UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

High iron content glasses: an alternative in the use of electric arc furnace dust Ionescu, Denisa V. 1996

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

Item Metadata


831-ubc_1996-0002.pdf [ 8.13MB ]
JSON: 831-1.0078674.json
JSON-LD: 831-1.0078674-ld.json
RDF/XML (Pretty): 831-1.0078674-rdf.xml
RDF/JSON: 831-1.0078674-rdf.json
Turtle: 831-1.0078674-turtle.txt
N-Triples: 831-1.0078674-rdf-ntriples.txt
Original Record: 831-1.0078674-source.json
Full Text

Full Text

H I G H IRON C O N T E N T GLASSES - A N A L T E R N A T I V E IN T H E USE O F E L E C T R I C A R C F U R N A C E DUST by D E N I S A V . I O N E S C U Metallurgical Eng., Polytechnical Institute of Bucharest, 1980 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF A P P L I E D S C I E N C E in T H E F A C U L T Y OF G R A D U A T E S T U D I E S ( Department of Metallurgical Engineering ) We accept this thesis as conforming to the required standard T H E UNIVERSITY O F BRITISH C O L U M B I A October, 1995 © Denisa Ionescu, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date 0% tfow r DE-6 (2/88) 11 A B S T R A C T Electric arc furnace baghouse dust ( E A F ) represents an environmental problem, both in terms o f temporary storage and, ultimately, safe disposal. Vitrification of the dust so that hazardous components (such as zinc, lead, chromium, cobalt, barium, and arsenic) are incorporated into the glassy structure is one option for environmentally benign disposal which is currently being explored commercially in silica-based glasses with levels of iron up to 15 wt%. However, the ability of vitreous materials to resist leaching of hazardous components may be substantially reduced by components which promote crystallinity, in particular Fe, Z n , Cd, and other transition elements. In the present work, the impact of F e 2 0 3 on the vitrification process has been explored experimentally, with the objective to generate glasses o f high iron content. Several compositions were prepared in the system Fe20 3 - Z n O - CaO - M g O - N a 2 0 at 50 wt% S i 0 2 and 1500 °C. Amorphous products were obtained with F e 2 0 3 content varying from 15 to 35 wt%. The influence of Zn , Cd, and Pb in the silicate melt was also studied, with emphasis on the effect of these elements on high iron content glasses. The addition levels were within the range of variation of these elements in E A F dusts. The suitability of high iron vitreous products to retain hazardous metallics is normally tested under standard leaching tests designed by special wastes protocols. Si l ica glasses with zinc and glasses containing both iron and zinc, as well as their crystallized samples, were tested in order to determine both the leaching resistance and the influence of crystallinity on the leaching process. Satisfactory results in the leaching tests show the versatility of silica glass to incorporate hazardous elements such as zinc, render the products non-toxic and hence create an alternative to expensive disposal costs associated with E A F dust. i i i Table of Contents A B S T R A C T i i Table of Contents i i i Lis t of Tables v Lis t o f Figures v i A C K N O W L E D G M E N T S v i i i 1.0 I N T R O D U C T I O N 1 2.0 P R O C E S S I N G OF E L E C T R I C A R C F U R N A C E D U S T S 4 3.0 C L A S S I F I C A T I O N 10 3.1 Sil ica Glasses 13 3.2 Composition Limitations of Iron in Glassmaking 13 3.3 Zinc, Cadmium, Lead and Sodium in Silica Glasses and Glass-ceramics 16 3.4 Capture of Hazardous Elements by Glass Systems 18 3.5 Leaching Behaviour of Silica Glasses 21 4.0 S C O P E A N D O B J E C T I V E S 24 5.0 E X P E R I M E N T A L W O R K 26 6.0 R E S U L T S 30 6.1 The Influence of Iron on Glassforming 30 6.2 The Influence of Zinc on Glassforming 40 6.3 The Combined Influence of Iron and Zinc on Glassforming 45 6.4 The Influence of Cadmium and Lead on Glassforming 47 6.5 Stability of the Glass Structures 50 6.5.1 Devitrification of Glasses with Zinc 50 6.5.2 Devitrification of Glasses with Iron 59 6.5.3 Devitrification of Glasses with Zinc and Iron 65 6.6 Leaching Behaviour of Glasses 69 7.0 C O N C L U S I O N S 74 8.0 R E C O M M E N D E D W O R K 77 R E F E R E N C E S 9IV Table of Contents ( cont. ) A P P E N D I X I 83 A P P E N D I X II 87 List of Tables Table I - Comparative Chemical Composition of E A F Dust 2 Table II - High Temperature Metals Recovery Processes ( H T M R ) 5 Table III - Direct Recycling Processes 6 Table I V - Solidification and Stabilization Processes 7 Table V - Additional E A F Dust Treatment Processes 8 Table V I - T C L P Tests Results 8 Table V I I - Chemical Analysis of the Leachate Solutions 70 Table VIII - Leaching Results for Z n O Glasses Crystallized at 1180 °C, 2hr 72 v i List of Figures Figure 1. X-ray diffraction pattern for sample with 35 % F e 2 0 3 , water quenched 31 Figure 2. X-ray diffraction pattern for sample with 35 % Fe20 3 , air quenched 32 Figure 3. X-ray diffraction pattern for sample containing 40 % Fe20 3 , water quenched 33 Figure 4. Ternary phase diagram for the system S i 0 2 - CaO - Fe20 3 35 Figure 5. Quaternary phase diagram of the system CaO - M g O - F e 2 0 3 - S i 0 2 35 Figure 6. S E M of glass sample with 35 % F e 2 0 3 , fractured unpolished sample, 36 (a) 60um, (b) 10 urn Figure 7. S E M of sample with 40 % F e 2 0 3 37 Figure 8. Sample of glass containing 35 % F e 2 0 3 39 Figure 9. Sample of glass containing 40 % Fe20 3 39 Figure 10. X-ray diffraction pattern of sample with 35 % Z n O 41 Figure 11. S E M of glass sample with 30 % Z n O 42 Figure 12. X-ray diffraction pattern of sample with 30 % Z n O 43 Figure 13. Sample containing 30 % Z n O 44 Figure 14. Sample containing 35 % Z n O 44 Figure 15. Region of glass formation in F e 2 0 3 - Z n O - CaO - N a 2 0 system at 46 5 0 % S i O 2 a n d 1500 °C Figure 16. S E M of glass sample with 30 % F e ^ - , and 5 % Z n O 48 Figure 17. S E M of glass sample containing iron, zinc, lead and cadmium 48 Figure 18. X-ray diffraction pattern of sample with 30 % ZnO, (a) after 1 hr, 52 (b) after 2 hr, (c) after 3 hr at 900 °C Figure 19. X-ray diffraction pattern of a glass with 30 % Z n O after 3 hr at 1000 °C 53 v i i List of Figures (cont.) Figure 20. S E M of 30 % ZnO sample after 2 hr at 1180 °C 54 Figure 21. X-ray diffraction pattern of sample with 30 % Z n O after 2 hr at 1180 °C 55 Figure 22. Ternary diagram of the system CaO - Z n O - S i 0 2 56 Figure 23. Ternary diagram of the system M g O - Z n O - S i 0 2 56 Figure 24. Sample of glass material containing 30 % ZnO(left) and same material 57 crystallized at 1180 °C for 2 hr(right) Figure 25. X-ray diffraction pattern of sample with 35 % F e ^ after 2 hr at 1080 °C 60 Figure 26. Ternary phase diagram of the system CaO - F e 2 0 3 - S i 0 2 61 Figure 27. Ternary phase diagram of the system CaO - M g O - S i 0 2 61 Figure 28. S E M of 35 % F e 2 0 3 sample after 2 hr at 1080 °C 62 Figure 29. S E M of 35 % F e 2 0 3 , crystallized sample 63 Figure 30. Sample of glass material containing 35 % Fe 2 0 3 (r ight) and same material 64 crystallized at 1080 °C for 2 hr(left) Figure 31. Sample containing 30 % F e 2 0 3 and 5 % ZnO, (a) amorphous, 66 (b) crystallized Figure 32. S E M of 30 % F e 2 0 3 and 5 % Z n O sample crystallized at 1080 °C for 2 hr 67 Figure 33. X-ray diffraction pattern of sample containing 30 % F e ^ and 5 % Z n O 68 after 2 hr at 1080 °C Figure 34. Level of Z n in the leachate solution as a function of % Z n O in the glass 71 or glass-ceramic Figure 35. Leaching behaviour of crystallized glass samples containing Z n O 72 Vll l A C K N O W L E D G M E N T S The author would like to express her gratitude and thanks to her supervisors, Dr. T.R. Meadowcroft and Dr. P. V . Barr for their advise, expertise and encouragement during the course of the project. Thanks are also extended to the faculty, staff and fellow graduate students in the Department of Metals and Materials Engineering. Financial assistance from N S E R C is gratefully acknowledged. 1 1.0 INTRODUCTION Electric arc furnace ( E A F ) dust is an inevitable by-product of the steelmaking process since, during the steelmaking process, the temperatures are generally about 1600°C and under these conditions metallics such as Zn , Pb, Cd and some M n are volatilized. A s the metal vapours exit the furnace they condense into physically and chemically complex, microscopic agglomerates which form on condensed nuclei such as fugitive dust. These agglomerates are collected as fine material in the baghouse system. This dust contains not only valuable amounts of iron and zinc, which are potentially recoverable, but also lead, cadmium, hexavalent chromium and sometimes selenium. Unfortunately the water solubility o f many of these elements is substantial and E A F dust is therefore classified as a hazardous waste since the heavy metal components can be leached by ground water [4]. The main hazardous constituents in the dust are hexavalent chromium, lead, zinc and cadmium. Commonly, E A F dust is transported to a hazardous waste site for disposal but there is considerable cost in doing this. Wi th environmental regulations continuously evolving and changing, the regulations for E A F handling are expected to become ever more strict, and there is real potential for a ban on shipping of the dust from the steelmaking sites. Future liabilities and increasing dumping costs represent the driving forces for steelmakers in adopting alternative, environmentally sound and economic disposal methods. In the operation of a typical electrical arc furnace approximately 1 to 2 % o f the charge is converted into dust [4]. In the steelmaking industry dust is generated at rates of 2,000 to 20,000 tonnes of dust/year/plant, with an estimated total of 600,000 tonnes/year in the U S A and 77,000 2 tonnes/year in Canada [1]. A 1993 report on samples from 52 steel companies provides the average chemical analysis for E A F dusts shown in Table I : Table I - Comparative Chemical Composition of E A F Dust |T] Chemical 1982* 1992** Carbon steel dust Carbon steel dust (wt%) (wt%) Fe 35.1 28.5 Z n 15.4 19.0 C d 0.028 <0.01 Pb 1.5 2.1 Cr 0.38 0.39 CaO 4.8 10.7(CaO+MgO) * Lehigh, 1982 ** Arthur D . Little Survey for C M P , 1993 From this tabulation it is evident that the main components of E A F dust are iron and zinc. However the chemical composition, particularly with regards to the minor components, can vary widely due to fluctuations in the composition of the scrap feed. For example, i f scrap contains substantial stainless steel, then the dust can have much higher concentrations of Cr and N i , or i f it contains galvanized steel, then the Z n content can be much greater than indicated above. In typical operations the iron content can vary over a wide range between 10 and 45 %, while zinc can be anywhere between 5 and 40 %. The majority of iron occurs as magnetite (Fe 3 0 4 ) , but some metallic iron and hematite may also be present. Around 80% of the zinc is present as zinc oxide (ZnO), the remainder being mostly zinc ferrite (ZnO F e 2 0 3 ) or spinel. However, small amounts of zinc may also be present as sulfides, silicates or aluminates. The lead is present in the form of lead oxide and occurs as 3 small discrete particles. The distribution of cadmium is not well understood, but it might be expected to behave in a similar manner to zinc [4]. Reflected light microscopy ( R L M ) , scanning microscopy- energy dispersive spectroscopy ( S E M - E D S ) and cathodoluminescence microscopy ( C L M ) have all been employed to examine the mineralogy of E A F dusts. The results indicate that the most predominant mineralogical phases are iron spinels; magnetite (Fe 3 0 4 ) , zinc ferrite and jacobsite ( M n F e 2 0 4 ) , and solid solutions of these three spinels with zincite (ZnO) and fayalite ( F e 2 S i 0 4 ) . In the presence of high zinc concentrations the prevalent phases were found to be zincite and zinc ferrite [5]. The presence of magnetite-franklinke-jacobsite (iron spinels) solid solution in a Ca-Fe-Si glass together with hematite and zincite has also been reported [7]. According to a study done at Lehigh University [2], each particle contains a diversity of compounds. In this light E A F dust can be thought of as agglomerated collections of microfine and chemically complex particles. In terms of physical characteristics, E A F dust particles are approximately spherical, ranging in size from 0.1 to 10 um. The specific area of the dust is estimated at 2.5 - 4.0 m /g and the density is in the range of 1.1 to 2.5 g/cm . The residual moisture can vary from 5 to 50 % and is dependent on the amount of lime in the dust [4]. 