content of gypsum in the materials was increased. However, there is no difference between the leaching characteristics of solidified products from the washed fly ash and the raw fly ash. The results also demonstrated that hydrothermal products of the washed fly ash displayed superior stability in the acidic environment at the pH examined. The leaching experiments of the MSWI fly ash samples exhibited a significantly higher mobilisation of heavy metals in the pH range 5 to 8 than the TCLP limits for landfill leachate, indicating that the materials are hazardous. The heavy metal leaching of the different fly ash samples investigated were as expected, completely different. The results also showed that a variability of the metal leaching characteristics for the raw fly ash samples collected at different electrostatic precipitator locations in the same MSWI plant (FA-A2, FA-A3, and FA-A4). The incineration process, combustion conditions, and mineral compositions have a marked effect on the final release of metals (Sabbas et al ., 2003). It is well established that the matrix compositions of the fly ash sample could substantially control the leaching of the metals. This can be observed in part for the higher leachability of Cd which is probably related to the dissolution of NaCl and KCl

(Brunori et al ., 2001). For the FA-A1 sample, gordaite, zinc chloride (ZnCl 2) and

K2ZnCl 4 might be responsible for the leaching of zinc, while a part of lead leaching was possibly influenced by the dissolution of caracolite. Furthermore pH is the main experimental condition influencing the heavy metal release; the lower the pH, the greater the heavy metal release. Zn and Pb are amphoteric with hydroxy complexation, exhibiting increased solubility under strongly acidic conditions (Sabbas et al ., 2003). The present study confirms the findings of the

192 CHAPTER 5

experiments that the acidity of the leaching solution had an impact on the solubility of metals, especially zinc and lead. The results also show that a significant amount of cadmium was dissolved in the acidic leaching solution and this agrees very well with the expected behaviour reported in the literature (Youcai et al ., 2002). Therefore the acidic solution can be used for a partial extraction and recovery of lead and zinc from the fly ash, but not for ultimate detoxification of the fly ash. It is suggested that leaching tests over wide ranges of pH values with the use of extraction fluid will be completed to fully understand the complexity of leaching resistance in these materials. In contrast, the concentrations of heavy metals in the leaching solution from the washed fly ash were higher than those in the solution from the raw fly ash samples. This may be attributed to the washed fly ash being enriched in heavy metals after the washing process (Mangialardi et al ., 1999). However, As was not detected in the leached washed fly ash samples. Moreover, the metal concentrations in the leachate were considerably higher than the TCLP limits; therefore the washed fly ash is still considered as hazardous waste. The pozzolanic solidification and hydrothermal treatments of the raw and the washed fly ash samples had an evident impact on the reduced leaching of metals, in that all metals showed a marked reduction in the leaching concentrations (Tables 5.3 and 5.4). It is supposed that the reduction in mobilisation of heavy metals examined is caused by the incorporation of metals in phases with low solubility. For example, Zn, Pb and Cd may possibly be incorporated in the CSH phase (solidified product) and tobermorite-11Å (hydrothermal product), respectively. However, the hydrothermal method may be more effective for the development of chemically stabilised products. Indeed, the pozzolanic solidification could reduce dramatically the leachability of the fly ash, if this material could be properly prepared by extracting the heavy metals to some extent with chemical agents such as sodium hydroxide or ethylenediaminetetraacetic disodium salt (EDTA) (Youcai et al ., 2002). Nevertheless, from the economic of point of view, the cost for both pozzolanic solidification and hydrothermal treatments should be relatively lower than that for high temperature volatilisation and evaporation of heavy metals or vitrification.

193 CHAPTER 6

CHAPTER 6 CONCLUSION

6.1 Characteristics of MSWI Residues This thesis has examined the characteristics of German MSWI residues (bottom ash and fly ash) from two local MSWI plants A and B (MHKW Iserlohn and MHKW Essen), respectively, located in the Ruhr-industrial area. Chemical and mineralogical phase compositions of bulk and individual residue particles, particle morphology and size, mineralogical reactions during aging, and water-extraction experiments of the fly ash have been investigated.

Chemistry and Mineralogy The major findings and conclusions from the characterisations of bulk chemical and phase compositions, particle morphology and size of the bottom and fly ashes can be summarised as follows: (i) XRF analysis indicated that the bottom ash samples from two incinerator plants A and B mainly consisted of less volatile oxides of Ca, Si, Al and Fe with the high melting points (> 1000 OC), while more volatile chlorides and sulfates of Ca, Na, K, Zn, Pb and Cd in addition to oxides of heavy metals ( e.g ., Zn, Pb and Cd ) with low melting points (< 1000 OC) were concentrated in the fly ash. Specifically, the fly ash samples from incinerator A were enriched in Cl, Zn and Pb compared to those present in the fly ash specimens from incinerator B. This may be attributed to the composition of feed municipal solid waste (MSW) burn in the incinerator. (ii) Optical and SEM observations indicated that all bottom ash samples consisted of large particle aggregates (< 2 mm in size) with angular shape. In contrast to bottom ash, the fly ash samples have a small particle size (< 200 µm). The fly ash has two general components: (i) an extremely fine-grained polycrystalline (<1 µm) platy material and (ii) globules of aluminosilicate glass (<100 µm). The platy

materials contained many volatile elements Cl, K, Zn, Na, SO 3 and Pb. The aggregate coated the spherical aluminosilicate glass particles. (iii) The bottom ash contained two major groups; refractory waste and melt products according to the macroscopic characterisation. The refractory waste products

194 CHAPTER 6

consisted of fragments of glass, metal, minerals and lithic fragments that survived during combustion. The melt products included glasses (slags) and crystalline phases. (iv) The XRD Rietveld analysis confirmed that quartz, corundum, melilite and spinel group minerals as well as a large amount of glassy materials (>30 wt.%) are the most abundant phases in the bottom ash. Likewise, the fly ash samples were dominantly amorphous (> 40 wt.%), except for sample FA-A4, in addition to the relatively high contents of sulfate and chloride minerals. Small amounts of spinel and melilite minerals were also found in the fly ash materials. (v) The certain polluting elements (Zn and Pb) were more concentrated in the crystalline phase than amorphous phase according to quantitative XRD in comparison with XRF. For the fresh FA-A1, more than 50 % of Zn was present as

K2ZnCl 4 and ZnS, while over 80 % of Pb was found to be associated with minium

(Pb 3O4), massicotite (PbO) and caracolite [Na 3Pb 2(SO 4)3Cl]. Importantly, gordaite

[NaZn 4(SO 4)(OH) 6Cl(H 2O) 6] was formed in the water-treated FA-A1. In XRD analysis of fly ash samples from incinerator A, mineral compositions of the heavy

metals Zn, Pb and Cu were detected such as hydrozincite [Zn 5(OH) 6(CO 3)2],

contunnite (PbCl 2), cerussite (PbCO 3) and digenite (CuS). However, they were not found in the fly ash from incinerator B, because of having lowered these metal bulk concentrations. Further As, Cd, Cr-, and Ni-bearing phases could not be determined by XRD, presumably because of the low contents of these pollutants in the fly ash.

