WATER TREATMENT TECHNOLOGIES FOR REMOVAL OF ACID, SULPHATE AND METALS A.J. GELDENHUYS M.lng Thesis submi for the degree Philosophiae Doetor in Chemical Engineering at the North-West University (Potchefstroom Campus) Promoter: Co-Promoter: Prof. F.B. Waanders Dr. J.P. Maree 2004 Potchefstroom DECLARATION I, Andries Johannes Geldenhuys, hereby declares before a Commissioner of Oaths: 1. That, all the material submitted for the degree PhD in Chemical Engineering at the North-west Universrty (Potchefstroom Campus) has not previously been submitted for a degree at any other universii 2. That, this submission takes place with due recognition given to my copyright where applicable n AP el L .......................... ..... Signed at P.FTWH. on this 13. day of 2005 A J Geldenhuys . ,. ~ ,,. . ...... S~." .. *,*, .. ,- ,.. - - . ,/ .,,.. :._/ >,,~~ ,,.. . . i> ,~, ,~ J. ._:, ..- --*Orl* .1.01 ?* as a,,., n.. I:,. .,*w6,.,, ,*; xl",,e: ...... .... ....... ..., ,*,,.,m ../.I._ **Fr.'"" 'n.*V.. a.o,r. .,nq !,'(n;Llg,.Ys :Iy ........... ...... : . - : rl~t~ms .me inas ~trt~,,,~.;~ rs,; a1,4mrl -,>. ... ; .. ,.> ....... .- .- mwmu n; iwrn !cia*.,mn qotoa ,m sre 16y+l, .'""",' :;,- '.,,".i.,,." . *. .," . . , . ... . , - *...,+" .-.. ..T cylOl;.ieirnwknhurnnaln, wacoiereathsraQr n r. *I rr.s.nn - I .- ............... ......... .......... 'FNAME EN VAN IN ORuKSFRlF IRNAME IN CCI I P~FM ..................... -- I ABSTRACT A great deal of research effort has been undertaken to find an effective solution to the problem of acid mine drainage. Indeed, South African legislation requires mining companies to respect environmental regulations by minimising water intake from local municipalities and providing a rehabilitation plan. In order for the South African mining industry to remain competitive, the proposed solutions have to be not only efficient but also economic. This is the reason for the use of a waste material being attractive for water treatment and an integrated treatment technology being developed to treat water to different quality levels. The main objectives of this study were to develop more cost-effective treatment processes speclfic to the needs of the mining industry in southern Africa and to investigate the technical and environmental feasibility of utilising an alkaline waste product from the local paper industry as stabilising agent for acid mine residues. All the research and development work was carried out on laboratory and pilot scale plants. Five papers, with the present author as principal contributor, will form the basis of this thesis, of which one has been published and four are being peer- reviewed, presented at international conferences (locally and overseas) and published in the proceedings of the various conferences. Another five papers, with the author as co-author, have also been presented at international conferences and are being published in the proceedings of these conferences, and are included in this thesis. The results of the laboratory and pilot scale studies have been incorporated into the design and implementation of the following full-scale plants: A limestone handling and dosing system to supply slurried limestone of constant density to the neutralisation plant was constructed and commissioned during 2001 at Navigation Section of Landau Colliery, W~bank A limestone handling and dosing system, including a fluidised-bed limestone neutralisation plant, was constructed and commissioned during 2001, at Ticor, Empangeni An iron(ll)-oxidation and fluidised-bed limestone neutralisation plant was constructed and commissioned during 2002 at BCL, Selebi-Phikwe, Botswana A limestone handling and dosing system, to supply slurried limestone of constant density to the neutralisation plant, was constructed and commissioned during 2003, at Kromdraai Colliery, Witbank An iron(l1)-oxidation and fluidised-bed limestone neutralisation plant was constructed and commissioned during 2004, at the Navigation Section of Landau Colliery, Wibank These plants consist of specific units (stages) of the completely integrated process, developed by the CS1R:Environmentek over the past four years. These stages are: Heating unit: Production of CaO (quick lime) and C02-gas from burned coal and precipitated CaC03 (limestone) Limestone neutralisation and partial sulphate removal to a level of 1 900 mg/e Ca(OH)2 (hydrated lime) stage: CaO contacted with the acid water to produce Ca(W2 Lime treatment stage: Partial sulphate removal as CaS04 (gypsum) to below 1 200 mgle, and full removal of magnesium and other metals pH adjustment stage: C02 from the heating unit applied to reduce the pH to 8.6 while CaC03 precipitates Barium sulphide treatment or biological sulphate removal treatment: Removal of sulphate to below 200 mg/e Production (regeneration) of barium sulphide: Heating barium sulphate from the above stage Stripping of H2S either from the barium sulphide or the biological sulphate removal processes. H2S is contacted with Fe(lll)-rich water for elemental sulphur production. 'n Uitgebreide navorsingsondersoek is geloods om 'n effektiewe manier van waterbehandeling te vind vir suur mynwater en probleme wat daarmee gepaardgaan. Suid-Afrikaanse wetgewing vereis dat myne die omgewing met versigting hanteer deurdat die aankoop van water vanaf die munisipalieit beperk word tot 'n minimum en 'n rehabiliasieprogram in plek moet wees. Vir die Suid-Afrikaanse mynbou- industrie, om koste-effektief en kompeterend te wees, moet die behandelingsprosesse effektief en ekonomies wees. Daarom is die gebruik van 'n afvalproduk as alternatief tot alkaliese medium, so 'n aantreklike opsie en het gelei tot die ontwikkeling van 'n geintegreerde behandelingstegnologie. Hierdie tegnologie behels die behandeling van swak kwalieit water, afkomstig vanaf myne, tot spesifieke vlakke, afhangend van wat die eindbestemming van hierdie water is. Die hoofdoel van hierdie proefskrif was om 'n meer koste-effektiewe manier van waterbehandeling te ontwikkel wat voldoen aan die behoefles van die mynbou- industrie in Suid-Afrika. Ook is daar intensief gekyk na die tegniese en omgewingsvatbaarheid van die gebruik van kalksteen, 'n afval produk vanaf die plaaslike papierindustrie, as alkaliese medium tot neutralisering van suur mynuitvloeisels. Alle navorsingswerk en ontwikkeling is gedoen in 'n laboratorium en loodskaalaanleg. In die hoedanigheid as hoofouteur, is vyf artikels gelewer, waarvan een gepubliseer is. Die ander vier artikels is geproeflees en aangebied op internasionale konferensies, plaaslik sowel as oorsee, asook gepubliseer in die verrigtinge van die onderskeie konferensies. Hierdie vyf artikels vorm die basis van die proefskrif terwyl 'n verdere vyf ander artikels, waarby die skrywer as mede-outeur opgetree het, as aanvullende materiaal aangebied word. Die resultate van die laboratorium- en loodskaalaanleg studies het gelei tot die ontwerp, oprigting en inwerkstelling van die volgende volskaal aanlegte: 0 Kalksteen behandelings- en doseringsisteem om vloeibare kalksteen met konstante diitheid te lewer aan neutraliseringsaanleg - opgerig en inwerking gestel gedurende 2001 te Navigation seksie van Landau Colliery, Wibank Kalksteen behandelings- en doseringsisteem en gefluidiseerde bed kalksteen neutraliseringsaanleg - opgerig en inwerking gestel gedurende 2001 te Tiwr, Empangeni Yster(l1)-oksidasie and gefluidiseerde bed kalksteen neutraliserings aanleg - opgerig en inwerking gestel gedurende 2002 te BCL, Selebi-Phikwe, Botswana Kalksteen behandelings- en doseringsisteem om vloeibare kalksteen met konstante digtheid te lewer aan neutraliseringsaanleg - opgerig en inwerking gestel gedurende 2003 te Kromdraai Colliery, W~ank Yster(l1)-oksidasie and gefluidiseerde bed kalksteen neutraliserings aanleg - opgerig en inwerking gestel gedurende 2004 te Navigation seksie van Landau Colliery, Wfibank Bogenoemde aanlegte bestaan uit een of meer van die volgende stadiums van die geintegreerde proses sws ontwikkel oor die afgelope vier jaar deur die WNNR: Verhittingseenheid - produksie van CaO en Codas deur steenkool verbranding, CaC03 presipitasie Kalksteen neutralisasie en gedeeltelike sulfaatverwydering (tot 1 900 mgle) Ca(OH)2 stadium - CaO word gekontak met suur water om Ca(OH), te produseer Kalkbehandelingstadium - gedeeltelike sulfaatverwydering as CaS04 (gips) tot 1 200 mg/!. volledige verwydering van magnesium en ander metale pH verstelling na neutraal - C02 afkomstig vanaf verhiiingseenheid verlaag die pH tot 8.6 terwyl CaC03 presipiteer Barium suffied behandelingstadium of Biologiese sulfaatverwydering - sulfaat word verwyder tot 200 mg/e en laer Produksie van barium suffied - vemit barium sulfaat vanuit bogenoemde stadium Stroping van HS vanaf die barium suffied proses of die biologiese sulfaat verwyderingsproses. H2S word dan in kontak met Fe(lll) ryke water vir swawel produksie gebring Increasing exploitation of the natural water resources in southern Africa will necessitate widespread exploitation of nonconventional sources, such as municipal and industrial wastewater. The direct use of wastewater, as well as being an economically attractive option of utilising supplies, represents prevention of pollution and eutrophication and exploits a guaranteed source of supply which can be tailored to meet specific requirements. The advanced treatment options described in this thesis may be utilised to further these objectives. This thesis reports the results of research conducted at the Council for Scientic and Industrial Research, in the Division for Water, Environment and Forestry Technology (CSIR: Environmentek), situated in Pretoria, on newly developed water treatment technologies during the past four years. The author was personally involved in the research and development of these technologies and processes as well as in the construction and running of laboratory and pilot scale units and the commissioning of full scale plants. The purpose of this thesis is to present basic principles and general process guidelines based on operational experience with the various processes employed for the production of reusable water from mining effluents, resulting from coal mines and the mining of heavy minerals. Numerous full scale plants, based on specific stages of the developed technology, have been designed by Wates, Meiring 8 Barnard and constructed by Thuthuka Project Management, both of which are industrial partners of the CSIR. These plants were operated by the CSIR under the supelvision of the author as process engineer for a period of 12 - 24 months in order to optimise the processes. He also provided on-site training to mine operators in order to ensure efficient running and understanding of the technology. These full scale, operational plants are listed in the Abstract of this thesis. GLOSSARY Acid mine drainage: Acid water that is rich in iron and is produced when pyrites (Fe2S) is oxidised in water due to the presence of air and iron oxidising bacteria Fluidised-bed reactor: A column type reactor pa&ed with solid material (e.g. limestone). A gas is moved through the reactor at a high enough rate to expand the volume inside the reactor Limestone: Ore containing primarily CaC03 Slaked lime: Ca(OH)2 SRB Sulphate Reducing Bacteria Unslaked lime: CaO CONTENTS Abstract Uittreksel Preface Glossary Chapter I: Literature Overview 1.1 Mining of Heavy Minerals 1.2 NickelCopperCobalt Mining 1.3 Coal Mining 1.3.1 Introduction to Coal 1.3.2 Coal Formation and Composition of Coal Macerals 1.3.3 Mineral Matter 1.3.4 Coal Mining and the Environment 1.3.5 Origin of Acid Mine Drainage 1.3.6 Geochemistry of Acid Mine Drainage 1.4 Conventional Treatment Options for Acid Mine Drainage and Acidic Solutions I .5 Alternative Neutralisers 1.6 Further Treatment 1.7 Novel Integrated Technology for Treatment of Acid Mine Drainage and Effluents I .8 Legal requirements Chapter 2: Papers 1 to 8 2.1 Introduction to Papers Paper I : Paper 2: Geldenhuys,A.J., Maree. J.P., de Beer, M. and Hlabela, P. 2003. An integrated limestonellime process for partial sulphate removal, The J. South African Institute of Mining and Metallurgy, 103(6). 345 - 353. Geldenhuys, A.J. & Maree, J.P. Synthetic organic polymers (PAC6 and 3095) as coagulantslRocculanh for optimisation of an integrated limestone/lime neutralisation process for partial sulphate removal, Pmeedings 9 Annual Industrial Water Management and Treatment Symposium, 1 5 - 1 6 May 2002, Johannesburg, South Africa. Paper 3: Geldenhuys, A.J., Maree. J.P., Strobos, G., Smit, N. and Buthelezi, B. Neutralisation and partial sulphate removal of acid leachate in a heavy minerals plant with limestone and lime, Proceedings fjrn lnternational Conference on Acid Rock Drainage, 12 - 18 July 2003, Cairns, Australia. Paper 4: Geldenhuys, A.J., Maree, J.P., Fourie, W.J., Bladergroen, B.J. and Tjati, M. Acid mine drainage treated electrolytically for recovery of hydrogen, iron(ll) oxidation and sulphur production, Proceedings 8m lnternational Congress on Mine Water and the Environment, 19 - 22 October 2003, Johannesburg, South Africa. Paper 5: Maree, J.P., de Beer, M., Geldenhuys, A.J., Strobos, G. 92 Greben, H.. Judels, C. and Dreyer. Comparison of the combined limestonellime and combined limestonelbiological sulphate removal process for treatment of acid mine water, Proceedings 8m Hard Rock Mining Conference: Issues Shaping the Industry, 7 - 9 May 2002, City Colorado. USA. Paper 6: Adlem, C.J.L.. Geldenhuys, A.J., Maree, J.P. and Strobos, 95 G.J. Examining the implementation of limestone neutralisation technology in the mining and industrial sector to neutralise acid and reduce sulphate pollution, Proceedings 9 Annual Industrial Water Management and Treatment Symposium, 15 - I6 May 2002, Johannesburg, South Africa. Paper 7: Maree, J.P., Hlabela. P., Geldenhuys, A.J., Nengovhela, R., 11 1 Mbhele, N. and Nevhulaudzi, T. Treatment of mine water for sulphate and metal removal using barium sulphide, Proceedings Waste Management, Emissions & Recycling in the Metallurgical .S Chemical Process Industries, 18 - 19 March 2004, Johannesburg, South Africa. Paper 8: Maree, J.P., Netshidaulu, I., Strobos, G., Nengovhela, R. and 127 Geldenhuys. A.J. Integrated process for biological sulphate removal and sulphur recovery, Proceedings Water Institute of Southern Africa (WISA) Biennial Conference & Exhibition, 2 - 6 May 2004, Cape town, South Africa. Concluding Discussion Acknowledgements Bibliography Appendix A List of Additional Papers and Posters Appendix B List of Confirmations CHAPTER 1 : LITERATURE OVERVIEW The generation of acid mine drainage (AMD), from working or abandoned mines, and its discharge into the surrounding environment is a cause of serious environmental pollution. At present, AMD is becoming a problem as increasing numbers of mines are facing closure, which will finally lead to the shutting down of whole coalfields. Pumps, which currently keep these mines dry, are removed and consequently the groundwater returns to its pre-mining levels leading to AMD. The treatment, or the prevention, of such pollution by current means is costly and the legal requirement to treat it is also likely to become a more pressing requirement. The problem with treatment is that the available technologies for dealing in an environmentally friendly way with AMD. Until recently, the standard practice was to treat AMD with lime. This produces a ferruginous (iron bearing) waste material which is often too variable in quality to represent a useful source of ochre. Such waste has to be disposed of in a tailings dam if possible or to landfill. The many technologies proposed for treatment of mine drainage are usually expensive and complex. Liming is also not sustainable because of the requirement for lime and the need for disposal space. This research looked into the feasibility of replacing lime treatment of AMD with other technologies which, not only offer a more sustainable solution, but also cost effective answers to water issues that may become major problems. Without this industry accepting responsibility and realising the extent of the pollution by mine water, detrimental effects on the environment and its water resources will result, especially in a semi-arid country like South Africa. In this section, an in-depth overview is given of the available literature on specifically coal mining. The history of coal and the geochemistry of coal mine drainage in Southern Africa will be discssed. The mining of heavy minerals in South Africa also adds to the list of many polluters of ground water resources on which one paper will focus. A neighbouring country of South Africa, Botswana, also very arid in climate, produces acid water, resulting from a nickel-copper-cobalt mine, operating as BCL Limited at Selebi-Phikwe. In general, focus will be on coal mining, which is one of South Africa's biggest role players in the mining industry. Overall, the focus will be on the origin and background of these dint mining operations: how they contribute to water pollution and how their adverse impact on the environment can be mitigated, assisted and guided from a legislation point of view. The responsibility for treating AMD is a crucial issue. As it was not foreseen, when the pumping of water in mines began, that there would be a problem of AMD, there were no funds set aside to meet the considerable financial implications. When mining started, there was also lile or no concern about potential environmental problems that might result from such industrial activity. AMD events are more pernicious than, for example incidents involving nitrate or oil, because the pollutant is not broken down in the environment. Whilst nitrates may be utilised by aquatic organisms and oil may eventually be broken down to carbon dioxide and water, metal pollutants will remain in the environment in one form or another. Metals may be concentrated or dispersed in the environment and without treatment, there will be no control over where these concentrated or dispersed metals will deposit. In the meantime there will be an extended period in which the local environment will suffer the effects of the pollution. However, AMD is not a new problem. The mining industry in South Africa is therefore under pressure to find solutions for the seriously degradation of the aquatic habitat and quality of water supplies for which they are responsible. Academic and industrial partnerships have investigated a range of mine water treatment technologies to assist in water treatment and remediation. There is currently no consensus on what is the ideal solution, and it may be that each AMD case will require its own treatment solution. I MINING OF HEAVY MINERALS Although not as prominent and well-known as the mining of coal, gold and diamonds in South Africa, Ticor Limited (TOR) is involved in the mining, processing and smelting of mineral sands. TOR in South Africa is a heavy minerals sand mining and processing operation, near Richards Bay (http:llau.bi.yahoo.comlpMor.ax.html). The operation is based on three alluvial, high grade, placer, mineral sand deposits, namely Hillendale, Fairbreeze and Gravelotte, with reserves of 16 million tonnes of valuable heavy minerals. These deposits yield ilmenite, zircon, rutile and leucoxene which are primarily sources of titanium dioxide feedstock for the paint, paper and plastic's industries. The company produces about 200 000 tonnes of chloride grade and about 50 000 tonnes of sulphateable feedstock per year and will manage about 11% of the world's titanium feedstock supply. Approximately 250 000 tonnes of titania slag and 140 000 tonnes of pig iron are also produced (http:l/au.biu.yahoo.comlp/t/tor.ax.Mml). In the mining and processing operations where these minerals, high in pyrites and low in calciteldolomite are processed, acid is generated which requires neutralisation in the processing plant where acid is leached into the wash water. This water needs to be treated to a quality suitable for re-use in the metallurgical process, or to a higher quality rendering it suitable for discharge into the Empangeni sewage system, 200 km north of Durban, Kwazulu-Natal. For re-use the water should be neutral and under-saturated with respect to gypsum. For discharge into the sewage system the sulphate concentration should be less than 500 mgl! (as Sod. Acid mine water is generally neutralised with lime. Disadvantages associated with lime are the costs and maintenance of the slaking equipment as well as hazards, associated with handling. The cost of powdered limestone (CaC03) in South Africa, a by-product from the paper industry, is 60% lower than conventional lime. Lime has been successfully replaced by limestone as neutralising agent while no compromises have been made on the quality of the discharge water or that of the final products. The legislative requirements governing effluents resulting from industries are set out in the Water Act (Act 36 of 1998) which is discussed later in this introductory review. 1.2 NICKELCOPPERCOBALT MINING Botswana's rapid economic growth, which began in the 1970's, continued into 1997. Much of the growth is attributed to the country's successful program of mineral exploration and development. The mining sector, mostly on the strength of diamonds, accounted for about 33% of the gross domestic product. Nickel and copper also played significant roles in the national economy. BCL Limited is operating nickel- coppercobalt mines and a smelter at Selebi-Phikwe, about 350km northeast of the capital, Gabarone, and processes 45Wday of ore (Van Tonder eta/., 2000) At BCL, Mine waste discard, that contains pyrites, is produced during mining operations and poses problems of acid leachate. This leachate contains high concentrations of acid, sulphate and metals. The operations consist of underground mining, concentration of the copper and nickel components of the ore by means of flotation, and smelting of the concentrate to produce copper and nickel. The main flows of water into the underground workings include cooling water (with high NaCl content from the ice plant), groundwater (fissure water) and water recycled with the coarse waste backfill. These streams are currently mixed and returned to surface where the combined stream of 350 m3/h water is neutralised (Van Tonder et a/., 2000). Central to the water network is the Mill Return Water Sump (MRWS). The used-water streams are recycled to the MRWS, from where the concentrator circuit is supplied with water. Lime is used to adjust the pH of the return water to 8.5 in the MRWS. This water is used in the concentrator circuit as transport medium and to facilitate separation. The pH of the water is the main quality consideration for the concentrator as high salinity levels do not pose a problem. In the copper-nickel concentration processing plant, solid waste material containing 5% pyrite is produced. The coarse fraction of the solid waste material is discarded underground as backtill, while the fine waste is discharged onto a tailings waste dump. These wastes give rise to acidic leachate due to pyrite oxidation. Lime is used to neutralise 350 m3/h of underground mine water (with an acidity of 235 mglt as CaC03) and 60 m3/h of tailings dump seepage (with an acidity of 5000 mg/t as CaC03). Excess water is used for the cooling and granulation circuit in the smelter. The smelter intake water chloride concentration should be limited to 5 mg/t to prevent pitting corrosion in the smelter cooling jackets. For this purpose surface water is piped from a local dam (Van Tonder et a/., 2000). BCL currently experiences the following water-related problems: Neutralised water is discharged into a public stream at a rate of 300 m3/h. The effluent quality does not meet the permitted level of 500 mg/t sulphate. . The neutralisation cost is high due to the use of imported lime. Excessive acid seepage has resulted in deterioration of the land area adjacent to the tailings dump. The water intake of 300 - 400 m3/h is expensive. A modelling exercise was carried out during 1999 to audit and simulate the water network of BCL with the aim to identi the optimum water management strategy (Van Tonder, eta/, 2000). It was found that discard leachate could be neutralised with limestone to minimise chemical cost. It should be treated, before being mixed with less polluted streams, to achieve maximum sulphate removal through gypsum crystallisation and precipitation. The latter will result in reduced gypsum scaling in the metallurgical plant. 1.3 COAL MINING 1.3.1 Introduction to coal Coal has been described and classified by many scientists. Grainger et al. (1981) delineated coal as an organic sedimentary rock, formed by the action of temperature and pressure on plant debris. Coal is a complex mixture of organic matter consisting of mainly carbon, hydrogen and oxygen, together with some small amounts of nitrogen, sulphur and trace elements. Sanders (1996) referred to coal as a generic term which belongs to a family of solid fossil fuels with a wide range of physical and chemical compositions. Coal is actually a heterogeneous rock composed of diierent kinds of organic matter which vary in their proportions in different coals. He also noted that no two coals are absolutely identical in nature, composition or origin and proposed that 'coal is a compact stratified mass of metamorphosed plants which have, in part, suffered arrested decay to varying degrees of completeness". According to Grainger et al. (1981). the rank of coal can be described as the degree of metamorphism to which coal has been subjected after burial. It then results in the transformation of the original peat swamp through the progressive stages of brown coal (lignite), subbiiuminous, and bituminous coals to anthracite and the meta- anthracites. The rank is then defined as the level to which coal has reached in this coalification series (Falcon et. el., 1981). Rank also refers to the degree of maturity or metamorphism or coalification achieved by a coal through the process of time, temperature, and pressure as a result of depth of burial or proximity to heat following peat accumulation. The progressive change from peat into coal passes through a number of stages, namely: Peat + Lignite -+ Biuminous + Semi-anthracite + Anthracite + Graphite (Falcon, 1988). 1.3.2 Coal formation and composition of coal macerals Neavel (1981) stated that "Macerals" are organic substances derived from plant tissues and cell contents that were variably subjected to decay, incorporated into sedimentary strata, and then altered physically and chemically by natural processes. Each of the materials recognised as belonging to a specific maceral class has physical and chemical properties that depend upon its composition in the peat swamp and the effects of subsequent metamorphic alteration. For applications in coal utilisation it is often sufficient to group the macerals together as vitrinite, exinite (or liptinite) and inertinite (Grainger et a/.. 1981). In South African coals, a fourth maceral group has been identified, i.e. semi-fusinite. By means of optical examination, the different macerals in coal can be distinguished. Macerals may be differentiated from one another on the basis of morphology, relief, size, shape, colour. reflectance and, origin in some cases (Falcon eta/., 1986). 1.3.3 Mineral matter Table 2 lists the most abundant and common mineral groups found in South African coals, namely: clays, carbonates, sulphides, quartz and glauconite (Falcon & Snyman, 1986). Table 2 Distribution of the common minerals in South African coals (Falcon 8 Snyman, 1986) pup of I Minerals I minerals I I Clay minerals I Kaolinite Montmorillonite I I Dolomite I 1 Aragonite Siderite Sulphides Marcasite I Silicates I Quartz Mineral particles are evident in coal sections and form a major portion of the ash (Grainger et a/., 1981). Falcon et a/. (1986) stated that the forms in which minerals occur in coal fall into two major categories: one of which includes the intrinsic inorganic matter which was present in the original living plant tissue; a second, which includes the extrinsic or induced forms of mineral matter. Intrinsic inorganic matter is trapped in coal in the form of submicroscopic mineral grains and as organo-metallic complexes. The extrinsic mineral may be primary or syngenetic, and arises from the accumulation by means of wind and water or precipitation in situ (Falcon et a/., 1 986). According to Stach et al. (1982), the inorganic matter in coal can be classified into three groups: lnorganic matter from the original plants; lnorganic - organic complexes and minerals which formed during the first stage of the coalification process; or which were introduced by water or wind into the coal deposits as they were forming; Minerals deposited during the second phase of the coalification of the coal, by ascending or descending solutions in cracks. Coal has a significant inorganic material content, varying from less than 3-4°h(m/m) to more than 40%(m/m). There is general agreement in the literature that clays, sulphides, carbonates and quartz are the most common minerals in coals (Alpern et a/. , 1 984). 