4 2.0 PROCESSING O F A R C F U R N A C E DUSTS There are a variety of potential processing options for treating E A F dust including both hydrometallurgical and pyrometallurgical techniques. Hydrometallurgical methods can currently be used for recovery of Zn , Cd and Pb. However, since Z n in the dust is tied to the iron, forming a ferrite or spinel, chemical removal by known reagents is far from complete. Cadmium is more amenable than zinc to removal by almost all the reagents, and Pb has a very narrow spectrum of solvents [6]. From a technological aspect, the recovery of the metals and the recycling of the reagents is currently possible but at the present time metal recovery by hydrometallurgical routes is not financially attractive. The existing pyrometallurgical processes can be grouped into four categories: 1 - High Temperature Metals Recovery Processes (HTMR) - Several high temperature processes have been developed wherein the E A F dust is processed in a furnace or reactor with a reducing agent (coke or coal) in order to recover the zinc, lead and cadmium as metals or oxides. A by-product of the process is an iron-rich slag which is non-hazardous. The main H T M R process is the Waelz K i l n . However, since other processes recover a higher value of zinc metal and can be carried out on-site, electrotechnologies, such as the I M S Plasma Process [1] might be attractive in the long term. A summary of some current commercial E A F processing by H T M R is given in Table II. A l l these processes recover zinc and leave behind an iron rich slag containing hazardous elements in nonleachable form. However, they are also perceived by steelmakers as being economically unatractive. 5 Table II - High Temperature Metals Recovery Processes ( H T M R ) [1] Process Company Product Waelz K i l n Horsehead Development Co. Zinc oxide and slag H T R Himeji Tekko Refine Co. Zinc metal and slag Inclined Rotary Reduction Z i a Technology Zinc metal and slag System Flame Reactor Horsehead Development Co. Zinc oxide and slag High Temperature Metal Elkem technology Zinc metal,iron and Recovery slag H i Plas Technology Davy M c K e e Ltd. Z inc metal, iron and slag IMS Plasma International M i l l Services Zinc metal and slag Sirosmelt B i g River Minerals Zinc metal and slag Pyromet-Mintek Plasma Mintek Zinc metal and slag Extended Plasma A r c Ticron Ecological Corp. Zinc metal and slag Reduction S K F Scandust S K F Engineering Zinc metal and slag Inmetco Intern't Metals Reclamation Processes stainless Co. steel E A F dust ISP Imperial Smelting Zinc metal and slag 2 - Direct Recycling within the Steelmaking Process - Since E A F dust is rich in iron, under some circumstances it can be recycled upsteam in the steelmaking process. This generally involves pelletizing or briquetting for subsequent charging to the blast furnace. However, the high levels of zinc and sodium present in the dust severely restrict the amount o f dust which can be recycled to the blast furnace. Examples of processes for reforming dust and subsequent recycling to the blast furnace are given for in Table III: 6 Table III - Direct Recycling Processes [1] Process Company Method Dereco Ca lx-Br iq ISI E U S Dereco Inc. Process Calx Inc. International Solidification Elwood Udderholm Steel Briquetting Briquetting with lime Solidification of mixed materials Briquetting 3 - Solidification/Stabilization - Solidification and stabilization are chemical fixation technologies which involve the mixing of specialized additives or reagents with waste materials to reduce, either physically or chemically, the solubility or mobility of the contaminants. Stabilization involves chemically modifying the contaminant to form a less soluble, mobile, or toxic form without necessarily changing the physical characteristics of the waste. Solidification involves changing the physical form of the waste to produce a solid structure in which the contaminant is mechanically trapped [1]. Regulations for landfill disposal require that all hazardous wastes meet the Toxicity Characteristic Leaching Procedure ( T C L P ) test prior to disposal in landfills. The T L C P procedure developed by the U . S . Environmental Protection Agency assesses the potential for leaching of hazardous contaminants from landfill wastes to groundwater. Environmental regulations give the leachate concentrations for a variety of heavy metals and organic compounds that are used in defining a waste material as hazardous. Several companies have developed chemical fixation processes which meet this test and some of the processes currently used are listed in Table IV: 7 Table I V - Solidification and Stabilization Processes T i l Process Company Stabl - Stat CeTech Resources Heckett technology Roanoke Electric Steel Conversion Systems. Inc. Quality Environmental Systems Solifix R T T S Super DeTox R T M A n alternative process [6], stabilizes heavy metal bearing wastes such as Pb, Cd, Zn , As , N i and Cr by adding additives to produce less soluble mineral phases. The process is reported to produce a stable, non-toxic compound, which is sufficiently resistant to natural rainwater leaching to pass T C L P tests. In general stabilization techniques have not found substantial application. From a technological point of view the production of non-hazardous materials containing E A F dusts can be accomplished, but economically these processes are still costly. 4 - Recycling Outside the Steelmaking Process - Additional technologies are currently being developed for the treatment of dusts outside the steelplants. These include using E A F dust in the production of mineral wool, ceramic products and glass beads. Since these products generally pass T C L P tests no further processing is required [1]. Several processes can be mentioned (Table V ) , however, their production costs are still considered high and the products have limited applications. 8 Table V - Additional E A P Dust Treatment Processes |"1] Process Company Comment Enviroscience Enviroscience, Inc. Mineral wool Inorganic Recycling Inorganic Services Corp Ceramic Materials Classification Classification Intl., Ltd. Glass beads Fertilizer Several M a y be disallowed in future As mentioned in the above table, glassification has some potential for recycling E A F . Several processes treating E A F dusts via glassification are commercially available. The types of glasses and their performance in terms of leaching are described below. The process developed by Oregon Steel M i l l s [8] is similar to commercial fabrication of glass in that the E A F dust is combined with other glassmaking materials and melted at ~ 1400°C in a glassmaking furnace [9]. The molten product exits the furnace and it is quenched to obtain glass granules. The product passes environmental tests, including T C L P , (Table VI ) and the company is reportedly achieving some success in marketing the vitrified products as roofing materials, ceramic glazes and colorants, abrasive media and synthetic slag fluidizers. Table V I - T C L P Results for Glassified E A F Dust T91 Contaminant Arsenic Barium Cadmium Chromium Lead Mercury Selenium Silver Regulatory level (mg/1) 5 100 1 5 5 0.2 1 5 Detection limit (mg/1) 0.50 0.50 0.050 0.050 0.10 0.00050 0.50 0.050 Test 1 (mg/1) 0.12 0.0042 Test 2 (mg/1) 0.16 0.57 0.0060 Test 3 (mg/1) Test 4 (mg/1) 0.12 0.0021 Test 5 (mg/1) 9 Another vitrification process for treating E A F dust has been developed by C A N M E T [14]. The amount of iron oxide in the product is 10 wt%. The basic technique is to vitrify and densify the E A F dust mixture at about 1340 °C, immobilizing the hazardous elements in the final product. Leaching tests of the products have confirmed compliance with Ontario regulations. Potential uses for the resulting product include aggregate for concrete blocks and raw material for bricks and tiles, as well as simply disposal in standard landfill sites. Bethlehem Steel [15] has also reported the development of a stabilization technology which renders E A F dust non-hazardous. The process involves a series of chemical reactions to convert the metals to their least soluble states by combining metals as metallic-silicate polymers [15]. The leachate levels of the final product comply with the environmental regulations. Although the product is not completely amorphous, the concept of obtaining metallic-silicate polymers is similar to the one used in silica-based glasses, since silica glass can be cosidered a polymeric material formed by the polymerization of S i 0 2 tetrahedra. 10 3.0 GLASSIFICATION Glasses are known to provide a possible solution for trapping hazardous elements due to their ability to incorporate a variety of elements in a nonleachable form. Certain general considerations w i l l further be discussed before moving on to the specific topic of silica glasses containing iron which are of primary interest in the present work. Glass is a valuable and versatile materials and there are about 700 different commercial glasses. The constitution of glass can be explained by the Zachariasen network theory [17]. The theory postulates the formation of networks of oxygen tetrahedra such that, although there is no periodicity of structure which would yield a sharp X-ray pattern, the statistical arrangement is similar to the ordered arrangement of the crystalline state. In slags the net-forming cations are silicon and aluminium which link together by sharing the oxygen atoms of a tetrahedra. Calcium and magnesium ions occupy the relatively large voids in the network structure and are known as network modifiers. Aluminium ions can act as a network former or modifier, but they can only replace Si ions to a limited extent and in doing so they increase the interionic distance between the cation and the oxygen ions. This increases the specific volume that can accomodate more of the larger Ca ions in the voids in the structure. For tetrahedral-type stable glasses the ratio of the network forming cations to oxygen should lie between 0.33 and 0.50. Another theory of glass structure is the crystallite theory [18] which postulates a 'micro heterogeneity' in glasses. A glass is considered to contain short-range domains in which there is a certain degree of order, or, as they have been termed, 'embryos of crystallization'. These regions contain most of the cations and are linked together by amorphous regions formed by the 11 remaining anions. Langavant [19] has suggested that, in the liquid slag, there is a tendency for polymerised S i x O y groups to exist as a framework of chains or bands leaving a higher ratio of base to acid ions amongst the other consituents of the melt. Very hot slags cannot sometimes be chilled to a glass and this is attributed to fracture of the silica network with the loss of its inhibiting effect on crystallization. The difference between the network and crystallite theories of glass structure may well be reduced as more information becomes available but it is clear that the mineral compounds do not exist as such when the slag is in vitreous state and that there is some form of two-region structure. Glass-formers are those elements such as Si , B , P, As , that can be converted into glass when combined with O, S, Te or Se. In the liquid state, the silicon-oxygen bond is very strong so that a high activation energy is necessary to break the bond. Also , the molecular arrangement is conducive to formation of an intricate three-dimensional network of oxygen tetrahedra with a silicon atom in the middle bonded to each oxygen atom. Common glasses contain about 70 % S i 0 2 [16]. Glasses are regarded as supercooled liquids. It is a characteristic feature of silicate melts that on rapid cooling from the liquid state they tend to form a glass [16]. The passage from the liquid to a crystalline solid structure is accompanied by a rearrangement of the ions which take up a