Microanalysis of bottom ash and fly ash The outcomes of the EPMA analyses of morphological, chemical, and mineralogical phase compositions for the individual particles of the selected bottom ash and fly ash were as follows: (i) The bottom ash particles have Al-, Fe-, or Si-grains, while some of these grains chemically correspond to melilite, plagioclase, quartz and spinel, which have been also identified by XRD. The presence of Si-rich aluminosilicate glass and ettringite (mineralogical alteration product) was confirmed in the bottom ash particles.

195 CHAPTER 6

(ii) EPMA analyses indicated that the fly ash particles contained significant chemical and mineralogical variations. A large number of mineral phases (crystalline and amorphous phases) including metal alloys could be identified within the fly ash particles. Quartz, spinel, sulfates, chlorides and garnet group minerals were identified. The glassy phases were Ca-rich aluminosilicate compositions. Specifically, a high Fe-content was observed in some spherical glass particles.

(iii) Based on the very limited number of fly ash particles examined, Ca and SO 4 might be concentrated in the crystals attached to the surface of glass spheres and not within the glass itself. For all studied particles, a distinct preferential association of heavy metals ( e.g., Zn, Pb and Cd) with crystalline phases was not obvious. However, in very few cases, mineral phases enriched with Pb (PbO

and Pb 3O4) were observed. Fe-rich particles were frequently enriched in Cr and Ni.

Mineralogical reactions during aging The results of the XRD study highlighted the mineralogical alteration of the bottom and fly ashes as a consequence of the aging process. Mineralogical analyses were conducted on the bottom and fly ashes measured by XRD over time. The main findings were: (i) Bottom ash All bottom ash samples displayed the high chemical reactivity at the ambient conditions leading to the formation of new minerals after the quenching process of the hot bottom ash. The alteration products of the bottom ash during or shortly after quenching were anhydrite, calcite, ettringite, portlandite, and iron oxides. Ettringite was further altered by the aging process according to the dissolution and precipitation processes in part, to produce gypsum, while the remaining part reacted with chloride to form hydrocalumite.

196 CHAPTER 6

(ii) Fly ash The new minerals formed in the fly ash during aging included the hydrous sulfates syngenite and gordaite (the latter containing the zinc of the decomposed

K2ZnCl4) and hydrocalumite, a hydrous chloridic carbonate with calcium and aluminum. Gypsum and bassanite were also found in some of the aged fly ashes. The newly formed minerals are characteristic phases for growth at ambient conditions from the unstable high-temperature mineral assemblages formed during the incineration process in the combustion chamber.

Water-extraction experiments with fly ash The benefits of using water-extraction processes for recovery of heavy metals and recycling of fly ash as a resource were demonstrated. The simple water-washing and Soxhlet-water extraction processes were employed to determine the key variables controlling soluble chlorides, sulfates and heavy metals from the fly ash. Some of the significant findings of the experiments were: (i) The water-washing process at L/S equal to 10 removed more than 95 % of the Cl, and more than 50 % of the compounds containing elements Na and K from the fly ash examined. These results agreed very well with the previous studies on washing treatment of fly ash (Derie, 1996; Nzihou and Sharrock, 2002).

Correspondingly, the soluble salts such as NaCl, KCl and K 2ZnCl 4 were easily removed from the fly ash. However, the washing process produced heavy metal enrichment in the remainder of the fly ash, as a result of the high release of the water-soluble compounds and the low release of the heavy metals during washing. (ii) The washing process resulted in the formation of hydrate phases such as syngenite, gypsum, ettringite and gordaite. The present findings showed that the fly ash exhibited cementitious properties in the presence of water and hydration reactivity to produce ettringite. A rapid formation of hydrozincite in a fly ash sample was also observed. (i) Likewise the Soxhlet-water extraction provided an alternative method in removing significant amounts of alkali chlorides and sulfates from fly ash. Under special conditions, chlorides could be extracted from fly ash by water to

197 CHAPTER 6

produce aqueous solutions, which could be evaporated to yield practically pure

salts. Ca-and SO 4-bearing phases were partly extracted simultaneously leading

to the precipitation of gypsum and bassanite. The latter precipitated from CaSO 4 solution at 100 OC. Ettringite and monosulfate occurred as secondary precipitates as Al was also released from the fly ash during the Soxhlet extraction procedure. (ii) The Soxhlet-water extraction of fly ash yielded a material with higher contents of heavy metals such as Pb, Cu, and Cr than the original one. This is due to the effect of dissolution of the water-soluble phases during the extraction process. However, the partial extraction of Zn, Pb and Cd could be observed.

6.2 Pozzolanic and Hydrothermal Activations of the Raw and Washed Fly Ashes The fly ash/water interaction by imparting pozzolanic reaction and hydrothermal activation (alkali activation) for the five batches of German MSWI fly ash has been examined, under the controlled laboratory conditions. Based on the experiments of pozzolanic solidification and hydrothermal treatments conducted on the raw and washed fly ash samples, the following conclusions could be drawn.

Pozzolanic solidification The following conclusions were made from the evaluations of the pozzolanic behaviour of the raw and washed fly ash samples in contact with saturated Ca(OH)2 solutions: (i) The raw and washed fly ashes exhibited a cementitious property, in which ettringite and CSH were neo-formed phases giving a good degree of stabilisation of the solidified ash. (ii) The XRD Rietveld analysis indicated that the raw fly ash with a remarkable content of soluble chlorides and other compounds reacted with the saturated

Ca(OH) 2 solution caused the significant formation of gypsum and syngenite, but only small amounts of ettringite, hydrocalumite and CSH phase (modelled by tobermorite-14Å) were formed. Specifically, the high content of amorphous

phase in the raw fly ash could be considered an important source of Ca and SO 4 interacting with the aluminates to form ettringite, whereas the formation of

198 CHAPTER 6

hydrocalumite appeared to be due to the interaction between chlorides and aluminates. (iii) The XRD Rietveld analysis confirmed that small amounts of ettringite and CSH phase were formed in the washed fly ash after the pozzolanic solidification, presumably because of the insufficient content of the aluminate phases in this material. In addition, gypsum was the dominant hydrated phase formed in the solidified washed fly ash.