1.3.4 Coal mining and the environment By its very nature and scale, mining makes a marked and visual impact on the environment. Mining is, moreover, implicated as a significant contributor to water pollution, the prime reason being that most of South Africa's geological formations, which are mined, contain pyrites which oxidise to form sulphuric acid when exposed to air and water. The scarcity of water in South Africa is exacerbated by pollution of the surface- and ground- water resources. Typical pollutants of the aquatic environment include industrial effluents and acid mine drainage. Mine water in the Upper Oliants River Catchment in Mpumalanga (upstream of Loskop Dam) is at times discharged into local streams, resulting in local acidification and regional salination of surface water resources. Pollution of surface water can be prevented by collecting and treating mine water to a quality where it can be re-used without restriction (Cleanwater 2020 Initiative). Although mine water in the Oliants River Catchment currently amounts to only 4,6% of the total water usage, it contributes 78,4% of the sulphate load. Mine water in the catchments of the Wabank Dam and Middelburg Dam is rich in calcium, magnesium, sulphate and acid pH. This is due to oxidation of pyrites to sulphuric acid in the mined coal and coal waste, followed by neutralisation with dolomite that is also present in the mined coal. 1.3.5 Origin of acid mine drainage Coal mine drainage, also known as acid mine drainage (AMD) or acid rock drainage (ARD), is a natural consequence of mining adivlty where the excavation of mineral deposits (metal bearing or coal), below the natural ground level, exposes sulphur containing compounds to oxygen and water. Recently, it has become possible to mine ores substantially below the groundwater level which can cause surface waters to run over exposed ore seams and elicit similar chemical mechanisms and acid formation (Maree eta/. (1997)). Many have given definitions of what they understand under the term "AMD. AMD can be described as drainage resulting from, or caused by, surface mining, deep mining or coal refuse piles that are typically highly acidic with elevated levels of dissolved metals. The formation of AMD is primarily a function of the geology, hydrology and mining technology employed for a specific mine site. AMD is formed by a series of complex geochemical and microbial reactions that occur when water comes into contact with pyrite, amongst other iron disulphiie minerals, in coals, refuse or the overburden of a mining operation. The resulting water is usually high in acidity and dissolved metals. The metals remain in solution until the pH rises to a level where precipitation occurs. When mining began, it was only possible to mine those ores that were at or above ground level. As technology and mining engineering improved, it became possible to excavate horizontal shafts, adiis, leading away from the mine to drain groundwater into local low lying river valleys and provide access to lower levels. It later became possible to pump water from deep mines, atifcially lowering the groundwater level in the vicinity. When pumping ceases, groundwater floods the mine and will eventually approach the original groundwater level and may cause environmental problems. As the water rises it will eventually reach the levels where adits were built to drain the mine water into river valleys (Kuyucak, 1998). Oxidation reactions, often biologically mediated, take place which affect the sulphur compounds that often occur in coal seams. Whilst a mine remains dry these sulphur compounds normally generate sulphates in solid form. The metals occurring in these minerals are often incorporated into these salts. When water flows through the mine, they dissolve and this acidic, metal-containing mixture comprises the initial AMD discharge. AMD is a problem because the vast majority of natural lie only viable at a neutral or near neutral pH, i.e. 7. The drainage acidifies local watercourses and either kills or stunts the growth of river biota. Effects are even more pronounced on vertebrate lie such as fish than on the plant and unicellular life (Maree eta/., 1997). Metals contained in drainage are also of concern, particularly iron. Its presence in the water is a problem, due more to its physical properties than its toxic effects. Iron may be found in two forms, ferrous and ferric. When AMD is generated it will generally be in the ferrous form, but is changed in the presence of oxygen to ferric iron (Fe-) when it forms semi-solid particles, which are bright orange. Very small concentrations in water are capable of generating large volumes of ferric precipitates which cover the surfaces of land and streams close to the point of discharge. This effectively smothers the environment, prevents lie from flourishing, and coats the gills of vertebrate lieforms such as fish and causes fatalities. The metal is, however, not inherently toxic. Not all mine drainage is acidic. Some are close to neutral but the presence of ferric iron leads to the possibility of precipitation and causes environmental problems as outlined earlier. 1.3.6 Geochemistry of acid mine drainage The geochemistry of AMD has been the subject of numerous investigations with Rose 8 Cravotta (1998) being a general reference on the subject. The composition of coal mine drainage ranges widely, from acidic to alkaline and with typical elevated concentrations of sulphate (SO,), iron (Fe), manganese (Mn) and aluminium (Al) as well as common elements such as calcium, sodium, potassium and magnesium. With fewer intermediate or extreme values, the pH most commonly ranges either between 3-4.5 or 6-7. A key parameter is the acidity, which can be commonly described as the amount of base required to neutralise the solution. In coal mine drainage, major contributors to acidity are ferrous and ferric Fe, Al and Mn as well as free hydrogen ions. When pyrite is oxidised, it releases dissolved ~e", SO4'- and H' and is known as AMD. This process is followed by the further oxidation of Fez' to Fe3+ and the precipitation of the iron as a hydroxide ("yellow boy") or similar substances, producing more H'. Neutral mine drainage with high SO,", and possibly elevated Fe and Mn, forms with the neutralisation of acidic solutions by limestone or similar materials. If appreciable Fe or Mn is present, these neutral solutions can become acid on oxidation and result in the precipitation of the Fe and Mn. The rate and extent of AMD formation in surface coal mines are controlled by a number of factors. An increase in acidity of the drainage tends to be the result of more abundant pyrite in the overburden as well as decreasing particle size of the pyrite. Furthermore, acid formation is accelerated by iron-oxidising bacteria and low pH values. The presence of limestone or another neutraliser has an adverse effect on the rate of acid formation. The limiting factor in acid formation is the access to air which contains the oxygen needed for pyrite oxidation. Both access to air and pyrite surface exposure are promoted by crushing of the pyrite-bearing rock. The oxygen can gain access either by molecular diision through the air-filled pore space in the spoil, or by air flow which is driven through the pore space by temperature or pressure gradients. The complexity of these interactions and other factors results in the forecast and remediation of AMD to be site specific. Serious degradation of the aquatic habitat and the quality of water supplies, owing to the toxicity, corrosion, incrustation and other effects from dissolved constituents, can be ascribed to coal mine drainage, which can be either acidic or alkaline. AMD is a result of interactions of certain suKde minerals with oxygen, water and bacteria, as illustrated in Figure 3. Steps (a) to (d) correspond with reactions 2-5, respectively. Steps (d') and (d") represent the formation of iron-sulphate minerals (sources of acidity, fenic ions and sulphate). According to Davis (1981) and Hawkins (1984), the iron disulphide minerals pyrite (FeS2) and, less commonly, marcasite (FeS*), are the principle sulphide bearing minerals in bituminous coal. Upon oxidation, acidic solutions can also be generated from pyrrhotiie (FeS), arsenopyrite (FeAsS), chalcopyrite (CuFeS2) and other sulphide minerals containing Fe, Cu, As, Sb, Bi. Se and Mo. These minerals are, however, uncommon in coal beds. The stoichiometric reaction that best describes the oxidation of pyrite is commonly given as: FeS2 + 3.75 O2 + 3.5 H20 = Fe(OH)3 + 2 S02' + 4 HI + heat 111 C + 02 (b) (c) + FeS2 "F~"F~"'-so, salts" ct FeN -+ "Fell' oxyhydmxides" Overall reaction: FeS2 + 3.75 O2 + 3.5 H20 + Fe(OH)3 + 2 SO:- + 4 HI + heat [I] Steps: (a) FeS2 + 3.5 O2 + H@ + Fez' + 2 SO:-+ 2 H+ (b) Fez' + 0.25 O2 + HI + Fe3+ + 0.5 Hz0 (c) FeSZ 14 Fe3' 8 H20 + 15 Fez+ 2 SO4" + 16 H' (d) Fe3' + 3 H20 + Fe(OH)3 (s) + 3 H' (d') 2 Fe3' + Fez' + 4 SO:-+ 14 H20 + (d") 3 Fe* + K' + 2 SO:- + 6 HZO + KF~~~"(SO~)~(OH)~ + 6 HI Figure 3 A model for the oxidation of pynte. (Modified from Stumm 8 Morgan, 1981 by Rose & Cravotta. 1998) The oxidation of sulphur and iron (Figure 3, reactions a and b respectively) by gaseous or dissolved 02, can be mediated by various species of sulphur- and iron- oxidising bacteria (Thiobacillus spp.). According to others (Temple & Delchamps, 1953; Kleinman et a/. 1981; Ehrlich, 1990), these bacteria produce enzymes which catalyse the oxidation reactions and use the energy released to transform inorganic carbon into cellular matter. In reaction (c), (Figure 3) dissolved ferric iron (Fe3') from reaction (b) is the oxidising agent for pyrite and finally, part of the Fe precipitates as Fe(OH)3 (Figure 3, reaction d). Intuitively. pH best indicates the severity of AMD. However, acidity or total alkalinity of a solution probably outcompete this inclination. Acidity is the basic requirement of a solution in order to be neutralised and includes the requirement to neutralise acid generated by Fe-, Al- and Mn-hydrolysis. This is illustrated by reactions (b) and (d) in Figure 3 and the following two reactions: According to work done by Ott (1986), ~e~', ~e", MnZ', A?+ and H* are the main components of acidity in mine drainage from coal mines. Payne 8 Yates (1970) found that other species that precipitate as hydroxides or oxides, including Mgz', HZCO3 and HzS, can also contribute to acidity. Many methods express acidity as milligrams of CaC03 per litre of solution, based on the following relationship: From reaction [4], 2 moles (2.09) of H' are neutralised by I mole (100.lg) of CaC03. 1.4 CONVENTIONAL TREATMENT OPTIONS FOR ACID MINE DRAINAGE AND ACIDIC SOLUTIONS The use of lime to neutralise AMD and precipitate metals (active treatment system) is considered, in the context of this thesis, as the standard against which other methods are compared as it has been the conventional treatment choice for many years. Lime treatment is simple and robust, and the benefits and drawbacks are well known owing to long established usage. It does, however, present several environmental problems. The material produced after treatment with the lime is metal rich and usually contains a significant amount of water. The presence of metals means that it will often require special waste disposal fauliies that add to the costs of disposal. The water content increases the volume and mass of the waste which means that money is wastefully being spent to dispose of water, both in transport and landfill fees. The general methods to reduce the water content are often labour or energy intensive that also increases costs and, moreover, are often unable to keep pace with the flow of material from the treatment system. Alternatives must provide some advantage over the lime treatment either in the use of materials, the disposal of waste, or the production of usable materials. These questions are addressed in the research described in this thesis. In conventional lime neutralisation processes, acid is neutralised and metals and sulphate are precipitated in the form of metal hydroxides and gypsum (CaS04), respectively, as shown in Equation [5]. The mixture of precipitates is referred to as "sludge". Air is frequently used to oxidise the ferrous to ferric iron during precipitation to obtain sludges that are more chemically stable (MEND Report, 1994, Kuyucak, 1998). The sludge produced, is allowed to settle in ponds or darifiers/thickeners. The settled sludge is disposed of in specifically designed ponds for storage in perpetuity. Depending on site factors, lime neutralisation facilities range from the simple addition of lime to the tailings pipelines to facilities, and sludge dewatering equipment. The water strength (solid concentration) and the sophistication of the treatment process have been found by many to affect the sludge solids content. As a result, sludge densities may vary from 1 to 30% solids. The process parameters are set to obtain denser, less voluminous sludge. Major process parameters affecting sludge characteristics include: the rate of neutralisation; rate of oxidation; ~e'' to Fe" ratio; concentration of ions; ageing; recycle of settled sludge; temperature; and crystal formation (Kuyucak 8 Sheremata, 1995, Zinck 8 Griffih, 2000, Kuyucak eta/., 1999). 1.5 ALTERNATIVE NEUTRALISERS Under controlled conditions, limestone, in contrast to lime, can remove acidity and precipitate metals (e.g. Al, Cu and ~e~') producing higher density sludges. C02 gas is released as CaC03 (s) dissociates in AMD as illustrated in equations 6 and 7. C02 forms carbonate ions which act as a buffer system and sets an upper limit on the pH (maximum pH 6.5) and also affects the rate and amount of lime consumption. The precipitates may settle very slowly because of their small partide size. Removal of a broad range of metals and ferrous iron cannot be achieved since they require higher pH levels than 6.5. A combined-limestonellime process is suggested for removal of a wide range of metal ions. Magnesium hydroxide (Mg(0H)d usage can result in a lower volume of metal hydroxide sludge when it is properly applied due to the higher solubility of MgSO4 than CaS04. MG(OH)~ can also remove metals through surface adsorption. However, MG(OH)~ prevents the pH from exceeding 9. Depending on the pH requirements, it can be used in conjunction with lime. Limestone and other materials that produce alkalinity can affect the generation of AMD in two ways. If water contacting pyritic materials is alkaline, or if alkaline conditions can be maintained in the pyritic material, the acidgenerating reactions may be inhibited so that little or no AMD forms. Alternatively, once AMD has formed, its interaction with alkaline materials may neutralise the acidity and promote the removal of Fe, Al and other metals. Hence, water with high SO," and low Fe may be indicative of earlier AMD generation. The carbonate mineral, limestone (CaC03), is the main mineral providing alkalinity for the process to be described. Carbonate minerals may occur as layers of limestone in the overburden above coal, as cement in sandstone or shale, or as small veins cutting the rock. The initial reaction with an acid solution is: If a gas phase is present, the H2C03 may partly decompose and exsolve into the gas phase, i.e.: Upon further neutralisation of AMD with carbonate to pH values of greater than 6.3, the product is bicarbonate (HCO~): When it is necessary to lower the pH to between 6.5 - 8.5 in the final effluent, following treatment to a much higher pH, the pH is adjusted to the desired level with coz. 1.6 FURTHER TREATMENT Treatment technologies are commonly categorised as either passive or active. The main purpose of both types is to lower total acidity, raise pH and lower toxic metal and sulphate concentrations. Passive treatment approaches are economically attractive, but have some significant limitations. They are best suited to waters with low acidity (~800 mgll), low flow rates (<50 tlsec) and, therefore, low acid loads, where the key outcome is near neutral pH. Passive systems cannot handle acid loads in excess of 100-150 kg of CaC4 per day. When specific metal removal targets need to be achieved, as opposed to simple neutralisation, most passive treatment technologies are unsuitable. Although not limited by tight operational parameters, as in the case of passive systems, the unlimited chemical flexibility of active systems comes at a price, which proves to be one of the biggest challenges in the field of water treatment. Active treatment systems can be engineered to accommodate essentially any pH, flow rate and daily acid load. Economic considerations (i.e. capital and ongoing operational cost) play a big role in determining the viability of active treatment systems. A broad range of active treatment approaches is available for dealing with AMD. Physical, chemical and biological approaches include one or more of the following: pH control or precipitation Electrochemical oxidation Biological mediationlredox (sulphate reduction) Coagulation/Flocculation Crystallisation pH controVprecipitation with inorganic alkaline amendments is the most common and cost effective form of general purpose AMD treatment. A large variety of natural, manufactured or by-product alkaline reagents, are available, with their use generally dictated by availability and cost. Alkaline reagents treat AMD by increasing the pH and promoting the precipitation of heavy metals, generally as hydroxide complexes. The successful implementation and sustainability of 'pH control' active treatment systems requires the selection of a reagent appropriate for the treatment task and an efficient mixing and dispensing mechanism. Conventional alkaline reagents that are readily available in South Africa to treat AMD include hydrated lime and the carbonate mineral, limestone out of a list of reagents. Although the capital and operating costs of such systems are relatively high, they employ well established technology and are highly reliable. A key limitation of fixed plant systems is the need to deliver affected water, regardless of the number of discrete AMD sources. Mixing and dosing systems employing the CSIR's technology (Limestone Handling and Dosing System) provides the reagent dispensing capacity of a large fixed plant system. The Integrated LimestoneRime Treatment Technology was developed in response to the problems that passive treatment systems face in using limestone efficiently. Together with this treatment technology, the Limestone Handling and Dosing System has been developed to replace the conventional storage of the alkali in a hopper and automatic feeding with a screw-feeder. The Limestone Handling and Dosing System, which is the first technology of its kind to be built on full scale, can be designed to accommodate any load capacity. It consists of the following items: A concrete slab with a slope of 7' onto which the CaC03 powder is dumped and stored. The CaC03 powder is slurried with a water jet and collected in a slurry tank through gravity flow A slurry tank with stirrer which acts as a mixing chamber for the acid water and CaC03 A ball valve in the sluny tank to maintain the water at a specific level in the tank by dosing tap or clarified water. A CaC03- recycle slurry pump that withdraws some of the slurried CaC03 of higher density from the slurry tank or clear water through a water jet, passing through a density meter onto the CaC03 dump to keep slurried the CaC03 concentration constant. The slurried CaC03 is returned by gravity via the sloped concrete slab back to the slurry tank. The slurried CaC03 concentration is controlled by the density meter which activateslstops the recycle pump at preset lowlhigh values, respectively. A transfer pump to feed slurried CaCO, from the limestone make-up tank. The system provides the benefit of using environmentally benign, very low cost limestone aggregate that is locally available as a waste product from the paper industry. As the carbonate dissolves and neutralises the AMD, C02 builds up in the reactor and can be recirculated. Key benefits of systems in which lime is partially replaced by limestone, include the generation of high levels of alkalinity, partial sulphate removal and efficient use of low cost limestone. Electrochemical oxidation uses electrical techniques to oxidise Fez' to ~e~' in AMD while generating hydrogen (Hz) electrolytically to be utilised as energy later on in the treatment of AMD by means of the Biological Sulphate Removal Process. 6iobgicai mediationlredox (sulphate reduction) - Microbial Reactor Systems (MRS) are fully engineered and process controlled systems for harnessing chemical and biological processes to further remove sulphates in AMD to below 200 mgle. This process follows directly after the AMD has been fully neutralised and sulphates removed to below the saturation level of gypsum, i.e. 1 200 mgle, by means of the Integrated LimestonelLime Treatment Technology. These systems consist of a sulphate reducing bioreactor and H2S scrubbing process for sulphur recovery. The successful performance of MRS is reliant on the continued growth of sulphate reducing bacteria (SRB), which require temperatures between 25 - 35'~. CoagulationlFlocculation - Following neutralisation and partial sulphate removal with limestone and lime, fine particles (precipitates) in suspension need to be aggregated to improve soliiquid separation or sedimentation in clarifiers. Coagulation is a specific type of aggregation, which leads to the formation of compact aggregates, called flocs. The addition of coagulants, such as inorganic A* or ~e~' salts or organic 'polymers', help to electrically neutralise or destabilise electronegative colloids and bridge the neutral particles. Important parameters are the type of polymer and external stirring. Crystallisation - the lntegrated LimestoneILime Treatment Technology and Bas Treatment Technology offer new methods for lowering soluble sulphate concentrations in water that has already been subjected to lime treatment. It is possible to lower sulphate concentrations to below 200 mgle using these approaches. 1.7 NOVEL INTEGRATED TECHNOLOGY FOR TREATMENT OF ACID MINE DRAINAGE AND EFFLUENTS Both the environmental aspects and the economics of metal and coal mining operations worldwide are being affected by AMD. The latter can have significant impacts on the economics of a mining operation. This is due to the corrosive effects of acid water on the mine infrastructure, the limitations it places on water reuse and discharge and the expense incurred implementing effective closure options. While AMD minimisation and control must remain the focus of mine-site water management strategies, when acid generation is unavoidable, appropriate passive or active treatment technologies need to be implemented. As mentioned earlier, passive treatment systems are economically attractive but have some significant limitations. They are best suited to treating low flow rates and therefore low acidity. Newly developed technology resulted in active treatment systems that can accommodate any flow rate, pH and acid load and are not limited by operational parameters. Because every mine site is unique as are its water issues, these newly developed systems can be designed and engineered to cost effectively deliver the required water quality of such a site. The cost effectiveness is achieved by designing the system in such a way that the treated water can either be reused in the plant. thus decreasing the amount of water purchased, or the water can be utilised for irrigation purposes or even discharged into a water course. An integrated process (active treatment system), consisting of various treatment stages has been developed by CSIR: Environmentek in an effort to solve the current AMD situation and other acid water related issues, especially in the mining industry. Depending on the specific requirements, i.e. to what level of quality the mine needs to treat its water and the water quality and flow-rate that require treatment, this process can be adapted by omitting some of the stages. It offers huge cost benefk compared with existing processes, for example, over the conventional way of neutralising acid water with lime, sodium hydroxide or sodium carbonate. These chemicals have the disadvantage that they require accurate dosing to prevent under- or overdosage, which result in pH fluctuations. When pumped through pipelines of a mine water system, corrosion (low pH) or scale formation (high pH) will result which can adversely affect the whole system, necessitating shutdown for maintenance repair. The use of limestone as neutralising agent has the following bendts: Direct chemical cost savings, utilising limestone, a waste product from local paper industry No pHcontrol required, as limestone dissolution occurs mainly at below pH 7 Limestone is easy to handle and store as it contains 15% moisture which eliminates dust problems 0 Limestone is non-hazardous and environmentally friendly The completely integrated process has been recently developed and consists of the following stages, illustrated in Figure 4: Heating unit for the production of CaO (quicklime) and C02 from burned coal and CaC03 (limestone) precipitation. Limestone neutralisation and partial sulphate removal to 1 900 mgle Ca(OH)2 (hydrated lime) stage, where CaO is contacted with acid water to produce ca(0H)~. Lime treatment, to partially remove the sulphates as CaSO, (gypsum) to below 1 200 mgle and full removal of magnesium and other metals. pH adjustment stage where the C02 from the heating unit is applied to reduce the pH to 8.6 while CaC03 precipitates. Two options: Removal of sulphate to below 200 mgle as barium sulphate by means of barium sulphide treatment, or biological sulphate removal. Regeneration of barium sulphide for reuse by heating the barium sulphate produced from the preceding stage. Stripping of HS, either from the barium sulphide process or the biological sulphate removal process, to be contacted with Fe(lll) rich water for elemental sulphur production. Figure 4 Process flow diagram of the completely integrated process for neutralisation and removal of acidity, sulphate and metals from AMD In the completely integrated process, limestone is economically utilised to completely neutralise AMD and acid process water. The sulphate concentrations in these waters are lowered to 1 900 mg/e. Wth the addition of lime at this stage, the sulphate concentration is further reduced to below the saturation level of gypsum, i.e. 1 200 mgle. Metals are now also being fully removed from the water. C02, generated during the limestone roasting stage, is then used to adjust the pH to 8.5 and to achieve CaC03 crystallisation, which can be recycled to the limestone roasting stage. Either the biological sulphate removal stage or the Bas treatment stage can now be applied to achieve removal of sulphates to below 200 mg/e, i.e. the recommended sulphate level for discharge water. H,S gas, generated during both of these stages as a by-product, is stripped and contacted with Fe(lll) rich water to produce elemental sulphur which is a valuable product. The Fe(ll1) rich water results from the Fe(ll) oxidation stage prior to limestone neutralisation which is inevitable as acid will start forming without oxidation of Fe(l1). The Biological Sulphate Removal Process is an anaerobic treatment in which a reducing environment is produced and the proliferation of sulphate reducing bacteria (SRB) is encouraged. (Although this thesis contains a paper on biological treatment it only focuses on chemical treatment technologies. See Appendix A for Paper 9 on. "The Sustainability of Biologically Trwted NickeUCopper Mine Efffuent Suitable for lmmgation") These bacteria use sulphate in their metabolism, producing hydrogen sulphide which combines with metals such as copper, cadmium and zinc to form insoluble sulphides. Another product of SRB metabolism is the production of alkalinity, thereby raising the pH of the mine water. Naturally, these bacteria utilise sugar and ethanol as carbon source which is economically unfeasible. Electrolytically generated hydrogen has been successfully implemented as substiie for sugar and ethanol as energy source. Hydrogen is electrolytically generated on- site by means of stainless steel (type 316) electrodes in KOH (3%) as the electrolyte. An asbestos sheet of 3mm thickness serves as diaphragm between anodes and cathodes in the electrolytic cell as both hydrogen and oxygen are generated. In the Bas treatment stage, sulphates are removed from the water as BaSO, which is converted back to Bas for re-use. This conversion is achieved by heating the BaSO. to a specific temperature. The final, unsolicited waste product after either of these two stages is hydrogen sulphide gas (Hs), which is stripped and contacted with a small part of the Fe(lll) rich water, resulting from the Fe(ll) oxidation stage, that yields pure, elemental sulphur. 