definite orientation in the crystals. The viscosity of molten silicates near the freezing-point is sufficiently large that this rearrangement only takes place slowly. I f cooling is rapid, the ionic groups largely retain their irregular arrangement, and, the viscosity increasing rapidly as the temperature drops, the slag passes from the liquid state to one in which the rigidity approaches that of a solid, but without the development of a crystalline structure. Glasses are thus greatly undercooled liquids having a very high internal viscosity. Glasses have no definite melting-point, but soften and gradually pass into a liquid state on heating. A t ordinary 12 temperatures, a substance may persist in the glass condition, but, i f it is annealed at a temperature not too far below its melting point, devitrification may begin and continue more or less rapidly after cooling. Glass network structures tend to be covalent. The Si-O, P-O, B - O , A s - 0 bonds are all 50 percent covalent or greater [16]. Modifiers tend to bond ionically to the anions in the glassy network. They are used to alter properties; for instance, the addition of modifiers can decrease viscosity and increase both electrical conductivity and thermal expansion coefficients. In the manufacturing process, the most important factor controlling the workability is the viscosity. Small changes in composition may affect significantly the viscosity at a given temperatures. Strain point, annealing point, softening point and working point (all measures of viscosity) must be known before a given glass is blown, molded or cast into a rigid product. Seeds and blisters are small gas bubbles formed during the chemical reaction of the raw materials and by oxidation of impurities in the molten glass. They do not usually significantly reduce glass strength. Crystallized glasses, commonly referred to as glass-ceramics, are more than 50 % crystalline and depend on nucleation and crystal growth using a nucleating agent, which can be precipitated as a crystal or an immiscible liquid. Generally they are of two types: semi-crystalline articles, where only the surface areas of the article are crystalline and the interior portion is glassy and fully crystallized, where the whole body contains crystals generally randomly oriented. 13 3.1 Silica Glasses Sil ica glass is the most important of the single-oxide glasses. In many respects it is an ideal glass, resistant at high temperature, with a low coefficient of thermal expansion and very low ultrasonic absorption. It is an excellent dielectric and is highly resistant to chemical attack. Virtually pure silica glass is very expensive to produce. Frondel [21] stated that amorphous silica is not truly amorphous but consists of local atomic order, or crystals of extremely small size, which by careful X R D examination appear to have a cristobalite structure. However by ordinary diffraction procedures this amorphous material gives only a broad band, with no wel l defined peaks which are ordinarily obtained from crystalline phases [20, 21]. Sil ica glasses have a high ability to incorporate a wide variety of elements in their structure. This is the reason why they are used as a base for inglobing hazardous wastes. However, there are certain limitations regarding the amount of crystal producing elements, such as iron and other transition elements, that can be inglobed in a silica structure. 3.2 Composition Limitations of Iron in Glassmaking The literature gives little information on the fabrication of glasses with iron and the role of iron in glassmaking, however, several works can be mentioned as attempting to determine to what extent F e 2 0 3 can be incorporated into a glass and the type of glasses, solubility problems and melting temperatures. 14 Iron is normally considered as a crystal producing element; however, when it is present as F e 2 0 3 , in the tetrahedral form it acts as a glass former, while in octahedral form it has a natural tendency to crystallize [30]. When present as F e 3 0 4 in tetrahedral form it is again a glass former. According to Corning Glass [23], F e 2 0 3 is a compound which is not conducive to glass formation when present in large amounts. In general, as the iron content increases, the glasses devitrify uncontrollably upon cooling from the melt or glasses cannot be formed. Recent work [22] describes glass batches in the barium silicate system containing up to 40 wt% iron oxide. Glass granules were obtained when melting the mixtures of oxides at 1600°C in platinum crucibles, followed by water-quenching. Another work was carried out by Corning Glass Works [23] regarding the fabrication of glasses in the system P b O - F e 2 0 3 - S i 0 2 which were found to yield ferrite upon suitable heat treatment. The general composition range of these glasses is 20-40 mole % PbO, 10-20 % F e 2 0 3 and 40-65 % S i 0 2 . Melt ing was done at 1500°C in an air atmosphere. The major constituent of these glasses is F e 2 0 3 . The glasses are all semiconductors and exhibit magnetic effects, being further crystallized to obtain ferrites. Dumbaugh [24] describes glasses, not containing conventional glass-forming oxides such as silica or boric acid, but comprising 33-68 % PbO, 2.5-27 % CdO, 10-30 % Fe^O;,, 4-28 % T1 2 0. The key oxide in the forming of the glasses is iron oxide (Fe 2 0 3 ) . Although this oxide is not normally considered to be a glass-forming oxide, it does have a strong stabilizing influence on the lead oxide. CdO is used as modifying oxide, improving glass-forming and working characteristics. A smelting process has also been developed [25] for the treatment of solid hazardous wastes. Wi th a tap temperature of 1500°C, the process produces metal, fumes and slag. Slags are rich in iron, nickel and chromium and when slowly cooled, present more crystallinity than when 15 water-quenched. Slags with high iron content, such as 38-45 % F e 3 0 4 , 24-42 % S i 0 2 , 1-5 % N i O , slow cooled or water-quenched exhibit dendritic and plate-like growth of the precipitated phase. Slags with lower iron content in the range of 20 % F e 3 0 4 quickly cooled do not exhibit phase precipitation. The slag, due to its siliceous structure, was tested to be leach resistant. These slags exhibit magnetic properties. Another study on glasses in the system F e 2 0 3 - S i 0 2 - N a 2 0 , melted at 1400°C in platinum crucibles using an electric furnace with S iC heaters, reports an iron content in the glass of 10 wt% F e 2 0 3 [26]. A n investigation developed at the University of Leeds [27] shows that the addition of F e 2 0 3 (5-15 wt%) and subsequent heat treatment (600-950°C/24 hr) of a L i 2 0 - M g O -A l 2 0 3 - S i 0 2 results in the development of a glass-ceramic with a silver and gold coloration and a brilliant lustre. The study of glasses in the system N a 2 0 - F e 2 0 3 - S i 0 2 and especially the ones with the composition of aergerite has also been reported [28]. Pure aergerite is an alkali mineral of the pyroxene group and a rock-forming component of volcanic alkali rocks containing 52 wt% S i 0 2 , 34.6 wt% F e 2 0 3 , 13.4 wt% NajO and various impurities such as K 2 0 , M g O , A 1 2 0 3 , T i 0 2 , etc. The mineral is chemically stable, but has no industrial uses as a raw material. Glass formation in the above mentioned system and crystallization of glasses with the composition of aergerite were investigated in this study. Laboratory tests performed at 1350°C rendered glasses containing up to 25 wt% F e 2 0 3 . The glasses were found to be resistant to water. Goethite, an iron rich toxic waste that is generated in the roast-leach-electrowin process for producing zinc concentrates, is another material that is considered a hazardous waste on the basis of heavy metal content. A process has been developed for the production of glass-ceramic materials from goethite waste, reportedly fabricating products with properties suitable for commercial use. In this process the goethite is mixed with other industrial wastes and raw 16 materials, melted, vitrified and crystallized by thermal treatment [10]. This study may have originated with current research trends aimed at trapping toxic and radioactive wastes in a glass or glass-ceramic matrix. For the work, melting was carried out in an alumina crucible at 1400-1450°C and quenching was performed on graphite moulds. Four base compositions were tested with an F e 2 0 3 content between 15 and 25 wt%. The evaluation of the glass-ceramic product consisted of X-ray diffraction to verify that no crystallization occured and determination of the amorphous and crystalline fractions. Phases identified in the X-ray diffraction spectra were: clinopyroxene, olivine, magnetite and titano-magnetite. Crystalline phases and crystalline/amorphous phase ratio were similar to those of basaltic glass-ceramics obtained from fused rocks [10]. Another literature source mentions that optical properties of glasses with up to 35 wt% F e 2 0 3 were investigated by Faust and Peck [29]. A study on glasses with low content of iron (0.001-1.0 %) shows that it is possible to produce glasses with iron in a silicate melt when F e ^ 3+ and F e 3 0 4 are present and not when FeO is present. The explanation is that Fe can have 2+ tetrahedral or octahedral coordination and can be linked to silica tetrahedra, while Fe has a sixfold coordination, leaving non-bridging oxygens in the silicate structure [30]. 3.3 Zinc, Cadmium, Lead and Sodium in Silica Glasses and Glass-ceramics On the basis of the network hypothesis by Zachariasen, most alkaline-earth oxides are assumed to behave as network modifiers because of their large cationic radius [17]. Hence the incorporation of a CaO molecule in a silicate glass would result in the formation of two non-bridging oxygen atoms. M g O and ZnO, however, are believed to be intermediates between 17 network formers and network modifiers, depending on the exact glass composition. Glasses containing 10 % Z n O were obtained in the sodium-silicate system (70 % S i 0 2 ) . CaO, M g O and Z n O behaved as network-modifiers. In the presence of A 1 2 0 3 , however, Z n O and M g O are 2 2 stabilized as Z n 0 4 " and M g 0 4 " tetrahedra forming part of the glass network. A proposed mechanism suggests that non-bridging oxygens play an important role in promoting corrosion. Consequently glasses containing Z n O and M g O in combination with A 1 2 0 3 are expected to be highly corrosion resistant [31]. Another work [22] describes glass batches in the strontium silicate system containing up to 25 wt% ZnO. Cadmium is an essential ingredient in glass production. The addition of cadmium produces glasses colored in red, orange and yellow that are being used in traffic control and warning lights [32]. Toxic properties of Cd are reflected in legislation of several countries. The leachability of C d from these very stable glasses has been measured and the release makes an insignificant contribution to pollution. Leaching by ground water in disposal sites have been shown to be small compared with normal background levels of C d in soil. N o information is given regarding the amount of Cd added to the glasses. Glasses with 2-68 wt% PbO are reported in various studies [23, 24, 33]. Lead oxide is used as a modifying oxide. The glasses are obtained between 1400-1500°C and are reported to be very stable. Lead and cadmium act as stabilizers, increasing resistance to chemical attack [16, 23]. Sodium is added in glasses to lower the melting temperature and decrease the viscosity. When added in high amounts the glass becomes less chemically-resistant [16, 26, 28]. A l k a l i silicate glasses are prepared from silica glass and fluxes. These fluxes are added to reduce the high viscosity inherent to silica glasses. The most usual is soda, N a 2 0 . The addition of alkali 18 softens the glass by breaking S i - 0 bonds. As soda concentration is increased, it becomes impossible to quench without the formation of crystalline silicates. The addition of alkali also increases the solubility [16]. 