Hydrothermal treatment A novel synthetic hydrothermal method for the conversion of the raw and washed fly ashes into a mixed product containing Al-substituted tobermorite-11Å, katoite and zeolite-like materials has been uniquely demonstrated. From the experimental results discussed, the following conclusions can be drawn: (i) The treatment of the raw fly ash in the alkali solutions NaOH and KOH at various temperatures and times produced a variety of mineral products. Hydrocalumite was firstly formed during the hydrothermal treatment at 90 OC. The hydrothermal reaction occurred more rapidly at elevated temperatures with the significant formation of tobermorite-11Å and katoite. The Al-substituted tobermorite-11Å became the major neo-formed phase under the hydrothermal conditions of 180 OC, for 48 h with 0.5 M NaOH. (ii) The XRD Rietveld analysis was employed to quantify changes in the compositions of quartz, amorphous reactants and the hydrothermal products as a function of temperature and fixed time of 48 h. Apparently, the amorphous alumina-silica gel precursor, which is expected to form in hydrothermal products, did not completely crystallise to form tobermorite-11Å or zeolitic materials, and therefore it was still present in the hydrothermal products. In addition, results on the synthesis of tobermorite-11Å by direct conversion demonstrated that the most important criteria for the selection of the powder

precursor are the ratios of CaO/[SiO 2+Al 2O3] and Al/[Al+Si], which should be adjusted in the compositions to within the optimal ranges as suggested by Coleman and Brassington (2003).

199 CHAPTER 6

(iii) The water-washing treatment of fly ash was also essential to successfully prepare zeolitic compounds such as hydroxylcancrinite and analcime, but no significant difference in yielding tobermorite-11Å and katoite from either using the raw fly ash or the washed fly ash could be gained. (iv) After the hydrothermal treatment, the heavy metals were either absorbed or physically encapsulated. Therefore, the hydrothermal treatment might have reduced the deleterious effect on (trace) element mobility in a zeolite-based system. (v) The transformation of fly ash into tobermorite-11Å and zeolitic compounds could be considered of great significance, opening new opportunities for this type of waste, which can be used among other for immobilising other toxic and radioactive wastes in cement, asphalts etc.

6.3 Leaching Resistance of the Raw and Washed Fly Ashes and the Stabilised Materials The main conclusions from the leaching resistance of the raw and washed fly ashes including the stabilised materials in the acidic environment are summarised. Concluding remarks from the TCLP tests and evaluations of the toxic potential of those materials are also presented.

Leaching resistance The following conclusions were made from the leaching test of the materials examined in the acidic solution. The XRD observations were performed for the samples measured before and after leaching. The main findings were: (i) The leaching of the raw fly ash appeared to be governed by the dissolution of

certain soluble minerals ( i.e., NaCL, KCL and K 2ZnCl 4) in the contact solution, while the leaching behaviour of the washed fly ash was controlled by the diluting effect of some phases such as ettringite, hydrocalumite and gordaite which were unstable in the acidic environment. (ii) The hydrate phases ( e.g ., syngenite, etrringite and hydrocalumite) formed in the raw fly ash did not resist to the acid and had disappeared after leaching. Likewise ettringite and CSH phase developed in the pozzolanic solidification of the washed fly ash were dissolved in the leaching solution. No other positive

200 CHAPTER 6

evidence could be established in contrast to the reported benefits of using the washed fly ash for the development of chemically stabilised products, which are resistant to the acidic environment. (iii) The main hydrothermal products ( e.g ., tobermorite-11Å, katoite, analcime and hydroxylcancrinite), obtained from either the raw fly ash or the washed fly ash remained stable under acidic conditions. Importantly, the findings showed that hydrothermal treatment produced stabilised products, which could be proven to be non-hazardous. This could be explained by the association of heavy metals in the more stable minerals such as tobermorite-11Å, katoite, analcime and hydroxylcancrinite.

Leachability of heavy metals The leaching experiments yielded information on the mobility of the heavy metals in the raw and washed fly ashes as well as in the stabilised products. The main concluding remarks were: (i) The leaching test results showed that the raw and washed fly ashes had a substantial high leachability of certain heavy metals. The TCLP leaching for Zn, Pb and Cd was beyond the respective regulatory limits. (ii) A dramatic reduction in the heavy metal leaching could be achieved by pozzolanic solidification and hydrothermal treatments of the materials. Results of the TCLP tests verified that leaching of heavy metals such as Zn, Pb and Cd stabilised in the hydrothermal products of both the raw and washed fly ashes did not exceed the regulatory limits, and are in good agreement with the hypothesis of fixation of some heavy metals in the hydrothermal product. Other metals (As Cr, Cu, and Ni) may be fixed in alloys and amorphous phases. (iii) The hydrothermal treatments could represent a promising solution to the reduction of environmental impacts for the fly ash as deposits and open opportunities for giving value to the fly ash as resources.

6.4 Suggestions for Further Research The study has shown that there remain unanswered questions concerning the recycling and recovery of heavy metals, pozzolanic solidification and hydrothermal

201 CHAPTER 6

treatments of fly ash, and the leaching resistance of the stabilised materials. Suggested future work is as follows: (i) Development of the improved fly ash processing method based on particle size classification for the efficient removal of heavy metals such zinc, lead and cadmium from the fly ash. (ii) Included here should be an expanded study on the use of chemical extraction device for recovery of some heavy metals by using appropriate chemical extraction agents (acids, bases or complexants). (iii) Further EPMA analysis in interpreting chemistry and mineralogy in terms of a wide range of particle sizes should be conducted to provide detailed information for the heavy metals association with certain fractions of particles. (iv) Systematic study of the pozzolanic solidification of fly ash with aqueous

Ca(OH) 2 solution, which particularly concerns the product stability and the low release of heavy metals in the leaching solutions. Included here should be an expanded study of appropriate additives to enhance the pozzolanic solidification process and to further exploit the pozzolanic property of fly ash. (v) Development of improved hydrothermal methods for optimising the yield of tobermorite-11Å, and analytical methods for assessing the heavy metal sorptive properties of tobermorite-11Å. (vi) Long term leaching experiments on the hydrothermal products over wide pH ranges are required before accepting them as environmentally products for use in waste treatment applications.