1.8 LEAGAL REQUIREMENTS The treatment processes that are widely applied mainly to acid water, have one goal in mind: to comply with legislative requirements before discharge to the receiving water body. The legislative requirements imposed on industrial effluent derive primarily from the National Water Act (Act 36 of 1098), as laid down by the Department of Water Affairs 8 Forestry in consultation with the SABS and as published in the Government Gazette. The Act states that the ultimate aims of water resource management are to achieve the sustainable use of water for the benefd of all users and to recognise that the protection of the quality of water is necessary to ensure sustainability of the nation's water resources in the interests of all water users. In this Act it is required that any person who uses water for industrial purposes shall purify or otherwise treat such water in accordance with requirements prescribed in the Government Gazette. Before a pennit for discharge of water is granted, all efforts should be made to ensure effective utilisation of the water through recycling or alternative applications. One specific alternative would be to pass the water on to a responsible local authority who then can treat it for use. Certain criteria are prescribed to be met before discharge water is accepted by such an authority. In Chapter 4 of the National Water Act (Government Gazette (Parliament of the republic of South Africa), iW8), the use and discharge of water are dealt with in detail in order to down the basis for regulation. Water use is defined broadly, and includes pumping and storing water, activities which reduce stream flow, waste discharges and disposals, controlled activities which impact detrimentally on a water resource, removing underground water for certain purposes, and recreational use of water. In general, a water use must be licensed under a general authorisation and before any permit for discharge is granted, all efforts should be made to ensure optimum use of water through recycling or alternative processes. The Department of Water Affairs 8 Forestry uses a so-called Waste Load Allocation to lay down allowable discharge parameters from some major industries. In theory, a Waste Load Allocation is the amount of tolerable discharge to a water body whilst monitoring the water quality at a usable level for the designated users. CHAPTER 2: PAPERS 1 - 8 2.1 INTRODUCTION TO PAPERS This thesis describes research and development work conducted on laboratory and pilot scale plants, situated at numerous mines across South Africa and one in Botswana. The results have been utilised in the design of full scale plants at some of these sites of which a few have already been constructed and commissioned. Others are in the design process and being finalised. Eight of the nine papers, which comprise this thesis, were arranged in chronological order to conform to the requirements of the University of the North-West (Potchefstroom Campus) for the degree of Philosophiae Doctor. It is required that "the work should clearly demonstrate advanced original research and/or creative work which must also constitute a real and major advance to the technology of engineering science or practice". All of the papers have been presented at local or international scientific conferences are fully referenced in this chapter. Papers 1 to 3 describe the development of the fluidised-bed limestone neutralisation process during laboratory and pilot scale studies. It was also demonstrated that iron(ll) needs to be oxidised to iron(lll) upstream of the limestone neutralisation stage as direct treatment of iron(l1)-rich water results in scaling of the limestone particles with gypsum and ferric hydroxide. Paper 4 describes an innovative and cost-effective way of generating hydrogen to be utilised as an energy source for SRB in an anaerobic biological treatment process. Hydrogen was generated on-site which eliminates the need for purchasing it from a local supplier and resulting in the process not being cost-effective. To date, sugar and ethanol were utilised as energy source for these bacteria. Paper 9, which describes the biological process, has been placed under Appendix A, as the other eight papers describe research that is chemically based. Paper 5 compares the chemical neutralisation and partial sulphate removal from AMD, to the biological process that has the same aim. Paper 6 presents practical information on the implementation of a full scale limestone neutralisation process which replaces the existing method of neutralising AMD with lime. The process has been successfully implemented at a mine site in South Africa and has been in continuous operation for almost two years to date. Paper 7 describes an alternative process for removing sulphates from AMD to below 200 mgle by means of chemical treatment, i.e. initially the AMD is neutralised with lime resulting in the removal of sulphates to below 1 200 mgle. The sulphates are then further removed to a lower level of < 200 mg/L Paper 8 proves that hydrogen, generated electrolytically (Paper 4), is the most cost- effective energy source for SRB. It also addresses other issues like energy utilisation efficiency of feed water with hot gas, rate of sulphate removal by SRB and the effect of biologically related issues. Paper 0 was included in this thesis for completeness. The paper addresses the treatment of effluent from a copper mine and is included in Appendix A. Papers 1 - 8 involve the chemical treatment of mine effluent and the utilisation of AMD as a medium for generating hydrogen as a useful by-product and energy source for SRB, while Paper 9 concentrates on the biological treatment of mine effluent. PAPER 1 r Geldenhuys,A.J., Maree, J.P., De Beer, M. and Hlabela, P. 2003. An integrated limestone/lime process for partial sulphate removal. J.S.A. lnst. Mining Metallurg.., 1 O3(6), 345 - 353. In the investigation of lime being largely replaced with limestone in order to achieve neutralisation and partial removal of sulphates from acid mine water, discharged by a coal mine near Wtbank in Mpumalanga, the main objectives were: J To determine the effect of limestone on the chemical composition of the coal processing water before and after treatment J To determine the effect of various parameters on the rate of gypsum and CaCO, precipitation J To determine the characteristics of the gypsum and CaC03 sludge produced The following findings resulted from the investigation: J Acid mine water can be neutralised effectively from a pH of 2.1 to 7.7 and the sulphate concentration being lowered from 3 000 mgle to 1 900 mgle J Wi lime treatment, as a follow-up stage to limestone neutralisation, the sulphate concentration was further reduced by means of gypsum crystallisation to below the original target of 1 200 mgle, i.e. 1 100 mgk 4 Wth lime treatment, pH values of 12 and higher were reached and magnesium was fully removed by gypsum crystallisation. 4 To lower the high pH, C02 sparging resulted in CaC0, precipitation was recycled to the limestone neutralisation stage and utilised as additional alkali J For design purposes, a contact time of 1 hour was needed for the neutralisation stage with limestone and 2 hours for the gypsum crystallisation stage. The surface areas of the limestone and lime played a major role in the rate of neutralisation and crystallisation J When replacing lime with limestone for neutralisation purposes and partial sulphate removal, an alkali cost saving of up to 62% could be achieved. Lime was only added to lower the sulphate concentration to below the saturation level of gypsum, i.e. 1 500 mgll, to ensure complete removal of sodium and magnesium thereby preventing scaling of pipelines Paper 1 was peer-reviewed and published in The Journal of the South Afn'can Institute of Mining and Metallurgy Paper I was also presented as a poster by J P Maree at the Hard Rock Mining 2002 Conference, Westminister, Colorado, USA (See Poster 1, Appendix A). An integrated limestonellime process for partial sulphate removal An integrated limestone/lime process for partial sulphate mwal Addrccdv*ahmthcpbntaedndllmetmfrom auh-hrombn aod dosing systen we pumped ido a Une (Q(OH)1) was fed w a semod -, rmninog the neaaalipdoraterdtbconaplintion-htkm - An integrated%nestone/lime prooess for partial sulphate removal Fci the gnstun hdge. Lhc Mling me Lnrcawd from 0.101 w 2.185 Rm ap ihc&im damad from Zm w An inlegrated limcstonc/ltme prmecs for parUal sutphak nmmval addMm~ oia coaguent and Bandant wy ubtained within r ~~~sodhomlhnc1kabwe optimm~m,untd~spdlicmmbinmon W been -dclBlnmed -- NarmuanDnandsuphlelemalhvv~amalwel %cr mniwdlpvelrhul Cam-,a/hi&qwllt;ran &be wand then wded w lhcFmaone neurralinrion ~neutralni& I g!Ridicy by ur$q iimmne ad tinle. IS illupNed in hble WI. 11 can be mid ibr acid water ran k .An inlegrated limestonellime process for partial sulphate removal ~mtheeElwnlsdthenmuS~dthepmgs lo enrjure fas ed sealine d the sluk, is inwltable. uearedwanwasakogrer)yrrpowdinⅆuy. Chriry unprowrrnl db3.92 and 9% can be vhievrd for the o&l&warer of the seamdarv dne sluk in the pram by theaddition of very & Gpdymrd subslanm lothesludge% SdlEng drhse sludges wa b Acid water m k neunalizpd efTmivdy vith limesone i& dlii. h addiion. subhate a rewed IO I 900 @I (arso,) the onginal aim of 1 200 rqgll by means of lime mtmenc to a pH lewl 01 12 and hi@. When pH vahrsof12a~l~mreahcd.gvpwnayaaC andtkwarerdependsontheslrhcearadthe llmmne and the lime. Fa dedgn purposes. a mnM rmraadydbyridaastreampnl* t k~tkverybw~~dAlinthetreared w~erdthepraessafw&sageoftmment.it an be conduded that the odvmers w!d fa ranoval t A contm lime d 1 rninuIe befwffn polyma addition la coegulationl8ocmhcion reasons was frmod to be Nmsent ra effmive senling oithe seamdary eWuenrs dthe three amsecutive stdges in the p-. a ~.Rr~~c.m.~~#lgwrr 8~&~mrd~. UniMSlaa .mn&bormc Fvbbhna. 1976. p. U3. 9. Ax"a.CJLiksax#srCCf*~~i,ktk~ ruPr,Idrpaar, Y-DSStfUh. 1497. p. 107. lo. Sh-.F. W~~qmbw(md~1. UniM %IS. Mffimwufl. led. 1m.~ 11%. 11. .u"UWadb. wccCawm-: RIla--(mdm lar UnW%a~la.M~-Hi3rdadi1Bn, 1491. p. IM. PAPER 2: Geldenhuys, A.J. & Maree, J.P. Synthetic organic polymers (PAC6 and 3095) as coagulants/flocculants for optimisation of an integrated limestonellime neutralisation process for partial sulphate removal, Proceedings gh Annual Industrial Water Management and Treatment Symposium, 15-1 6 May 2002, Johannesburg, South Africa. This investigation arose from a need to improve the poor settling rates of gypsum- and CaC0,-sludge in the combined limestonellime process. These settling rates were very slow and caused the overflow water from each stage to have a high degree of turbidity and colour. The effectiveness in terms of cost and water overflow quality, using two polymeric substances to improve settling of the sludges, was extensively investigated in order to formulate design criteria for settling tanks (clariiiers) in full scale neutralisation and sulphate removal plants. The main objectives were: J To determine the effect the polymers had on the chemical composition of the water, before and after treatment J To determine the optimum dosage of polymer and contact time between polymer and sludge for each stage in the combined limestonellime process J To evaluate the economic efficiency of the polymeric substances J To produce design criteria for settling tanks in each of the three stages of the combined limestone/lime process Type 3095 flocculant, which is poly-acrylamide in granule form, was effectively utilised to serve as flocculant after both the lime treatment stage (gypsum crystallisation) and the CO, treatment stage (CaCO, precipitation). The CaC03 sludge, resulting after C02 treatment in the third stage of the process, requires the use of a coagulant prior to the flocculant to enable the growth of aggregates into sizeable flocs. PAC6, which is a poly- (aluminium-hydroxy) chloride and only available in solution, was successfully introduced to the process as coagulant. The gypsum sludge, resulting from the first stage of the process (limestone neutralisation), did not need any polymeric addition for improvement of sludge settling, although the suitability of PAC6 as flocculant was evaluated at this stage and presented in this paper. For design purposes, the results are such that the required settling rates of the various sludges, the essential clarity of the overflow water, and the cost of the amount of additive are listed. From these results, a full scale limestone neutralisation and iron(l1)-oxidation plant has been designed, constructed and commissioned in January 2004 to treat toe- seep from a coal discard dump near Witbank. Paper 2 was presented orally by A J Geldenhuys at the gh ~nnual Industrial Water Management and Treatment Symposium SYNTHETIC ORGANIC POLYMERS (PAC6 AND 3095) AS COAGULANTS/FLOCCULANTS FOR OPTIMISATION OF AN INTEGRATED LIMESTONEJLDIE NEUTRALISATION PROCESS FOR PARTIAL SULPHATE REMOVAL mmm A J GELDENHUYS and J P MAREE CSIR Division of Water, Environment and Forestry Technology, P 0 Bm 395, Pretoria, 0001, Smth Africa (Tel no: +27 I2 841 41 78, Fax no: +27 12 841 2506, Email: ajgeldenhuys@csir.co.za) KEYWORDS: Limestone neutralisation, gypsum crystallisation, CaCa precipitation, acidic effluents, partial sulphate removal ABSTRACT Limestone and lime treatment is the most cost-effective technology for neub;llisation and partial sulphate removal of acidichlphate-rich water to sulphate levels of less than 1 200 mgK due to precipitation of magnesium and removal of the associated sulphate fraction (through gypsum crystallisation). Neutralised mine water of this quality may be suitable for irrigation or reuse in mine applications. The overflow water 6om the clarifier of each of the above stages generally contains a wide variety of colloidal impurities that may cause the water to appear turbid or may impart colour. This results in very slow sludge settlimg rates that causes overflow water with a high degree of turbidity and colour. There are a number of successive or simultaneous stages involved in the agglomeration of particles. This paper describes the use of two of these stages in an innovative process for the neutralisation and partial sulphate removal of acid streams produced during coal mining and processing. In this process, the integrated lime and limestone has been developed to treat water for removal of magnesium, metals and sulphate (to less than 1 200 mg/t ). The process consists of the following stages: 1. Limestone (CaCO,) neutralisation to raise the pH to 7 and C02-production. 2. Lime (Ca(OH)2) treatment to pH 12 for Mg(OH)2 precipitation and gypsum (CaS04) crystallisation. 3. pH adjustment with C02 recovered from Stage 1 to enable CaCO, precipitation for re-use in Stage 1. In an earlier study it was found that the rate of gypsum crystallisation is directly proportional to the surface area of the gypsum. To get optimum neutralisation of the acid water and partial sulpate removal to below the saturation level of gypsum (i.e. 1 200mg/t), maximum sludge recovery is needed. The higher the concentration of seed crystals the larger the surface area. Therefore, effective removal of suspended and colloidal matter from the overflow of each stage are required and can be achieved by coagulation and flocculation. The polymers, PAC6 and 3095 were effectively used as coagulant and flocculant respectively. The addition of these polymers resulted in a clear overeflow in each of the above stages with a very low turbidity. 1 INTRODUCTION Lime clarification is applied throughout the world in water treatment. Its most extensive use is in the purification of surface waters for domestic and industrial use. Limestone and lime clarification were effectively introduced to the water reclamation field at the Navigation Colliery near Witbank. The integrated limestonellime treatment process (Figure 1) comprises three consecutive stages which may be carried out seperately in different pieces of equipment (reactors, clarifiers, and coagulant/flocculant mixing tanks). Limestone neutralisation Lime treatment co, treatment rm Figure 1: Block diagram of integrated limestonellime treatment process The clarification system at the Navigation treatment plant consists of the following items: I. A limestone reactor for mixing the limestone and acid feedwater thoroughly during limestone neutralisation, and has a retention time of 1 hour. A flocculant storage- and feed tank for dosing flocculant: 3095 into the overflow from the reactor to a clarifier. In the clarifier, the clarified effluent is seperated from the sludge. The sludge produced in the clarifier is recirculated to the reactor by means of a variable speed peristaltic pump. The sludge density is controlled by the recirculation and draw-off rates. 2. A lime reactor, in which the neutralised water from Stage 1 (limestone neutralisation) is thoroughly mixed with lime (retention time: 2.72 hours) to achieve Mg(OHh precipitation and gypsum (CaS04) crystallisation. A flocculant storage- and feed tank for dosing flocculant: 3095 into the overflow from the reactor to a second clarifier. For handling the sludge and clarified effluent of 51 ------- \J I ,J r- Reactor Clarifier Coagulation: Flocculation: Overflow Sludge PAC6 3095 recycling the lime treatment stage as well as the next stage (C02 treatment), the same applies as for limestone neutralisation. 3. A mixed reactor for pH adjustment by decrease the pH of the treated water resulting from Stage 2 to enable CaC03 precipitation for re-use in Stage 1. A retention time of 30 minutes was used. A coagulant and flocculant storage- and feed tank for dosing coagulant: PAC6 and flocculant: 3095 in serie into the overflow from the reactor to a third clarifier. Considerable attention has come to be directed at the removal of small particles by combining them into larger aggregates by coagulation and flocculation. These particles are mainly of chemical origin as limestone and lime are added to the system which result in the production of Mg(OHh, CaSOd and CaCO,. Water contains many compounds, which can be classified into three categories: Suspended solids - These products may be mineral in origin (sand, silt, clays, etc.) or organic (products resulting from the decomposition of plant or animal matter, humic or fulvic acids, for example). Added to these compounds are microorganisms such as bacteria, plankton, algae and viruses. These substances, in particular, are responsible for turbidity and colour. Colloidal particles (less than I micron) These are suspended solids of the same origin as the above but of smaller size and with a settling rate that is extremely slow. They are also responsible for turbidity and colour. Dissolved substances (less than several nanometres) - These are usually cations or anions. Part of the organic matter is also dissolved. Gases are also present (02. C02. HZS, etc.). Colloidal particles that cause colour and turbidity are difficult to separate from water because the particles will not settle by gravity and are so small that they pass through the pores of most common filtration media. To be removed, the individual colloids must aggregate and grow in size. Aggregation is not only by the small size of the particles but more importantly by the fact that physical and electrical forces keep the particles separated from each other and prevent the collisions that would be necessary for aggregation to occur. Thus colloids are particles that cannot settle naturally and for which surface area factors are most important. These factors determine the stability of colloidal suspensions. In fact, colloids are subject to two major forces: 52 J Van der Waals attraction, which relates to the structure and form of colloids as well as to the type of medium, J The electrostatic repulsive force, which relates to the surface charges of the colloids. In order to destabilise the suspension, it is necessary to overcome the energy barrier. To accomplish this and, thereby, promote the agglomoration of colloids, it is neceassary to reduce the electrostatic forces. This destabilisation is brought about by coagulation (Benefield et al.'). Chemical agents can be used to promote colloid aggregation by destroying the forces that stahilise colloidal particles. Mechanisms responsible for destabilisation of inorganic clay colloids have been identified through extensive research studies and are well understood. The process of destroying the stabilising forces and causing aggregation of clay colloids is referred to as chemical coagulation (Benefield et al.'). It can also be referred to as the destabilization of colloidal particles brought about by the addition of a chemical reagent known as a flocculant. Flocculation can be adressed as the agglomeration of destabilized particles into microfloc, and later into bulky floccules which can be settled called floc. The introduction of another reagent, called a flocculant may promote the formation of the floc. According to ont ti us" the process of coaggulation as practised in water treatment can he considered as three separate and sequential stages: coagulant formation, particle destabilisation, and interparticle collisions. Coagulant formation and particle destabilization occur in rapid-mixing tanks; interparticle collisions occur predominantly in flocculation tanks. The major factors affecting the coagulation/flocculation processes in diverse ways are: (1) coagulant/flocculant dosage; (2) pH; (3) colloid concentration, often measured by turbidity; (4) anions or cations in solution; and (5) mixing effects. Synthetic organic polymers can be effective as coagulants. These polymers are long-chain molecules comprised of many monomers. Polymers typically have a helical molecular structure comprised of carbon chains with ionising groups attached. When the groups are ionised in solution, an electrical repulsion is created which causes the polymer to assume the shape of an extended rod. As the ionised groups become attached to colloidal particles the charges are neutralized and the polymer starts to coil and form a dense floc with favourable settling properties. The main objectives of this investigation were: 0 To determine the effect of the polymers on the chemical composition of the water before and after treatment To determine the optimum dosage of coagulant!flocculant for each stage of the treatment process (concentration) To determine the efficiency of the polymers (as coagulant/flocculant) on the chemical composition of the coal processing water before and after treatment in all stages of the treatment process (settling rate of sludge and clarity of overilow water) To determine the economic feasibility of the polymers used as coagulant/flocculant 2 MATERIALS AND METHODS 2.1 Feed Water Acidic leachate hm a waste coal dump, was used as feed water for batch studies on laboratory scale in beakers and continuous studies on pilot plant scale. The relevant chemical composition of this leachate is listed in Table I. Table I. Chemical composition of acidic mine leachate as feedwater Parameter Acid feed water PH 2.10 so:- (mg/e ) 3 000 ca (mde ) 420 Mg (mg/f ) 160 Na (mgM 41 Mn 17 cr @g/e ) 16 Acidity (mdt ) 3 000 Waste powder limestone (CaCOJ 6om paper industries was used in the limestone neutralisation stage of the process. For the gypsum crystallisation stage, unslaked lime (Ca(0H)J was used. The limestone was analysed for its calcium, magnesium and alkalinity content. Calcium and magnesium were determined with EDTA, while the alkalinity content was determined by dissolving it in a stoichiometrically excessive amount of hydrochloric acid. This excess amount of hydrochloric acid was titrated with sodium hydroxide (see Tables Il & III for chemical composition of powder lime and limestone respectively). Table 11. Chemical composition of powder lime (Ca(0Hk) Bulk Available Available Total CaO MgO A1203 + SiOz Acid density Ca(0m CaO (ma) (max%) Fe203 (nu%) insolubles Particle size: Fine powder, 100% passing 90 micron Table 111. Chemical composition of powder limestone (CaC03) CaCO, WOHk Ca Mg Na K Moism (%) (%) (we ) (we ) (we ) ) w) 97.02 3.87 354.08 5.66 11.17 0.43 24 Percentages based on dried basis The polymers that were used as coagulant and flocculant respectively are PAC6, which is Poly-aluminium- hydroxy chloride with a specific gravity of 1.3 (only available as solution) and 3095 (granules), Poly- acrylamide which is a co-polymer of acrylamide and acylic acid with a molecular weight of about 15 000 000 &ole. These chemicals were supplied by Montan Chemicals, Germiston. The coagulant was used as received while a stock solution of 2.5g of 3095 flocculant per litre distilled water was made up on a weekly basis. For dissolving the flocculant in distilled water, a small amount of methanol was first added to the granules to assure effective granular dissolution. 2.3 Batch studies and continuous studies Batch studies were conducted in the laboratory in lt volumetric cylinders at atmospheric pressure and ambient temperature to study the efficiency of the two polymers, PAC6 and 3095 as coagulant and flocculant respectively. These polymers were used to enhance settling of the various sludges produced in the three consecutive stages of the limestone/lime treatment process (Figure I). The polymers were added separately to the effluents of these three stages as follows: Primary effluent (limestone neutralisation): 3095 + flocculant Secondary effluent (lime treatment for gypsum crystallisation): 3095 + flocculant Tertiary effluent (COz treatment for CaCO, precipitation): PAC6 --t coagulant 3095 + flocculant Each effluent was stirred at a medium stirring rate for lminute whereafter increasing amounts of polymer were injected into the solution (effluent). To assure thorough mixing of the polymer and solution, it was further stirred for 1 minute at the same rate. For the tertiary eflluent, an additional stirring of 1 minute between the coagulant and flocculant addition was allowed. At various time intervals, the interface of the settling region was marked and 30mE samples were withdrawn from the the top part of each cylinder (representing overflow) and analysed for turbidity until a settling time of 30 minutes have been reached. These measurements enable quantification of the influence of the mentioned polymers as effective coagulants/flocculants by means of polymer concentration, settling rate and clarity. Cylindrical tests should be used only for the evaluation of the operation, chemical dosage and removal of contaminants. It is not valid for evaluating the size of the clarifier or to simulate a reactor-clarifier but serves merely as guidelines. Based on the studies conducted in the laboratory, the optimum amount of polymer was used in each consecutive step of the process at pilot plant scale to demonstrate the effective settling of the various sludges and clear overflow water on larger scale. 2.4 Equipntent and procedure A mobile pilot plant at Navigation Coal Mine near Witbank, with a capacity of 10m3/day was used for on- site lreahnent of the acid mine water. The design and operational parameters for the mobile plant are summarised in Figure 1 and Table N. Table IV. Design parametem for the three consecutive stages of the treatment process as illushllted in Figure 1 Parameter Limestone Limestone Lime Lime c02 co2 reactor clarifier reactor clarifier reactor clarifier Diameter (m) * 1.220 1.220 0.600 1.220 * Base area (m2) a t 1.169 1.169 0.283 1.169 Height (m) t * 0.850 2.100 0.800 2.100 Volume (m3) * * 0.994 2.821 0.226 2.821 Hydraulic retention 1 2.72 - 0.5 time (HRT, hours) * Full scale plant in operation In all three stages, the polymer(s) was added to the effluent in the overflow between the reactor and clarifier for rapid mixing to disperse the polymer(s) homogeneously for enabling good contact between the effluent and polymer@). In each clarifier, slow speed agitation was provided by recycling a part of the sludge back to its reactor by means of a variable speed pump to enable the growth of the aggregates into voluminous flocs by collision and entrapment of the suspended matter. Another reason for recycling a partial of separated solids (the sludge) is to provide nuclei for better coagulation and flocculation. The rest of the sludge is disposed of. 2.5 Programme 2.5.1 Batch studies on coagulatiod/flocnrlation The concentration of polymer added was varied to evaluate and optimise the efficiency of the polymers as coagulant/flocculant by measuring the clarity and settling rate of each hial. 2.5.2 Continuous studies on cogulatiodflocculation The quality of the water before and after treatment for the various stages of the process were studied at the pilot plant on a continuous basis at the hydraulic retention times listed in Table 4. 2.6 Analytical The samples were collected regularly and filtered through Whatman No 1 filter paper. Determinations of pH and sulphate were carried out manually according to procedures described in Standard Methods (APHA~). Calcium and magnesium were recorded with the ICP Method for metals. Alkalinity was determined by titrating with sodium hydroxide to pH 7.0. Clarity was determined with a spectrophotometer, measured in NTU (Nephelometric Turbidity Unit). The measurement of turbidity is based on comparison of the intensity of light scattered by a sample as compared to the light scattered by a reference suspension under the same conditions. 3 RESULTS AND DISCUSSION 3.1 Water quality The chemical composition of the acid mine feed water and the water qualities after each of the three stages of treatment are listed in Table V (samples were taken in each reactor and after each polymer addition) Table V. Water qualities of feed water and treated water of each of the three wnsecutive stages of the liestoneiliie treatment process (see key notes in Figure 1) To optimise the process, settling of the various sludges in the three stages were improved by the addition of the polymers PAC6 and 3095 as coagulant and flocculant respectively to the effluents. The addition of limestone, lie, C02 and polymeric substances to the secondary effluents in the various pH ranges had the following effects on its inorganic composition, as illustrated in Table V: a) Limestone (CaC03) can be used effectively in the integrated process for neurralion of acid water. When discard leachate was treated with limestone, the pH of the water was raised hm 2.71 to 6.20 in the neutralisation reactor. To achieve further reduction of sulphate to the level of 2 200mg/t, lie was added to the neutralised effluent. The pH and Ca content increased to 11.71 and 994 mg/t respectively, due to gypsum crystallisation in the second stage of the process. The high 59 pH of 11.71 after lime addition to the water was corrected by bubbling C02 through the treated water until a pH of 8.50 bad been reached. b) The acidity of the water after limestone neutralisation was initially increased owing to the addition of hydroxide ions, but was subsequently reduced by the precipitation of carbonate as CaCO, and hydroxide as Mg(OH)>. The Mg concentration was not affected by the addition of lime per se, but was reduced as a result of the precipitation of Mg(OH)2 at pH levels of >10.8 c) The presence of Al ions in the various secondary effluents played an important role in the achievement of optimum clarification during limestone-, lime- and C02-treatment. This is due to the formation of gelatinous Al-precipitated compounds that is illustrated in the decrease of Al concentration, from 556 mglt to zero, in each of the three stages of the process. It can therefore he concluded that the Al content in the polymers used as coagulant or flocculant did not affect the final quality of the treated water. 3.2 Optimum polymer dosage It is known that particles in the relatively dilute solutions did not act as discrete particles but coalesced during sedimentation (Metcalf & Eddy4) . As coalescence or flocculation occurs, the mass of the particle increases and it settles faster. The extent to which flocculation occurs is dependent on the opportunity for contact, which varies with a number of variables. One of these variables are the concentration of the particles. The settling rates for various concentrations of particles (polymer) which were added to the three consecutive stages of the process individually, are photographic illustrated in Tables VI, VII and VIII. Table VI. 1 .. '.,JI Photographic illustration of the settling of the sludge in the limestone neutralisation stage (stage 1) of the limestone/lime treatment process with increasing concentration of polymer 3095 as flocculant (including values for clarity and cost) u 2 .u u u . _0 T: 111 I T: 173 R - Cost (c/m3) SR -Settling rate (m/h) T:237 61 -- - -- - -- --- I :al 8i .... . . - -- T:264 I T:247 IT: 283 - ... .. U .. 1 "'Ir a..,1f 1..,/1 11.5- < . _I --err ji" . I .. 8L I \_.0_ I-=-" . - .. .iI;...._-:::.:...... --- - Polymer:3095 as fIocculant Settling time: 0 minutes - direct, C: 0.125 C: 0.250 C: 0.500 C: 1.000 C: 1.250 C: 1.500 T: out of spec. T: out of spec. T: out of spec. T: out of spec. T: out of spec. T: out of spec. SR: 4.620 SR: 5.100 SR: 5.040 SR: 5.370 SR: 5.790 SR: 5.910 R: 0.275 R: 0.550 R: 1.100 R: 2.200 R: 2.750 R: 3.300 3__ U 4 t'>""I.! -:... -po 'r81":[ orTr ,r Table VII. Photographic illustration of the settling of the sludge in the lime treatment stage (stage 2) (including values for clarity and cost) of the limestone/lime treatment process with increasing concentration of polymer 3095 as flocculant 't r.. 0_ \ Polymer. 3095 as floccu/ant Sett/in time: 0 minutes - direct, C: 0.250 C: 0.500 T: out of spec. T: out of spec. SR: 0.672 SR: 0.840 R: 0.550 R: l.l 00 3 .. o.~ 'I"'IIr' u D o ,r I 1 ~ L..-- I --- ---- - -- - - --- - ---- C: 1.250 T: out of spec. SR: 2.580 R: 2.750 C: 1.750 T: out of spec. SR: 3.440 R: 3.850 11 T: 72.7 T:22.6 " 1;15-01' --,- - T: 10.7 I T: 10.5 R - Cost (clmJ) SR - Settling rate (mIh) ... IT.. 0_ .I!:. C: 2.000 T: out of spec. SR: 5.610 R: 4.400 v , ~'0":J'f -,p ...1- I~ -'...,,-..I T: 11.5 ~l ., I 1~ ..."i-, ~ - .. T:7.86 62 1 V U 2 U 11 11 H -oJ1 0.15." ".2S"':Ik 1."1S'\III' 1'°"3" - - "";p ....---.. 