3.4 Capture of Hazardous Elements by Glass Systems Many agencies in the United States have developed defence and commercial high-level waste ( H L W ) glasses. The basic task is to formulate a glass that can be processed with an acceptably high waste loading and the best possible chemical durability [34]. When certain elements are introduced at higher than acceptable levels, problems can appear, such as formation of crystalline sludges on the melter floor, increased off-gas system corrosion and decreased product chemical durability. Most of the glasses produced are of an alkali-borosilicate-based composition. This type of glass is proposed for both light-water reactor ( L W R ) and breeder fuels. In-situ vitrification is a new technology used for treating radioactive, organic and inorganic contaminated soils. In this process, electricity is applied through electrodes buried in the contaminated soils to melt it and form an environmentally stable glass-like solid. In full-scale operation, the I S V process can treat 4 to 6 tons of soil per hour. The hazardous wastes are immobilized in the glass-like solid [35]. Several aspects have to be considered in the production of H L W glasses, such as waste glass properties and glass-making problems related to specific waste components. The composition must have acceptable high-temperature properties for efficient melter performance and acceptable product properties for storage, transportation and disposal, in order to ensure the formation of a liquid at a reasonable glassmaking temperature (below 1500 °C), an adequate 19 fluidity, tolerance and also the glass should not undergo significant crystallization. Crystallization is the condition that limit the glass waste loadings. Crystallization of the final product during cooling is of concern only i f it significantly degrades leaching performance. Crystal ingrowth could cause microcracking of the glass, increasing the surface available for leaching. Because of its amorphous structure, silicate-based glass can incorporate a wide variety of elements. This attribute is one of the major advantages of glass as a nuclear waste form. However, waste glasses are limited in their ability to incorporate a number of elements. Aluminum, iron and transition metals are the most common sources of significant dense phase formation in glass melts [34]. For example, calcine (a fixed mixture of H L W oxides containing A l , Fe, M n and N i ) increases the amount of crystallinity in the glass. In glasses with 12-17 % Fe, and showing no crystal formation, the addition of 1 wt% C r 2 0 3 produces a few weight percent crystallinity after short times at 1050 °C. The maximum waste content of glasses is determined by the crystal-producing elements. Elements that either participate in crystal phase or displace crystal-forming elements from the glass structure, include Fe, other transition metals, A l , rare earths, Zr , Nb , M o , A g , Cd, Sn, Re, Pb and the actinides. The sum of these elements roughly indicates the maximum possible waste loading of the glass. It has been reported that waste glasses can accomodate 20-25 wt% of the oxides of these elements. Above this level, significant crystallinity occurs [34]. A l k a l i metal oxides are essential components in silicate glasses for nearly all practical applications. They serve as network modifiers, lowering the processing temperature. Alkal is are the main components in determining processing temperature and chemical durability. Sodium has the largest effect on decreasing glass melt processing temperature, but also on decreasing 20 chemical durability, as glass leachability increases. Viable waste glass formulations contain 10-20 wt% of alkali oxides. A recent investigation [36] has achieved the incorporation of up to 10 wt% A s into "glassy" silicate slag with very low levels of arsenic leaching. The slags considered are based on the system S i 0 2 - F e x O y - C a O - A l 2 0 3 . Experiments were performed at 1400 °C with compositions up to a maximum 60 wt% F e 2 0 3 (normalized) and a S i 0 2 / C a O ratio between 0.9 and 2.6. Variation of S i 0 2 / C a O ratio and the iron oxide content of the slag appear to have no significant effect on arsenic retention in the slag. Leaching results show that, in general, the levels of arsenic that are leached from the slowly cooled (crystalline) slag samples are high compared with those leached from water-quenched (glassy) slag samples. The results of the leaching tests established that the major influence on the leaching behaviour of the final slag is its glassy or crystalline nature when solid. Microstructural examination showed that slags of low iron oxide content (e.g. within the pseudowollastonite primary phase field) are very glassy when quenched in water. These quenched slags exhibit very low levels of arsenic leachability. The same is true even for slags of high iron oxide content, in which, even when water-quenched, primary crystallization of wustite or magnetite often occurs. The quite low levels of arsenic leaching that were observed may be due to the fact that the crystallization of a relatively pure iron oxide phase (principally magnetite) results in a silica-rich residual melt that subsequently forms a glassy phase when quenched. It appears that the incorporation of arsenic in silicate slags with subsequent water-quenching represents an environmentally acceptable route for the disposal of many arsenial residues. 21 The long-term stability of such slags is unclear and it must be further investigated. A study carried out by another group of researchers [37] shows that long-term aqueous exposure (7 years) of slags containing < 10wt% arsenic resulted in low levels of arsenic extraction (<2 mg/1), whereas slags with A s >18 wt% showed much higher extractions (6-9 mg/1). 3.5 Leaching Behaviour of Silica Glasses Glass-crystal composite (GCC) wastes are currently being developed for eventual disposal of nuclear and hazardous waste materials. These wastes forms are an alternative to the traditional vitreous and cementitious waste materials currently explored for disposal purposes. The glass fraction of G C C serves as a binder for the crystalline phases and also incorporates elements excluded from the crystal structures [39]. Experiments have shown that the majority of the U , N i , Cr, Ce can be incorporated in crystals with minor amounts in the glassy matrix. Leaching tests of these glass-crystal materials give results comparable to other wastes materials, including nuclear glasses. This study shows that although some crystallinity is present, the durability and corrosion-resistant properties of the material are not greatly affected. Leaching behaviour of individuals ions in a glass have been studied in a borosilicate glass containing Si , B , Na , A l , Cs, Sr, Ba, Fe ( 15 wt% ) , Z r and M o [40]. Monovalent ions, L i , N a and Cs, increase the rate of release, divalent ions, Sr, Ba, slightly improve and tri and tetravalent ions, A l and Fe, considerably improve the stability against chemical attack. Particularly, A 1 2 0 3 is very effective in reducing the total leach rate. N o particular mutual interaction was observed and the rate of release can be predicted by additives amount, although further experiment is needed. The mechanism of leaching is a complex process of corrosion and diffusion. Leaching varies 22 with time and a suppressive effect operates as times passes. S E M observations has revealed that a heterogeneous layer is formed at leached surface and its thickness is time dependent. X-ray line scan profiles of leached surfaces indicate that ions with a high leach rate, such as L i , Na , Cs and M o are depleted in the layer and ions with low leach rate are retained in the layer. Similar results were obtained by Altenheim and co-workers [41]. Divalent ions, Sr, Ba , are retained in the beginning and then gradually are lost in time. Si is accumulated in the layer of the glass. The fact that the layer is increasing its thickness with time and attains a saturated value suggests that the layer is a protective coating which impedes further leaching of glass, and the degree of protection depends upon the composition of the layer. In conclusion, the rates of release of constituents and stability are mainly governed by species and the contents of tri and tetravalent ions, particularly A l , Zr , Fe in the glass form. The rates of leaching decrease with time due to the formation of a protective surface layer and the degree of protection depends on the composition of the layer. Monovalent ions transfer rapidly by diffusion or solution, divalent ions transfer by diffusion from the gel layer, and tri and tetravalent ions are retained in the layer and transfer slowly by corrosion or crumbling of the layer. Another study on the interaction of silicates glasses with aqueous solutions [42] shows the formation of a hydrated silica-rich leached layer on the surface of the glass and its interface with the solution. Studies revealed that, unless buffered, aqueous solutions in contact with glass suffer substantial changes in their p H within seconds of contact due to contamination with the reaction products. A study carried out on solid residues from geothermal power generation [43], containing mainly silica, aluminium and iron oxides and hazardous elements such as As , Pb, Zn , C u and Cr, shows that when submited to leaching tests the samples release low levels of these hazardous elements. X-ray diffraction show the geothermal residues were poorly crystalline and 23 predominantly composed of amorphous materials. Common phases in the samples were quartz, magnetite, hematite, cristobalite, chalcopyrite, pyrite. Results of the regulation leaching procedure did not classify the geothermal residues as hazardous since their leachates are below the standard values. 24 4.0 SCOPE AND O B J E C T I V E S A t the present time, our knowledge base for the incorporation of E A F dust in a glass matrix extends only to a level of about 15 % iron oxide. Since E A F contains up to 40 % F e 2 0 3 this means that a very substantial addition of silica must be made to the processing furnace i f the 15 % iron oxide limit is to be observed. The economics of recycling E A F dust v ia glassification can be expected to improve as the silica addition declines since less overall mass must be melted. In addition, the size of the furnace w i l l also go down (per tonne of E A F processed) which reduces capital costs. Currently it is unkown i f 15 % F e 2 0 3 represents the actual limit of iron level in a silica glass or simply the limit of our knowledge. Therefore the overall objective of the project is to extend our knowledge o f the effect of Fe and Z n , both individually and in combination, onto higher levels o f concentration in a silica -based glass and determine the onset of concentrations which trigger the substantial crystalline structure. Since crystallinity is known to increase the leachability levels of hazardous elements present in a glassy material these limits are of significance with regards to rendering the dust environmentally benign. The work was essentially experimental and involved the preparation of synthetic, silica-based mixtures containing iron and zinc in quantities within the range of variation to be expected in the E A F dusts. The mixes were then melted and water quenched and the solidified product was examined by X R D , S E M and leaching techniques. Considering that zinc is one of the main hazardous elements present in the E A F dust, the experimental work also examined the leaching behaviour of silica glasses containing zinc, both 25 on its own and in combination with iron. Special emphasis was also given to the influence of crystallinity on the leaching process. For this purpose several glass samples were annealed, thus converting these to glass-ceramics. These were later subjected to leaching tests and the results compared to those of the amorphous samples. 