202 REFERENCES Akiko, K., Yukio, N. and Teruji, I. (1996). Chemical Speciation and Leaching Properties of Elements in Municipal Incineration Ashes . Waste Management, 16 , 527-536. Alba, N., Gasso, S., Lacorte, T. and Baldasano, J.M. (1997). Characterisation of Municipal Solid Waste Incineration Residues from Facilities with Different Air Pollution Control Systems . Journal of the Air and Waste Management Association, 47 , 1170-1179. Andreola, F., Barbieri, L., Corradi, A., Lancellotti, I. and Manfredini, T. (2001). The Possibility to Recycle Solid Residues of the Municipal Waste Incineration into a Ceramic Tile Body. Journal of Materials Science, 36 , 4869-4873. ASTM, (1992). Standard Test Method for Shake Extraction of Solid Waste with Water . Annual Book of ASTM Standards, 4, D3987-85. Barbieri, L., Corradi, A., Lancellotti, I., Scarinci, G. and Hrenglich, S. (2000). Controlled Crystallization of Glasses coming from Incinerator Solid Residues . Glass Science and Technology, 73 , 383-386. Barrer, R.M. (1982). Hydrothermal Chemistry of Zeolites . Academic Press, London. Baur, I., Ludwig, C. and Johnson, C.H. (2001). The Leaching Behaviour of Cement Stabilized Air Pollution Control Residues: A Comparison of Field and Laboratory Investigations . Environmental Science and Technology, 35 , 2817-2822. Baur, I., Keller, P., Mavrocordatos, D., Wehrli, B. and Johnson, C.A. (2004). Dissolution-Precipitation Behaviour of Ettringite, Monosulfate, and Calcium Silicate Hydrate . Cement and Concrete Research, 34 , 341-348. Belevi, H. Stampfli, D.M. and Baccini, P. (1992). Chemical Behaviour of Municipal Solid Waste Incinerator Bottom Ash in Monofills . Waste Management and Research, 10 , 153-167. Belevi, H. and Moench, H. (2000). Factors Determining the Element Behaviour in Municipal Solid Waste Incinerator . 1. Field Studies . Environmental Science and Technology, 34 , 2501-2506. Blanco, F., Garcia, M.P. and Ayala, J. (2005). Variation in Fly Ash Properties with Milling and Acid Leaching . Fuel, 84 , 89-96.

203 Boccaccini, A., Köpf, M. and Stumpfe, W. (1995). Glass-Ceramics from Filter Dusts from Waste Incinerators . Ceramics International, 21 , 231-235. Boccaccini, A., Petitmermet, M. and Wintermantel, E. (1997). Glass-Ceramics from Municipal Incinerator Fly Ash . American Ceramic Society Bullettin, 51 , 75-78. Breck, D.W. (1984). Zeolite Molecular Sieves: Structure, Chemistry, and Use . John Wiley and Sons Inc., New York. Brunori, C., Balzamo, S. and Morabito, R. (2001). Comparison Between Different Leaching Tests for the Evaluation of Metal Release from Fly Ash. Fresenius J Anal Chem, 371 , 843-848. Caglioti, G., Paoletti, A. and Ricci, F. P. (1958) . Choice of Collimator for a Crystal Spectrometer for Neutron Diffraction. Nuclear Instrumental Methods, 35 , 223- 228. Chang, E.E., Chiang, P.C., Lu, P.H. and Ko, Y.W. (2001). Comparisons of Metal Leachability for Various Wastes by Extraction and Leaching Methods . Chemosphere, 45 , 91-99. Cheng, T.W., Ueng, T.H., Chen, Y.S. and Chiu, J.P. (2002). Production of Glass- Ceramic from Incinerator Fly Ash . Ceramics International, 28 , 779-783. Chimenos, J.M., Fernandez, A.I., Nadal, R. and Espiell, F. (2000). Short-Term Natural Weathering of MSWI Bottom Ash . Journal of Hazardous Material, B79 , 287-299. Coleman, N.J. and Brassington, D.S. (2003). Synthesis of Al-Substituted 11 Å Tobermorite from Newsprint Recycling Residue: a Feasibility Study . Materials Research Bulletin, 38 , 485-497. Comans, R.N.J. and Meima, J.A. (1994). Modelling Ca-Solubility in MSWI Bottom Ash Leachates . Edited by Goumans, J.J.M., van der Sloot, H.A. and Aalbers, Th.G. Environmental Aspects of Construction with Waste Materials, Studies in Environmental Sciences, Elsevier Science Publisher, Amsterdam, 60 , 103-110. Conner, J.R. (1990). Chemical Fixation and Solidification of Hazardous Waste . Van Nostrand Reinhold, New York. Deer, W.A., Howie, R.A. and Zussman, J. (1992). An Introduction to the Rock-Forming Minerals . 2 nd Edition. Pearson Education Limited, Essex/England. Derie, R. (1996). A New Way to Stabilise Fly Ash from Municipal Incinerators . Waste Management, 16 , 711-716.

204 Dyer, T.D. and Dhir, R.K. (2004). Hydration Reactions of Cement Combinations Containing Vitrified Incinerator Fly Ash . Cement and Concrete Research, 34 , 849-856. Eighmy, T.T., Eusden, J.D., Krzanowski, J.E., Dominggo, D.S., Stampfli, D., Martin, J.R. and Erickson, P.M. (1995). Comprehensive Approach Toward Understanding Element Speciation and Leaching Behaviour in Municipal Solid Waste Incineration Electrostatic Precipitator Ash . Environmental Science and Technology, 29 , 629-646. Erol, M., Kücükbayrak, S., Ersoy-Mericboyu, A. and Övecoglu, M.L. (2001). Crystallization Behaviour of Glasses Produced from Fly Ash . Journal of the European Ceramic Society, 21 , 2835-2841. E.U .(2000). Directive 2000/76/EC of the European Parliament and of the Council of 4 December 2000 on the incineration of waste . The Council of European Union. Eusden, J.D., Eighmy, T.T., Hockert, K., Holland, E. and Marsella, K. (1999). Petrogenesis of Municipal Solid Waste Combustion Bottom Ash . Applied Geochemistry, 14 , 1073-1091. Feldman, R.F. and Sereda, P.J. (1970). A New Model of Hydrated Cement and Its Practical Implications . Engineering Journal of Canada, 53-59. Fermo, P., Cariati, F., Pozzi, A., Demartin, F., Tettamanti, M., Collina, E., Lasagni, M., Pitea, D., Puglisi, O. and Russo, U. (1999). The Analytical Characterization of Municipal Solid Waste Incinerator Fly Ash: Methods and Preliminary Results . Fresenius J Anal Chem, 365 , 666-673. Fermo, P., Cariati, F., Pozzi, A., Tettamanti, M., Collina, E. and Pitea, D. (2000). Analytical Characterization of Municipal Solid Waste Incinerator Fly Ash: Part II . Fresenius J Anal Chem, 366 , 267-272. Ferreira, C., Ribeiro, A. and Ottosen, L. (2003). Possible Applications for Municipal Solid Waste Fly Ash . Journal of Hazardous Materials, B96 , 201-216. Forestier, L.L. and Libourel, G. (1998). Characterization of Flue Gas Residues from Municipal Solid Waste Combustors . Environmental Science and Technology, 32 , 2250-2256. Geelhoed J.S., Meeussen, J.C.L., Hillier, S., Lumsdon, D.G., Thomas, R.P., Farmer, J.G. and Peterson, E. (2002). Identification and Geochemical Modelling of