10{;i) .... Table VIII. Photographic illustration of the settling of the sludge in the C02 treatment stage (stage 3) of the limestonenime treatment proce ss with increasing concentration of polymers PAC6 and 3095 as coagulantlflocculant respectively (including values for clarity and cost) 1 ~ a_41'11U.' ..~- .,f.. , . O_..f, ,.,. , ,JJ~t 'r'n1' 1\, 2r J1-.:' ;~.r !i v If _"'~I.U t I - T..:o-. I -/_. _ - .- . __ J Polymer:PAC6 as coagua/ntand 3095 as ffoccu/ant Sett/in time:0minutes- direct a ercoo 'occuIanJdosae C: 0.400 C: 0.500 C: 0.625 C: 1.000 T: 1891 T: 1891 T: 1891 T: 1891 SR: 19.980 SR: 20.040 SR: 20.100 SR: 22.080 R: 4.130 R: 4.350 R: 4.625 R: 5.450 3 I! ¥_.-, 4 ~.,..- -,,- .~ i~ C: 1.250 T: 1891 SR: 20.340 R: 6.000 C: 1.500 T: 1891 SR: 20.160 R: 6.550 v -" - "iL.... If 1_J;",,,U ~~.11- 'IF ...- J ~ .. I' I . T: 0.115 The photo's listed in Tables VI, VII and VIII, are an illustration of the liquid's tendency to move up through the interstices of the contacting particles. As a result, the contacting particles tended to settle as a zone (or 'blanket'), maintaining the same relative position with respect to each other. As the particles in this region settled, a relatively clear layer of water was produced above the particles in the settling region. In the case of stages 1 and 2, an identifiable interface developed between the more or less clear upper region and the 63 -- - -- - - - - - - -- - -- ----- - - --- ---- hindered settling region, as illustrated in Tables VI and VII. The rate of settling in the hindered-settling region is a function of the concentration of solids and their characteristics. As settling continued, a compressed layer of particles began to form on the bottom of each cylinder in the compression-settling region. The particles in this region apparently formed a structure in which there is close physical contact between the particles. As the compression layer formed, regions containing successively lower concentrations of solids than those in the compression region extend upward in the cylinder. In the case of stage 3, such a small concentration of sludge was formed and settling of the sludge with an addition of a coagulant and flocculant was obtained within a few seconds. Inefficient coagulation/floccualtion resulted from an overdose of polymer to the system or from intense or prolonged agitation. If excessive polymer is added, the segments may saturate the surfaces of colloidal parfticles so that no sites are available for the formation of polymer bridges. This can restabilise the particles and mayor may not be accompanied by charge reversal. A very narrow optimum exists for the polymer and overdosing or underdosing will result in restabilisation of the colloids. Intense or prolonged mixing may destroy previously formed bridges and lead to restabilisation. An inverse relationship exists between the optimum polymer dosage and the concentration of colloids to be removed. This can be explained as follows: at low colloid concentrations a large excess of polymer is required to produce a large amount of precipitate that will enmesh the relatively few colloidal particles as it settles. At high colloid concentrations, coagulation/flocculation will occur at a lower chemical dosage because the colloids serve as nuclei to enhance precipitate formation. Depending on the settling rate of the sludge, the clarity of the overflow water and the cost of polymer added for sufficient sludge settling, the optimum concentration of polymer can be determined. Dosages of only 0.50 to 1.0 mglCare needed for effective settling of the sludge in the first stage of the process with a settling rate of 5.04 to 5.37 rnlh. Nearly 33% of the sludge had already been settled after 1 minute of polymer dosage. In the second stage of the process only 0.75 to 1.25 mglCof the polymer are needed to achieve a settling rate of 1.635 to 2.58 rnlh.For the third stage, the addition of a coagulant first, followed by a flocculant are necessary to achieve good settling of the very fine CaC03 sludge. Only 0.005 mE/C of the coagulant PAC6 and 0.4 64 mglf of flocculant 3095 were needed for effective sludge settling. A specific combination of coagulant:flocculant exists and tTomthere the above optimum amount of this specific combination had been experimentally determined. For design purposes, the photographic illustrations and values for the above variables are listed in Tables VI, VII and VIII for the three consecutive stages of the process. 3.3 Polymerefficiency The efficiency of the polymers used in each of the three consecutive stages is directly related to the settling rate of the sludge and the clarity of the overflow water. The effect of polymer concentration on sludge settling rate and clarity of the overflow water of the three stages are illustrated in Figures 2, 3 and 4. a) 1. Polymer concentration (mg/1 0.5 1.5 b) , Polymer concentration (mg/1I25 1.5 Figure 2: Illustration of the effect of polymer concentration on clarity of the overflow water (a) and sludge settling rate (b) for stage 1 65 - --- ------- ----------- 9XI 800 I I U ----- f I I 511J< 700 - : . Ii 800 , - 500 ! 3!--H I 11111 III 1M .. . Of :; 400 C :i i! .. 300 . . too 200 100 . ! III i a) Po/yIMr concentration (mgI'II b) Figure 3: b) Illustration of the effect of polymer concentration on clarity ofthe overflow water (a) and sludge settling rate (b) for stage 2 22.5 22 l' 21.5 0.5 b) 21 Settling rate (m/h) 1.25 0.5 Polymer concentration (mg/1 1.50.' 0.825 PoIyrMrconcenlraUon(mg/1I 1.25 1.5 Figure 4: Illustration of the effect of polymer concentration on clarity of the overflow water (a) and sludge settling rate (b) for stage 3 The turbidity of the clarified effluents are a parameter that are very sensitive to the mode of operation of the limestone-, lime- and COrtreatment units (i.e. chemicals used, overflow rate and retention time and sludge recirculation rate) and can be manipulated by the use of suitable coagulant or flocculant aids. A reasonable relationship exists between turbidity and suspended solids for the settled secondary effluents ftom this process. From Figures 2, 3 and 4, it can be noted that: 66 10D0D0 9001IO 11IIII- 7ODOD eoooo l: 5001II ;; :e 4DQDD 3DIIDO ZIIOO 101D) With increasing polymer concentration, the color of the overflow water from the settled secondary effluent in the first and second stage of the process, changed sequentially from dark brown to transparent. For the third stage of the process, the very fine CaC03 particles settled immediately after polymer addition and the overflow water had an almost perfect clarity of 3.83 NTU (turbidity) within I minute's time. For stages 1 and 3, the settling rate of the sludges remained fairly constant with increasing polymer concentration. For stage 3, a linear increase in settling rate resulted from increasing polymer concentration. Economic feasibility Neutralisation and sulphate removal from water to a level where the water quality is suitable for re-use in the process or discharge into the sewerage network, can be achieved with limestone (CaC03) and slaked lime (Ca(0H)d. It can he noted from Table IX that the addition of polymers to the effluents of the various stages of the process, to ensure fast enough settling of the sludges, are inevitable. Table IX. Effed of polymer concentration on the settling rate of the sludges produced in the three consecutive stages of the treatment process Stage 1: Limestone neutralisation Polymer 0.000 0.125 0.250 0.500 1 .000 1.250 1 .500 (we) Settling rate 1.30 4.62 5.10 5.04 5.37 5.79 5.91 Cost(dm3) 0.00 0.28 0.55 1.10 2.20 2.75 3.30 Stage 2: Gypsum crystallisation with high lime treatment Polymer 0.000 0.250 0.500 0.750 1.250 1.750 2.000 concentration (we) Settling rate 0.21 0.67 0.84 1.64 2.58 3.44 5.61 (d) Cost (c/m3) 0.00 0.55 1.10 1.65 2.75 3.85 4.40 Stage 3: CaC03 precipitation with C02 treatment Polymer 0.000 0.400 0.500 0.625 1.000 1.250 1.500 concentration (mm Settling rate 0.00 19.98 20.04 20.10 22.08 20.34 20.16 The limestoneflime neuhalisation process will be overall more cost effective, because of the higher settling rates that will not only save on capital cost, as much smaller clarifiers are needed for sludge settlii, but will also produce overilow waters hm secondary effluents of a much higher quality in terms of clarity. Clarity improvement of 63, 92 and 99% can be achieved respectively for the overilow water of the secondary settling sludges in the process by the addition of vq small quantities of polymeric substances to the sludges (see Tables VI, VU and VIII). Settling of these sludges were improved by 74, 90 and 1Wh in the three consecutive stages of the process (see Table M). 4 CONCLUSIONS Due to the very low concentration of Al in the treated water of the process after each stage of treatment, it can be concluded that the polymers used for coagulationJflocculation had no effect on the chemical composition of the treated water after limestone and lime treatment for neutralisation and partial sulphate removal. A contact time of I minute between polymer addition for coagulationJflocculation reasons was found to be sufficient for effective settling of the secondary effluents of the three consecutive stages in the process. The settling rates of the secondary sludges in the process were increased by 74.2, 91.9 and 100% respectively for the three stages by the addition of PAC6 as coagulant to the third stage and 3095 as flocculant to all three stages of the process. Clarity of the secondary effluents were improved by almost 100% for all three stages of the process, as the water clarity before polymer addition for the various stages were >99 999. Depending on the required level of sludge settling and effluent clarity of the treated water after each stage in the process, large capital cost savings can be achieved by the addition of a specific amount of flocculant to the first two stages and a coagulant and flocculant to the third stage of the process. 5 REFERENCES 1. Benefield, L.D., Judkins, J.F. and Weand, B.L. Process Chemistry for water and wastewater treatment. United States. Prentice-Hall, Inc., Engelwood Cliffs, N.J. 1982. Pp. 510. 2. Pontius, F.W. Water qualify and treatment. United States. McGraw-Hill, Inc. 4' ed. 1990. Pp. 1 195. 3. APHA, Standard methods for the examination of water and wastewater treatment. 12Ih ed. American Public Health Association. New York, McGraw-Hill, 1985. 4. Metcalf & Eddy. Wastewater Engineering: Treatment, disposal and re!-use. United States. McGraw-Hill, Inc. 3* ed. 1991. Pp.1 334. PAPER 3: Geldenhuys, A.J., Maree, J.P., Strobos, G., Smit, N. and Buthelezi, B. Neutralisation and partial sulphate removal of acid leachate in a heavy minerals plant with limestone and lime, Proceedings 6'h lnternational Conference on Acid Rock Drainage, 12- 18 July 2003, Cairns, Australia. Ticor SA produces zircon, rutile and ilmenite from mining sand dunes on the Natal coast. The plant produces 85 m3/h of acid water that needs to be treated for re-use in the metallurgical process leaching circuits. A portion of the process water has to be blended with the Empangeni sewage system (discharge water) to prevent build-up of soluble ions. The stream is required to have a sulphate content of less than 2 200 mg/e. Initially, during 2001, a combined limestone/lime treatment plant was designed, constructed and commissioned to treat acid water, resulting from the process. It comprised a primary neutralisation stage utilising limestone and a secondary neutralisation stage employing lime. With this combined process, the sulphate concentration in the water was lowered to 1 150 mgk. During an attempt to optimise the process and further save on alkali cost, the lime treatment stage was abandoned and only limestone utilised to neutralise the acid water (pH 6.5) and lower the sulphate concentration from 5 100 mg/e to 2 200 mg/Z to comply with the above mentioned minimum requirements from the local municipality. During this investigation, the following findings were made: J Aeration of the acid water after the primary neutralisation stage resulted in C02 removal. Consequently lime was completely replaced by limestone for neutralisation of the acid water resulting from the process J Sulphate concentration in the acid water was lowered from 5 100 mgle to 2 200 mgl! to render the water acceptable for re-use in the plant or for discharge into the local sewerage system J The quality of the treated water could be controlled and any build-up of ions in the water avoided J Savings of R1 M per annum were achieved by replacing lime with limestone J Paper has been presented as a poster by A J Geldenhuys at the 6'h lnternational Conference on Acid Rock Drainage. (See Poster 2 in Appendix A). Neutralisation and Partial Sulfate Removal of Acid Leachate in a Heavy Minerals Processing Plant With Limestone and Lime A J Geldenhuys1. J P Maree 1. G Strobos 1. N Smn2 and B Buthelezi2 ABSTRACT In mining and processing operations where minerals, high in pyrite and low in calcitcldolomite are proccsscd, Kid is generated. which needs to be ncutraliscd. licor at Empangeni in South Africa (SA) produces rutile (Ti9!%), basadongaschranatagaphicre~ubfmm(heSABSwhereUmgassamplesmn,analysed for hyhogen pvity. Fa+- A 326 326 5 32 215 J 2.43 mdardc LmAkD A 513 3.13 5 33 233 -f 3.53 Anh& 'Nk30YKOn B 13.78 13&3 5 3.8 243 4 0.00 Aniaic .Nl+XKM B 13.4 1355 5 6 250 4 0.00 hid NiprsbtAMD C 3.49 3.49 5 6 4.W J 0.00 AnioriO - C 2M 245 2 4.3 168 4 0.00 Anionic Sabd+MmC245~1 3.7 0.n J 0.00 Anionic 6th lnfsmatbnal Congrass on Mine Weer 8 the Envimrnenl. Johannesbwp, So& AMca In category A, hydrogen was produced at the cathode whle the anode started to dissolve due to the anodic reaction. The main reactions are given by: Anode: FeZn -t ~e~*/~n~' + 2e- I21 Cathode: 2H20 + 2e' + H2 + 20H- [31 Although hydrogen was generated at a fairly high rate at the cathode by using FeZn electrodes, the potential became increasingly high due to the decreasing surface area that resulted in a much higher resistance and electrode inhibition by electrode products. For every litre of hydrogen produced. 1.13 gram of Fe or 1.52 gram of Zn is needed. The costly effect of the sacrificial nature of FeRn in AMD will be presented in section 3.4. In category B, hydrogen was generated at the cathode while oxygen was generated at the anode. Not only was hydrogen produced at a higher rate and at an almost ten times smaller potential, but the electrodes were totally unaffected by the KOH used as electrdyte. The amount of gas production can be increased by increasing the current density. This will. however result in an increase in production cost as the voltage will also increase. The reactions at the anode and cathode can be given as: Anode: 20H' + H20 + 'h02 + 2e- 141 Cathode: 2H20 + 2e -t H2 + 20K [51 After 72 hours of running the set-up continuously, the diaphragm developed microholes as a result of the caustic nature of the KOH (30%). This phenomenon however abated when a more diluted solution of KOH (3%) was used. From the volt-amperometric results in category C, fairly high volumes of hydrogen can be generated, using AMD as electrolytic medium. The electrodes (Ni and Stainless Steel) and membrane were totally unaffected by the AMD that was used as electrolytic medium. It has been proved that the higher the current density, the higher is the amount of gas that is generated electmlytically. This also results in an increase in production cost as a result of higher voltages. The membrane was unaffected by the AMD. 3.2 WENT OF IRON(1I) OXIDATION One of the main benefits that resulted from the use of electrolysis in order to generate hydrogen economically, was the oxidation of iron(ll) at the anode in using stainless steel as electrode material in AMD as electrolytic medium. This means that, while producing hydrogen, iron(ll) oxidation as a pre-treatment stage to AMD can be applied. The half-cell reactions for the stainless steelIAMD cell are illustrated by: Anode 2Fe2* -t 2~e" + 2e- PI Cathode 2H30* + 2e + 2H2 + H20 [7l Figure 3 illustrates the relevant reactions at the anode and cathode. The oxidation of imn(1l) to Fe(l1l) can be proved by the transparent, ocher coloured AMD that was transformed to a dark brown solution. Precipitates that formed from these solutions, onto the anode, were analysed by means of Mossbauer spectroscopy. Signatures of P-FeOOH (akaganeite) were quite distinct in these spectra and appears to have evolved paltially to a-FeOOH (goethite). 8th International Congress on Mine Water & the EnvIronment, Johannesburg, South Africa ~ Electrolyte: AMD U t '., 'I' , 't '.1 I Stainless steel anode Stainless steel cathode :. Cathode Anode ~igure3. IUustrationof electrolytic set-up: stainless steel electrodes in AMD 65 - - --- - ------ 85 -- The ox!ddm of imn(ll) is illusbaled in Figure 4 over a pemd of 55 minutes An lncrwse in adaty (8 4M) r@t b 11 200 IT@/) resuted from caMaibn of the iron(ll) (4 245 n?@/ imn(ll) lo 279 mglf imn(1l)). Dunng oxidatbn. lhe pH aisa dm+ from 310 to 265. mi drq~ in pH and increase in dily can be ascribed tn Us, higk axidallrn state ol the iron Ion Fon(ll) b Mill)). These rnwb are 4lustmW in Table IV. 8th International Congress on Mine Water & the Environment, Johannesburg, South Africa The effect of various other parameters on the rate of oxidation, as Illustrated in Agures 5 to 8, have been tested. --- ---- 1.00 2.00 3.00 4.00 5.00 6.00 7.00 Time (hrs) I · 0.5AITp-":::'-1Arrp 5 AITp] Figure 5. Effect of cutrent density on the rate of iron(ll) oxidation - --- 7000 6000 ~ 5000 at E - 4000 d c 8 3000 - ;F2ooo 1000 o 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 Time(hrs) ~-. pH1.5 --- pH2.0 pH2.5 ~ pHi] --- --- Fl!Jure 6. Effect of pH on the rate of iron(") oxidation 67 ------------- ----------- --- - - ---- 87 9000 8000 .§. 7000 c .2 6000 - 5000- c 8 4000 c 8 3000 §: 2000 c ,g 1000 0 0.00 8th International ~ on Mine water &the Environment,Johannasburg, South Africa Time (hrs) 5 g/I ton(l) _10 gJI~on(II) _ 15 gJI~on~1}---*"'- 20 gn 1ron(1)J Figure 7. Effect of initial iron(ll) concentration on the rate of iron(JI) oxidation I I ---- 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8. Time (hrs) -+-10 1Tg/lM'1___ 100 rrgIIM'1 1000 rrgIIM1 -*-1 rrgIIM'1 Figure 8. Effect of Mn concentration on the rate of iron(lQ oxidation From Figures 5 to 7, it is clear that the current density, pH and initiallron(lI) concentration had no effect on the rate of oxidation. The Mn concentration in the AMD as electrolytic 68 --- --- ---- --- - - ---- - - --- 88 ------ .---- I 30000 25000 :.1 a. 20000 I - - .. n - .§. ci c 15000 0 u - - 10000 CI u. 5000 y y , . . . . . . . . 0 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.001 6000 5000 "..J' - c. E 4000 - U 3000 u §: 2000 QI U. 1000 0 I 0.00 I I I I L -- - ..- -- ----- 8th International Congress 011Mine Water & the Environment, Johannesburg, South Africa solution, however reached an optimum at 10 g/i of Mn, resulting in the highest oxidation rate for iron (II). 3.3 CONVERSIONOF HYDROGENSULPHIDETO ELEMENTALSULPHUR The presence of sulphide, produced as a waste product after the Biological Sulphate Removal Process, may affect the sulphate reducing bacteria (SRB) in several ways. Because of its detrimental characteristics, it is forbidden to drain sulphide containing effluents to sewer pipes or surface waters. Since a number of physicochemical methods require large investment and operational costs, e.g. high temperatures, high pressures or special chemicals, the continuing search for more economical methods has led to partly investigating this Issue of purifyingH:zS. The oxidation of the iron(lI)-rich effluent (AMD) as a pre-treatment stage prior to limestone/lime neutralisation, produced an iron(lII)-richwater that can be reduced back to Iron(lI)by contacting it with the waste product, H:zSof the biologicallystage, as illustrated by reactions 8 and 9. The result of reactions 6 and 7, combined with reaction 8, proves the production of sulphur from H~ gas and illustrated in reaction 9. H:zS+ 2Fe3+ _ S + 2Fe2+ + 2H+ 2Fe2++ 2H20 ~ 2Fe3++ H2+ 20H- H:zS -7 S + H2 [8] [9] A major advantage of this new process Is that a potentially valuable end-product is produced without large cost implications. No additional energy, e.g. pressure or temperature, needs to be applied. The produced iron(lI)-richwater can now be recycled back to the oxidation stage for re-use in the production of hydrogen_ During an experiment, it has been found that at a specific pH of 2.5, the maximum amount of sulphur can be produced from the H:zSgas, and is illustrated in Photo 1. Table V contains the values for the amount of sulphur, produced per litreof iron(lII)-richwater, when contacted withthe stripped H:zSgas. Photo f Effect of pH on the amount of sulphur produced when contacting itDn(IlI)-dch water with stripped Hz!>gas 69 89 Mass S pmduced (g) 0.014 0.122 0.217 1.565 0.089 3.4 COST ANALYSIS No mass lost occurred when slainh st& (Ssteel) electmdes were usad, provig that the elechaks are Fesistant to axmsion. Wi AMD (which k a waste podud) senred as el-, the he content in the AMD sen& as a reducing agent and therefore limitad amcunt of wen was produced. Stainless steal in AMD as elecbdytic medium was found to be the cheapest way d generating hydmgen eWmMidy and was therefore used as benchmark to calculate the cost eRciency of the oher possible ekbubtic &ups. The electmlytic poduction cost of hydrogen is mainly inlluenced by the voltage in the eledrdyk cal whlch is dmdy related to the resistance in the cell setup, and the cost of elecbidty. The Fm the cost analysis, as liW in TaWa V, it can be nded that iron(ll) oxidation is an excallent benefit added to the ekddytk pod- of hydmgen. It wlY however not pmduce enough hydrogen as anergy source to the sulphate radudng baderia to remove all the sulphates in the water that needs to be Wed biically. For example. an AMD stream mntaining 4 580 mglt imn(ll) and 9 150 mgK (see Table I) wid wly be able to debver 0.08 mdeR of hydrogen which is only emugh bactetM enargy to redum 2 OM) mgll~0,~. The mmblnafion of ddtei (Ni) et&mdes in an electrdytic medium d KOH (30%) will therefore be an economlcdly dtema(hra way of pmdudng sxba hydroBen dddytblly to remove wlphate mncenbabons higher than 2 WO mgit 8th intemalimal Congress on h4me Wafer & the Env~mrnmt, Johannesburg, South Africa 4. CONCLUSIONS Hydrogen can be produced electrolytically, 52% cheaper than buying it in bulk wrnrnercially. Stainless steel electrodes in acid mine drainage as electrolytic medium were found to be the most cost effective way to generate hydrogen electrolytically. The use of a membrane has the advance that the purity of this hydrogen will be of such a standard that it can be used as energy source to sulphate reducing bacteria in a biological sulphate removal process. . If higher volumes of hydrogen is needed, it can be generated 42% cheaper than buying it from industry, making use of nickel electrodes in an electrolytic medium of KOH (30%). One major benefit in the generation of hydrogen electrolytically by means of stainless steel electrodes in AMD was the oxidation of iron(ll) to iron(ll1). This reaction is benefcial to the down stream processes as iron(ll) preapitates at a lower pH than imn(lll) and will not form acid further down in the process. During neutralition, the formation of an Fe(OH)3 layer onto the calcium carbonate partides affects the dissolution rate of the calcium carbonate negatively which will have a direct effect on the rate of neutralisation. It is therefore vital to assure oxidation of iron(ll) in AMD as pre-treatment to neutralisation of the effluent by means of limestone and lime. Afler the AMD has been lreated biologically to remove the sulphates to below 200 mgll. unwanted and toxic HzS gas was produced as waste product. An alternative way of treating this gas economically had been developed by contacting the HS gas with part of the imn(lll)-rich effluent, resulted from electrolytic oxidation of iron(ll) as pre- treatment to neutralisation. This resulted in a low wst and uncomplicated conversion of HzS gas to a valuable end-product, i.e. elemental sulphur. REFERENCES 1. Barnes. H.L. & Rombewer. SB. Chem&ie ofaddminedWnage. J. WPCF.Vd. 40(3). 1~x0. pp. 371-384. 2. W PIeeS L.A. Odendwl. J.P.. Maree. J.P. and PmMnby, M. BakgicaiRem~~Id~ale bm indushiaIERluenls UsingRduosr Gas a Eneqy Soum Envimnmental Tedao(ogy RSA Vd. 13(9), 1992. p. 875. 3. Van Houtm. R.T., Van Spl. H., Van der Aebl. A.C., Hulshoff Pol. L.W. and Le*nga, G ewcglcaf solfale rsductan usingsyn~J~aseg~ asenergy endcarlron~~~m,. Bbtechmtogyand Bioengmmmhg. Vd.Eq2). 1996, pp. 138144. 4. OudeElferink. S.J.. Vnset. W.H.. Hulrhdt L.W. and Slams. AJ.M. SMaleredwSm h me~icbioreactorr FEMS Micrabiigy Reviem. Vd. 15.1994, pp. 119-136. 5. %en4 M. 8 Martin, L. Adivatedn~ekmlainmg elecbode and its use particubdyfor wsterslRbDt,s(r US Patent 4.363.707: W. 18 June 1980:Acc. 14 December 1982. 6. Ovshimky, S.R.. Sapru. K and rang. G. EnqymversM dev&es. US PaDsnt 4,537,674: MI. 19 July 1982: Aac. 27 August 1985. 7. Geldenhuys, A.J.. Maree, J.P.. DeBew. M. and HiMa, P. Anin~ratedlim8hwwn~~-~part,allsulphaDs rmv& Conferewe a, EnviraMantaUy Responsble Mming in Saumem Afrb. September 2W1. PAPER 5: Maree, J.P., De Beer, M., Geldenhuys, A.J., Strobos, G., Greben, H., Judels, C and Dreyer, J. Comparison of the combined limestonellime and combined limestonelbiological sulphate removal process for treatment of acid mine water, Proceedings gh ~ard Rock Mining Conference: Issues Shaping the lndustfy, 77- May 2002, Colorado, USA. In this paper the Combined LimestonelLime process to treat acid mine water was compared to the Combined LimestoneIBiological process for the same purpose. In the Combined LimestonelLime process, acid water (Berkeley Pit water), containing iron and aluminium, was neutralised (pH 6) using limestone, followed by lime treatment (pH 12) for complete removal of metals. Sulphates were partially removed through gypsum crystallisation to less than 1 200 mgle as a result of lime addition. To adjust the high pH of the water, it was contacted with C02 (generated during limestone treatment stage) to precipitate CaC03 that is available for re-use as alkali in the limestone stage. Aeration can be applied in the limestone stage for iron(ll) to be oxidised to iron(lll) and precipitated as Fe(OH),. This results in a very stable sludge that cannot form acid because Fe(lll) is in its highest oxidation state. A disadvantage of this process is the high volume of sludge that requires disposal. The same acid water was neutralised with limestone, followed by biological treatment (SRB), using either ethanol or hydrogen as energy source, to produce a sulphide-rich stream. Sulphide is biologically generated from sulphate. This process offers the benefit that the amount of sludge that is generated is much less than for the Combined LimestonelLime process. Sulphate was only removed to 2 500 mgle. The sludge generated was unstable as it forms acid when exposed to the leading to the oxidation of sulphide precipitates. Paper 5 was presented as a poster by J P Maree at the gh Hard Rock Mining Conference COMPARISON OF THE COMBINED LIMESTONEnIME AND COMBINED LIMESTONEBIOLOGICAL SULPHATE REMOVAL PROCESSES FOR TREATMENT OF ACID MlNE WATER J. P. hfaree (presenter), M. de Beer, A. Geldeabnqq G. Strobes, H Greben, C. Jude and .I. Dreyer Environmentek CSIR: Water, Environment and Forenry Technology P.O. Box 395, Pretoria, 0001, South Africa Phone: +27-012-841-2285: Fax: 127-015841-2506 hail address: jmaree(a~csir.co.7a Waste ore contammg pntcs results lo leshates d acid wlzr when cxposcd to the etmosphcre and prec~pltatam (as ran) Acld mme nater contains blah concn!mhons of dssolved metals (e Q Iron alum~ruurn. manganese. \ - -, - copper and MC) and ntlpherte, and can have pH values as low as 2 0 The parpose of this investigation was to compare. the following hvo ~ntegrated pmcessa for the tmatment of acid mine water (e.g. Berkeley Pit water): 6 Inteerated Limestonefljme vmcess for uactial suluhste removal in this process the free acid and acid xsociated with iron and aluminium are neutralized with limestone to pH 6, followed by lime treahnent to pH 12 for complete rcmoval ofmetals (e.g. manganese, &, copper and magnesium) and pwal removal of sul~~hate thmurh ewm avstallimiion to less than 1.200 me.4 (as SO,). After sludge seuaratlon - -, . (kbnih settling or filtration) me pH is adjusted to 8.6 by &tad& &e he treated water 2 p~'12 with COz mat is prcdwced in the limestone treatment stage. The WOX pmipitates out ad is ~cycfed for nwtralizahon of furlha acid water r; Integrated Limestone/Biolanical suiphate removal vmcess. In this pnuzsr, limestone irea~ment~ as desCnbed above, is fotloued by biological treatment of a side-- for sulphide produchou. The sutphide-I%% side s#rem is my& to the CaM meat stsge for precipiion of the me& x metal sulphides. Snlpbide is produced &om sutphate through biological treatment using eih ethaool or hydmgen as the energy sow. A spreadsbeet-based, water flow and chemical mass balance model was developed to determine various parameten (chemical composition of the watcr after each treatment stage, chemical consumption, size of various remr units, capital and mmhg eost) far the two treatment options. Laboratmy s&uiii were alu, canied out to determine the chemical composition of the water afta various treatment stages and the reaction rates of the slow chemical reactions (bi0logic;il sulphate reduction, gypsum crysWzrdion and calcium cartmnate wystalliation). The foltowing findings were made ficm modeling, beaker studies in the laboratoff and continuous pilot-scale studies m the field: Sulphate was removed to less than 1,200 mgl (as SO,) and complete removal of the metals including magnesium and aluminium wx achieved (Table 1). If aeration was applied in the limestone stage, in@) was oxidized to imnw) and pmIpirated as Fe(0Hh. In this case a stable sludge was produced and cannot form acid water due to further oxidation of the md. This profess p.oduces good quality water Howeyer, large sludge volumes need to be disposed of LimestoneBiological sulphate remowl: Sulphale was moved to 2,500 mg/l by =psum qstallization and complete metal iomoval (Table 1). Aluminium was precipitated as Ai(OH), and the other metals as sulphdes. Magnesium remained in sol&on This pmxss offers the benef~t that less sludge is pmduced bemuse less sulphate is removed from the water and becaw metals are precipitated as snlphides and not as hydroxides. The sludge produced, however, is unstable as sulphide precipitafes will oxidize and acid will be generated when exposed to atmosphere. Table 1. Chemical camposition of feed and treated water PAPER 6: Adlem, C.J.L., Geldenhuys, A.J., Maree, J.P. and