26 5.0 E X P E R I M E N T A L W O R K The experimental work involved the preparation of samples using laboratory grade oxides which were then melted in crucibles using a high temperature muffle furnace and subsequently quenched. Actual arc furnace dust was not used due to the presence of minor components which would exert an influence on the test results. For all experiments the S i 0 2 component was fixed at 50 wt%, since this level is known to give a high probability of glass formation and also allows a relatively broad range of variation for the iron oxide and other network modifiers. In all tests, 2 wt% sodium carbonate was included in order to lower the melting point, fluidize the melts and facilitate the removal of the sample from the crucible. Although not really a factor in the test program, an additional benefit o f N a 2 C 0 3 addition is an increase in the chemical resistance of the glass. Sodium, as wel l as calcium and magnesium, were included in the raw material blends at levels which are typical for E A F dusts. However, their influence on the glassification process was not explicitly considered. In addition the influence of basicity was not explored and for most trials a C a O / M g O ratio of 1.79 was maintained. This enabled the incorporation of less calcium in the melt, thus allowing for the addition of other oxides of interest. For the test program a uniform sample size of 50 g was employed. Crucibles were standard fireclay containing approximately 50 % silica and 50 % alumina. A l l samples were heated to 1500 °C in a muffle furnace and held for 60 minutes. Although this is 100 to 200 °C higher than the theoretical liquidus temperature for the melts, the superheat was necessary to 27 achieve adequate pourability in the acidic S i 0 2 melts and is fairly typical of melt temperature in commercial glassmaking operations [6]. Quenching is an important step in the fabrication of glasses, since, by definition, glasses are product of fusion cooled to a rigid condition without crystallizing. Various quenching methods were considered; i.e. air, water, a copper chil l plate and a graphite mould. In order to finalize the quench method, samples containing 15 % F e 2 0 3 , a level known to be easily incorporated into a glass, were melted and quenched using each of these possibilities. Subsequent X R D examination of the product spectra indicated that water quenching generated the most consistently an amorphous product owing to the high cooling rate achieved, therefore the standard technique employed was to quench samples by pouring directly into bath of 7 °C (approximately) running water. After quenching the fragmented product was dried and granulated in preparation for microscopic analysis, X-ray diffraction analysis and leaching tests. The X R D examination was performed on a computerized Rigaku Rotating Anode X-ray diffractometer which is designed specifically for measurement of powders and polycrystalline thin films. To determine possible mineral phases present, the peaks from the spectra were analyzed using a search and match program with respect to the JPCDS index. In order to examine the microstructure and chemical composition, samples were mounted and polished to a 1 micron grit, carbon coated and analyzed using S E M and E D X . Generally, glass surfaces are difficult to observe under a microscope, since in a completely amorphous structure no crystallinity w i l l be visible. Therefore some of the samples were not polished in order to obtain an image of the fractured surface o f the glass. X-ray diffraction analysis and S E M analysis were the principal tools used to determine the presence of 28 crystallinity in the samples and also identify the mineral phases present, i f any. Included in the analysis of the test results was a comparison with a reference commercial glass containing 15 % F e 2 0 3 , as wel l as the use of phase diagrams to explain the possible phases formed when exceeding the solubility limits or when annealing the glasses. The experimental program also included an evaluation of the leaching behaviour for the samples containing zinc (the main hazardous element present in the E A F dust). The influence of crystallinity on the leaching process was also assessed and, for this purpose, several samples were annealed in order to obtain crystalline materials or glass-ceramics. The leaching tests were performed according to the British Columbia Leaching Extraction Procedure ( L E P ) [46], protocol which is similar to the Toxicity Characteristic Leaching Procedure ( T C L P ) , and is designed to simulate underground conditions experienced in landfilling. Since the amount of sample generated in the tests was relatively small (40 g) compared to the 100 g specified by L E P , the analytical methods were scaled down appropriately and also modified with respect to particle size ranges. The standard specifies particle sizes below 9.5 mm mesh opening. However, leaching is enhanced by the specific surface [9] and the products which might reasonably be marketed after glassification of E A F dust (for example sandblasting grit or roofing) can be expected to be of somewhat smaller sized particles (and hence have higher specific surface area) than the L E P standard. Therefore two particle size ranges were employed for the leaching tests, (- 3.35 mm + 1 mm), which represents the real size of a potential product to be made out of these glasses, and - 200 mesh (-74 \xm). These test conditions are more aggressive than the standard. In accordance with L E P procedures, samples were subjected to continuous leaching for 24 hours, while maintaining the p H constant at 5.0 by the addition of 0.5 acetic acid. The leachate was vacuum filtered through a 0.45 | im pore size paper prior to analysis. The 29 concentrations of zinc in the leachate solutions were analyzed by atomic absorption. The results were compared to leachate quality standards. According to the Waste Management A c t (Special Waste- Schedule 4) any waste material exceeding 500 mg/1 Z n in the leachates, is considered a special waste. However, there is no restriction for Fe content. 30 6.0 RESULTS The experiments were designed to examine the role of F e 2 0 3 on glassmaking ability and particularly to assess the highest possible level of F e 2 0 3 which could be incorporated in a glass. A secondary aim was to determine the effect of key minor elements such as Zn , C d and Pb on the ability to form amorphous product structures. The compositions of the mixtures, the solubility limits o f iron and zinc in a silica-based glass, as well as the possible incorporation of cadmium and lead in the glass are discussed in subsequent sections. 6.1 The Influence of Iron on Glassforming The economics for vitrification of E A F dust improve as the silica addition drops since the total weight of material to be melted declines accordingly. This implies F e 2 0 3 levels wel l in excess of the 15 % used in commercial glasses, since the F e 2 0 3 content of dust is in the range of 10 to 45 %. Therefore during the experimental campaign the F e 2 0 3 content was systematically varied from 15 wt% to 40 wt% in order to determine the solubility l imit of iron in a silica glass. The compositions of the synthetic melts are shown in Appendix I, Table I, which indicates the initial mixtures, i.e. prior to melting. A s noted earlier, a 50 wt% S i 0 2 level was chosen in order to ensure the formation of a silica glass structure and also allow for a relative broad range of variation for iron oxide and other network modifiers. Wi th the aim of introducing less variability in the compositions to be tested, C a O / M g O ratio was maintained between 1.78 and 4.00, since these oxides are network modifiers and can introduce additional solubility problems in the glass. 31 3000 -p 2500 --^ 2000 -0 -( 1 1 1 1 1 1 1 H 1 1 1 1 1 ' 7 11 16 20 25 29 34 38 43 47 52 56 61 65 70 Angle (2 Theta) Fig. 1 - X-ray diffraction pattern of sample with 35 % Fe 20 3, water quenched 32 The X-ray diffraction pattern for the 35 wt% F e 2 0 3 sample is shown in F ig . 1 in which the amorphous hump, which is characteristic of glassy materials, occurs within the 2® angle range of approximately 16° to 38°. The noise seen in the curve is attributed to the fact that particles submitted to the test were fine and also because the sample holder was not rigorously firm in place while the sample was rotating. The cooling rate of the sample is critical in the formation of the glass. When slow cooled in air, crystallinity is detected in the form of distinct peaks in the diffraction pattern associated with the interatomic distances in crystalline structures. A search and match program gives the probability o f matching the interatomic distance or the corresponding 2 0 angle to a certain known crystalline structure thus identifying the crystals. For example a diffraction pattern of the 35 % F e 2 0 3 melt slow cooled shows distint peaks that can be attributed to the formation of crystals of M g F e 2 0 4 (magnesioferrite) and S i 0 2 (Fig.2). 10 16 22 28 34 40 46 52 58 64 70 Angle (2 Theta) Fig . 2- X-ray diffraction pattern for sample with 35 wt% F e 2 0 3 , air quenched In contrast to the material obtained with iron oxide concentrations of 35 wt% or lower, at 40 wt% F e 2 0 3 the product begins to show substantial crystallinity. The distinct peaks shown in F ig . 3 indicate the presence of cristobalite, magnetite, hematite and magnesioferrite crystals. 33 Fig. 3 - X-ray diffraction pattern for sample containing 40 wt% Fe203, water quenched where : a = Fe304, b = Si02, c = Fe 20 3, d = MgFe204 34 According to the ternary phase diagram for the system S i C V C a O - F e j C ^ (Fig.4), at 40 % F e 2 0 3 , 50 % S i 0 2 and 10 % (CaO + M g O ) the first crystals to be formed upon cooling w i l l be cristobalite. The X R D analysis indicates the presence of cristobalite, as well as magnetite, hematite and magnesioferrite. The presence of magnetite and hematite might be explained by the fact that when crystallization of cristobalite occurs in the range of 1400-1500 °C, an iron enriched solid solution is left behind which, with continued cooling, generates crystals of magnetite and hematite around 1300-1400 °C. The presence of magnetite and hematite as possible crystals in high iron content glasses was also reported in an investigation carried out for silicate glasses with 60 % F e 2 0 3 (normalized) and arsenic incorporation [36]. According to the quaternary diagram CaO - M g O - F e 2 0 3 - S i 0 2 - shown in F ig . 5 magnesioferrite is also one of the possible crystalline phases to be formed. S E M analysis confirms the presence of crystallinity when the iron oxide content exceeds 35 %. For example, the micrograph of the sample containing 35 % Fe20 3 (Fig. 6) shows a glassy structure incorporating Fe, Si , Ca, and M g , while the ones corresponding to 40 % F e 2 0 3 (Fig. 7) show a dendritic crystalline area containing a solid solution of Si , M g and Fe (light area) surrounding a glassy structure containing mostly Si (dark area). The S E M analysis of the sample containing 35 % Fe 2 0 3 was performed on an unpolished surface of the glass. F ig . 6(a) shows the fractured glass surface at a 500x magnification (60 |xm), while F ig . 6(b) shows the same sample at a 3000x magnification (10 p.m), where small white particles can be observed. White areas seem to be just dust particles on top of the surface and not crystallites, since an E D X 'spot' analysis of the particles give the same chemical composition as the rest of the sample and also, under a microscope, these appear on a different plane than the sample itself. When the same sample is polished, the S E M analysis gives no visible structure even at higher magnification (3 Fig. 5 - Quaternary phase diagram of system CaO - MgO - Fe 20 3 - Si0 2 [47] (a) 6 3 1 2 6 2 S 0 K V 'k I ! I 0 K " i 6 ". 8 U ;n (b) Fig. 6 - SEM of glass sample with 35 % Fe 20 3 , fractured unpolished surface, (a) 60 urn, (b) 10 urn Fig. 7 - SEM of sample with 40 % Fe 2 0 3 38 |j,m). A t the macroscopic level, samples with 35 % F e 2 0 3 (Fig. 8) have a typical glass luster, while at 40 % F e 2 0 3 (Fig. 9) this characteristic is lost. In summary, the experimental work strongly indicates a 35 % Fe20 3 solubility limit for iron in silica glasses containing 50 % S i 0 2 , at least for the test conditions. This result is in agreement with studies showing that although iron generally is regarded as a crystal forming element, when present in the F e 2 0 3 form and in a silica matrix it can be glassified [10, 24, 25, 30, 35, 36]. Fig. 9 - Sample containing 40 % F e 2 0 3 40 6. 