205 Processes Controlling Leaching of Cr(VI) and Other Major Elements from Chromite Ore Processing Residue . Geochimica et Cosmochima Acta, 66 , 3927- 3942. Glasser, F.P. (1993). Chemistry of Cement-Solidified Waste Forms In: Chemistry and Microstructure of Solidified Waste Forms . Edited by Spence, R.D. Lewis Publishing. Boca Raton, FL, pages 1- 40. Glasser, F.P. (1997). Fundamental Aspects of Cement Solidification and Stabilization . Journal of Hazardous Materials, 52 , 151-170. Grutzeck, M.W., Kwan, S., Thompson, J.L. and Benesi, A. (1999). A Sorosilicate Model for Calcium Silicate Hydrate (C-S-H) . Journal of Material Science Letters, 18 , 217-220. Gilliam, T.M. and Wiles, C.C. (1996). Stabilization and Solidation of Hazardous, Radioactive and Mixed Wastes , Vol.III, ASTM STP 1240, ASTM, Philadelphia, PA. Hamernik, J.D. and Frantz, G.C. (1991). Physical and Chemical Properties of Municipal Solid Waste Fly Ash . ACI Materials Journal, 88 , 294-301. Hill, R.J. and Howard, C.J. (1987). Quantitative Phase Analysis from Neutron Powder Diffraction Data using the Rietveld Method . Journal of Applied Crystallography, 20 , 467-474. Hong, K.J., Tokunaga, S., Ishigami, Y. and Kajiuchi, T. (2000a). Extraction of Heavy Metals from MSW Incinerator Fly Ash using Saponins . Chemosphere, 41 , 345- 352. Hong, K.J., Tokunaga, S. and Kajiuchi, T. (2000b). Extraction of Heavy Metals from MSW Incinerator Fly ashes by Chelating Agents . Journal of Hazardous Materials, B75 , 57-73. Hjelmar, O. (1996). Disposal Strategies for Municipal Solid Waste Incineration Residues . Journal of Hazardous Materials, 47 , 345-368. Hjelmar, O., Birch, H. and Hansen, J.B. (1999). Development of a Process for Treatment of APC Residues from MSW Incinerator prior to Land Filling . Proceedings of Sardinia 97, Seventh International Landfill Symposium, Volume I, 543-548.

206 Inorganic Crystal Structure Database (ICSD). (1999). FIZ Karlsruhe and Gmelin- Institute, Germany. International Centre for Diffraction Data (ICDD). (2002). The Powder Diffraction File (PDF). International Soil Reference and Information Centre (ISRIC) (1995). Procedures for Soil Analysis . Technical Paper 9, Food and Agricultural Organisation of the United Nations, Rome, 9.1-9.13. Jennings, H.M. (2000). A Model for the Microstructure of Calcium Silicate Hydrate in Cement Paste . Cement and Concrete Research, 30 , 101-116. Jiang, W. and Roy, D.M. (1992). Hydrothermal Processing of New Fly Ash Cement . American Ceramic Society Bullettin, 71 , 642-647. Johnson, C.A., Kersten, M., Ziegler, F. and Moor, H.C. (1996). Leaching Behaviour and Solubility-Controlling Solid Phases of Heavy Metals in Municipal Solid Waste Incinerator Ash . Waste Management, 16 , 129-134. Katsuura, H., Inoue, T., Hiraoka, M. and Sakai, S. (1996). Full-Scale Plant Study on Fly Ash Treatment by the Acid Extraction Process . Waste Management, 16 , 491-499. Kersch, C., van der Kraan, M., Woerlee, G.F. and Witkamp, G.J. (2002). Municipal Waste Incinerator Fly Ash: Supercritical Fluid Extraction of Metals. Journal of Chemical Technology and Biotechnology, 77 , 256-259. Kirby, C.S. and Rimstidt, J.D. (1993 ). Mineralogy and Surface Properties of Municipal Solid Waste Ash . Environmental Science and Technology, 27 , 652-660. Kirkpatrick, R.J., Yarger, J.L., McMillan, P.F., Ping, Y. and Cong, X. (1997). Raman Spectroscopy of C-S-H, Tobermorite and Jennite . Advanced Cement Based Materials, 5, 93-99. Kosson, D.S., van der Sloot, H.A. and Eighmy, T.T. (1996). An Approach for Estimation of Contaminant Release during Utilization and Disposal of Municipal Waste Combustion Residues . Journal of Hazardous Materials, 47 , 43-75 LAGA (der Länderarbeitsgemeinschaft Abfall) (1994). Merkblatt über die Entsorgung von Abfällen aus Verbrennungsanlagen für Siedlungsabfälle . Erich Schmidt Verlag, Nummer 19. Lamers, F.J.M. and Born, J.G.P. (1994). Upgrading Techniques for the Quality Improvement of Municipal Waste Incineration Residues . Edited by Goumans,

207 J.J.M., van der Sloot, H.A. and Aalbers, Th.G. Environmental Aspects of Construction with Waste Materials, Studies in Environmental Sciences 60, Elsevier Science Publisher, Amsterdam, pages 645-654. Li, C.T., Huang, Y.J., Huang, K.L. and Lee, W.J. (2003). Characterization of Slags and Ingots from the Vitrification of Municipal Solid Waste Incineration Ashes. Industrial Engineering of Chemistry Resources, 42 , 2306-2313. Li, M., Xiang, J., Hu, S., Sun, L.S., Su, S., Li, P.S. and Sun, X.X. (2004). Characterization of Solid Residues from Municipal Solid Waste Incinerator . Fuel, 83 , 1397-1405. Lide, D.R. (1997). CRC Handbook of Chemistry and Physics: A Ready-Reference Book of Chemical and Physical Data . 78 th Edition. CRC Press, Boca Raton, Florida.