Strobos, G.J. Examining the implementation of limestone neutralisation technology in the mining and industrial sector to neutralise acid and reduce sulphate pollution, Proceedings ifh Annual lndustrial Water Management and Treatment Symposium, 151 6 May 2002, Johannesburg, South Africa. This paper focussed, in a very practical way, on the implementation of a limestone neutralisation system for AMD, to replace the current lime system that is in place at various mine sites in South Africa. A new limestone handling and dosing system has been developed for dosing powdered limestone instead of lime. Dust problems that are part of the old-fashioned lime treatment technology are avoided when using limestone instead as it has a content of 15% moisture. A single stage process, the Integrated LimestoneJLime Treatment Process, for the simultaneous oxidation of Fe(ll), neutralisation of acid and the crystallisation of gypsum from the treated effluent, has been developed. Where lime is largely replaced with powdered limestone. A small amount of lime is dosed after limestone neutralisation to ensure sulphate removal to below the saturation level of gypsum, i.e. 1 500 mgle. The reactor is a fluidised bed that runs under high suspended solids conditions with air and acid water being fed to the system. The limestone-based approach has been shown to effectively compete with lime neutralisation in terms of efficiency (utilisation) and effectiveness (rate of reaction). The benefits are that two aspects of neutralisation technology are successfully combined: J Limestone treatment for neutralisation of acid mine water and partial sulphate removal is cost-effective and generates carbon dioxide, and J Lime treatment to high pH (>12) reduces sulphate concentrations to below the saturation level of gypsum Paper 6 was presented orally by C J L Adlem at the Sh Annual lndustrial Water Management and Treatment Symposium Examining the Implementation of Limestone Neutralisation Technology in the Mining and Industrial Sector to Neutralise Acid and Reduce Sulphate Pollution C J L Adlem, A J Geldenbuys, J P Mame and G J Strobos Division of Water, Environment and Forestry Technology, CSIR P 0 Box 395, Pretoria 0001 INTRODUCTION The CSIR has developed the biological sulphate removal process (since 1981) and the limestone process since 1991. Both were developed in response to the acid water problem in South Africa. The purpose of this document is to describe the implementation of the limestone neutralisation based technologies in the mining and industrial sector. Acid water is a problem at many of South Africa's mines. The water needs treatment, e.g. neutralisation, before releasing it into the environment. Untreated it pollutes water sources, threatening aquatic life and safe human consumption. Re-use of these waters is also not possible without treatment. Most of the ores and coal mined in South Africa contains pyrite. Pyrite oxidises to form sulphuric acid in the presence of water and air. This reaction is often faster due to bacteria that catalyse the oxidation of the Pyrite. The neutralisation technology has reached the stage of full-scale implementation and the biological sulphate removal technology is currently evaluated for full-scale implementation. EXTENT OF PROBLEM To illustrate the kind of problems that need to he addressed by mines with acid water problems some data from a case study at a Witbank coal mine will he discussed. CSIR have been involved with the mine for several years and recently implemented the limestone neutralisation process at the mine. Column studies carried out on typical coal discard from the mine showed acid was produced consistently at 1140 gltlweek for 23 weeks (Figure I). It shows that acid water must be handled both as a short term and a long term problem. Many mines and associated industries must handle large volumes of similar water daily. Looking at a water balance (Table 1) for the coalmine near Witbank one sees the extent of water quality problems caused by pyrite oxidation. Their water circuit receives 17.3 t/d (as CaC03) acid of which 15.6 tons is from the pyrite in the coal. The sulphate received from the pyrite oxidation process is 16.8 t/d (as SO4). The water with this load needs to be neutralised to protect equipment pipes etc. as the mine re-uses the water in their circuit. From the neutral water about 4.Wd of the sulphate crystallises in the coal processing plant. Scaling causes an increase in maintenance and operating costs One example is the precipitation of gypsum on magnetite particles. A higher fraction of the magnetite is wasted instead of being recovered due to the loss of magnetic properties. 0 2 4 6 8 10 12 14 16 18 20 22 24 Week - Aeldii - Fe (total) Figure 1. Acid Mine Drainage produced by a typical coal discard. Table 1. A water balance for a typical coal colliery near Witbank so4 STAGE - - INPUT I 17.3 ACID 24.1 Discard OUTPUT Effluents NeutralisationICrvstall~sation The gypsum saturation index (OSI) (Table 2) shows that neutralisation of acid waters formed by pyrite oxidation (sulphuric acid) causes over-saturation with regard to gypsum. As this crystallisation of gypsum is slow it causes scale in the coal processing plant. t/d CaC07 Neutralisation Plant Coal Processing Plant (CPP) Fines Discard dam t/d SOL Inflows 9.6 17.3 96 7.7 11.0 24.1 16,l 8 .O 1.7 69 0.0 1.7 0.2 4,5 3 -2 7.3 Table 2. The gypsum saturation index for a coal colliery near Witbank Colliery Place OSI Neutralisation plant out I 1.07 1 Fines discard dam 0.99 Neutralisation ~lant feed 1.02 1 Coal ~rocessing ~lant Thickener overflow LIMESTONE NEUTRALISATION TECHNOLOGIES 1.47 1.23 In response to these acid water problems and similar problems in the mining and industrial sector CSIR developed a suite of technologies. The aims of these developments can be summarised as follows: Reducing scaling of equipment Reducing neutralisation cost Treating effluent to meet discharge specifications Technologies developed were: Limestone handling and dosing system Limestone neutralisation process Integrated process for Fe(I1)-oxidation, limestone neutralisation and gypsum crystallisation Integrated limestoneAime treatment process Biological sulphate removal process Lime is currently used to neutralise the acid water. It is a fine powder that effectively and efficiently neutralises acid at a fast rate as it is relatively soluble (0.15%) (Maree et al, 1992). Metal hydroxides e.g. iron and aluminium hydroxides precipitate. Sulphate also crystallises as gypsum to a saturation level of about 2000 mglt dependent on the presence of other metals and salts in the water treatment. Slaking equipment is also known for maintenance costs and problems. The most serious disadvantage of lime is the hazards that operators can be exposed to when handling the chemical. Lime is more costly than limestone (2 to 4 times). Smaller sludge volumes are produced with limestone than lime. However, in contrast with lime, limestone reacts relatively slowly and inefficiently with acid mine water. Systems developed and implemented before the CSIR process lead to wastage of unutilised limestone. A new limestone handling and dosing system for dosing powdered limestone instead of lime has been developed by CSIR. Handling it, using the new system, the limestone is wetted as part of the procedure, preventing dust problems often found when handling lime or limestone. Maree et al (1998) developed a single stage process for the simultaneous oxidation of Fe(II), neutralisation of acid and the crystallisation of gypsum from the effluent treated. The reactor is a fluidised bed that runs under high suspended solids conditions with air and acid water being fed to the system. Sulphate is removed to the saturation levels of gypsum. Maree et a1 (1992) showed that a fluidised bed reactor and its entrenched principles can establish a limestone based approach that effectively competes with lime neutralisation, in efficiency (utilisation) and effectiveness (rate of reaction). The integrated LimestonelLime Treatment Process, can be used for neutralisation of acid mine water and partial sulphate removal to below the saturation level of gypsum. The benefits of this process are that two aspects of neutralisation technology are combined successfully: The cost-effective limestone treatment for neutralisation of acid mine water, and partial sulphate removal, that produces carbon dioxide, and lime treatment to high pH (>12), that can reduce sulphate to below gypsum saturation levels. Limestone Handline and Dosina System The dosing system is an inclined concrete slab and a slurry tank. The limestone used is stored on the slab. From there, it is washed into the slurry tank with water. A load cell controls the density of the slurry in the slurry tank. The weight is measured while the tank is kept at a constant volume. Low and high weight settings activate and deactivates the spray of the recycled slurry onto the slab (Figure 2). Some optimum operating parameters were found with experimental pilot scale studies: Recycle slurry to make-up water ratio of a minimum of 4: 1 (Figure 3), a slurry density of approximately 14% in the slurry tank, and the need for nozzles to slurry the limestone e.g. an open ended pipe. The load cell was found to be an effective control for the density of the slurry in the slurry tank under continuous operating conditions. Studies showed that the linear velocity of the water on a gaslab would be enough to wash off particles on a smooth concrete surface. The benefits that this system has in comparison to a lime neutralisation plant set-up is: 0 No silo for storage required Reduced use of lime slaker (where applicable) Reduced alkali cost CaC03 safe to handle No dust Using this data a full-scale plant was designed and constructed at the Witbank coal colliery. Within one year from development, the first full-scale plant was constructed last year. I I I lndi surface I Make up tank Figure 2 Limestone dosing system -$ver Row (slurry) 1 t to tank L- A 0.00 -- 0.00 5.00 10.00 15.00 20.00 25.00 Ratio recycle to feed Figure 3 The influence of feed to recycle ratio Case studylfull-scale implementation at Witbank coal colliery A full-scale makeup and dosing facility Limestone was shown to be a viable substitute for lime in the neutralisation of acidic waste water (Maree er al., 1996). A limestone by-product from the pulp and paper with a particle size of -6 pn was more reactive than previous limestone products. If aeration was carried out in the same reactor basin iron oxidation could be achieved simultaneously (Maree et al., 1999). It was agreed by S A Coal Estates (SACE) to substitute lime for limestone at a neutralisation plant that treats 4 to 6 Mtlday of acid water with an average acidity of 450 mglt as CaCO?. A Full-scale limestone dosing system was designed to dose limestone slurry to a converted neutralisation plant at the colliery. The limestone dosing plant has an inclined wncrete slab (14.56 x 10.4 m) with a fall of 5% or 23, a 10 m3 makeup tank in an excavated wncrete sump to enable the slurry to decant into the tank, a 40 m3h recycle pump and a 1,5 m3h dosing pump. The recycle stream is controlled by pneumatic valves to be diiected either onto the slab or back into the tank Signals from the load cell control the valves. A simple ball valve dose make-up water and wntrols the level of the tank (Figure 4). Figure 4 Photo of full scale limestone makeup and dosing plant A policy of minimal operator intervention is practiced by operating the plant automatically. The utilisation of the plant is currently bemg expanded to supply limestone to the coal plant where about 1 200 to 1 600 m3h of water, with an acidity of 500 mg/t as CaC03 have to be neutdised. Conve~sion of lime nculmIhotion pht to an integrated iron(#) oxidatiodimestone neuholisafion plant Limestone effectively replaced lime as neutralion agent for acid mine water. The lime neutralition plant that was converted to a limestone neutralisation plant is briefly described below: The neutralisation plant consists of three stages, a conditioning tank with a volume of 30 m3 (3.5 x 3.5 x 2.5m), an aeration tank with a volume of 80m3 (5.7 x 5.7 x 2.5m) and a clarifier tank (also referred to as a turbo circulator). With a floe of 200 m3/hr, the retention time 10 minutes in the conditioning tank and 25 minutes in the aeration tank. Limestone, the feed water and the recycled sludge flows to the conditioning taok The slurry then overflows into the aeration tank. Here the slurry is aerated so that neutralisation and iron(1I) oxidation occm simultaneously. If lime dosing is necessary, it is dosed in the overflow of the aeration box. In Figure 5 (Strobos et al., 2001) it can be seen that the pH of the feed was raised 6om 3 to 6.5 in the treated water during December 2001. 1 t Feed I +Conditioning tank a 4.0 - -- -- Aeration tank 1 ' + Turbo circulator - - 2.0 ! ~- - - -~ - Date Figure 5 pH of the feed and treated water Figure 6 shows the acidity for the feed water and the stages of treatment. As expected the acidity was removed from about 450 mg/t to below 50 mgle. - -~.A~- ~- - - ~- - -+Feed water +Conditioning tank Aeration tank ! Turbo circulator ~ - -- -- EFF?- 01 -Dec 06-Dec 11 -Dec 16-Dec 21 -Dec 26-Dec 31 -Dec Date Figure 6 Acidity of the feed and treated water The results show that free acid and acid associated with metals was removed through this neutralisation process. In contrast lime neutralisation removed less than 25% of the acid associated with metals. Aluminium is also removed from the water, but magnesium and manganese is not removed. Financial savings due to the introduction of limestone The Witbank coal colliery saved about R20 000,00 per month by the replacement of lime with limestone at their neutralisation plant. Table 3 shows the relevant data for a six month period. Approximately R120 000,OO was saved due to the introduction of limestone. In contrast to known science before the development of the CSIR's limestone process, comparison of before and after scenarios show that lime is commonly utilised at an efficiency of 60% and limestone at an efficiency of 90% at the neutralisation plant. Table 3 Calculation of cost saving due to usage of limestone for July to December 2W2 Parameter I I I I Fluidised bed limestone neutralisation process low rate (m3/d) - Acidity (mglt) Total acid load (t/d CaC03) Molecular mass (g) A high limestone concentration and a small particle size favoured the rate of neutralisation, and the reactor configuration (fluidised bed) ensured utilisation efficiencies of almost 85% over a particle size range of 0.3 to 1.7 mm. Powdered limestone of >O.3 mm should be dosed directly to a mixing tank with the acid feed, to prevent the loss of particles that can wash out of a fluidised bed. Iron(III), aluminium(II1) and fluoride are effectively removed, while iron(11) retards the neutralisation process. Sulphate is removed to about 26W mgle from a typical acid mine water if sufficient gypsum crystals are present. PH increases from about 2.2 to 7.0. As the solubility of limestone reduces to almost nothing at a pH of 7.5, no pH control is necessary for a basic process using limestone alone. Theoretical To solve the potential difficulties in treating water containing iron(II), a biological iron oxidation stage has been developed to use with the limestone process where applicable. Bacterial oxidation of iron(I1) to iron(1II) occur in a separate reactor. Many neutralisation plants are also equipped with aerator type systems to chemically oxidize their effluents. The use of either will be determined by the levels of Fe(I1) and the specific needs of the potential user. Current Unslaked lime 4037 45 3 1 .8 56 The capital costs of a limestone neutralisation and lime neutralisation plant were shown to be similar, making the new limestone process an attractive alternative to lime neutralisation. Unslaked Lime Powder Limestone 4037 1 4037 1 4037 Integrated Iron(I1)-oxidation and Limestone Neutralisation Process Powder Limestone 453 1.8 100 The technological and kinetic aspects was intensively studied in labscale beaker studies and then upscaled to a pilot plant. It was found that iron(I1) oxidation rates is a function of iron(II), hydroxide, oxygen and suspended solids concentrations. The resultant rate equation developed from that of Stumm is given below. Chemical oxidation of iron(I1) dominates the reaction at a pH between 4.5 to 5.5. I d[Fe]/dt = ~[F~~+I.[OH-~~.P~~ (Stumm) 453 1.8 56 -d[Fe]ldt = k[~e"l~~.[OH-]' 5.[~2J05.[~~]05 (This investigation) 453 1.8 100 The reactions occurring in this process are: The pilot scale plant could treat 24 Eh of iron(I1) rich acid water with a residence time of typically 6.5 h in the fluidised bed reactor. Typical results for the treatment of discard leachate treated in the pilot scale plant is shown in Table 4. The process generated a sludge with a typical concentration of 550 g/E that compares well with the 200 g/t when the High Density Sludge process (a lime neutralisation process) is applied. Disposal of sludge is an important cost parameter for most mines and industries. Table 4 Water Quality of Discard Leachate treated in the integrated oxidation/neutralisation pilot plant Parameter PH Acidity Iron (11) Maximum sulphate removal Cheapest alkali No pH control High density sludge Feed 1.8 Sulphate - Ortho phosphate Case studylfull-scale implementation at BCL Ltd (Botswana) Treated 6.6 7 300 2 500 BCL of Botswana approached the CSIR and its implementation partners for the design of an integrated limestoneliron oxidation neutralisation plant. They discharge 300 m3h of effluent with sulphate concentrations of > 500 mg/C. In addition, acid seepage occurs at a tailings dump and they have to deal with land deterioration. Another cost is the raw water intake of 300 to 400 m3/h. 100 -. Limestone Neutralisation Gypsum aystallisation Neutral water Acid water Iron oxidation Figure 7. Schematic diagram of the neutralisation/crystallisation plant at BCL Completion of the plant is expected in July 2002. CSIR will then be involved in the commissioning of the plant. Integrated limestone/lime treatment process A pilot scale plant was constructed to evaluate the process described. Waste, powdered limestone (CaCO) trom the paper industry was used in the limestone neutralisation stage of the process and slaked lime (Ca(OHh) in the gypsum crystallisation stage. Each stage of the process consists of a reactor and clarifier. All sludges produced were recycled back to the respective reactors. Figure 8 shows the process flow diagram with the water quality of each stage in the treatment process. Limestone neumlisation Lime treatment Ca(OH)z co, treatment pH 2.10 sol- 3 000 ea1+ 420 M!f' 160 Na+ 41 Mn1+ 17 cr 16 AIle -3000 Figure 8: Flow diagram indicating quality of the water before and after treatment in the integrated LimestonelLime Neutralisation Process When acid mine water (discard leachate) was treated with limestone, the pH of the water was raised trom 2.10 to 7.68 in the neutralisation reactor. Sulphate was reduced trom 3 000 to I 900 mglt. The decrease of96.6% in the acidity ofthe water after limestone neutralisation is a result of the following reaction: The stability of the treated water with respect to calcium carbonate is determined by the pH, calcium and alkalinity values of the treated water. The alkalinity of the treated water was 100 mglt (as CaCO) owing to the escape of C02 trom the solution as seen in the following reaction: 106 -- - _ _ 1....__ - - - - - - ---- '11 \if pH 12.26 pH 8.50 SO2- 1900 sol- 1094 sol- 1219 of '" , ea1+ 829 ea1+ 542 Mi' 147 . M!f' 0 Mi' 3.03 Na+ 40 Na+ 47 Na+ 47 Mn1+ 13 Mn1+ 0 Mn1+ 0.01 cr 17 cr - cr AIle 100 AIle 940 AIle 50 U (mtYl) I I (mtYl) I I I (mtYl) From the results in Figure 8, it can be seen that sulphate in the neutral water can be reduced effectively, with lime treatment during gypsum crystallisation, to levels below the original target of 1200 mglt solo, by raising the pH to 12.3. Maximum sulphate removal was achieved by raising the pH to 12 and higher for magnesium precipitation and removal of sulphate associated with magnesium through gypsum crystallisation. The magnesium concentration at pH 7.68 was 140 mglt while at pH 12.26, the magnesium content dropped to zero, due to magnesium hydroxide precipitation. The calcium concentration increased ITom636 to 829 mglt due to the dissolution oflime (Ca(0H)2) as seen in the following reaction: MgS04 + Ca(OHh -+ Mg(OHh + CaS04 The pH of the treated water that was under-saturated with respect to gypsum was adjusted, using C02 ITomthe neutralisation stage, ITom 12.26 to 8.50. Due to CaC03 crystallisation the calcium content was also reduced ITom 829 to 542 mglt. A slight increase in the sulphate content was ascribed to some of the gypsum possibly having been washed out to the third stage of the process (CaC03 precipitation). Gypsum concentration was found to have a major influence on the rate of sulphate removal during gypsum crystallisation (Figure 9). This concurs with the findings of Maree et af (1998), that the rate of crystallisation is influenced by the concentration of gypsum seed crystals. It can be concluded that the rate of gypsum crystallisation is directly proportional to the surface area of the gypsum. The higher the concentration of seed crystals the larger the surface area. Therefore, effective removal of suspended and colloidal matter ITomthe overflow of each stage is required and can be achieved by coagulation and flocculation. In order to optimise this process, the settling rates of the sludges produced during the three different stages were increased dramatically. The polymers, PAC6 and 3095 were effectively used as coagulant and flocculant respectively. The addition of these polymers in very low dosages resulted in clear overflow in each of the above stages with very low turbidities. -. =t 6000 1 ~ 5000 - o ~ 4000 - i 3000 u 5 2000 u .! 1000 ftI -a. 0 ~ 0 -- --- -- 50 100 150 Time (min) 200 250 300 -+-Sludge conc.=1gn Sludge conc.= 50g/1 - Sludgeconc.=10gn ~ Sludgeconc.=200gn Figure 9: Effect of sludge concentration on rate of gypsum crystallisation at pH 12 The combination of limestone neutralisation with lime treatment for neutralising acid mine water achieves complete magnesium removal and sulphate removal to less than 1 200 mglt. The primary neutralisation is done using limestone (much cheaper than lime), followed by lime treatment for metal removal and partial sulphate removal. Treatment of the high pH water with C02 lowers the pH of the water to near neutral levels while a high quality CaC03 is produced. It can be recycled to the limestone neutralisation stage. The treated water will be suitable for re-use in the process or discharge into the sewerage network. A comparison of the costs of treating acid mine water with 107 ---- -- --- - ----- limestone neutralisation, combined with further neutralisation using lime and sulphate removal via gypsum crystallisation is shown in Table 5. Table 5. Comparison of chemical costs for neutralisation of acid mine water and sulphate removal through gypsum crystallisation in the integrated limestone/lime process or by lime treatment Chemical Minimum sulphate Chemical cost (Rlt) Chemical usage Cost of sulphate removal process level (mg/t) (tlt sulphate) (Rlt sulphate) Limestone 1 900 120 1.04 124.80 Slaked lime l I00 675 0.77 5 19.75 US$l = 7ARll-65 Using PAC6 (Montan Chemicals) and 3095 (Montan Chemicals) as coagulant and flocculant, respectively, for improved settling of the sludges, produced in the three consecutive steps of the process, amount to 1.10c/m3 for Stage I, 2.75c/m3 for Stage 2 and 4.13c/m3 for Stage 3. Case study1 full-scale implementation at TICOR (Natal) TICOR approached the CSIR for the design of an integrated limestonellime neutralisation process for their new metal beneficiation plant at Empangeni in Natal. TICOR produces, rutile (TiOz), titanium (Ti), zircon (Zi) and illminite (Ti02 plus iron compounds). The plant also produces 80 m3h of acid water. This water must be heated to a quality suitable for re-use in the metallurgical process or for discharge into the Empangeni sewage system. Neutral water that is undersaturated with respect to gypsum can be re-used. The sulphate needs to be less than 500 mg/t (as SOa) for discharge into the sewage system. The neutralisation plant consists out of the following stages (Figures 10 and 11): CaC03-neutralisation Lime treatment and gypsum crystallisation Filter press and COrtreatment for CaC03-precipitation Figure 10 CaC03-neutralisation (right column) and lime treatment/gypsum crystallisation (left column) stages. Figure II COrtreatment stage. Construction and initial commissioning of the plant was finished at July 200 I. The cheapest alkali available for the neutralisation plant is waste limestone ITomMondi. The CSIR is currently cooperating closely with TICOR to fully optimise and commission the plant. Ongoing involvements also led to a decision to upgrade the limestone dosing system to the newly developed limestone handling and dosing system. Initial results show that pH is controlled effectively at pH 5 and 10 respectively. Sulphate levels of 8000 mglt is reduced to 2000 mglt. CONCLUSIONS A full-scale limestone dosing system was constructed and commissioned successfully to treat acid water for re-use. Lime was replaced by limestone as neutralisation agent saving the company as much as RI20 000 a year (e.g. full-scale plant at Witbank Colliery). Acid water can be neutralised effectively with limestone. In the tluidised bed, limestone, and integrated limestone and iron oxidation processes, sulphate is removed to the saturation levels of gypsum. Lime neutralisation plants can be successfully converted to a limestone neutralisation plant (e.g. conversion of the lime plant at the Witbank Colliery). The integrated limestone/lime process was proven technologically and economically to reduce sulphate of effluents to less than the saturation levels of gypsum producing a water that can be re- use by the mine or industry (e.g. full-scale plant at (TICOR). 109 -- - ------ - - - - - --- - The limestone neutralisation technologies developed at the CSLR are proven, effective and cheaper alternatives to lime neutralisation. REFERENCES Adlem CJL, Strobos JG AND Maree AJ (2001) Cost effective feeder for neutralisation plants, Internal seed report of the CSIR, March Maree, JP, de Beer, M, Strydom, WF and Christie, ADM (1998) Limestone neutralisation of acidic effluent, including metal and partial sulphate removal, IMWA. Proceedings of the 1998 Symposium of IMWA, Johannesburg, South Africa. Maree JP, Strydom WF and De Beer M (1999) Integrated iron(I1) oxidation and limestone neutralisation of acid water. Water Sci. Tech. 39 (10-11) 231-238 Maree JP, Van Tonder GJ, Millard P and Erasmus TC (1996) Pilot-scale neutralisation of underground mine water. Water Sci Tech. 34 (10) 141-149 Strobos JG, Maree JP, Malatji N, De Beer M and Adlem CJL (2001) Progress report No 6- Limestone handling and dosing system. Prepared for: The Mine Manager, December PAPER 7: Maree, J.P., Hlabela, P., Geldenhuys, A.J., Nengovhela, R., Mbhele, N. and Nevhulaudzi, T. Treatment of mine water for sulphate and metal removal using barium sulphide, Proceedings Waste Management, Emissions & Recycling in the Metallurgical & Chemical Process Industries, 18-1 9 March 2004, Johannesburg, South Africa. High volumes of mine water are generated in the Gauteng region and also in the Olifants River Catchment in Mpumalanga. Several processes are under consideration to remove sulphate from the water, e.g. the Biological Sulphate Removal Technology, SAVMIN, Etringite, EcoDose, reverse osmosis and electrodialysis. The barium process can also be used and offers the following advantages: J Sulphate can be removed to specific values due to the solubility of barium sulphate J The soluble barium salt, barium sulphide, can be regenerated from the product, barium sulphate The specific aims of the work were to: J Demonstrate that sulphate can be removed to less that 200 mg/e by means of barium sulphide treatment J Determine the optimum process conditions for the following stages: Partial sulphate removal through lime pre-treatment Reduction of barium sulphate to barium sulphide H2S stripping and processing J Estimate the running cost of the process The process consists of the following stages: J Lime pre-treatment for partial sulphate removal J Removal of sulphate as barium sulphate to below 200 mgle by means of barium sulphide treatment J H,S-stripping with C02 gas J Stripping of C02 and crystallisation of CaC03 4 Regeneration of barium sulphide by heating the barium sulphate, resulting from the process The following conclusions were drawn from the results, obtained during the investigation: J After lime treatment, the sulphate concentration in the water was reduced from 2 800 mg1P to less than 1 200 mgle due to gypsum crystallisation. Metals were precipitated as metal hydroxides. J The sulphate concentration was further reduced to below 200 mg/e after BaS04-precipitation J Sulphide was removed from 333 mu! to less than 10 mgle, using C02 gas for stripping J Due to CaC03-precipitation after C02-stripping with air, of the water following lime treatment, a lowering of the alkalinity resulted, from 1 000 mgle to I I 0 mg/e J Stripped H2S gas was contacted with an iron(ll1)-solution to produce elemental sulphur J Following sulphur production, iron(ll) was re-oxidised to iron(lll) using an electrolytic step J For a removal of 2 glP of sulphate, the running cost of the barium sulphide process amounts to I32.12/m3 Paper 7 was presented orally by P Hlabela at the Waste Management Symposium for Emissions & Recycling in the Metallurgical & Chemical Process Industries TREATMENT OF MINE WATER FOR SULPHATE AND METAL REMOVAL USING BARIUM SULPHIDE J P Maree, P Hlabela, A J Geldenhuys, R Nengovhela, N Mbhele and T Nevhulaudzi Division of Water, Environment and Forestry Technoiogy. CSlR. P 0 Box 395, Pretoria, South Aflica (email: jmaree@csir.co.za) ABSTRACT ~ining is a significant contribulor to water pollufion. Emuents need to be treated for sulphate and metal remvai. The barium pmcess can meet these repuhments. The plvpose of this My was to demonstrate the perfomance of tbe lntegaied barium sulphide process, consisting of the foilowing stapesr Suiphate precipitation as barium sulphate using Mum sulphide, HS-skipping, Crystallization of CaC02 and recovery of barium sulphide from bm'um sulphate. me foHowlng wndusions were made from the investigation: 1. During lime tieafment sulphate was removed fmm 2 800 mgN down to less than 1 200 mg/l due to gypsum nystallizatfon and metab were precMated as mefa1 hydroxides). 