2 The Influence of Zinc on Glassforming The Z n O content of E A F dusts can also vary substantially (anywhere from 2 to 40 wt%) and therefore a second series of experiments were carried out to determine the solubility limit of Z n O in the glassy structure. The compositions of all samples are summarized in Table II, Appendix I. In this series, Z n O content was varied between 15 to 35 wt% while the S i 0 2 level was again held at 50 wt% and CaO/MgO was maintained at 1.79. N o Fe20 3 was included in these trials. Again the samples were heated to 1500 °C and held for 1 hour prior to water quenching. The X R D results for the progressively increased Z n concentrations show essentially amorphous structures for 30 wt% and lower. However at 35 wt% (Fig. 10) sharp peaks appear which can be attributed to the formation of ZnO, Z n 2 S i 0 4 , S i 0 2 and M g O (periclase) crystals. A n S E M micrograph shows a homogeneous glassy structure for a content of 30 wt% Z n O (Fig. 11), while the X-ray diffraction indicates the glassy state (Fig. 12). Figures 13 and 14 provide a comparison o f the glassy material containing 30 % Z n O to the crystalline material obtained at 35 % ZnO. One aspect of glassification which has important ramifications on leachability of hazardous components is the stability of the amorphous structure over extended periods. Therefore, several X-ray analysis were repeated 6 months after production o f the glasses to verify that devitrification did not occur. Thus, for short periods of time, the amorphous structure o f glasses containing zinc appears to be stable. 140 4-12 17 23 28 33 38 44 Angle (2 theta) 49 54 59 Fig. 10 - X-ray diffraction pattern of sample with 35 % ZnO where : a = Si02, b =MgO, c = Zn2Si04, d= ZnO 4 2 Fig. 11 - SEM of glass sample with 30 % ZnO 43 Fig. 12 - X-ray diffraction pattern of sample with 30 % ZnO Fig. 14 - Sample containing 35 % ZnO 45 6. 3 The Combined Influence of Iron and Zinc on Glassforming Having established the individual influence of Fe20 3 and Z n O on crystallinity in the silica melt, the next series of tests was designed to assess the effects when both are present in the system. Three basic levels of total Fe20 3 + Z n O were employed for these trials; i . e. 40, 35 and 30 wt% as shown in Tables III through V , Appendix I. These levels were chosen with the aim of establishing a region of glass formation in the system F e 2 0 3 - Z n O - C a O - M g O - N a 2 0 , starting with a total of 40 %, representing the level of iron oxide at which the material is not a glass anymore. A s in the previous series of tests, the S i 0 2 content and C a O / M g O ratio were held constant at 50 % and 1.79, respectively. For each level of total F e 2 0 3 + Z n O the ratio Fe:Zn was systematically varied as indicated in the tables. The ternary diagram shown in F ig . 15 summarizes the glassmaking potential for all the mixtures tested (including the S i 0 2 - F e 2 0 3 and S i 0 2 - Z n O trials). In the system F e 2 0 3 - Z n O - C a O -M g O - N a j O at 50 % S i 0 2 , the solubility limit of Fe^O-j is 35 wt%, while the solubility limit of Z n O is 30 wt%. The ternary diagram also indicates that at higher iron levels there is a tendency to produce crystallinity. The results are in agreement with experimental data reported for glasses in the F e 2 0 3 - S i 0 2 - N a 2 0 system, with high silica content (55-85 %) and fabricated at 1350 °C. This previous study reveals the possibility of obtaining glasses containing 30 % Fe20 3 , with a chemical composition similar to volcanic alkali rocks [28]. The tendency to crystallization is also likely to increase when the sum o f F e 2 0 3 and Z n O is around 35 %. Studies done on waste glasses have shown that the maximum waste content of glasses is usually determined by the combination of crystal-producing elements. Elements that Fig. 15 - Region of glass formation in the F e 2 0 3 - ZnO - CaO - MgO - Na 2 0 system at SO % S i 0 2 and 1500 °C. 47 either participate or displace crystal-forming elements from the glass structure include Fe, other transition metals, Cd and Pb. The total of these elements roughly indicates the maximum possible waste loadings of the glass. It is believed that waste glasses can accommodate 20 to 25 % o f heavy metal oxides. Above this level significant crystallinity might occur [34]. Our study however, shows that the silica glass (50 % S i 0 2 ) can incorporate up to a total of 40 % of these elements in a mixture containing 20 wt% Fe20 3 and 20 wt% Z n O and a total of 35 % for a combination of 30 wt% Fe20 3 and 5 wt% Z n O (Fig. 16). Combinations of 35-5 or 30-10 produced crystallinity. Their diffraction patterns are given in F ig . 1(a) and F ig . 2(a), Appendix n. 6. 4 The Influence of Cadmium and Lead on Glassforming Considering that E A F dusts contain Pb and Cd, several melts were prepared with PbO and CdO in order to determine i f these elements can be incorporated into a glassy structure. Their addition to the mixture was done in quantities equal to the maximum content normally present in electric arc furnace dusts. The range of test conditions is summarized in Table V I . A s in all previous tests, melting was carried out at 1500°C for 1 hour followed by water quenching. A l l compositions produced glasses. CdO is a network modifier, and should contribute to crystallinity. Since the CdO addition was actually small (0.08-0.97wt%), its influence on the crystallization process was not clearly established. On the other hand, PbO is a glass former [16], and no problems were expected regarding its incorporation into the glassy structure. Figure 17 shows an S E M image of a glass containing iron, zinc, lead and cadmium, indicating no crystallinity at 10 jam. On the unpolished fractured surface small particles of dust can be seen. Fig. 16 - SEM of glass sample with 30 % Fe 2 0 3 and 5 % ZnO Fig. 17 - SEM of glass sample containing iron, zinc lead and cadmium 49 The X-ray diffraction curves for glasses with Cd and Pb are presented in the Appendix II, F ig . 3(a) and F ig . 4(a), together with an X-ray diffraction and a S E M image corresponding to a reference commercial glass sample fabricated with E A F dust and having a composition of 15% iron, 10% zinc, 3% lead and 0.08 % cadmium oxides in F ig . 5(a) and F ig . 6(a). 50 6.5 Stability of the Glass Structures The long-term stability of glasses incorporating hazardous elements is an important issue. Although one study on silica glasses with high iron content and containing arsenic [37] suggests that chemical stability is maintained, the possibility of devitrification is not considered. In the present work several samples of glasses tested by X R D for devitrification showed the amorphous structure was retained 6 months after their production. For practical reasons, glasses were not tested for longer times, but rather were submitted to accelerated aging by means of annealing. The main objective was essentially to assess the influence of aging (slow reversion to the stable crytalline state) on the ability of the glass to retain hazardous components under leaching conditions. Several samples of amorphous material containing Z n and Fe were crystallized for this purpose. The conditions for crystallization as wel l as the crystalline phases formed in the process are reported in detail in the following subsection. 6.5.1 Devitrification of Glasses with Zinc. Studies done on the devitrification of glasses show the process occurs by nucleation and growth of thermodynamically stable crystalline phases in the glass at temperatures sufficiently high that ionic mobility is adequate to support it. Generally, the temperature for devitrification is reported to be 200-300 °C below the melting temperature. A literature review on glasses with zinc reveals a study of the devitrification of PbO - Z n O - B 2 0 3 glasses [45] concluding that higher content of Z n O (> 10 %) promotes devitrification. The exact mechanism o f nucleation 51 and growth of crystallites is not well understood, although it is believed to be occurring first at the surface. In order to determine the temperature at which glasses with zinc begin crystallizing samples o f glasses containing 30 wt% Z n O were submitted to annealing at 900 and 1000 °C for 1, 2 and 3 hours. After annealing, the X-ray diffraction patterns (Fig. 18 and Fig . 19) still retain the characteristic amorphous hump of a short range order which characterizes amorphous structures [44]. This result suggests that glasses with high zinc content have good stability at high temperature and are also stable in time, since the annealing process can be regarded as an accelerated ageing of the glasses. However when subjected to 1180 °C, evidence of crystallinity begins to appear at 2 hours. This is shown in the S E M micrograph (Fig. 20) as gray platelike crystals evolving from a black glass matrix. The analysis of the X-ray diffraction pattern (Fig. 21) indicates the sample is not completely crystalline showing peaks that arise from the short range order hump and can be attributed to the formation of S i 0 2 , willemite ( Z n 2 S i 0 4 ) , C a 2 S i 0 4 , calcium zinc silicate ( C a 2 Z n S i 0 7 ) , zincite (ZnO) and rankinite ( C a 3 S i 0 7 ) . It is very likely that the solid solutions might contain these phases, since the ternary diagram for the system CaO -Z n O - S i 0 2 (Fig. 22) for compositions of 50 % S i 0 2 , 30 % Z n O and 16 % CaO indicates a eutectic at 1170 °C with the possibility of formation o f these silica compounds. The possibility of obtaining willemite is confirmed again by the ternary diagram M g O - Z n O - S i 0 2 (Fig. 23), for a composition of 50 % S i 0 2 , 30 % Z n O and 4 % M g O at aproximately 1200 °C. A t a macroscopic level, the presence of crystallinity can be easily detected, since devitrified glass samples change their appearance from a transparent green to an opaque white-green as shown in F ig . 24. 52 300 28 35 42 49 Angle ( 2 Theta ) 56 63 70 (a) 300 i 250 4-= 200 o — 150 >. « 100 -f c <L) •S 50 •4- •4-14 21 28 35 42 Angle ( 2 Theta ) 49 56 63 70 (b) 300 Angfe ( 2 Theta ) 4 9 56 63 70 (C) Fig. 18 - X-ray diffraction patterns of sample with 30 % ZnO, (a) after 1 hr, (b) after 2 hr, ( c) after 3hr at 900 C Fig. 19 - X-ray diffraction pattern of a glass with 30 % ZnO after 3hr at 1000 C 54 Fig. 20 - SEM of 30 % ZnO sample after 2 hr at 1180 C 55 200 a 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 Angle ( 2 theta) Fig. 21 - X-ray diffraction pattern of sample with 30 % ZnO after 2hr at 1180 C where : a = willemite (Zn2Si04), b = Ca2Si04, c =Si02l d =calcium zinc silicate ( Ca2ZnSi207) e = zincite (ZnO), f = rankinite (Ca 3Si0 7) CaO-ZnO-SiOj CaO ZnO Fig. 22 - Ternary diagram CaO - ZnO - Si0 2 [47] M g O - Z n O - S i 0 2 MgO ZnO Z n 2 S i 0 4 S i 0 2 Mot. • / • Fig. 23 - Ternary diagram MgO - ZnO - Si0 2 [47] 57 Fig . 24 - Sample of glass with 30 % ZnO (left) and same material crystallized at 1180 °C for 2 hours (right) 58 Additional glass samples containing 15 to 25 % Z n O were devitrified at 1180 °C for 2 hours in order to be used in further leaching tests. 59 6.5.2 Devitrification of Glasses with Iron Preliminary tests done during this project indicated that glasses with iron are less stable than glasses with zinc and that they devitrify at 1080 °C, 100 °C lower than the temperature required for the devitrification of zinc glasses. A sample of glass containing the highest level o f Fej 0 3 (35 wt%) was annealed at 1080 °C for two hours. The X-ray diffraction curve reveals several distinct peaks. The peak search and match of the JPDCS index indicates the phases with higher probability of formation are magnetite (Fe 3 0 4 ) , magnesioferrite ( M g F e 2 0 4 ) and periclase (MgO). Some peaks, with lower degree of confidence, can be attributed to the presence of crystals of wollastonite (CaS i0 3 ) , larnite (Ca2Si0 4 ) and maghemite ( F e ^ ) as shown in F ig . 25. Possibilities o f obtaining these phases are in agreement with ternary phase diagrams for the systems CaO - Fe20 3 - S i 0 2 and CaO - M g O - S i 0 2 shown in Fig . 26 and Fig . 27, respectively. S E M micrographs also confirm the crystallization of the samples in Fig . 28 and Fig . 29 showing calcium silicates (light area) amid glassy areas (dark area). The devitrification can be detected immediately after taking out the sample from the furnace, as the material loses the brightness and luster typical to a glass surface, becoming dim and shaded, and changing its color from black to brown-gray ( F ig . 