Ma, W. and Brown, P.W. (1997). Hydrothermal Reactions of Fly Ash with Ca(OH) 2 and

CaSO 4·2H 2O. Cement and Concrete Research, 27 , 1237-1248. Macphee, D.E. and Glasser, F.P. (1993). Immobilization Science of Cement Systems . Material Research Bulletin, 18 , 66-71. Maenami, H., Shin, H., Ishida, H. and Mitsuda, T. (2000). Hydrothermal Solidification of Wastes with Formation of Zeolites . Journal of Materials in Civil Engineering, 12 , 302-306. Mangialardi, T., Paolini, A.E., Polettini, A. and Sirini, P. (1999). Optimization of the Solidification/Stabilization Process of MSW Fly Ash in Cementitious Matrices. Journal of Hazardous Materials, B70 , 53-70. Mangialardi, T. (2003). Disposal of MSWI Fly Ash through a Combined Washing- Immobilisation Process. Journal of Hazardous Materials, B98 , 225-240. Meima, J.A. and Comans, R.N.J. (1997). Geochemical Modeling of Weathering Reactions in Municipal Solid Waste Incinerator Bottom Ash . Environmental Science and Technology, 31 , 1269-1276 . Meima, J.A. and Comans, R.N.J. (1999). The Leaching of Trace Elements from Municipal Solid Waste Incinerator Bottom Ash at Different Stages of Weathering . Applied Geochemistry, 14 , 159-171. Miguel, A.F., Lluis, M., Merce, S., Jose, C.G. and Ferran, E. (1992). Behavior of Heavy Metals in the Combustion Gases of Urban Waste Incinerators . Environmental Science and Technology, 26 , 1040-1047.

208 Miyake, M., Niiya, S. and Matsuda, M. (2000). Microwave-assisted Al-substituted Tobermorite Synthesis . Journal of Materials Research, 15 , 850-853. Miyake, M., Tamura, C. and Matsuda, M. (2002). Resource Recovery of Waste Incinerator Fly Ash: Synthesis of Zeolites A and P . Journal of the American Ceramics Society, 85 , 1873-1875. Moreno, N., Querol, X., Ayora, C., Alastuey, A., Pereira, C.F. and Jurkovicová, M.J. (2001). Potential Environmental Applications of Pure Zeolitic Material Synthesized from Fly Ash . Journal of Environmental Engineering, 127 , 994-1002. Murayama, N., Yamamoto, H. and Shibata, J. (2002a). Zeolite Synthesis from Coal Fly Ash by Hydrothermal Reaction using Various Alkali Sources . Journal of Chemical Technology and Biotechnology, 77 , 280-286. Murayama, N., Yamamoto, H. and Shibata, J. (2002b). Mechanism of Zeolite Synthesis from Coal Fly Ash by Alkali Hydrothermal Reaction . International Journal of Mineral Processing, 64 , 1-17. Neuhoff, P.S., Hovis, G.L., Balassone, G. and Stebbins, J.F. (2004). Thermodynamic Properties of Analcime Solid Solutions . American Journal of Science, 304 , 21-66. Nzihou, A. and Sharrock, P. (2002). Calcium Phosphate Stabilization of Fly ash with Chloride Extraction . Waste Management, 22 , 235-239. Park, Y.J. and Heo, J. (2002a). Nucleation and Crystallization Kinetics of Glass derived from Incinerator Fly Ash Waste . Ceramics International, 28 , 669-673. Park, Y.J. and Heo, J. (2002b). Vitrification of Fly Ash from Municipal Solid Waste Incinerator . Journal of Hazardous Materials, B91 , 83-93. Pfrang-Stozt, G. and Schneider, J. (1995). Comparative Studies of Waste Incineration Bottom Ashes from Various Grate and Firing Systems, Conducted with Respect to Mineralogical and Geochemical Methods of Examination . Waste Management and Research, 13 , 273-292. Piantone, P., Bodénan, F., Derie., R. and Depelsenaire, G. (2003). Monitoring the Stabilization of Municipal Solid Waste Incineration Fly Ash by Phosphation: Mineralogical and Balance Approach. Waste Management, 23 , 225-243. Polettini, A. and Pomi, R. (2004). The Leaching Behaviour of Incinerator Bottom Ash as Affected by Accelerated Ageing . Journal of Hazardous Materials, B113 , 209- 215.

209 Pomiès, M.P., Lequeux, N. and Boch, P. (2001a). Speciation of Cadmium in Cement . Part I. Cd 2+ uptake C-S-H. Cement and Concrete Research, 31 , 563-569. Pomiès, M.P., Lequeux, N. and Boch, P. (2001b). Speciation of Cadmium in Cement . 2+ Part II. C 3S Hydration with Cd Solution. Cement and Concrete Research, 31 , 571-576. Powers, T.C. (1960). Physical Properties of Cement Paste . In Fourth International Symposium on the Chemistry of Cement, Washington, D.C., National Bureau of Standards, 577-609. Prince, E. (1993). Mathematical Aspects of Rietveld Refinement . The Rietveld Method. Edited by Young, R.A. International Union of Crystallography, Oxford, New York, 43-54. Querol, X., Alastuey, A., Lopezsoler, A., Plana, F., Andres, J.M., Juan, R., Ferrer, P. and Ruiz, C.R. (1997). A Fast Method for Recycling Fly Ash: Microwave-Assisted Zeolite Synthesis . Environmental Science and Technology, 31 , 2527-2533. Querol, X., Moreno, N., Uma ňa, J.C., Alastuey, A., Hernández, E., López-Soler, A. and Plana, F. (2002). Synthesis of Zeolites from Coal Fly Ash: an Overview . International Journal of Coal Geology, 50 , 413-423. Reimann, D.O. (1994). Menge, Beschaffenheit und Verwertungsmöglichkeiten von MV- Schlacke, Gesamtübersicht , Beihefte Müll und Abfall, 31 , Entsorgung von Schlacken und sonstigen Reststoffen. Reed, B.E. and Nonavinakere, S.K. (1992). Metal Adsorption by Activated Carbon: Effect of Complexing Ligands, Competing Adsorbates, Ionic Strength and Background Electrolyte . Separation Science and Technology, 27 , 1985-2000. Rémond, S., Bentz, D.P. and Pimienta, P. (2002). Effects of the Incorporation of Municipal Solid Waste Incineration Fly Ash in Cement Pastes and Mortars II: Modeling. Cement and Concrete Research, 32 , 565-576. Ricou, P., Lecuyer, I. and Le Cloirec, P. (1999). Removal of Cu +2 , Zn +2 , and Pb +2 by adsorption onto Fly Ash and Fly Ash/Lime Mixing . Water Science and Technology, 39 , 239-247. Rietveld, H.M. (1969). A Profile Refinement Method for Nuclear and Magnetic Structures . Journal of Applied Crystallography, 2, 65-71.