2. Owing Bas tieatmeot, suiphaate was removed down to less than 200 mgV due to Basor precipitation. 3. sulphide was removed from 333 down to iess than 10 mg7(as S) in the shipping stage, using CO? gas for strippbgng 4. The shipped HS-gas was contacted with an ilvn(1NJ-solution and converted quantitatively to elemenfal sulphur. 5. The alkalinkl of the calcium bicahnate rich water was reduced from 1 WO to If0 mg/f (as CaCOJ aEer CO?.shipping with au due to CaCOrprecipitatbn. 6. imn(llJ, affer sulphur pmdudbn, was ~~-~xidned to iron(1IIJ using an ekdmylic step. 7 me ~nnlng cost of the barium suiphide pmcess amounts to RZ. 1%' fw the removal of 2 PA. of sulphate. I INTRODUCTION Mmng rs a slgnlficant contributor to water pollution. the pnme reason being that many geological formatfons mat are mined contain pyrites which is Oxlalsed lo sulphunc acid when exposm to air and water. Mine water also has high wkntrations of metals such as magnesium.. his water is receiving major attention due to its poor quality and large volumes which impad on the qualtty of surface water. The volume of mine water generated in Gauteng is estimated at 200 Mud while in the Olifants River Catchment in Mpumalanga at 50 Mud. Legislation requires rwnovai of sulphate to less than 500 mglt Several pmcesses are currently considered for sulphate removal, e.g. biological sulphate remwal pmcess. SAVMIN. Etringiie, ecoDose. revem osmosis and eledrodialysis. The barium process can also be used for sulphate removal and offers the following advantages: Sulphate can be remwed quanlcatively to specific values due to the low solubilly of barium sulohate. ~hb soluble banum salt banurn sulphlde can be recovered tm the product. Darlum sulphate Kun dudled the removal of sulphate with banum carbonate and obtamea good results [I] However, he ldentifled three problem areas, namely. the requirement of a long retention time, high concentrations of soluble banum in the treated water when more barium carbonate is dosed than doichiometricaily reqllired. and the high cod of the barium carbonate. Volman ovemame the ood problem by demonstrating that the barium sulohate omduced could be reduced efficientiv and ecnnomicaflv with mai under thermic condtions ~~ ~~-~~ lo pmduce'banuh sulphide [a. This compouna can de used directly fo;the -ss or &nierled to banum carbonate Wllsenach demonstrated the economlc viablny by calculating the cosl of producing wrium sulphide from barim sulphate 131. Trusler et a1 developed a barium carbonate method in a two-staged fluidised bed reactor system to overcome the other pmbiems identified by Kun, namely, long retention time and the high barium concentration in the treated water [4],[1]. However, the balium carbonate process was found to be unsuitable for water containing metals, as is the case with some mine waters The barium carbonate crystals became inactive when coated with metal hydroxide precipitates. Maree et al. also saw a disadvantage of the barlum carbonate pmcess in the separation of barium sulphate and calcium carbonate, which co-precipitate [q. The purpose of this study was to demonstrate the performance of the integrated barium sulphide process. consisting of the following stages (Figure 1): . Lime pre-treatment for paitial sulphate removal. . Removal of sulphate as barium sulphate to below 200 m@t by means of barium sulphide treatment. . H,Ssriooina with Codas. - - . sinppln$'of & and crystall~at~on of CaCO, . Produdlon of banum sulphlde by heatlng the bar.um sulpnate produced from the above sage The specific aims were to: . Demonstrate that sulphate can be removed to less than 200 mgn with barium sulphide treatment. Determine oDtimum conditions for the following process-staaes: partial suiphate removal through ~mepre-treatment: Reduction of banum sulphate to barium sulpnide. o H,S-stripping and processing. - Estimate the running cost of the process. Feed water Bas r Figure 1: Process flow diagram for the barium sulphide process (pre-treatment M lime not shown). 2 MATERIALS AND METHODS 2.1 Feedstock Artificial feed water containing 2 500 m@t sulphate and 292 mg/t magnesium (1.96 g/t MgSO+7H2O, and 1.01 mUe H2S04) was used during continuous pilotscale studies. Lime (supplied by Ume Distributors) and barium suiphide (supplied by G&W Base Minerals) were used for pH adjustment and sulphae remwal, respectively. For H2Sstripping studies, a synthetic sulphide rich feed water, with sulphide concentrations between 700 to 800 mgR was used. C4 gas (supplied by Afrox) was used for H2S-stripping. An Fe2(S0& solufion (1 1 gl)? as Fe was used for absorption of the stripped H~S-gas. During thermlc studies, BaSOd pmduced during continuous pilot-scak studies and chemically pure BaSO, were used for Bas recovery studies. 22 Equipment Figure 1 shows the laboratocy-scale plant thaf was used for suiphate removal with barium sulphide. Figures 2 and 3 show the Iaboratoly-scale plant that was used for H2Sstripping and H2S-processing to elemental sulphur. Table 1 shows the volume and dimensions of the various reactors depicted in Figures 1. 2 and 3. Photo 1 shows the tube furnace that was used for themic studies. The tube had a diameter of 40 rnm and was 530 mm long. In mwn 8, the sulPhide solulla, was reoycled lhmu@ me ven(uri sydem in batch mod., and ;Illoved fa Wppinp, by .%cWl h COrpaor gas, hence EontadinO a wah ths Yrlphide-641 watw (m -1. The JblW -rich e8s was psaed UmuOh Um psdced bed-csador to whhh the hon(lll>sollianwasmnUMwslyfedfarPllphUfwn. Tabb 1: Volume and cUmensms of various mdors. Feed water (2 500 mglf SO,; 83 mumin), lime slurry (10%. 3 Mmin), barium sulphiie siurry (57.7 glc; 3 Mmin) and Flocculant 3095 (3 Mmin) were fed to the system shown in Figure 2. Sludge was recycled from the under flow of the two clarifiers to the completeiy-mixed reactors at a rate of 83 Mmin. Sludge was withdrawn periodically to maintain the solids content in the he treatment reactor at 40 glt and in the BaS-lreatment reactor at 32 @L Batch studies Batch studies were carried out on the barium sulphide treated water for H2S-stripping and softening. H2S- stripping was achieved by bubbling C02 through while soflening was achieved by dosing 5 glf CacQ and stripping C02 with air. Configuration A Configuration B Figure 2: H2S-stripping and processing Figure 3: H2S-stripping and processing Hs-s6iMng and pmcessing In configuration A (NaS fed continuously to the packed bed-reador and contacted with an iron(1ll)-solution which was passed through a venturi system in batch mode), Na& was contacted with Cotgas with varied concentrations and flow rates. Suiphide removal was monitoredin the feed water afler the feed pipe at the inlet of the packed bad-reactor and the treated water. The imn(l1)-cmcentration was monitored in the imn(1fI) solution that was handled in batch mode. In conftguratlon B (an imn(lli>solution was fed to the packed bed-reactor and contacted with stripped H~S- gas from the venturi system), iron(ill) solution was fed at various flow rates. The resulting ironol) was monitored in the feed and treated slreams of the imn(lll)-solution. The sulphide concentration was monitored in the sulphide solution that was handled in batch mode. Themrsl studies BaSO, (industrial grade and Pure Bas04 and coal were mixed and readed at elevated temperature in the tube fumace and the muffle fumace for various reaction periods. Sola samples were Wllected and analyzed for mass loss, sulph.de content and abili to remove sulphate. 2.3.2 Experimentalprogram The following parameters were investigated: Bas-treatment stage Water quality (feed and treated) HzS-stripping and processing stage Reactor type (packed bed-reador and venturi system) . C9-concentration (20% to 100%) C@ : Sulphide ratio . Feed rate of C02 rich stream (0.2 to 1.0 Umin) = Retention time of sulphide solution (Feed rate of sulphide rich stream (0.5 to 2 Umin)) Efficiency of sulphide reaction with imn(li0-solution Themtic studies stage c BaS0,-rauo (2. 2 5 an0 3) Type of furnace (rune fumace and Mme furnace) - Tem~erature (1000 T 1050 'C and 1100 T) ~ea&ion time-(15 min, 30 min, 60 min and 120 rnin). 2.3.3 Analytical Samples were wileded regulariy and fiitered through Whatman No 1 filter paper. Suiphate, sulphide. alkalinity, caldum, iron(ll), mixed liquor suspended solids (MLSS), volatiie suspended solis (VSS), acidity and pH determinations were carried out manually according to procedures described in Standard Methods 161 . Caldum was assayed using atomic absorption spectrophotometry. Addily was determined by Mrating the solution to pH 8.3 using NaOH. Sulphide (in the product from thermic studies) was determined by mortaring the produd, and analyzing for wlphide with the iodine method in a 0.5 gI100 rMBaS solution. 3 RESULTS AND DISCUSSION 3.1 Water Quality Table 2 shows the chemical composition of the feed water before and afier treatment wilh lime and Bas. n was noted that: During pretreatment sulphate was removed from 2 650 mglt to 1 250 mglL This low sulphate concentration was achieved due to the high caldurn wncentration as a resun of lime addition and the solubiliy pmduct of gypsum. Magnesium and other metals were wmpletely removed. During BaStreatment, sulphate was removed according Lo the stoichiometric Bas-dosage (I 000 mgq. - During H2Sstripping with Was. sulphide was removed from 320 down to c 20 mglt(Faure 4). During the saflening stage. 890 mgt CaCG (993 - 103) precipitated due to COrstripping with air (Figure 5). Figure 4: H2Sstripping with C02. Figure 5: COrstripping with air and CaC0,-precipitation. Table 2: Chemical composition of feed water and after treatment with Bas. Parameter I Concentration 1 Feed 1 Lime Bas HzS- I Sonening 32 HSStripping and Processing Stage Sulphide can be removed through COTstripping from 300 to less than 20 mglt (Figure 4). In order to obtain engineering design criteria forfull-scale implementation, the effeds of various parameters on the rate of US- strlmim were determined. using confiauration A ffiwre 2). Bv fsedina a sodfum sulphide solution and a .. " C-as stream, counter current, on a &ntinuous bails Mmughto a packed bed reador, it was noted that: Number of stages. By passing the sulwie solution thrwgh two stages in series at a C4 feed loadMalS feed load of 1.4 g CWg S, sulphide was rehved from 834 down to 376 mgt in stage 1, and to 77 mgR (as S) in stage 2 (TaMe 3). By providing a third stage sulphide could have been removed to less than 20 mgn. The aim, however, with this 'investigation was to identify optimum process mnditions to allow the minimum number of pmcess stages for wmplele sutphMe remwal. Sulphlde is quaniitative!y converied to sulphur as indicated bythe wmspandence between the adual and theorefical values for the ratio: load of iron(l1) pmducedlload of sulphlde remwed (3.65 versus 3.49). More CG was consumed than theoreticallv rewired 11.75 versus 1 38). The hmher C4 consumption can he ascribed to the patdial solubiliiy hf both C& and H2S in the bH rangei to 7. The theorebwl ratios for load of CQ consumedrload of sulphle remwed (1.38) and load of iron(l1) pmducednoad of sulphide remwed (3.49) were calwlaed from Readions 2 and 3. Efled of HRT. Table 4 shows the effea of the NazS feed rate on the suiphide removal. Better sulphide remvai was achieved with increasd Na2S feed rate (lower HRS). An amount of 217 mgl sulphle was removed at a feed rate of 0.5 Umin (HRT = 59 min), compared to only 154 mgit at a feed rate of 2 Umin (HRT = 15 min). The results in this experiment were however negathe in Ule sense that the pH of the treated water was higher than 8 (9.0), despite of the fad that the adual value of dCWdNa,S of 14.53 was higher than the Meoretical value of 1.38. The negative result can be ascribed to the fact that C4 was not completely uiilized due to too link contact time between CO, and the Na,Ssolution. In the next experiments this problem was solved by allowing sumciern contad time by passing the Na,S/CO, mixture through a 5 m pipe with a diameter of 20 mm (volume = 6.3 0. At a feed rate of 1 Umin, the hydraulic retention fkne in the 5 m pipe amounted to 6.3 min. Eflea of COI flow rate. Table 5 shorn the efled of CG flow rale on the sdlphlae rernoval at a constant Na7S flow rate of 1 Umm. By maeastng lhe CO, h rate from 0 19 Umfn to 0 83 hnm, the sulphlde removal increased fmm 342 to 474 mqt and resldual sulphlde n Mluuon decreased from 134 to o mgit (as S) The corresponolng rabos of CO1 feed losolNaiS few load Increased from 0 78 to 3.46. he stoichiometric value required for this ratio is 1.38 (Reaction 3). This demonstrates that complete suiphide removal can be achieved by dosing 2.5 times more c'& than stoichiwnetrically required. Excess CO, gas would be available in many applications. During barium treatment. C02 is produced at the rotary kiln where barium sulphide is recovered from barium sulphate. Weh the biological suiphate removai process C02-gas is produced by the heating unit. During limestone neutralization of add water, C4 with a high concentration is produced due to dissolution of CaC03. - tnen of C0,-concentrallon Table 6 shows the elfen of the C& concentrarlon on sulphlde removal By lncreasmg the CO, concentrabon from 20% to 100%. Ihe sulphioe removal increasea from 278 to 387 mglt (as S) In this case sulphlde was not removm to 0 mpll as the dC02/dNa,S feed ntlo was less than the value of 3.46 as determined above empiricatly. Effect of gas recycle rate. TaMe 7 shows the effect of the gas recycle rate on the sulphide removal. By increasing the gas recycle rate from 9.1 to 19.6 Umin. the sulphide removal was impmved from 304 to 666 mgK In this experiment sulpMde was also not removed to 0 mQ/l as the dC02/dNa2S feed ratio was less than the value of 3.46. It was aemonstrated above that a packed bed-reactor (configurat~on A) (ngure 2) can Ce used fw sulphme sfnpplng In lhs wnllgurat~on it appeared that the absorption stage. where Hsncn gas was contaclec wlfh an iinflli) solution in a venturi system, was effective, due to gwd contact between gas and liquid phase. Wlh the apparent good performance of the venturi system for HIS akxirption, kt was decided to evaluate the suitability of the venturi system also for &Sstripping. The same equipment that was used for configuration A (Figure Z), was used for configuration B (Fgure 3), except that the venturi systMl was used for sulphide stripping in batch mode, and the packed bed-reador was used for H,Sabsoption into an iron(l1O solution under continuous conditions. Table 8 and Figure 6 show the effea of iron(lil) feed rate on sulphide removal. It was noted that Better sulph.de removal was achieved with increased feed rate of iron(lli). T-ds can be ascribed to only partial absorption of &S at low iron(l1l) feed rates in the closed circuit of Configuration 2. This is an indication that the packed bed-reactor does na fundion as well as the venturi system for the absorption of Ha into an iron(lll>solution. The experimental (actuaO dFe/dH,S ratio was similar to the theoretical value of 3.49 (Read~on 2) This resun shows that all iron(lll) that was introduced to the packed bed reactor was consumed for H&absorption through the readion. The results showed that HS-stripping and MS-abso@ion is favoured by iniensive mMng. Intensive mMng supporn mass transfer of H2S from liquid to gas phase in the case of H2S-stripping and from gas to liquid phase in the case of W-absorptmn. It was demonstrated that the venturi device was more eftkient than the packed-bed reactor. This could be ascribed to the high pressure (300 kPa) and the high velodty (50 m/sec) of gas and liquid particles. Based on this observation. B is recommended that the Turbulator be used during a scale-up version. The Turbulator exists out of a motor which directly (no gear box) drives a dic via a hollow shaft. The Turbulator allows mixing between the gas and liquid phase by sucking in air through the hoiluw shaft that PnateS at 2000 rm. The velocity at the outer limit of the disc is 15 m/sec @ia = 0.15 m; rpm = 2 000). Table 3: Sulphide removal in two stages in series. Parameter CO2/Na2S-feed ratio (9 COdg S) Na,S feed rate (Umin) C02 feed rate (Umin) C02-concentration (56) HRT (min) Gas recycle rate (Umin) Sulphide in feed (mglt S) Sulphide afler pipe (mg/lS) Sulphide in treated water (mg/t S) Suiphide removed (mglo pH in feed pH afler pipe pH in treated water dC02/dNa2S ratio (g COdg S) . Theoretical Actual dFe/dH2S Theoretical Actual state 1 Table 4: Effect of retention time (Na,S feed rate) on sulphide removal. NaS inside table change Na2S Parameter COrcDncentratian (%I HRT Wn) Gas recycle nb (Umn) Sulphide in feed (mgl(SJ SUIphtde aner pipe (mg/l S) sulphide in tmted water (mg/( s) SdpMde removed (mg/l) pH in faR( pH sftrr pip pH in lreated water mber 2 1.40 0.90 0.24 100 32.7 22.9 Table 5: Effect of the COZ feed rate on the sulphide removal. Parameter CO,INa,feed ratm (g COdg S) NaS fad rale (Umin) cormtcenV.tion (%) HRT (mn) Gas recycle rate (Urnin) sulphlde in feed (m@ S) Sdphlde alter plp (mglt s) SulpWs tn treated water (mgn S) Sulphide removed (mglt) M in feed pH met pipe pH in hated wstm CO feed rale Urnin sl=%iFk Table 6: Effect of CG-concentration on sulphide removal pH in feed p~ ancr pips pH in beatcd Wlsr Table 7: Effect of gas recirculation rate on sulphide removal. pH in feed pH aner pipe pH in Oeetcd rrster 7.64 Parameter COJNbl fd r=&a (g Cod9 S) N+S r& ra(e (Umin) CG feed rate (mn) Gconcenoaeon (%) HRT (min) Gas recyUe rsle (Umin) Table 8: Effect of the iron(ll1) flow rate on the sulphide removal I I dFeidH2S Theoretical 3.48 3.49 3.49 I 3.49 4duai 2.87 , 3.49 8.75 3.12 C%concenlra ism 0.78 1 0.26 100 23.5 19.8 ram 1.54 1 0.48 1W 29.5 13.1 Figure 6: Effect of imn(ill) flow rate on the sulphida removal. 3.3 Thermal Studies Table 9 shows the effed of various parameten on the Bas yieM during the thermic conversion of BaSO, to Bas. It is noted that: - The conversion efficiency reduce with time when a Mme fumace was used (Exp 1). This is ascribed to the large volume of air surrounding the readion vessel. Initially Bas01 is converted to Bas due to reducino conditions created bv the convenion of wal to CO and C6. When the cartmn IS exhausted. the Bas pmduccd IS coniacted with oxygen af the hlgh tempeiature Wlch allows oxidabon of BaS to BaSO, Therefore, all funher studies (ExpenmeMs 2 to 8) wen, camed out in a tube fumace. The air was also purged wilh ntiogen to eliminate oxidation condffions. A short reaction time of 15 min ns suftident to obtain a high yield of Bas at a temperature of 1050 'C (Exp 2). Figure 7.a shows the conversion of BaSO, to Bas as a fundion oftime. me reaction starts at a temperature of 900 "C and reaches an opWnum at 1050 'C (Exp 3 and Figure 7.b). The minimum CBaSO. mole ratio required for complete reduction is 2 which indicates that the reaction is as folrows (Exp 4 and Figure 7.c. 2C + Bas01 -r 2C4 + Bas Both adivated carbon and coal can be used for BaSO. redudion (Ewpts 5 and 6). SiightJy better values are achieved with activated canon than with coal which wuld be asuibed to impurities (e.g. FeS) in coal. Botn anaiyilcai grade and indbslrial grade B~SO. provided good yields of Bas (Exp 7) MQ(OH)~ does not interfere mfh the reduclion readion of BaSO, to Bas (Drp 8). The methods used for conversion measurements(mass loss, sulpide and sulphate precipitation) wmpares well. The sulphide values were lower than the mass loss values and can be ascribed to sulPhide losses during the dwlution stage. Such losses are confirmed by a sulphide odor that was picked up. me product was also tested fwthe ability to remove sulphate (Exp 5). Although the sulphate method is nof as accurate as the other methods, a value of the same order was achieved. 3.4 Running Cost The running wsl of the banum suiphlde pmcess amounts to ~2.1~m' for the removal of 2 g/t of sulphate and the cost of waier (~21m) ano by-produas, sulphur (RO 3011713 and calaum camonate 4 CONCLUSIONS The following wndusions were made from the investigation: 1. During lime treatment suiphate was removed from 2 800 mg/t down to less than 1 200 mu1 due to gypsum crystallization and metals were precipitated as metal hydroxides). 2. During Bas treatment, sulphate was removed down to less than 200 mgN due to BaSOc precipitation. 3. Sulphide was removed from 333 down to less than 10 mg/l (as S) in the stripping stage, using CO, gas for stripping. 4. The stfipped Haas was contacted with an iron(lll)-solution and converted quantitaiively to elemental sulphur. 5. The alkalinity of the calcium bicarbanate rich water was reduced from 1 OW to 110 mglt (as CaC03 afler CO~striwino with air due to CaCOrDrecioitaiian. 6 lron(ll), iner &lpiiur producbon was reix~d~zed to imn(lil) uslng an eieclrolyllc step 7 The runnmg cost of the banm sulphlde Process amounls to R2 1~m' for tne removal of 2 glt of sulphate. Table 9: Effect of various Darameten on the lhermic conversion of BaSO. to BaS Parameter Value Time (min) Time (min) 15 I 20 Temperature ('C) 900 950 1 i! CAaS0,-mole ratio 3 Carbon Activated Coal Carbon Activated Conversion % dass Sulphids Sulphate i itatlon 80.6 37.5 35.5 75.6 74.4 industrial 0 Tube coal ; 1050 I 1 1050 Activated Pure 0 Tube 1050 Activated industrial 0 Tube 1050 Activated Industrial 0 Tube 20 1050 Aclivated Industrial 0.7 Tube 20 1050 Activated industrial 1.7 Tube 20 1 1050 Activated 1 Industrial I 4.3 1 Tube 1 1 Temperature (deg C) c a, 0. 0 0.5 1 1.5 2 25 3 CIBaS04-male ratio c. C/8aSOcmoie ratio Figure 7: Effect of valious parameters on the conversion of BaSO, to Bas. 6. REFERENCES 111 Kun, L.E. 'A repori on the reduetion ofthe suiphate content of add mine drainage by precipitation with barium carbonate'. internal report of Anglo American Research Laboratories. Projecl No. Dm?. (1972). 121 Volman. R.. The use of barium sulphide to remove suiphate tmm industrial effluents.' Thesis presented for the degree M.Sc. (Chem. Eng). University of Steilenbosch. (1984). 131 Wsnach, i.T. 'Cosl estimate for barium sulphate reduetion.' Internal report of the Division of Water Technology. CSIR. Pmjed No 620/2616/6. (1986). (41 Trusler. G.E.. Edwards, R.I.. Brouckaeri. C.J. and Buckley, C.A.. "The chemical removal of sulphates." Proceedings of the 5th National Meeting of the S.A. institution of Chemical Engineem, Pretmia, W3-0 -W-11. (1988). IS] J.P. Maree, J.P., Bosman, D.J. and Jenkins, G.R. 'Chemical remwai of sulphate, calcium and heavy metals from mining and power slation effluents.' Proceedings of the first biennial Conference of the Water Institute of Southem Africa. Cape Town. March. (1989). 161 APHA.. Standard Methods lor the Examination of Water and Wastewater. Twelfth Editjon. American Public Heaith Association, New Yo*. (1985) PAPER 8: Maree, J.P., Netshidaulu, I., Strobos, G., Nengovhela, R. and Geldenhuys, A.J. Integrated process for biological sulphate removal and sulphur recovery, Proceedings WlSA 2003 - Biennial Conference & Exhibition, 2-6 May 2004, Cape Town, South Africa. The Biological Sulphate Removal Process has been used to remove sulphates in industrial effluents, rich in acid, sulphate and metals, to less than 200 mglC. Previously, ethanol was employed as energy and carbon source. Because of an unexpected increase in the price of ethanol, it became imperative to explore the possibility of finding an alternative source of energy and carbon for the sulphate reducing bacteria (SRB). In this paper, the specific aims were to investigate: J The energy utilisation efficiency when feed water is contacted directly with hot coal gas (SRB thrive between 30 - 35OC) J The rate of sulphate removal in the anaerobic reactor J The effect of reactor type, C02-feed rate, COJHZS-ratio and the efficiency of H,S-stripping ./ The cost of electrolytic hydrogen generation (to replace ethanol as energy source) From the investigation the following conclusions were made: J The Biological Sulphate Removal Process allows sulphates to be removed from industrial effluents to below 200 mgfe, using ethanol as the carbon and energy source J With the use of a venturi system, H2S can be stripped to a concentration below 20 mgfe (as S) with C02. A number of aspects influenced the efficiency of H,S-stripping, namely, H,S feed rate, C0,-concentration, COJH,S load ratio and absorption efficiency of H2S in an iron(lll) solution ./ Hydrogen is a cost-effective energy source for the SRB (R1.53lkg of sulphate removed) Paper 8 was presented orally by J P Maree at the WlSA 2003 - Biennial Conference & Exhibition INTEGRATED PROCESS FOR BIOLOGICAL SULPHATE REMOVAL AND SULPHUR RECOVERY J.P. Maree, I. Netshidaulu, G. Strobos, R. Nengovhela and A.J. Geldenhuys Dwism of Water. Envrronment and Forestry Technolqy. CSIR. PO Box 395. Pretoria. 0001. South Africa Tel. (012~ 841 2285. Fax. (012) 841 250E E-mail. :;nareEScsirco.za ABSTRACT industrial effiuents rich in subhate acid and metals are produced when sulphuric acid is used as a raw material and when pyrites is ox~dised due :o ekposure to the atmosphere, e.g in the mining industrj B~ologica! sulphate removal can be used to treat tndustnai effluents to achieve, in addifion to sulphate. metal removal and neutrabsation Sulphafe can be removed as elemental sulphur via sulphide as an intermediate product when an eoergy source is provided The biologics: sulphate removal process has developed over the past 15 years to the stage where it can compete successfully with other sulphate remova! techno!ogies for full-scale treatment of mine and other ~ndustnal effluents. The aim of this mvestigation was to demonstrate the performance of the integrated process consMng out of hydrogen generation, sulphate reduction wfth hydrogen as energy sourca. HzS-stripping and stabrlization stages. The following conclusions were made from this investtgation. . The biological sulphate removal process can be used for removal of sulphate to less than 200 mgtl using ethanol as the carbon and energy source H,S can be stripped to below 20 mgA (as S) with GO2 by using a ventun as a str(pp1ng device. The efticiency of MS-stripping is influenced by H2S feed rate. CO~wncentration. COa2.S load ratio and absorption efficiency of H2S in iron(ll1) . Hydrogen is the most cost-effective energy source The cost of hydrogen amounts to R1 53/m3 per g.4 of sulphate removed /R1.53/kg SO4 removed). INTRODUCTION Industrial effluents rich in sulphate. acid and metals are produced when sulphuric acid is used as a raw material. and when pyrites is oxidised due to exposure to the atmosphere. e.g. in the mining industry. Acid mine waters contain high concentrations of dissolved metals and sulphate. and can have pH values as low as 2.5. Acidic industrial effluents require treatment prior to discharge into sewage networks or into public watercourses. In water-rich countries the main causes of concern are the low pH and metal content of acidic effluents. Salinity is not a problem due to dilution with surplus capacity of surface water. In water-poor countries. e.g. South Africa, the high salinity associated with acidic industrial effluents is an additional concern. Biological sulphate removal can be used to treat industrial effluents to achieve. in addition to sulphate, metal removal and neutralisation. Sulphate can be removed as elemental sulphur via sulphide as an intermediate product when an energy source is provided. Desalination is achieved by effecting calcium carbonate crystallisation affer sulphate removal. Metals are completely removed by precipitation as sulphides. Alkalini!y is generated in quantities stoichiometrically equivalent to the amount of sulphate removed. which allows direct treatment of acid water^ The btological sulphate removal process has developed over the past 15 years to the stage where 11 can compete successfully w~th other sulphate removal technologies for full-scale treatment of mine and other ~ndustrial effluents Maree and Strydom showed that sulphate can be removed In an anaerob~c packed-bed reactor using sucrose pulp mill effluent or molasses as carbon and energy source (1) Metals lrke nickel cadmium and lead were completely removed due to precipitation of metal sulphides, Maree and Hill showed that a three-stage process can be applied for sulphate removal, using molasses as carbon and energy source in an anaerobic packed-bed reactor (2). Sulphide can be stripped with a mixture of C02/N2 from the effluent of the anaerobic reactor in a H2S-stripping stage and residual COD and CaCO3 can be removed in an aerobic final treatment stage. Maree et al. showed that when molasses is used as carbon and energy source it can either be utilised in the fermented or unfermented form (3). When molasses is allowed to ferment. acetic acid is the main carbon and energy source for the sulphate reducing bacteria. When molasses is kept sterile in the storage tank. sucrose is Me main carbon and energy source with acetic acid as the metabolic end-product. With this information in mind it was conduded that by running two anaerobic sulphate removal reactors in series; sucrose can be fermented to lactate in the first reactor and via acetate to CO, in the second reactor. Du Preez et al. were the first to demonstrate that producer gas (mixture of Hz, CO and C02) can be used as carbon and energy source for biological sulphate reduction (4). Hz and CO were utilised as energy and carbon sources. respectively, Visser investigated the competition between sulphate reducing bacteria (SRB) and methanogenic bacteria (ME) for acetate as energy and carbon source in an upflow Anaerobic Sludge Blanket (UASB) reactor (5). He found that at pH values less than 7.5, SRB's and MB's are equally affected by the presence of H2S. while at higher pH values SRBs out-compele MB's. Van Houten showed that sulphate can be reduced to H2S at a rate of 30 g SOJ(1.d) when H2IC02 is used as carbon and energy source and employing pumice or basalt particles to support bacterial growth in a fluidised-bed reactor (6). He found that the pH should be between 6.5-8.0; the temperature should be between 20 - 3S°C; the H2S concentration should be less than 450 mgll; the system should be completely anaerobic; the biomass should be immobilized and the retention of the active biomass should be high: the gas should be in the ratio: H2:C02, 80%:20%: the hydrogen mass transfer should be maximized; there should be a high gas hold-up (through the system recycle) and there should be a low bubble diameter (small gas bubbles). Geldenhuys et al, demonstrated that hydrogen can be generated on-site to provide it cost-effectively (7). Eloff at al. showed that a venturi device can be used to introduce hydrogen gas into the system as the energy Source, while geotextile (a coarse. fibrous material. used in road construction) can be used as a support material for SRB growth (8). The aim of this investigation was to demonstrate the performance of the integrated process consisting out of a heating unit, anaerobic reactor for sulphate reduction. HzS-stripping and stabilization stages (Figure 1). The following reactions take place in the various stages: 4H2 + SO‘" + C02 + Ca2+ + HS- + HCOi + 3H20 + Ca2+ M" +HS + MS+H' HS +C@ + Hz0 + fizS+ HCO,' H~S + ZF~~' 3 S + 2Fe2+ + 2H' 2Fez+ +%O2 + 2H+ + 2FeJ+ + H20 2HC03- + Ca2+ + CaC03 + COz + H20 The following specific aims were investigated: The energy utilization efficiency when feed water is contacted directly with hot coal gas. * Determination of the sulphate removal rate in the anaerobic reactor. The effect of reactor type. C02-feed rate. COZIHZS-ratio and the efficiency of H2S-stripping. The cost of electrolytic hydrogen generation. Sulphur Coal Sugar i Treated water L - - - Ca(HG032 - C02-nch offgas Fgure 1 Process :low d~agram of Integrated bboioglcal sulphate removal process (CSIRosure process) MATERIALS AND METHODS Feedstock The various stages of the process as shown in Figure 1 were operated in parallel using dtfferent feed waters. The biological sulphate removal stage was fed with effluent from a lime neutralization plant at a rate of 8 to 16 m3/h (hydraulic retention time of 10.3 to 5.2 h), while 0.1 to 0.2 g sugar/! mine water. 