30). 60 100 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 Angle (2 theta) Fig. 25 - X-ray diffraction pattern of sample with 35 % Fe 2 0 3 after 2 hours at 1080 C where : a = magnetite (Fe 30 4), b = magnesioferrite (MgFe 20 4), c = periclase (MgO), d = wollastonite (CaSi0 3), e = larnite (Ca 2Si0 4), f = maghemrte (Fe 2 0 3 ) 61 CaO-Fe203-Si02 -—System CaO-FeiOr-StO, in air at subsolidus temperatures. (A) for temp, slightly above 1155°C, (B) for temp, slightly below 1155°C. Hem. = hematite. • Bert Phillips and Arnulf Muan, J. Am. Ceram. Soc, 42 [9] 422 (1959). Fig. 26 - Ternary phase diagram of the system CaO - Fe 2 0 3 - S i0 2 [47] Fig. 27 - Ternary phase diagram of the system CaO - MgO - S i0 2 [47] 8 31c 89 2 8 KV k3!86K' 10!0u1 Fig. 28 - SEM of 35 % Fe 20 3 sample after 2 hr at 1080 C 0 3 12 8 3 £ 0K V k 3 @8K ' 11! 0 u m Fig. 29 - SEM of 35 % Fe 2 0 3 crystallized sample 64 Fig . 30 - Sample of glass material containing 35 % F e 2 0 3 (right) and same material crystallized at 1080 °C for 2 hours (left) 65 6.5.3 Devitrification of Glasses with Zinc and Iron Samples of glasses containing both iron and zinc were also submitted to annealing. For example, one o f the sample of glass containing 30% Fe20 3 and 5 % Z n O was devitrified after being held for 2 hours at 1080 °C. The macroscopic observation o f the sample in F ig . 31 denotes the presence o f crystals through a change of color from lustrous black to brown gray. A t microscopic level, S E M micrographs show intertwined ramificated crystals (Fig. 32), while distinct peaks appearing in the X-ray diffraction pattern (Fig.33) confirm the sample devitrified. The evaluation of the peaks suggests the presence of magnetite (Fe 3 0 4 ) , cristobalite (S i0 2 ) , zincite (ZnO), wollastonite (CaSi0 3 ) , larnite (Ca2Si0 4 ) and calcium zinc silicate (Ca2ZnSi 2 0 7 ) . A validation of these phases by using a complex diagram is not possible. The literature does not provide a diagram for a complex system such as S i 0 2 - F e 2 0 3 - Z n O - CaO - M g O - N a 2 0 or even a simpler one such as S i 0 2 - F e 2 0 3 - ZnO, thus the results had to be interpreted according to phase diagrams with either iron or zinc oxide and not a combination of both. In this light the crystalline phases identified by X R D are in consonance with the ones obtained for glasses containing either iron or zinc oxide described in earlier subsections. 00 (b) Fig. 31 - Sample containing 30 % F e 2 0 3 and 5 % ZnO, (a) amorphous, (b) crystallized Fig. 32 - SEM of 30 % Fe 2 0 3 and 5 % ZnO sample crystallized at 1080 C for 2 hr 68 1600 -p 1400 -1200 -| 1000 -7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 Angle (2 theta) Fig. 33 - X-ray diffraction pattern of sample with 30 % Fe 20 3 and 5 % ZnO after 2 hours at 1080 C where : a = Fe 30 4, b = Si02, c = ZnO, d = CaSi03 l e = Ca2Si04, f = Ca2ZnSi207 69 6.6 Leaching Behaviour of Glasses Recent studies on waste glasses show that silica glasses are capable of incorporating hazardous elements and, when subjected to leaching tests, produce leachates with low concentrations o f toxic elements [7,9,36]. Glasses developed in the present work contain Zn , Pb and Cd , all o f which are designated hazardous by the current environmental protocols. Leaching tests were carried out to confirm the suitability of using silica-based glass as a stabilizer for zinc in particular, since zinc represents one of the major hazardous elements present in the E A F dust. The results of the leaching tests are presented in Table VII . The standard limit for rendering the glass as non-hazardous is 500 ppm of Z n in the leachate solution. Atomic absorption analysis o f the leachates gave concentrations below this limit, which means that glasses containing 15 to 35 wt% Z n O are capable of stabilizing this element. Even the samples that are not completely amorphous complied with the standard. The amount of Z n in the leachate was found to increase when the percentage o f Z n O in the glass was augmented (Fig. 34). This tendency was also apparent in the glasses containing both Z n O and Fe20 3 (Table VII -B) . The leaching process is enhanced by an increased surface area exposed to the leaching agent. This is confirmed by those tests performed on fine particles. Fine particle samples below 200 mesh (-74 |xm) produced leachates with higher concentrations of Z n than coarser samples in the 1-3 mm size range. It was also observed that the completely amorphous samples (glasses) have lower levels of Z n in the leachates than the partially crystalline ones (glass-ceramics). Table VI I - Chemical Analysis of the Leachate Solutions A . - Glasses with Z n O additions Sample Sample(-200 mesh) Sample (-3.35mm+lmm) No. ppm Zn ppm Fe ppm Zn ppm Fe Z n 15 48.41 - 12.36 -Z n 20 39.51 - 19.00 -Z n 25 90.17 - 3.52 -Z n 3 0 ( l ) 127.90 - 12.69 -Z n 30(c) 155.50 - 11.82 -Z n 35(c) 260.50 - 7.78 -Z n 35(c) 174.10 - 13.99 -B . - Glasses with Z n O and F e 2 0 3 additions Sample Sample(- 200 mesh) Sample (-3.35mm+lmm) No. ppm Zn ppmFe ppm Zn , ppm Fe FeZn 05-35 241.80 1.04 10.91 0.05 FeZn 10-30 208.00 1.82 10.92 0.13 FeZn 15-25(c) - - 44.86 0.11 FeZn 20-20 77.36 1.81 24.76 0.11 FeZn 25-15(c) 43.60 1.62 10.24 0.06 FeZn30-10(c) - - 35.03 0.05 FeZn35-05(c) - - 27.21 0.09 (c) crystallinity present 71 LEACHING BEHAVIOUR OF ZnO GLASSES •a £ o re <u E a. a. c N # Glass - 200mesh s Glass ceramic - 200mesh ; : Glass - 3.35mm +1 mm x Glass ceramic - 3.35 mm + 1 r 15 20 25 30 35 ZnO(%) in glass or glass ceramic mixture 40 Fig . 34 - Level of Z n O in the leachate solution as a function of % Z n O in the glass or glass-ceramic Leaching tests were also performed on samples o f glasses containing zinc after being annealed at 1180 °C. Table VIII, as well as F ig . 35, indicate that the materials showing crystallinity display higher levels of zinc in the leachate solution than the same samples in amorphous state. These results are in agreement with experimental work done on silica glasses with iron and incorporating arsenic as hazardous element [36], which indicates that the major influence on the leaching behaviour of the material is its glassy or crystalline nature and which also reports lower leaching levels for glasses compared to crystalline materials. 72 Table VIII - Leaching Results for Z n O Glasses Crystallized at 1180 C, 1 hr Sample Original Glass Crystallized sample Z n 15 12.36 34.14 Z n 20 19.00 27.11 Z n 25 3.52 33.09 Z n 3 0 11.82 24.24 50 40 30 i 20 10 0 Leaching behaviour of crystallized glass samples containing ZnO 10 15 20 25 30 35 Glass Crystallized sample Fig . 35 - Leaching behaviour of crystallized glass samples containing Z n O 73 A l l glasses or glass-ceramics evaluated in the present work, when submitted to leaching, display zinc concentrations below the 500 ppm limit established by the standard for special wastes for rendering a material as non-hazardous. The levels are low compared to the standard, however it is interesting to point out again that there is a clear tendency of enhanced leaching levels in the presence of crystallinity. Leaching results for glasses with Pb and C d are not available. However, considering the low levels of the corresponding oxides in the glass and the fact that Z n and Fe glasses have proved leach resistant, no significant amount of Pb and Cd in the leachate solution is expected. This speculation finds support in recent reports of leaching tests performed on silica glasses containing Pb and Cd which rendered metal concentrations below standard limits [16, 23, 32]. In general the results are in agreement with previous studies carried out on waste or nuclear glasses incorporating hazardous materials [36,39] which report low levels of contaminants in the leachate solutions. One explanation to this phenomena is the fact that glasses undergo hydration immediately after mixing with water and a protective layer is formed on its surface, which inhibits the water penetration to the glass particle and the further dissolution of ions from the glass as perceived by previous research works [41, 42, 43]. 74 7.0 CONCLUSIONS Electric arc furnace dust currently represents a substantial economic and environmental liability to steelmakers. Since several hazardous components of the dust are in leachable form, the dust is classified as hazardous and ever more stingent regulations and costs can be expected for disposal. One option for resolving the problem is encapsulating the dust in a glassy, silica structure so as to render the hazardous components essentially unleachable. Since the glass products have some potential as marketable products, this provides a recycling option which could be economically attractive to steelmakers. However, the economics of vitrification improve as the amount of silica added to the dust decreases. The present work consisted of an experimental campaign to determine the envelope for iron and zinc incorporation into a silica glass matrix, as well as the influence of structure (amorphous versus crystalline) on leaching characteristics. Several conclusions can be drawn with respect to iron and zinc limits in silica-based glasses and the influence of crystallinity onto the leaching resistance o f these materials: (i) Silicate glasses in the system Fe20 3 -CaO-MgO at 50 % S i 0 2 can incorporate up to 35 wt% Fe20 3 . Above this level the product becomes partially crystalline due to the presence of magnetite, hematite, magnesioferrite and cristobalite. (ii) In the system Z n O - C a O - M g O at 50 wt% S i 0 2 , the zinc oxide solubility l imit is 30 wt% ZnO. A n increased amount of Z n O w i l l produce crystals of zincite, willemite, periclase and cristobalite. 75 (iii) When analyzing the combined influence of iron and zinc on glassforming in the system Fe 2 03-ZnO-CaO-MgO-Na20 , the total of Fe20 3 and Z n O that can be incorporated into a glass is 35 % (30 wt% F e ^ and 5 wt% ZnO). However, for lower amounts of iron oxide the summation of F e 2 0 3 and Z n O can reach 40 % with a combination of 20 wt% F e 2 0 3 and 20 wt% ZnO. (iv) The addition of 0.08 - 0.97 % CdO and 0.97 - 2.99 % PbO to mixtures containing 25 % F e 2 0 3 and 5 % Z n O did not contribute to crystallinity, rendering amorphous products. C d O addition was done in relatively small amounts and its influence on the crystallization process was not observed. On the other hand, PbO is a glass former and it is known to be one o f the oxides that easily can be incorporated into a glass structure. A n increased amount of PbO in the silica glass should facilitate the formation o f a glassy structure. (v) Leaching levels for silica glasses containing Z n O up to 35 wt% are below 500 ppm Zn , which represents the limit established by environmental regulations for rendering a material as non-hazardous. Even when the surface area exposed to leaching is higher, as is the case of samples with particle sizes < 74 um, the levels of zinc are below the standard. These samples leach out 40 - 260 ppm compared to 4 - 20 ppm for coarser samples with particles size between 1 and 3 mm. Leaching tests performed on amorphous and crystalline samples indicate that crystalline samples leach out almost twice the amount of zinc compared to the amorphous ones (12 vs. 34 ppm Zn). Nevertheless for both crystalline and amorphous samples, the zinc level is below the maximum limit established by the standard. For samples containing both iron and zinc 76 the same tendency is observed. The presence of iron does not seem to influence the leachability o f zinc. (vi) This study also suggests the possibility of incorporation of other hazardous elements in the silica glass, such as Pb and Cd. Leaching results are expected to be comparable to the ones obtained in glasses containing zinc, since both elements are present in low amounts which ensure their efficient incorporation into the silica glass structure. This research indicates the possibility of obtaining silica-based glasses with high iron contents which implies higher loadings of E A F dust into the glassmaking furnace and a decreased amount of additives with consequent economical benefits. Satisfactory results in the leaching tests show the ability of silica glass to incorporate hazardous elements such as Zn , Cd and Pb and hence create an alternative to expensive disposal costs associated with E A F dust. 77 8.0 R E C O M M E N D E D W O R K Research has still to be conducted toward a less costly vitrification process o f the E A F dust and also toward establishing new potential applications for this product. In this respect, it would be of interest to fabricate glasses at temperatures lower than 1500 °C. The addition o f more than 2 % sodium carbonate to the melt might be a solution for a lower working temperature. Glasses fabricated with real E A F dusts samples would have to be contemplated in order to determine the possibility of using this material in higher loadings in the mixture. Special care might have to be given to solubility problems related to chromium and nickel present in the dust. Previous studies show that 1 % of each of these elements might promote significant crystallinity with correspondent leaching levels higher than the limit established by environmental regulations. To find new applications for these types of materials, it would be useful to perform tests for the characterization of their thermal and magnetic properties. Furthermore the fabrication o f glasses with 0.005 - 0.01 % selenium ( some samples of E A F dust already contain low amounts o f Se) might give products with different colors ranging from pink to brown, which might enlarge the possibility of using the glasses as ceramic tiles and decorative objects. 