210 Rimstidt, J.D. and Barnes, H.L. (1980). The Kinetics of Silica-Water Reactions . Geochimica et Cosmochima Acta, 44 , 1683-1699. Rincon, J.M.A., Romero, M. and Boccaccini, A.R. (1999). Microstructural Characterisation of a Glass and a Glass-Ceramic obtained from Municipal Incinerator Fly Ash . Journal of Materials Science, 34 , 4413-4423. Rudnick, R.L. and Fountain, D.M. (1995 ). Nature and Composition of the Continental- Crust-A Lower Crustal Perspective . Reviews of Geophysics, 33 , 267-309. Sabbas,T., Polettini, A., Pomi, R., Astrup, T., Hjelmar, O., Mostbauer, P., Cappai, G., Magel., G., Salhofer, S., Speiser, C., Heuss-Assbicher, S., Klein, R. and Lehner, P. (2003). Management of Municipal Solid Waste Incineration Residues . Waste Management, 23 , 61-88. Sauve, S., Hendershot, W. and Allen, H. (2000). Solid-Solution Partitioning of Metals in Contaminated Soils: Dependence on pH, Total Metal Burden and Organic Matter . Environmental Science and Technology, 34 , 1125-1131. Scarinci, G., Brusatin, G., Barbieri, L., Corradi, A., Lancellotti, I., Colombo, P., Hreglich, S. and Dall’Igna, R. (2000). Vitrification of Industrial and Natural Wastes with Production of Glass Fibers . Journal of the European Ceramic Society, 20 , 2485-2490. Schneider, J., Vehlow, J. and Vogg, H. (1994). Improving the MSWI Bottom Ash Quality by Simple in-Plant Measures . Edited by Goumans, J.J.M., van der Sloot, H.A. and Aalbers, Th.G. Environmental Aspects of Construction with Waste Materials, Studies in Environmental Sciences 60. Elsevier Science Publisher, Amsterdam, pages 605-620. Siemens. (1986). X-ray Spectrometers and Diffractometers . Catalogue MP 42.1986, Federal Republic of Germany, order number E86060-K6042-A101-A2-7600- P2/2.

SØrensen, M.A., Koch, B.C., Stackpoole, M.M., Bordia, R.K., Benjamin, M.M. and Christensen, T.H. (2000). Effects of Thermal Treatment on Mineralogy and Heavy Metal Behaviour in Iron Oxide Stabilized Air Pollution Control Residues . Environmental Science and Technology, 34 , 4620-4627.

211 Song, G.J., Kim, K.H., Seo, Y.C. and Kim, S.C. (2004). Characteristics of Ashes from Different Locations at the MSW Incinerator Equipped with Various Air Pollution Control Devices . Waste Management, 24 , 99-106. Sparks, D.L. (1995). Environmental Soil Chemistry . Academic Press Cop, San Diego, CA. Speiser, C., Baumann, T. and Niessner, R. (2000). Morphological and Chemical Characterisation of Calcium-Hydrate Phases Formed in Alteration Processes of Deposited Municipal Solid Waste Incinerator Bottom Ash . Environmental Science and Technology, 34 , 5030-5037. Speiser, C., Baumann, T. and Niessner, R. (2001). Characterisation of Municipal Solid Waste Incineration (MSWI) Bottom Ash by Scanning Electron Microscopy and Quantitative Energy Dispersive X-ray Microanalysis (SEM/EDX) . Fresenius J Anal Chem, 370 , 752-759. Sposito, G. (1986). Distinguishing Adsorption from Surface Precipitation . In: Geochemical Processes at Mineral Surfaces. Edited by Davis, J.A. and Hayes, K.F. American Chemical Society, Washington, pages 217-228. Stuetzle, R., Hagenmaier, H., Hasenkopf, O. and Schettle, G. (1991). Erste Erfahrungen mit der Demonstrationsanlage in der MVA Stuttgart-Muenster zum Abbau von PCDD und PCDF in Flugaschen . VGB Kraftwerkstechnik, 71 , 1038-1040. Steenbruggen, G. and Hollman, G.G. (1998). The Synthesis of Zeolites from Fly Ash and the Properties of the Zeolite Products . Journal of Geochemical Exploration, 62 , 305-309. Stumm, W. (1992). Chemistry of the Solid-Water Interface . John Wiley and Sons, Inc. New York. Stumm, W. and Morgan, J.J. (1996). Aquatic Chemistry . John Wiley and Sons, Inc. New York. TA-Siedlungsabfall. (1993). Dritte Allgemeine Verwaltungsvorschrift zum Abfallgesetz. Technische Anleitung zur Verwertung, Behandlung und sonstigen Entsorgung von Siedlungsabfällen . Bundesanzeiger Verlag. Nummer, 99a, Köln. Tan, L.C., Choa, V. and Tay, J.H. (1997). The Influence of pH on Mobility of Heavy Metals from Municipal Solid Waste Incineration Fly Ash. Environmental Monitoring and Assessment, 44 , 275-284.