0.7 to 1.0 mi ethanol B (75 % ethanol. 25 % propanol)il mine water were added as the carbon and energy source. Ammonium sulphate (25 mgll (as N)) and phosphoric acid (5 mgn (as P)) were added to maintain the C0D:N:P ratio at 1000:7:2. No trace elements, except for 3 mgll iron(ll). were added. as the mine feed water contains all trace elements which are required by sulphate reducing bacteria. For H,S-stripping studies a synthetic sulphide rich water was used as feed water containing a sulphide concentration between 700 to 800 mgll. C02 gas (supplied by Affox) was used for H&stripping. An iron solution (11 g Fe2(S0,)3/1 (as Fe)) was used for absorption of the stripped H,S-gas. The heattng un~t was fed wtth peas (25 mm coal) Equipment Anaerobic stage. The anaerobic stage (Photograph 2) was evaluated on-site. The anaerobic reactor consists of a completely mixed reactor (dia. = 4 rn. height = 8 m. volume = 105.5 m3) with a cone in the top of the reactor to allow for sludge separation (9). The feed-rate to the anaerobic reactor was 8 to 16 m3/h and 0.3 m3!h to the HSstripping stage. The reactor contents were stirred with a side entry stirrer positioned at the bottom of the reactor (260 rpm) and additional mixing was provided by a recyde pump (35 m3/h). The feed inlet pipe entered the reactor at the top from where it fed to the bottom. The reactor was inoculated with 10 m3 anaerobic digester sludge from Daspoorl Sewage Works Pretoria on 6 May 2000. The temperature was approximately 17'~. Heating unit. The heating unit consists out of the following items: coal bunker. speed control spiral feeder (100 kglh). heattng unit and fan and water spray reactor where feed water is sprayed through 318 inch spiral jet nozzles while hot air IS flowing upwards {Photo 2). Photo 1. Anaemb~c stage cf blologlcal sulphate removal process. Photo 2. Heatmg un~t. 4s-stripping and processing. Figures 2 and 3 show the laboratory-scale plant that was used for YS-stripping and processing to elemental sulphur. Table 1 shows the volume and dimensions of the various reactors. A packed bed reactor and a venturi system was used for H2S-stripping and MS-absorption into an iron(lll)-solution, using two different configurations. In configuration A (Figure 2). the sulphide solution was fed continuously to the packed-bed reactor (stripping stage), and allowed to drip down the packing material (25 mm dia. Pall-rings), while H+free COrgas. flowing from bottom to top. was recycled via the H&-absorption stage. In the H&-absorption stage. H& was contacted with an iron(lll)-solution at a pH of 2.5, to produce elemental sulphur (Reaction 2). The iron(l1l)-solution was replaced on a batch basis. In configuration B (Figure 3). the sulphide solution was recyded through the venturi system in batch mode. and allowed for H&-stripping. by sucking in COrpoor gas. hence contacting it with the sulphde-rich water (stripping stage). The stripped H2S-rich gas was passed through the packed bed-reactor reactor to which the iron(ll1)-solution was continuously fed for sulphur production. The packed column consisted of a 0.8-m randomly packed bed. with 25mm Pall-rings used as packing material. A 250-mm Pempex cylinder (adsorption) was used and Perspex plates were used for support of the packing and to aid in liquid distribution to the column. A Perspex plate with evenly distributed holes was installed at the top of the column to ensure adequate distribution of the liquid feed. The venturi was used for gas recirculation between the stripping and the absorption stages. A centrifugal pump was used (capaaty of lm3th) to recycle the iron (Ill) (configuration 1) or the sulphide (configuration 2) solution via the venturi. C& was transferred from a CGcylinder to a COZ float tank from where it was pumped at a set flow rate with a peristaltic pump at NTP (normal pressure and temperature) to the YS-stripping stage. Experimental Procedure .-111cren)hic Rcnclor m~d Iientirig f'nir A volume of 12 m3th of feed water was passed through the heating unit of which 4 to 8 m3lh was fed to the anaerobic reactor and the balance was returned to the feed water pond. The performance of the various stages (sulphate reduction and WS-stripping) were evaluated by determining Me chemical composition and temperature of the feed and treated water and the temperature of the in and off gasses durrng contrnuous operatton /f>s-.5/rip/)iflg 'ld f'fw<'~s.Yillg In configuration A (fed Na2S continuously to the packed bed-reactor arrd contact the stripped thS-gas with an iron(i1i)-solution which is passed through a venturi system in batch mode) Na,S was contacted with 60,-rich gas. while sulphtde removal was monitored in packed-bed reactor. The iron(ll)-concentration was monitored in the iron solution (that vras handled in batch nlodej In configuration B {fed ironllil)-solution lo the packed bed-reactor and contact it with stripped H,S-gas from the venturi system) and iron(llli-rich solution was fed at var:ous flow rates, lron(li) was monitored in the feed and treaied streanis of the iron(l1l)-solution. The sulphide concetltration was rrlor?itored in tne sulphide solution that was haridled iri batch mode. Experimental Program The followtng parameters were rnvesttgated Anaerobrc stage nydraulc retention tlme (5 h 10h 15 hj H2S-stripping and processing stage Reactor type (packed bed-reactor and venturl system). GO2-concentration (20% to 100%). GO2: Sulphide ratio. Feed rate of C02 rich stream (0.2 to 1.0 limin). Retention ttme of sulphide solutton (Feed rate of sulphide rich stream (0.5 to 2 llminji. Efficiency of sulphide reaction with iron(ll1)-solution. Analytical Samples were collected regularly and filtered through Whatman No 1 filter paper. Sulphate. sulphide. alkalinity. calcium. iron(il). mixed liquor suspended solids (MLSS). volatile suspended solids (VSS). acidity arid pH determinations were carried out manually according to procedures described in Standard Methods (10) Calcium was analysed usirlg atomic absorption spectrophotometry. Acidity was determined by titrating the solution to pH 8.3 using NaOH The COD samples were pre-treated with a few drops of E.i2S0, and N2 to strip off HZS gas. TaSre 1. Volume anc: almensions of vartous readofs. 25 vrn Pail-r!?os I I I Ventrrr reactor 40 300 800 item Anaerobic H&stripping/pmcessing Packed kc-reactor DaCked wth RESULTS AND DISCUSSION Sulphate Removal The feed rate of the reactor vaned between 8 and 16 m31h dur~ng the pertod from 1 September 2000 to 26 June 2001 wh~cl~ correspoilded with a hydraulic retent1011 tune of 5 2 to 10 3 h respedwely Table 2 shows the chetmcal cornpositlot? of the feed and treated wate~ Volume (I) 1 DO000 39 it was noted that: Sulptiate was renioved consistently down to 200 mgll. Thrs was achieved when sufficient carbon and energy source was provided. Ethanol was completely utilized for either sulptrate reduction or acetate production as indicated by measurelnent of ethanol and fatty acids. Almost nc forrnate and pr0p:onate were formed during sulphate ieduction. The sulphide concentration iri the effluent is stoichicmetricaliy eqillvalerit to the sulphate concentration. The high sulohide concentratiotls measured in the anaerobic reactor indicated that the suiphate-reducing bacteria can achieve high suiphate reduction rates despite the high suiphide concentrattons This rs contrary to the findings of b4cCartney and Oleskievricz (1 1) Diameter (mm) 4000 250 Height (mm) 8000 ROO who found that sulphate reduction is inhibited by suiphide concentrations higher than 300 mgil (as S). The sulphate removal rate increased to 12 g SOd{t.dj at a temperature of approximately 20°C. at an HRT of 6 h. This rate may Improve still further by increasing the temperature of the feed water. Good sulphate removal rates were achieved despite the fact that the most simple and cost-effective reaLYor type was used. By integrating a completely-mixed reactor with a settler. positioned in the lop of the reactor for sludge separation. the capital cost was reduced without compromising on the residence time. Sulphate removal could still be achieved withtn 6 hours. An AI~,~'SO4,-,,-ratio = 1.0 was measured, which corresponds well with the theoretical ratio of 1.04 iReacbon 3). Alkalinity vatues as high as 2000 mgil were measured with an equal reduction in the suiphate content. The biomass distribution was uniform, bonom to top, in the reactor. Durrng the slart..up period more sludge occurred in :he lower part of the reador due to the presence of heavy chemicals and gypsum. The biomass concentration increased horn 2 500 rngil on 1 September 2000 to 10 000 mgll where it stabilized. The specific biomass production was calculated to be 0.02 g biornassig SO+,-.-. F ~ure 2 Hfi-slrlppi?g and oiomsstng Feure 3 &S-strtpp~ng and processtilg Table 2 Chem!cal campositton of feed and mated >water ciunng Bloioglcai sulphare reciucitcn I Sulphtde Wig// SI fikalioitv (1w4 CaC03: Figure 4 shows the percentage sulphate removal and the COD,,dSO,,,,-ratio with time, It was rioted that sulphate removal increased from between 30 !% arid 50 % during the period before 18 October 2000 to above 75% after 18 October 2003 (Figure 4. Sulphate line). The improved sulphate removal wtth tme could be ascribed to the iticreased value for the C0Dia.d:S04ieod-rat~~ wth time (Figure 4. CODa,&04hd Itne). Before 18 October 2000. this value was lower than the theotettcai value of 0.67 jReact~on 4). Thereafter this value fincreased to between 1 and 1.2 .. . EfhanOl (msil) Acetate (ingl; Formate im*j;q Proplotiale (rngf~ volauie suspended wilds (mg!ij Mix& iauot suscendeo solids (ma!:! 690 i 0 0 0 0 0 21 8 5 3 9 DOD 'i3 Figure 4. Percentage sulphate removal during the perii 1 September to 26 June 2001 compared with the ralio of CODfeed to Wleed. The utilization efficiency of the energy source (ethanol and sugar) and its cost are calculated in Table 3. Table 3 Companson befween measured and calculated COD values. It was noted that: At a dosage of 0.1 gll sugar and 0.72 gll ethanol B (75 % ethanol + 25 % propanol). 2.0 gll sulphate was removed. This represented a utilization efficiency of 75%. The measured COD value of 1.75 gll (as 02) agreed well with the calculated value of 1.78 gll (as 4). The energy source cost assoaated with the removal of 2.0 k@m3 sulphate amounted to R2.ZZlm3. This cost could be reduced further should by partial replacement of ethanol and sugar with carbon monoxide. Carbon monoxide could be recovered when coal off-gas is used for heating of feed water. Heating of Feed Water The plant was operated at a flow rate of 9 m3/hr and an average water inlet temperature of 15%. The feed water was effectively heated from 15 "C to 30 'C. Heat transfer from the gas to the water was more Man 99% effective. The exit temperature of the gas was approximately equal to the inlet temperature of the water. The heat transfer from the coal to the water was approximately 75 to 90 %. Heat losses was due to incomplete combustion of coal and disposal of hot ash. Figure 5 shows the total heat transfer efficiency in the period August to September 2003. 17-Aug 22-Aug 27-Aug 01-Ssp OGSep 11-Sep 16Sep 21-Sep Date Figure 5 Total heat transfer eff~tency in the perlod 17 August to 20 September 2003 Table 4 gives a summary of the energy balance over the direct contact heat exchanger wlumn it can be seen that the spray column 1s a very effecuve configuration to establ~sh heat transfer Due to back mixtng the column allows complete heat transfer between the water and the gas (12) Table 4 Energy balance over the dlrectcontact heat exchanger wlumn HIS-Stripping and Processing Stage Sulphide can be removed through COz~stripping from 300 to less than 20 mgll (see Figure 6 and Maree (13) for a more detailed discussion on H2S-processing and stripping). In order to obtain engineering design criteria for full-scale implementation, the effect of various parameters on the rate of HzS-stripping was determined. using configuration A (Figure 2). By feeding a sodium sulphide solution and a COrgas Stream counter current on a continuous basis through to a packed bed reactor, it was noted that: Number of stages, By passing the sulphide solution through two stages in series at a C02 feed load/Na2S feed load of 1.4 g C02!g S. sulphide was removed from 834 to 376 mg:l in stage 1 and to 77 mil (as S) in stage 2 (Table 51. By providing a third stage. sulphide could have been removed to less than 20 mgil The aim. however, with this investigat~on is to identify optimum process conditions to allow the min~mum number of process stages needed for complete sulphide removal. Sulphide is quantitatively converted to sulphur as indicated by the correspondence between the actual and theoret~cal values for the ratio: load of iron(l1) producedlload of sulphide removed (3.65 versus 3.49). More Cot was consumed than theoretically required (1.75 versus 1.38). The higher COZ consumption can be ascribed to the partial solubility of both COz and H2S in the pH range 6 to 7. The theoretical ratios for load of C& consumedlload of sulphide removed (1.38) and load of iron(ll) producedlload of sulphide removed (3.49) is calculated from Reactions 5 and 6. 2C02 + S2- + 21-120 + 2HCOi + H2S [51 H~S + 2~2' 3 S+ 2~e"+ 2~* El Effect of HRT. An amount of 217 mgn sulphide was removed at a feed rate of 0.5 limin (HRT = 59 min). compared to only 154 mgll at a feed rate of 2 llmin (HRT = 15 min). . Effect of C02 Row rate. By increasing the C& flow rate from 0.19 llmin to 0.83 ilmin. the sulphide removal increased from 342 to 474 mgn and residual sulphide in solution decreased from 134 to 0 mgn (as Sf. The corresponding ratios of CO, feed loadlNa,S feed load increased from 0.78 to 3.46. The stoichiometric value required for this ratio is 1.38 (Reaction 5). This demonstrates that complete sulphide removal can be achieved by dosing 2.5 times more C02 than stoichiometricaily required. Excess GO2 gas would be available in many applications. Wth the biological sulphate removal process COrgas is produced by the heating unit. During limestone neutralization of add water. C4 with a high concentration is produced due to dissolution of CaCO,. Effect of CG-concentration. By increasing the COz concentration from 20% to 100% the sulphide removal increased from 278 to 387 mgll (as Sf. In this case sulphide was not removed to 0 mgn as the COnlNatS feed ratio was less than the value of 3.46 as determined above empirically. Effect of gas recycle rate. By increasing the gas recycle rate from 9.1 to 19.6 llmin. Me sulphide removal was improved from 304 to 666 mgll. In this experiment sulphide was also not removed to 0 mgn as the C02/Na2S feed ratio was less than the value of 3.46. It was demonstrated above that a packed bed-reactor (configuration A) (Figure 2) can be used for sulphide stripping. In this configuration it appeared that the absorption stage, where H2S-rich gas was contacted with an imn(ll1) solution in a venturi system, was effective, due to good contact between gas and liquid phase. With the apparent good performance of the venturi system for H2S absorption, it was decided to evaluate the suitability of the venture system also for HzS-stripping. The same equipment that was used for configuration A (Figure 3). was used for contiguration B (Figure 3), except that the venture system was used for sulphide stripping in batch mode, and the packed bed-reactor was used for HzS-absoption into an iron(lll) solution under continuous conditions. Figure 6 show the effect of iron(1ll) feed rate on sulphide removal. It is noted that: Better sulphide removal was achieved with increased feed rate of iron(1ll). This can be ascribed to only partial absorption of HzS at low iron(lll) feed rates in the closed circuit of Configuration 2. This is an indication that the packed bed-reactor does not function as welt as the venturi system for the absorption of H,S into an iron(ll1)-solution. The obtained experimental (actual) dFe/dH2S ratio was similar to the theoretical value of 3.49 (Reaction 6). This result shows that all iron(1ll) that was introduced to the packed bed reactor was consumed for H2S-absorption through 6~ The obtained results showed that HzS-stripping and H~S~absorption is favourd by intensive mixing. Intensive mixing supports mass transfer of 1-125 from liquid to gas phase in the case of H2S-stripping and from gas to liquid phase in the case of H2S-absorption It was demonstrated that the venturi device was more efficient than the packed-bed reactor. This could be ascribed to the high pressure (300 kPa) and the high velocity (50 mlsec) of gas and liquid particles. Based on this observation. it is recommended that the Turbulator be used during a scale-up verston. The Turbulator exists out of a motor which directly (no gear box) drives a disc via a hollow shaft The Turbulator allows mixing between the gas and liquid phase by suck~ng in air through the hollow shaft that rotates at 2000 rpm. The velocity at the outer limit of the disc is 15 mtsec (Dla = 0.15 m: rpm = 2 000). Table 5 Sulphlde removal in hM stages In senes Parameter I Stage number 1 I 2 I I I COdNa,S feed ratio ig CO,I~ S) I 0.771 140 ~efeed rate (11rn1n) I 080l 0.90 C4 feed rate (Ifmm) I 0.29) 0 24 COrconcentratlon (%) 1001 100 HRTirn~n) ! 32.71 32.7 Gas recvcR rate (Urnin) I 22.91 22.9 I H in feed H in treated water dC0ddNaS ratio (g COdg S) Actual 1.70 1.79 Teoatical I 3.49 Actual 3.601 3.70 Running Cost The energy source is the highest cost item of the CSlRosure process. Geldenhuys et al. investigated the electrolytic production of hydrogen and founded that hydrogen can be generated electrolytically at a cost of R14.57Ikg Hz (14) The cost for various energy sources for removing 2 g S04n are compared in Table 6. It is noticed that electrolytic generated hydrogen compares favorable with other energy sources. such as ethanol, propanol and sugar. Table 6 Cost comparlson of varlous energy sources for tnolog~wl sulphate removal Nde Dde dpi- - Jmusy2004 CONCLUSIONS The following conclusions can be made from thls lnvestlgatlon The b~ologcal sulphate removal process can be used for removal of sulphate to less than 200 mgll uslng ethanol as the carbon and energy source H,S can be stripped to below 20 mgll (as S) with CO, by using a venturi as a stripping device. The efficiency of H2S-stripping is influenced by MS feed rate, C&concentration. CWH2S load ratio and aborption efficiency of H2S in iron(ll1). Hydrogen is the most cost-effective energy source. The cost of hydrogen amounts to R1 .531m3 if 1 g/l of sulphate is removed (R1.531kg SO, removed). m 1 rmn FeOll) 3 VninFBOII) 4 imnFeOll) -Linear (4 rnin Feflll)) Linear (3imin FeQII)) 0 om4om80100 nmw Figure 6. Elfecl of iron(lll) flow rate on sulphlde removal. ACKNOWLEDGEMENTS Sincere thanks are due to the following organizations for their financial and logistical support of the research reported in this paper: Anglo Coal (Navigation Sectron of Landau Colliefy) who provided financial support. the necessary infrastructure at the mine and general assistance. The National Research Foundation (NRF) who provided funding through their Technology and Human Resources for Industry Program (THRIP) for CSlR projects on neutralization and sulphate removal. The CSlR who provided substantial financial support for the research program. REFERENCES 1. Maree. J.P. and Strydom. W.F . "Bilogical sulphale removal from a packed bed reactor. Water Research. 19(9). pp 1101-1106. (1985). 2. Maree. JP and Hill. E.. "Biological removal of sulphate fmm industrial effluents and concomitant production of sulphure. Water Sci. Technol.. 21, pp 265-276. (1989). 3. Maree. J.P. Hulse. G. Dods. D. and Schutte, C.E.. 'P~bt plant stud~es on blolcgical sulphate removal from induslnal efiluent'. Wat. Sw. Tech.. 23. pp. 1293-1300. (1991). 4 du Preez. LA Odendaal. J.P. Maree. J.P. and Ponwnby. M.. 'Biological removal of sulphale from mdusnal effluents uslng producer gas as energy source'. Env. Techn.. 13. 875-882. (1992). 5 Vlsser. A,. -The anaerobic treatment of sulphate containing wastewater-. Phd thesis Agricultural University Wageningen, The Netherlands, (1995). 6 Van Houten. R T.. 'Biological sulphale reduction with synthesis gas'. Ph.D lhesa. Wagenlngen Agr~cullural University. Wageningen. The Netherlands. (1996). 7. Geldenhuys. A.J.. Maree. J.P.. de Beer. M and Hlabela. P.. "An integrated limestonellime process for partml sulphate removal'. The Journal of The South African Institute of Mining and Metallurgy, July/August. (2003). 8. Eloff. E. Greben H.A. Maree. J.P. Radebe. B.V. and Gomes. R.E.; "Biological Sulphate Removal Using Hydrogen As The Energy Source". Prmeedings of the 8th International Mine Water Assacjation (IMWA) Congress. Johannesburg. 20 -25 Oct.. (2W3). 9. Maree. J P. Slrobos. G. Greben. H. Gunther. P and Christie A D M. Biological treatment of mtne water using ethanol as energy source. Conference on Environmentally Responsible Mining in South Africa. Muldersdrift. South Africa. 25 - 28 September. (2001). 10. APHA. 'Standard Methods for the Examination of Water and Waslewater". Twelfth Edition. American Public Health Associatton. New York. (1585). 11. McCartney. D. M. and Oleszkiewicr, JA.; Competition between methanogens and sulphate reducers. effect of COD suiphate ratio and accl~mabon, Wat. Environ. Res.. 65, 655664. (1953). 12. Piass S.B. Jambs H R and Boem R F. Operational characteristics of a spray column type direct contact preheater. Heat Transfer-San D~ego. Vol. 75 pp 227-234, (7575). 13. Maree. J.P.. Hlabela. P.. Geldenhuys, G J.. Mbhele. N and Nevhiiaudzi. T., "Treatment of mine water for suiphate and metal removal imng banum suiphide". Waste Management. Emissions and Recycimg in the Metallurg~cai and Chemical Process Industries Conference, Mintek Conference Centre, Randburg. 18-19 March. (2004). 14. Geldenhuys A.J.. Maree, J.P.. Fourie. W.J., Smit. J.J.. Bladergroen. B. and Tjat~. M.. "Acid Mme Drainage Treated Electrolytically for Recovery of Hydrogen. lrontll) Oxidation and Sulphur Production'. Proceedings of the 8 th International Mine Water Associabon (IMWA! Congress, Johannesburg. 20 -25 Oct, (2003). CONCLUDING DISCUSSION Acid mine drainage is the term used in this thesis to define drainage that occurs as a result of natural oxidation of sulphide minerals contained in rock and exposed to air and water. The principal ingredients in the acid mine drainage process are reactive sulphide minerals, oxygen and water. The oxidation reactions are often accelerated by biological activity. The chemical and biological reactions yield low pH water that has the potential to mobilise any heavy metals contained in the rock. Acid mine drainage, besides sulphate, contains high concentrations of dissolved heavy metals, including iron(ll) and sulphate, and can have pH values as low as 1.9. Since water and oxygen are essential components to form a strongly acidic solution, exclusion of one or the other will prevent this reaction from occurring. However, much larger quantities of oxygen are needed compared to water. Unless neutralised, such waters can have a detrimental impact on water quality in the environment, e.g. public water courses. The objective of acid mine drainage treatment is, therefore, the elimination of acidity and precipitation of heavy metals. Neutralisation of acid mine drainage. The most common acid consuming mineral is calcium carbonate (calcite), the major constituent of limestone. Although several minerals are capable of removing acidity, including carbonates of iron and magnesium and hydroxides of iron and aluminium, a limited number will increase the pH from relatively low values, such as 1.9, to an acceptable, approximately neutral level. Whether the drainage from a mine is acidic depends on a number of factors and site-specific conditions. The most important factors are probably the balance between sulphide and neutralising minerals and their relative reactivities. The rate at which neutralisation occurs, is dependent on a number of physical and chemical factors. An important consideration is the susceptibility of carbonates to the forming of surface coatings of precipitates, such as gypsum and iron salts. These can cause "blinding" of the carbonate and result in a decrease in reaction rates. Previously lime was used for neutralisation. Replacing lime with limestone, the cost of neutralisation could be reduced significantly. Table 3 compares the cost of neutralisation with limestone, slaked- and unslaked lime. Table 3 Cost comparison between limestone (CaC03), unslaked lime (CaO) and slaked lime (Ca(OH), as neutralisers of AMD Parameter I Limestone I Unslaked lime I Slaked lime pH after treatment 1 7.00 1 7.00 1 7.00 I I I Acidify (mgle CaCO,) 1 683.04 1 683.04 1 683.04 Flow (Melday) I I I Purity 1 69.00 1 85.00 1 85.00 I I I Utilisation efficiency (%) 1 90.00 1 90.00 1 90.00 75 Dosage (kg/m3) 11.10 1 0.50 1 0.66 I I I Usage (tonlday) 1 82.49 1 37.50 1 49.55 1 75 75 Other benefits of using limestone instead of lime for neutralisation are: Limestone is a by-product from the local paper industry Fluctuations in the flow of underground water do not affect the neutralisation process as the acid feed water is diluted by bleeding it into the recycle stream which passes through a fluidised bed containing excess limestone Limestone is easy and safe to handle and dust-free. It can be stored on an elevated slab of concrete and eliminates the need for silos or other expensive storage facilities as it already contains 15% moisture. Price (Wton) Treatment cost (Wmonth) Cost ratio With the limestone neutralisation process, the pH of the bulk of an acid stream can almost be neutralised (pH 6.8) with powdered limestone in a fluidised bed reactor. A small amount of lime is added for rapid iron(ll) oxidation at pH 7 and higher. Acidity can be reduced from 16 000 gle to almost zero, as CaCO,. Free acid and acid associated with Fe(lll)) will be completely removed and sulphate concentration lowered from 16 800 mgle to 1 900 mgle. A retention time of 1 hour is sufficient to achieve these results. The rate of neutralisation depends mainly on the type and particle size of the limestone. The chemical composition of both the acid water and limestone also play a role. The presence of iron(ll), aluminium or magnesium in the acid water 150.00 376 165 0.51 652.23 743 542 1 .OO 900.00 1 355 786 1.82 decreases the rate of neutralisation. The same rule applies to the concentration of magnesium in the limestone being used. Iron(l1)-oxidation Feed water containing iron(ll) concentrations in excess of 50 mgl! impaired the dissolution process, due to scaling, also known as blinding, of the limestone particles with ferric hydroxide and gypsum. It is essential that iron(ll) is converted to iron(lll) as a pre-treatment stage to the neutralisation of iron(l1)-rich acid water with limestone. Iron(ll) causes scaling as the oxidation rate increases with increasing pH values. Therefore, most of the reaction occurs on the surface of the limestone particles where the pH is higher than in the bulk of the water. Iron(ll) in the acid feed water can be oxidised either mechanically (aeration), chemically, biologically or electrolytically. For both chemical and biological iron(l1)-oxidation, it has been shown by Maree et a/. (1997, 1998) that the rate of oxidation depends on the surface area of the support medium and the suspended solids concentration. However, in this work, iron(ll) was oxidised electrolytically. This was done while generating hydrogen in an electrolytic cell for utilisation as energy source for sulphate reducing bacteria in the biological process. Oxidation rates as high as 110 g Fe/(e.day) were measured when performing the electrolytic generation of hydrogen with acid mine drainage (pH 1.9) as electrolytic medium and stainless steel (type 31 6) as electrode material. These results compared well with the rates of 80 and 66 g Fe/(e.day) for biological and chemical iron(l1)- oxidation, respectively, as reported earlier by Maree et a/. (1 997, 1998). These results were achieved at a current density of 1 ampere/dm2 and a potential of 4 volts and using asbestos plate (4mm thickness) as diaphragm to avoid contamination of the hydrogen with oxygen. Hydrogen was generated at the cathodes and oxygen at the anodes, according to the following reactions: It has been demonstrated that hydrogen can be generated electrolytically at a cost of 16% (R3.891kg of hydrogen) of the cost of purchasing hydrogen in bulk from industry (R25.00lkg). Cost figures for generating hydrogen electrolytically are listed in Table 4. Table 4 Cost comparison for various electrolytic hydrogen generation systems Electrode material 1 Electrolytic I Diaphragm I Production Stainless Steel I KOH (30%) I Anionic selective 1 14.57 Fe plate Zn plate I I I Stainless Steel I KOH (3%) I Anionic selective 1 22.39 I I I Stainless Steel I AMD (5 ampere) I None 1 21.20 solution AMD AMD Stainless Steel I AMD (5 ampere) I None 111.98 I I I Stainless Steel I AMD (5 ampere) I None 1 9.55 material Anionic selective Anionic selective I I I Stainless Steel I KOH (3%) I Asbestos 1 3.89 cost (Rlkg) 138.84 151.94 available in bulk I I I I Commercially It was concluded that iron(ll) oxidation is an excellent benefit added to the electrolytic production of hydrogen from AMD. This process can therefore be applied upstream of the neutralisation stage of acid mine drainage. Due to the fact that iron(ll) remains in solution up to pH7, it is beneficial to convert iron(ll) to iron(lll), which will precipitate at pH 3 and coat the carbonate particles which slowing down the neutralisation reaction significantly. 