78 Leaching tests for glasses with cadmium and lead w i l l give more information on the ability o f silica-based glasses in trapping hazardous elements. 79 R E F E R E N C E S 1. J.E. Goodwil l , R.J . Schmitt, " A n Update on Electric A r c Furnace Dust Treatment in the United States", 3 3 - Annual Conference of Metallurgists of C I M . Proceedings of the International Symposium on Resource Conservation and Environmental Technologies in Metallurgical Industries. Toronto, Aug . 20-25, 1994. 2. Lehigh University, Bethlehem, P A . "Characterization, Recovery and Recycling of Electric A r c Furnace Dust", Final Report. Prepared for U .S . Dept. of Commerce. Feb., 1982 3. C M P . Center for Materials Production, Carnegie Mel lon Research Institute, "Electric A r c Furnace Dust-1993 Overview", Report No.93-1 Arthur D . Little. Pittsburgh, P A . , 1993 4. J R . Donald, C A . Pickles, " A Review of Plasma-arc Processes for the Treatment o f Electric A r c Furnace Dusts ", Proceedings of the International Symposium on Resource Conservation and Environmental Technologies in Metallurgical Industries. 33— Annual Conference of Metallurgists o f C I M . Toronto, Aug . 20-25, 1994. 5. A . M . Haghi, R . D . Hagni, " Reflected Light and Scanning Electron Microscopic Study of Electric A r c Furnace Dusts ", Residues and Effluents - Processing and Environmental Considerations. The Minerals. Metals and Materials Society. Warrendale, P A , 1992, pp. 117-125. 6. K . Forrester, " Heavy Metal Stabilization Optimization Study ", Extraction and Processing for the Treatment and Minimization of Wastes. The Mineral . Metals & Materials Society. 1993. 7. A . M.Hagni , R . D . Hagni, " Significance of Mineralogy of E A F Dust in Pyrometallurgical Treatment to Render the Dust Non-hazardous ", Extraction and Processing for the Treatment and Minimizat ion of Wastes. The Mineral . Metals & Materials Society. 1993. 8. R . B . E k , J .K. Bray, J.O. Richartz III, " Glassification of Electric A r c Furnace Dust at Oregon Steel M i l l s , Inc.". Internal report. 9. J im Sundal, Kathi Futornick, Oregon Steel M i l l s , Inc., " Regulatory Issues Regarding the Recycling of Electric A r c Furnace Dust and its Use as a Slag Fluidizer ". Extraction and Processing for the Treatment and Minimization of Wastes; The Minerals. Metals & Materials Society. 1993, p. 637-644. 10. M.Pel l ino, C.Cantalini, F.Veglio," Crystallization o f Glasses Obtained by Recycling Goethite Industrial Wastes to Produce Glass-ceramic Materials ", Istituto Trattamento Mineral i , Roma, Italy; Journal of Material Science 29 (1994), p. 2087-2094. 11. P .W. M c M i l l a n . Glass Ceramics. Academic, London, 1979, p. 168. R E F E R E N C E S (cont.) 80 12. J.Hlavac. The Technology of Glass Ceramics. Elsevier. 1983. 13. M . Pelino, L . Marinacci, C. Cantalini and P.P. Boatini, Advanced Ceram.Glass 4 (1992) p.110. 14. J. Aota, S.A. Mikha i l , D.T. Liang and W . N . Howell , " L o w Temperature E A F Dust Vitrification Process", 33— Conference of Metallurgists. A u g 20-25, 1994, Toronto. 15. R . G . Hil ton and J.D. Lynn, " Application of the Super Detox Process to E A F Dust Treatment ", Extraction and Processing for the Treatment and Minimizat ion of Wastes; The Minerals. Metals & Materials Society. 1993, p. 579-594. 16. D . G . Holloway, The Physical Properties of Glass , 1973. 17. W . H . Zachariasen, Journal of Amer. Chem. Society 54.1932. p. 3841. 18. E . A . Porai-Koshits, Glastech. Berichte. 32, p. 450, 1959. 19. J.C. de Langavant, Revue Mater. Constr. Trav. publ. 401, 38 to 411, 425, 1949 and 448, 1 to 453, 175, 1953. 20. G . L . Peralta, J.W. Graydon, D . W . Ki rk , " Characterization of the Solid Residues from Geothermal Power Generation ", Extraction and Processing for the Treatment and Minimizat ion of Wastes, The Minerals. Metals & Materials Society. 1993. 21. C. Frondel, The system of mineralogy of Dana , John Wi ley and Sons Inc, 1962 N e w York, 7 t h Edition, V o l . 3, Sil ica Minerals. 22. R . L . Andrus, J. F. MacDowel l . , "Coated metal article ", United States Patent. No . 5,250,360, Oct.5, 1993, 23. P .C . Shultz, " Magnetic and Semiconducting glass-ceramics in the PbO-Fe^CvSiQ^ system", Proceedings of the 1969 Annual Meeting of the Intl. Commission on Glass. 24. W . H . Dumbaugh, Jr. Glass Technology - Recent Developments. Noyes Data Corporation, 1976, U . S . Patent 3,837,867, September 24, 1974. 25. M . J. M C G u i r e , " Non-traditional use of the Cupola Furnace ", Extraction and Processing for the Treatment and Minimization of Wastes. The Minerals. Metals & Materials Society. 1993. 26. N . A . Ghoneim, H . A . ElBatal , " Thermal Conductivity of some Ternary Silicate Glasses Containing F e 2 0 3 in Relation to Composition and Structure ", XIII Internationaler Glaskongress. 4 bis 9 Juli , 1983, Hamburg. 81 R E F E R E N C E S (cont.) 27. M L . Wang, R. Stevens, G . Hallam, P.F. Knowles, P. Knott, " The Role of Fe20 3 in a Transparent Glass-ceramic ", XIII Internationaler Glaskongress. 4 bis 9 Juli , 1983, Hamburg. 28. V . A . Sharai, N . N . Ermolenko, I.G. Lukinskaya, "Crystallization of Glasses of the Na20-F e 2 0 3 - S i 0 2 System ".The Structure of Glass. V o l . 5, Edited by N . A . Toropov and E . A . Porai-Koshits Consultants Bureau, New York, 1965. 29. G . W . Morey, The Properties of Glass , 2 n d Edition, Reinhold Publishing Corporation, 1954, Optical Properties, p. 406. 30. E.I . Swarts, L . M . Cook, " Photosensitivity of Reduced Silicate Glasses Containing Iron ", 7~ International Congress on Glass. June26-July3,1965. 31. B . M . J . Smets, " The Incorporation of Divalent Cations in Sodium Silicate and Aluminosilicate Glasses ". XIII Internationaler Glaskongress. 4 bis 9 Juli, 1983, Hamburg. 32. W . W . Fletcher, G.J. Copley, " Environmental Mearurements during the Production of Cadmium-containing Glasses ", Glastechnische Berichte, XIII Internationaler Glaskongress. 4 bis 9 Juli, 1983, Hamburg. 33. E . Greiner, " Study of Technology of Aventurine Glass ", Glastechnische Berichte. XIII Internationaler Glaskongress. 4 bis 9 Juli, 1983, Hamburg. 34. L . A . Chick, J .L. Swanson and D.S . Goldman,"Nuclear Waste Glass Composition Limitations", Fuel Reprocessing and Waste Management A u g 1984, Jackson, Wyoming Proceedings/American Nuclear Society Interational Topical Meeting. 35. B . Zhao, C. Tingand N . J . Themelis," Mathematical Model l ing o f in-situ Vitrification", Extraction and Processing for the Treatment and Minimization of Wastes, Edited by J. Hager, B . Hansen, W . Imrie, J. Pusatori and V . Ramachandran, The Mineral . Metals & Materials Society. 1993. 36. N . C . Machingawuta and C P . Broadbent, " Incorporation of Arsenic in Silicate Slags as a Disposal Option ", Trans. Instn. M i n . Metall.fSect. C: Mineral Process. Extr. Metall), 103, Jan-A p r i l 1994. 37. L . G . Twidwel l and A . K . Mehta, " Disposal of Arsenic Bearing Copper Flue Dust ", Journal of Nuclear Chem. waste. Manag.. 5, 1985, 297-303. 82 R E F E R E N C E S (cont.) 38. W.J . Crama, " A Network Model for Multi-component L iqu id solutions (slags) Including Amphoteric Oxides (A1 2 0 3 ) " Calphad X X I Meeting, 1992, Jerusalem, Israel.(Abstract) 39. D.J . Wronkiewicz, " The Development of Glass-crystal Composite Waste Forms for the Disposal of Nuclear and Hazardous Waste Materials ", Proceedings 1994 Annual Meeting o f The Geological Society of America. Seattle, Wa, Oct. 24-27, 1994. 40. K . O . Oka, K . Oda, T. Yoshio, K . Oda, N. Tsunoda, " Leaching Characteristics o f Constituting Ions from the Simulated Nuclear Waste Glass ", XII I Internationaler Glaskongress, 4 bis 9 Juli 1983, Hamburg. 41. F . K . Altenmein, W . Lutze, G . Malow. Ibid. Vol .3 , p. 363-370, 1981. 42. T . M . El-Shamy, A A . Ahmed, A . K . Abou-Seif, " Transfer of Cations across Glass-solution Interfaces XIII Internationaler Glaskongress. 4 bis 9 Juli 1983, Hamburg. 43. G . L . Peralta, J.W. Graydon, D . W . Ki rk , " Characterization of the Solid Residues from Geothermal Power Generation ". Extraction and Processing for the Treatment and Minimizat ion of Wastes. The Minerals. Metals & Materials Society. 1993. 44. E.Douglas, P. Mainwaring, M . Van Roode and R.T . Hemmings, " Determination o f Glass Content in F l y Ashes and Blast-furnace Slags Canmet Report 85-6E. oct., 1985. 45. D .R. Uhlmann, Glass : Science and Technology, vol.2, 1984, Academic Press Inc. 46. Ministry of Environment, B . C . , Waste Management Act . Special Waste. Schedule 4 . 1992, p. 72. 47. E . M . Levin , C.R. Robbins, H .F . McMurdie , Phase Diagrams for Ceramists , American Ceramic Society, 1969. 83 A P P E N D I X I Table I - Glasses with Iron (as-mixed compositions. wt% oxides ^ Sample F e l 5 Fe20 Fe25 Fe27 Fe30 Fe35 Fe40 15.00 20.00 25.00 27.00 30.00 35.00 40.00 CaO 22.40 19.50 16.00 16.00 16.00 9.60 6.40 S i 0 2 50.00 50.00 50.00 50.00 50.00 50.00 50.00 M g O 12.60 10.50 9.00 7.00 4.00 5.40 3.60 Table II - Glasses with Zinc ("as-mixed compositions. wt% oxides) Sample Z n l 5 Zn20 Zn25 Zn30 Zn35 Z n O 15.00 20.00 25.00 30.00 35.00 CaO 22.40 19.50 16.00 16.00 9.60 S i 0 2 50.00 50.00 50.00 50.00 50.00 M g O 12.60 10.50 9.00 4.00 5.40 85 Table III - Series of Compositions with ZfFe^O? + ZnO) ~ 40 wt% Sample FeZn 35-05 FeZn 30-10 FeZn 25-15 FeZn 20-20 FeZn 15-25 FeZn 10-30 FeZn 05-35 F e 2 0 3 34.58 30.00 25.02 19.90 15.03 9.76 4.94 CaO 5.79 5.67 5.44 5.46 5.44 5.53 5.79 S i 0 2 49.40 49.17 49.56 49.76 49.56 50.41 49.40 M g O 3.24 3.17 3.04 3.05 3.04 3.09 3.24 Z n O 4.94 10.00 15.03 19.90 25.02 29.27 34.58 N a 2 C 0 3 2.04 2.00 1.92 1.93 1.92 1.95 2.04 Table I V - Series of Compositions with ZfFeoO? + ZnO) - 3 5 wt% Sample FeZn FeZn FeZn FeZn FeZn FeZn 30-05 25-10 20-15 15-20 10-25 05-30 F e 2 0 3 30.12 25.47 20.30 15.37 10.17 5.17 CaO 7.48 7.16 7.19 7.16 7.36 7.02 S i 0 2 50.57 50.32 50.74 50.53 50.65 50.83 M g O 4.18 4.00 4.02 4.00 4.11 3.93 Z n O 5.01 10.53 15.22 20.42 25.11 30.58 N a 2 C 0 3 2.64 2.53 2.54 2.53 2.60 2.48 Table V - Series of Compositions with SfFeoO, + ZnO) ~ 30 wt% Sample FeZn 25-05 FeZn 20-10 FeZn 15-15 FeZn 10-20 FeZn 05-25 F e 2 0 3 25.84 20.88 15.10 10.78 5.81 CaO 8.79 7.89 8.63 7.80 8.23 S i 0 2 51.68 53.36 53.30 53.67 53.27 M g O 4.91 4.41 4.82 4.36 4.60 Z n O 5.68 10.67 15.10 20.64 25.18 N a 2 C 0 3 3.10 2.78 3.05 2.75 2.91 Table V I - Glasses with Iron. Zinc. Lead and Cadmium (as-mixed composition, wt% oxides) Sample FeZnPb FeZnPb FeZnCd FeZnCd FeZnPbCd F e 2 0 3 25.66 25.14 25.88 25.66 25.07 CaO 9.38 9.19 9.46 9.38 9.17 S i 0 2 51.03 50.00 51.49 51.03 49.88 M g O 5.24 5.14 5.29 5.24 5.12 Z n O 4.97 4.86 5.01 4.97 4.99 PbO 0.97 2.97 0.00 0.00 2.99 C d O 0.00 0.00 0.08 0.97 0.08 N a 2 C 0 3 2.76 2.70 2.78 2.76 2.70 87 APPENDIX H J L 89 91 92 Fig. 6(a) - SEM of reference glass sample 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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


Related Items