212 Targan, S., Olgun, A., Erdogan, Y. and Sevinc, V. (2003). Influence of Natural Pozzolan, Colemanite Ore Waste, Bottom Ash, and Fly Ash on the Properties of Portland Cement . Cement and Concrete Research, 33 , 1175-1182. Taylor, H.F.W. (1950). Hydrated Calcium Silicates. Part I. Compound Formation at Ordinary Temperatures. Journal of the Chemical Society, 3682-3690. Taylor, H.F.W. (1990). Cement Chemistry . Academic Press, London and New York. Taylor, J.C. and Clapp, R.A. (1992). New Features and Advances Applications of SIROQUANT TM : A Personal Computer XRD Full Profile Quantitative Analysis Software Package . Advances in X-Ray Analysis, 35 , 48-55. Tennis, P.D. and Jennings, H.M. (2000). A Model for Two Types of Calcium Silicate Hydrate in the Microstructure of Portland Cements Pastes . Cement and Concrete Research, 30 , 855-863. Theis, T.M. and Wirth, J.L. (1977). Sorptive Behaviour of Trace Metals on Fly Ash in Aqueous Systems . Environmental Science and Technology, 11 , 1096-1100. Thipse, S.S., Schoenitz, M. and Dreizin, E.L. (2002). Morphology and Composition of the Fly Ash Particles Produced in Incineration of Municipal Solid Waste . Fuel Processing Technology, 75 , 173-184. U.S. EPA. (1990). Part 261, Appendix II-Method 1311 Toxicity Characteristic Leaching Procedure (TCLP), Federal Register , Vol. 55, No. 61, March 29, 1990, Rules and Regulations, pages 11863-11877. Ubbriaco, P. and Calabrese, D. (2000). Hydration Behaviour of Mixtures of Cement and Fly Ash with High Sulphate and Chloride Content . Journal of Thermal Analysis and Calorimetry, 61 , 615-623. Ubbriaco, P., Bruno, P., Traini, A. and Calabrese, D. (2001). Fly Ash Reactivity: Formation of Hydrate Phases. Journal of Thermal Analysis and Calorimetry, 66 , 293-305. van der Sloot, H.A., Heasman, L. and Quevauviller, Ph. (1997). Harmonization of Leaching/extraction Tests . Elservier Science B.V. Amsterdam. Vaughan, D.E.W. (1978). Natural Zeolites: Occurrence, Properties and Use . Edited by Sand, L.B. and Mumpton, F.A. London Pergamon.

213 Wang, K.S., Chiang, K.Y., Lin, K.L. and Sun, C.J. (2001). Effects of a Water- Extraction Process on Heavy Metal Behaviour in Municipal Solid Waste Incinerator Fly Ash . Hydrometallurgy, 62 , 73-81. Warren, B.E. (1969). X-ray Diffraction . Addison-Wesley Publishing Company Inc., United States of America. White, A.F. and Brandley, S.L. (1995). Chemical Weathering Rates of Silicate Minerals , Mineralogical Society of America, Chelsea, MI, 31 , 407-461. Winburn, R.S., Grier, D.G., McCarthy, G.J. and Peterson, R.B. (2000). Rietveld Quantitative X-Ray Diffraction Analysis of NIST Fly Ash Standard Reference Materials . Powder Diffraction, 15 , 163-172. Yang, G.C.C. and Yang, T.Y. (1998 ). Synthesis of Zeolites from Municipal Incinerator Fly Ash . Journal of Hazardous Materials, 62 , 75-89. Yao, Z., Tamura, C., Matsuda, M. and Miyake, M. (1999). Resource Recovery of Waste Incineration Fly Ash: Synthesis of Tobermorite as Ion Exchanger . Journal of Materials Research, 14 , 4437-4442. Yoder, H. S. (1979) . Melilite-bearing rocks and related lamprophyres . Edited by Yoder, H. S. The evolution of the igneous rocks; fiftieth anniversary perspectives. Princeton University Press, Princeton, N.J. United States, pages 391-411. Youcai, Z., Song L. and Guojian, L. (2002). Chemical Stabilization of MSW Incinerator Fly Ashes . Journal of Hazardous Materials, B95 , 47-63. Young, R.A. and Wiles, D.B. (1982). Profile Shape Functions in Rietveld Refinements . Journal of Applied Crystallography, 15 , 430-438. Zevenbergen, C. and Comans, R.N.J. (1994). Geochemical Factors Controlling the Mobilization of Major Elements during Weathering of MSWI Ash . Edited by Goumans, J.J.M., van der Sloot, H.A. and Aalbers, Th.G. Environmental Aspects of Construction with Waste Materials, Studies in Environmental Sciences 60. Elsevier Science Publisher, Amsterdam, pages 179-194. Zevenbergen, C., Vander Wood, T., Bradley, J.P., van der Broeck, P.F.C.W., Orbons, A.J. and van Reeuwijk, L.P. (1994). Morphological and Chemical Properties of MSWI Bottom Ash with respect to the Glassy Constituents . Hazardous Waste and Hazardous Materials, 11 , 371-383.

214 Zevenbergen, C., Van Reeuwijk, L.P., Bradley, J.P., Bloemen, P. and Comans, R.N.J. (1996). Mechanism and Conditions of Clay Formation during Natural Weathering of MSWI Bottom Ash . Clays and Clay Minerals, 44 , 546-552.

215 APPENDICES

I. Lebenslauf

Name Athanasius Priharyoto Bayuseno

Geburtsdatum 20. Mai 1962

Geburtsort Ponorogo, Indonesien

Staatsangehörigkeit Indonesier

Berufliche 1989-2002 Dozent für Werkstoffwissenschaften, Tätigkeiten Institut für Maschinenbau der Diponegoro Universität Indonesien Schulbildung 1977-1981 Fachschule in Indonesien

Studium 1982-1988 Vordiplom (Bachelor Maschinenbau) an der Gadjah Mada Universität, Indonesien 1994-1995 Diplom (Angewandte Physik) an der Technischen Universität Curtin, Perth, Australien. 1995-1997 Master (MSc, Studiengang Physik) an der Technischen Universität Curtin, Perth, Australien. Thema: Thermal Shock Resistance of Spodumene-Alumina Ceramics Promotion Seit Doktorand am Institut für Mineralogie der Feb. 2003 Ruhr-Universität Bochum

Bochum, im Januar 2006

216 II. Publications from Thesis

1. Bayuseno.A.P., Schmahl.W.W., Trombach. U., Reinecke, T. and Müllejans.Th. (2005). Chemical and Mineralogical Characterisation of Municipal Solid Waste Incineration (MSWI) Bottom Ash. Neus Jahrbuch für Mineralogie. 2. Bayuseno.A.P., Schmahl.W.W., Trombach. U. and Müllejans.Th. (2006). Chemical and Phase Composition of Fly Ash from Municipal Solid Waste Incineration (MSWI). Submitted for publication in the Science and the Total Environment. 3. Bayuseno.A.P., Schmahl.W.W. and Müllejans.Th. (2006). Pozzolanic Solidification of MSWI Fly Ash-Mineralogical Phase Composition and Leaching Property Evaluation. In preparation. 4. Bayuseno.A.P., Schmahl.W.W. and Müllejans.Th. (2006).Hydrothermal Conversion of MSWI Fly Ash into Sable Mineral Phases: Characterisation and Leaching Property Evaluation. In preparation.

217