1 25.00 Gypsum crystallisation and CaC03 precipitation The primary product of acid mine drainage neutralisation using either limestone or lime is gypsum (CaSO4.2H2O), produced largely as a precipitate. Gypsum commonly forms scale in tanks and piping which then require periodic acid treatment and mechanical removal. Proper design can reduce the severity of some scaling problems. Heavy metal ions hydrolyse and precipitate as their respective hydroxides during neutralisation and any ferrous ion present in the slurry is oxidised to ferric iron and precipitated as ferric hydroxide. After neutralisation, sulphate concentration in water can be further effectively lowered with lime treatment to 1 200 mgle by raising the pH to 12 and higher. This is well below the saturation level of gypsum (1 500 mgle) and no scaling of pipelines will occur at a retention time of 2.7 hours. Magnesium was also fully removed as a result of the high pH. During gypsum crystallisation, the gypsum concentration was found to have a major influence on the rate of sulphate removal. It was concluded that the rate of gypsum crystallisation is directly proportional to the surface area of the gypsum crystals. Water, under-saturated with respect to gypsum, can be produced by adjusting the pH with C02 from the high levels of 12 and higher to pH8.5 where CaC03 has a low solubility. At this stage of the process, the treated water is suitable for re- use in a plant or for irrigation purposes. Sludges resulting from the various stages contain gypsum, heavy metal hydroxides and heavy metal sulphides. Polymeric substances, PAC6 and Type 3095, were successfully introduced to the process as coagulant and flocculant, respectively. The solids were recycled from the clarifier underflow to the feed, upstream of polymer addition, to build a high density sludge. Dosing these polymers in very small quantities to the various sludges, generated during the process, their settling rates were significantly increased. These properties are important in providing clarifiers of suitable size after each stage. Electrolytic hydrogen generation Several sulphate removal technologies are in place, amongst others the Biological Sulphate Removal Technology. For the treatment of these effluents, expensive organic material (e.g. ethanol or sugar) is used as the carbon and energy source for the SRB. The use of hydrogen as energy source provides a cheaper alternative for sulphate removal. Toxicity due to increased levels of sulphide and un-ionised hydrogen sulphide will not only lead to a diminished process performance of the SRB but will also become a health and safety hazard. Stainless steel (type 316) plates have been used effectively as electrode material with AMD as electrolyte for generating hydrogen in a cost-effective way. While generating hydrogen at the cathode, other benefits such as the oxidation of iron(ll) to iron(lll) at the anode and the conversion of hydrogen sulphide gas to elemental sulphur in a down-stream stage also occur. After the iron(l1)-rich acid water was treated electrolytically, oxidation resulted in an iron(lll)-rich product which is the feed stream to the limestone neutralisation stage. Part of this iron(ll1)-rich stream can be contacted with the waste H2S gas from the biological process and elemental sulphur will be produced. Thus, the waste product of the Biological Sulphate Removal Technology can be converted into a valuable product, elemental sulphur. The need for a membrane (diaphragm) is eliminated as no oxygen is generated at the anode. Hydrogen can also be generated using stainless steel (type 316) as electrode material in a high pH environment (3% KOH). Asbestos has been successfully utilised as membrane to prevent mixing of hydrogen and oxygen, generated at the cathode and anode, respectively. When comparing the cost of hydrogen purchased at R25.00 per kilogram, huge cost savings can be achieved by generating on site for use as energy source for the SRB. Pure hydrogen can be generated electrolytically at a cost of R9.55 per kilogram from AMD or R3.89 from a 3% KOH solution. Barium sulphide process After the lime- and C02 treatment stages, the water can be further treated with a Reverse Osmosis (RO) Process developed by The Chamber of Mines or the Biological Sulphate Removal Process developed by the CSlR (Maree, et a/., 1987; Maree & Hill, 1989) in order to be released into a river system (sulphate concentration required: <200 mgle) or to be treated to a level similar to that of municipal water. These processes are relatively expensive and therefore other means of desalination should be considered. Another promising process to treat sulphate-rich waters entails the chemical removal of sulphate by means of soluble barium salts such as barium sulphide and is illustrated in the following reaction: CaS04 + BaS -* BaS04 + CaS [I41 The barium sulphide process is attractive for large scale application for the following reasons: J Acid waters can be treated directly with barium sulphide J The process removes ammonia, magnesium, manganese and other heavy meals due to the high pH achieved prior to the softening stage J By-products like sulphur and NaHS can be produced from the H2S generated in the process and CaC03 from the softening stage J BaSOa can be converted to Bas on-site by heating to 1 050°C and re- used Closing assertion Several full-scale plants, as listed previously in this thesis, have been constructed over the past four years in South Africa, containing one or more of the above mentioned stages of the technology developed by CSIR: Environmentek. The Limestone Handling and Dosing System and the CSIR Density Meter were specifically designed by CSIR: Environmentek to add value to their chemical treatment technology. Because of the essential mechanical design of these two developments, they are not discussed in this thesis. Their significance for the treatment of acid mine water can, however, not be over-emphasized as they have a vital role in the successful treatment of acid mine water by means of the technology. The Limestone Handling and Dosing System and CSIR Density Meter are depicted in the following two photographs (Photos 1&2). The author was personally involved in the design, construction, testing and industrial application of these two developments. Continuous improvements have been effected to both systems in order to achieve optimum performance. Successful treatment depends on the selection of an appropriate technology for a task, as well as its correct implementation. pH adjustment, lowering metal loads and meeting specific discharge standards can require vastly different approaches requiring significantly different technologies depending on the aim and purpose. For most sites, successful treatment technologies will require site-specific installation and implementation to achieve maximum benefit. Correctly selected . treatment systems that are poorly installed or operated can be as ineffective as inappropriately chosen treatment systems. It is evident that, regardless of emerging technologies, pH control with cost- effective neutralisation agents will remain the most widely used and the most economical approach to both passive and active acid mine water treatment. Active treatment using calcium compounds, particularly limestone, is likely to remain the prime choice for neutralising acid mine water due to their non- proprietary nature, widespread availability, ease of application and cost effectiveness. Photo 1 Limestone Handling and Dosing System developed by CSlR Environmentek. Evident from this photograph is the powdered limestone stored onto concrete slab, slurried wlh water spraying system to flow into yellow sluny make-up tank in sump. Photo 2 Densrty Meter developed by CSlR Environmentek ACKNOWLEDGEMENTS I would like to express my gratitude to God who cares, loves and guides me and gives me the opportunity to glorify His name in what I do. I wish to direct a special thank to my dear parents for their love and loyal support in good and bad times and always being there to assist me. I love you. Thanks are due to the following people and institutions who were involved in this project. Without them this work would not have been possible: CSIR: Environmentek for the facilities to construct and run the laboratory and pilot scale units National Research Foundation (NRF) for financing the research during the past four years Dr J P Maree for his guidance and support as study leader and co-promotor Prof F.B. Waanders (School for Chemical & Minerals Engineering, North- West University) for acting as prornotor Colleagues at CSlR and University of the Western Cape that were involved in the projects and provided very valuable assistance BIBLIOGRAPHY References in this section are only applicable to the Literature Overview. Each paper has an independent bibliography. ALPERN B., NAHUYS J. and MARTINEZ L. 1984. Mineral Matter in Ashy and Non-washable Coals - Its Influence on Chemical Properties. Comun. Sew. Geol. Portugal, Vol. 70, No. 2, p. 299. DAVIS A,, 1981. Sulfur in coal: Earth and Mineral Sciences, Pennsylvania State University, University Park, Vol. 51, No. 2, pp. 13-21. EHRLICH H.L. 1990. Geomicrobiology (Pd Edition): New York, Marcel Dekker Inc., 646 p. FALCON L.M. & FALCON R.M.S., 1987. The Petrographic Composition of Southern African coals in relation to friability, hardness, and abrasive indices. J.S. Afr. Inst. Min. Metall, Vol. 87, No. 10, pp. 323-336. FALCON R.M.S., 1988. The Characteristics of Southern African Coals. J.S. Afr. lnst. Min. Metall., Vol. 88, No. 5, p 145. FALCON R.M.S. & SNYMAN C.P., 1986. An Introduction to Coal Petrography: Atlas of Petrographic Constituents in the Bituminous Coals of Southern Africa. pp. 3-4, 19. GOVERNMENT GAZETTE (PARLIAMENT OF THE REPUBLIC OF SOUTH AFRICA). 1998. National Water Act. Vol. 398, No. 36,200 p. GRAINGER L. & GIBSON J., 1981. Coal Utilization: Technology, Economics and Policy. pp 1-8. HAWKINS J.W., 1984. Iron disulfide characteristics of the Wayneburg, Redstone, and Pittsburgh coals in West Viginia and Pennsylvania. Morgantown, W.V., University of West Viginia, M.S. thesis, 195 p. PAYNE D.A. & YATES T.E. 1970. The effects of magnesium on acidity determinations. 9 Symposium on Coal Mine Drainage Research, National Coal Ass./Bituminous Coal Research, Inc., Pittsburgh, Pa., pp. 200-223. ROSE A.W. & CRAVOTTA C.A. 111, 1998. Geochemistry of coal mine drainage. Brady, K.B.C., Smith, M.W. and Schueck, J., eds. Coal mine drainage pollution prevention in Pennsylvania: Harrisburg, Pa, Pennsylvania Department of Environmental Protection, 5600-BK-DEP2256 8/98, pp. 1 .I-1.22 SANDERS D, 1996. Coal Characterisation in Marketing - an Elementary Approach. Workshop on Coal Characterisation - for Existing & Emerging Technologies, CRC for Black Coal Utilisation, Newcastle, Australia. STACH E., MACKOWSKY M.T.H., TEICHMULLER M., TAYLOR G.H., CHANDRA D. and TEICHMULLER R., 1982. Stach's Textbook of Coal Petrology. (3d Revised and enlarged edition). Gebruder Borntraeger, Berlin. STUMM W. & MORGAN J.J., 1981. Aquatic chemistry. Wiley Interscience, 470 p. TEMPLE K.L. & DELCHAMPS, E.W., 1953. Autotrophic bacteria and the formation of acid in bituminous coal mines. Appl. MicrobioL, Vol. 1, pp. 255-258. VAN TONDER, G.J., THERON, D.J. and MAREE, J.P. 2000. Cost optimisation of the water management strategy by steady-state modelling of the water network of a coppednickel mine and processing plant, Proc. of the WlSA 2000 Conference, Sun City, South Africa, 28 May-1 June. WARD C.R., 2002. Analysis and significance of mineral matter in coal seams. Int. J. Coal. GeoL, Vol. 50, pp. 135-168. ZlNCK J.M. & GRlFFlTHS W.F. 2000. An Assessment of HDS-Type Treatment Process - Efficiency and Environmental Impact. ICARD 2000, Vol. II, pp. 1 027- 1 035. APPENDIX A LIST OF ADDITIONAL PAPERS AND POSTERS POSTER 1: Poster to Paper 1 Geldenhuys,A.J., Maree, J.P., de Beer, M. and Hlabela, P. 2003. An integrated limestone/lime process for partial sulphate removal, The J. South African Institute of Mining and Metallurgy, 103(6), 345 - 353. CaCOa and lime treatment stage C02 treatment stage Division of Water, Environment & Forestry Technology SA ~CSIRI 154 The scarcity of water in South Africa is exacerbated by . ppllution of the surface- and ground-water ,sources. Typi~ pollutants are industrial effluents,mainlyacid ' water. Ticcif'SA' .>'; produ'bes z1rkon, rUtile" ~n.d illmj,nit:fT~2 'Plus 'ron compo(Jpds)from~ minfngSand dunes at .. the Natal Coastal ~~10n in ~o,uth Atri,PP. ''The plant pro.,d~ces 85nrlhr of"aciq vyater which need to be treated for~T~ufe 40the 'metallurgical process or to a higher quarrty, suitable for,discharge into the EflIean.geni sewage system " (2 200mg/f SOA)' 156 --- - -- - --- - - -- 1.82 15131 5558 . .___ n___. 5550 *:,< 6.10 12317 840 464 821 440 50 29 386 ..4';: Aim of study ~~i~'WQter at' aq;:l To biologically treat ,,' ' , 7MI/day of the produced AMD to remove the sulphate, acidity and metals through metal sulphide precipitation to render it suitable for irrigation of a citrus plantation. Material and Methods Feed Mine effluent: 504: 1500-2 200 mg/l, COD = 1500 mgll Macro (NH] + P04 ) and micro nutrients were added, HRT: 24-12hrs Reactor Systel\l Single -stage completely mixed reactor system (Figure 1) BiOlllass Obtained from Pilot Biological Sulphate removal Plant at No.vigation. Anglo Coal. Witbank. South Africa. r -- 10' Figure 1. Completely mixed single stage reactor 2!;00 2000 1500 1000 500 o o 40 6020 Tim. Cd) ~S04 treated water __ 504 feed __ CODtrealed !'i,. ~i9ure t Process Stobilnyof reactor system 1 l~" Sulp1Mrte ~;. tob utilization cmctsUlphide production (Table 1) Table 1. The chemical composition of feedund treated water DEl'ERMINANI) IVNlT I . ~ Feedrate ~dHRT hr Foad unit mg/I ,II >II 11 Metal Removal (Table 2) Table 2. The metal concentnrtion in the feed and treated water Discussion SulDhate removal/COb COIISUIIIDtion The results show that during the second period the sulphate was removed to concentrations <400 mg/l, at which time the HRT was ~ day. During the last two periods, it was noticed that the COD concentration decreased in the reactor, resulting in less sulpho.te removal. This COD consumption was ascribed to the presence of increased metal concentration in the feed (BCt. water), which was as high as 5 mg/l for Ni and 38 mg/( for Zn. The metals acted as trace minerals for the Methanogens, there by out-competing the Sulpho.te Reducing Bacteria. Metal removal All metals were removed to concentrations <1 mg/t, except for Mn during first two periods. When the reactor pH increased, the Mn was removed as well. Conclusions 1. Good sulphate reduction 2. Metal removal due to metal-sulphide precipitation 3. Initially COD utilisation by Sulphate Reducing Bacteria 4. Inereased metal cone. in feedwater stimulated the methanogens 5. Water after treatment suitable for irrigation 162 - -- -- - - - 1 I 2 3 4 15 I 30 30 15 24 1- 12 12 24 6.5 5.12 4.31 4.32 1600 1600 2100 2200 17V 1727 1742 1602 130 130 100 120 6.72 6.10 7.00 7.55 55D 375 1138 627 1019 1657 1099 488 903 1477 1305 1766 140 268 229 260 METAL IUNIT FEED TREATEDWATER PERIOD 1 2 1 2 3 4 Nicllcl mQ/I 5.86 5.50 0.15 0.17 0.14 0.61 Zinc """I 1.25 38.00 0.03 0.03 0.03 0.03 1--- mg/I 0.97 0.99 0.97 1.20 0.17 0.06 1Copper ""II! 0.01 0.05 0.03 0.03 0.03 0.03 POSTER 2: Poster to Paper 3 Geldenhuys, A.J., Maree, J.P., Strobos, G., Smit, N. and Buthelezi, B. Neutralisation and partial sulphate removal of acid leachate in a heavy minerals plant with limestone and lime, Proceedings 6'h International Conference on Acid Rock Drainage, 12-1 8 July 2003, Ciarns, Australia. PAPER 9: Greben, H., Geldenhuys, A., Maree, J., Strobos, G. and Hagger, M. The sustainability of biologically treated nickellcopper mine effluent suitable for irrigation, Proceedings 6'h International Conference on Acid Rock Drainage, 12 - 18 July 2003, Cairns, Australia. This paper focussed on the treatment of effluent from a copper mine in Botswana. Because Botswana is an arid country, it needs to treat the water either for re-use or for discharge into a river system. The results of this study showed that the effluent could be treated to a quality suitable for agriculture, i.e. for irrigation of newly planted citrus orchards. It has been shown by the Faculty of Agriculture at the University of Pretoria, that acid water that has been neutralised from pH 2.2 to 7.5, can be utilised for irrigation of citrus trees. The water discharged by this mine contains heavy metals that, when treated with the Biological Sulphate Removal Technology, the resultant H2S precipitates these as sulphides. This way the metals present in the effluent as well as the sulphides are removed. A single stage, completely mixed reactor was used with a clarifier to trap sludge that formed. All metals, except for magnesium, were removed after the biological treatment. The sulphates in the effluent were reduced to sulphide. Results indicated that an increase in metal concentration, especially zinc, in the feed water, stimulated the enzymatic activities of the methanogenic bacteria. Good sulphate reduction was also observed when 1 me ethanol 112 of feed water was added with a COD/S04 ratio of 1. Paper 9 was presented as a poster by A J Geldenhuys at the 6'h lnternational Conference on Acid Rock Drainage mag paseanu! uo!ruruaxm> m!z am pue qdmas p3aj ]sly aql ut se Iptq EB ~01 sem aldlues paaj poo3ar aql u! uoqelllia~uorr .('I? II~ ~f'o'nd!srl 81-OW 81;111ovo'uz d13fi 8P'O '!N I Sr0 X >:a wo WY d /a 882'0 '03 d lsfi 01~ 'ad 2 1% 001) 111m S" pappv aram slua!nnuom!m 'slelam PJU!BIIIM ~IBM auy 738 aqr Ipnomlv -d-areqdsoqd dig= $1 pw N-eye- liSm SL :EluaqnVOl3e1~ lg~\ pauamfddns ssm paaj au 1 - 'yo Lf318~u!xa~ddeJo O!IW *OS/a03 B U! Zuqlnsa~ 'amos Lama aql EB ppp scm 00~1- a03)@1dw I) IOU~W~ 'Mu oozz 01 OOSL mog pa!m uo!resuaxm ale~lns ?ql qqa P '~SM pj ss ~mnlya au!m 130 pa+31~ lo~e-3~ aqL H OREBEN, A GELDENHUIS, 1 MARE and M HAGGER TABLE 1 Experimental penpen& and conditions when sxscuting the suifate and nrdrpl removal BCL ordect Process stability Figure 2 shows the sulfate eonccnw.ti~n in tk feed and treated water, as well as the COD of the heated water aver a period of 46 days in the reactor. It was obsnved that sulfate was almost totally removed h the feed water after ten day9 to 50 mglf (as SO& The initial feed rate was 15 tld, which was doubled after day 10. From then on, the sulfate was reduced to a consistent 450 mglt in the treated water. When the sulfate concenhution in a new bateh of feed water increased to values of and over 2000 mgit, the sulfate eoncentration in the heated water increased to 1150 m#(. Due to the low pH value of the feed water, the reactor pH was low When the reactor pH was increased and the reactor vras operated in a batch mode (days 27 - 29). the sulfate was reduced to 50 mdt. A new fwd batch was innoduced to the reactor on dq 30. The feed rate was kepf at 30 (Id, resulting in a HRT of 12 h. However, the feed rate war reduced to 15 Ud on day 31, with the aim to nduce the sulfate concenhution to values lower than 200 mgif. The reactor pH was strictly monitored from day 30 onwards and when decreased to values lowcr than pH of 7.5, a rolutian of concentrated NsHCO, was added to the reactor Although the reactor was closely monitored hm day 30 onwards, the sulfate concentration in the reactor did not decrease to the required valuez laver than 5W mgit, until it was obwmed that the rcaeror COD concentration had decreased dramatically. When on day 43, additional ethanol was added to the reactor, the SO, concentration m the reactor was as law as 77 mvt. it can be FIG 2 The SO, and CM) concentration in the treated BCL mine water. 1032 Calms, QLD. 12 observed from Figurc 2 that the COD value decreased from the average value of abut 1750 mdf, to a concentration less than 100 mgit. This observation can possibly be ascribed to the metal concsntration in the reactor. It has been reported that mcthanagenesis is stirnulared by the addition of micronutzients, rrprcially the trace metals such as iron, cobalt and aickcl (Speecc, 1996). It can be hypothesised that the rnethsnogenr were stimulated and competed with the SRB for the available COD, due to the elevated nickol eoncmtration in the reactor From the data in Table 2, it can be noted thaf nickel was present in the feed at emcentrations of 5 - 6 mgl! and that the zinc concenhation in the feed varied from 1 - 38 mgtf. Sulfate reduction rates, COD uUiisaiion and volatile suspended sdids (VSS) The results of the %Ifate ternoval and sulfide production tognher with the sulfnte reduction rates are given in Table 3, while the metal removal i3 given in Table 4. When obsnving the results of the volumetric sulfate reduction rates nesting the BCL mine water during the four experimental periods, the rates varied fmm 1.05, to 2.45, to 1.92 and 1.62 g SO, (lid), during periods I, 2, 3 and 4, respectively. The highest date reduction rate (2.45 g SO, (Nd)) was obtained during the second expe"ments1 period. Uuhg this period, the feed rate ms 30 t14 the average feed sulfate concentration was 1M)O mgll and the reactor SO, concentration was an average of 375 mdt. The COD to the reactor was added in the form of ethanol (1 m!ll), which resulted in s reactor COD concentration of abut 1700 mglf Thc COD utilisarion for the sulfate reduction can be expressed os thc COD.,,,,&04,d ratio. This ratio as indicated in Tnble 3 raried during the four different periods. During the first period, the CODISO, ratio was 0.67, which corresponded with the theoretical COD,,JSO,,, ratio, durine the second rreriod however this exucrimenml raho was very ibv (0.06), f&which no explanation can be given. Duing the third period, the experimental ratio (1.06) was higher than the theoretical ratio. which can bc exdained by the increased aetivihi of the methano&s. bring the'founh period, the experimental ratio (0.69) was similer to the theoretical period. During this wriod an two occasions, additional COD was added to the ;caror m sttmuldre Iunhcr SO, ruluctlon, dllhough th: fred COD.SO, raw at (1.71 should lwr been adeqmce. 11 mJcd only the 'iKB had imltrcd lhr a\-ulablc COI) I1 ;an ilwrforr be assumed that competition for the availabtc COD with the methanogens occurred. As can be seen from Table 3, the VSS concentration throughout the four experimental periods remained more or less eonsmnr. The overall VSS concentration innrased with ten per cent from ahnt 3WO to 3300 mgif durins the total experimental period of46 days. Sulfate reduced, sulfide produced and metal sulfide precipitation It can be seen from Table 3 that during all four experimental was used to precipitate the metals present in the BCL mine water In addition, part of the formed sulfide oxidised, because of the ccculrenre of the sulfide oxidising bacteria (present when sulfide togetha with some air is available). The Cu, Ni and Zn concentration in thc tnated water was less than I !"dl, during the fim periods, while the Mn concentration did not decrease. However, during periods 3 and 4, thc %n concenlmtion in the ueatcd wmr war reduced to 0.1 7 and 0.06 mw'f, which might be pH related 8s the reactor ptt was higher at 7.00 and 755, resprctively. When the zinc concentration in the feed war as high -18JlAyZW3 6h CARD THE SUSI'AlhAHILITY OF HlOLOGICALLY TREATED NICKF.UCOPPER MINE EFFLL-NT SUITABLE FOR IRRIGATION m 38 mvt, no dnc was walyrd in the trea~ed watm, indicating that the ppraluced sulfide wsn used o precipirate the zinc During the fourth wind. the nickel coneenhation in the scared water was higher (0.6 ms%) than in the previous three pcrinds. This result cannot be pH related as with increasing pH, the solubility of Ni decreasff (Bhattacharrya el el. 1981). The residual sulfide in that priod was high enough at 260 m@?f to precipitare the available nickel. Therefore, no immediate explanation can be provided for the Ni coucenmtion increase dating pcriod 4. MBtals as trace elements Tnee clemcnta are present in low concolhalionr I" rocks, mils, water and in the atmosphere. Some Pace elements, such as depcndcnt ;pon'a hosl of factors, but in the case.of tram mctals, the lack of only one an severely lmir the own11 process. Especially iron, wbalt, nickel and zinc have heen showll to be stimulatw to the methanogcns, degrading acetatc (Speece, 1996). In the second mine water sample as obtained from BCL, the zinc concentration in she feed war as high as 38 mu(. Fmm Fl_rmre 2, it con he observed thar from ahut day 30, the COD concentration d-4 rapidly in Ihe reactor. This occurrence coincided with the feeding of the second mine water sample, containing a higher zinc coneenuation Ulan the previous mine (feed) water. It can be assumed thar a correlation exists betwren the hiaher zinc coneenhation in the feed water and the total COD dcmdahon in the reactor On two aessionn, during priod 4. ethanol was added directly to the reactor in order tc&vestigair whether sulfare reduction muld be observed when additional COD was added to the -or. During perid 4, the nickel commnation in the mated water was hiher than in the arevious pmods, which together with Zn is a &e element inn&mentnt for vcetate degradation in the bioreactor (Speccc, 1996). He dwctikd the importance of three mce elements, being nickel, imn and cobalt. When these three tmce elements were directly injected into thr reactor, acetate utiliwtion races -Id hs observed irmediately, ns could the gas pmduction. In the case of treating BCL mine water, nickel was present in the wstcr, thus 81 mi@t be possible chat the elevated nickel (and likely the zinc) conc~ntmtion played an important mlc in the cuymc activity of the methsnogenr, lhus enabling the methanogenr to use the available COD. It cat be assumed that the methanoems were 6th CARD cairns, mu. 12- 18NIy20CB 1033 H GREBEN, A GELDENHUIS, J MAREE and M IIAGOER CONCLUSIONS The results of this mdy indicated that st1 mcmls in the BCL mine water were removed alter the biological nulfate reduction to sulfide production. Bmausc the experimenml produced /sulfw-, ratios (0.13,0.22, 0.24 and 0.16) were lower t an the theoretical ratio of 0.33, it can bc nssumcd that the su'fid%w produced sulfide precipitated the memls in the feed water to meml sulfides (MeS). Good wlfate reduction was observed in period 2, resulting in a SO, remov-al rate of 2.5 g SO, (g.0 Dne m the sulfate coneenoation increase in the feed water. while the ethanol concentration remained the same, thc fd CODIwlfare ratio war il during periods 3 and 4. This might explain why the sulhte reduction during those periods uas not as efficient as throughout period 2. The lower sulfate reduction can possibly be ascrikd m the competition of the SRB and ihe merhanogens for the available COD. 7he results wen) m indicate that the increased feed Zn and Ni concentration stimulated the enzymatic activities of the mcthanogens, thereby degrading the available COD. REFERENCES Cams. QLD, 12 - 18 Jdy 2W3 APPENDIX B LIST OF CONFIRMATIONS WATER RESEARCH COMMISSION Errnail: inerim.~re~a plintsB.gx03 TB *ha FAX: (012) - (012) 337-2555 0031 SOUTHAFRlCA I~~lade +2712 Enqwies wxm: ~lkmwcongu, Telephme: Mr kT Geldenhuys PO Box 7253 1 LYNNWOOD RIDGE 0040 Dear Mr Geldenhuys ACID MINE DRAINAGE TREATED ELECTROLYTICALLY FOR RECOVERY OF HYDROGEN, IRON0 OXIDATION AND SULPHUR PRODUCllON by AJ Geldenhuys, JP Maree, WJ Fourie, JJ Smit, BJ Bladergroen M Tjati Thank you very much for your E-mail letter of 1 March 2004 enclosing the above paper to be considered for publication in WaterSrl. The paper will be refmed to referees and I shall again mmmuuicate with you as som as I have received the reports. 1G Buehau Ecliter: WofcrSA for CHIEF EXECUTIVE OFFICER From: Sent: To: Subject: imwa@shaftsinkers.co.za 04 August 2004 09:49 Geldenhuys.Andre Enquiry Dear Mr Geldenhuys Many thanks for your email regarding your paper which you presented at the IMWA Congress last October in Johannesburg. I can confirm to you that your paper Wed "Electrolytic Treatment of Acid Mine Drainage for Hydrogen Production" was reviewed and edited by Mr Don Armstrong, an Honorary President of the Executive Council of The International Mine Water Association and one of four paper editors for the IMWA 2003 Congress. Should you require further assistance please do not hesitate to contact me. For your information the official website for the lnternational Mine Water Association is www.imwa.de should you need to include this. Kind regards Mimi van Niekerk This footnote confirms that this e-mail message has been scanned for the presence of known computer viruses by the MessageLabs Virus Control Centre. However, it is still recommended that you use local virus scanning software to monitor for the presence of viruses. ...................................................................... Water Environment and Forestry Technology Beste Andre en Prof Waanders Ek gee graag hiennee stemming dat Andre Geldenhuys die ondergenoemde publikasies vir sy tesis mag gebmik. PAPER 1 : GeldenhuysJ.J., Maree, J.P., de Beer, M. and Hlabela, P. 2003. An integrated limestonellime process for partial sulphate removal, The J. South African Institute of Mining and Metallurgy, 103(6), 345 - 353. PAPER 2: Geldenhuys, A.J. & Maree, J.P. Synthetic organic polymers (PAC6 and 3095) as coagulants/flocculants for optimisation of an integrated limestondlime neutralisation process for partial sulphate removal, Proceedings 5th Annual Industrial Water Management and Treatment Symposium, 15-16 May 2002, Johannesburg, South Africa. PAPER 3: Geldenhuys, A.J., Maree, J.P., Strobos, G., Smit, N. and Buthelezi, B. Neutralisation and partial sulphate removal of acid leachate in a heavy minerals plant with limestone and lime, Proceedings 6th International Conference on Acid Rock Drainage, 12-1 8 July 2003, Ciarns, Australia. PAPER 4: Geldenhuys, AJ., Maree, J.P., Fourie, WJ., Bladergroen, B.J. and Tjati, M. Acid mine drainage treated electrolytically for recovery of hydrogen, iron(l1) oxidation and sulphur production, Proceedings 8th International Congress on Mine Water and the Environment, 19-22 October 2003, Johannesburg, South Africa. PAPER 5: Maree, J.P., de Beer, M., Geldenhuys, AJ., Strobos, G., Greben, H., Judels, C. and Dreyer, J. Comparison of the combined limestondlime and combined limestone/biological sulphate removal process for treatment of acid mine water, Proceedings Hard Rock Mining Conference: Issues Shaping the Industry, 7-9 May 2002, Colorado, USA. PAPER 6: Adlem. (:.J.I..; Gcldcnhuys. ,\..I., Marce. J.P. and Strohos. G.J. Examining the implementation of limestone neotraliratioo technology in the mining and industrial xctur tu neutrolise acid and reduce sulphate pollution. Pri~ceeclingi 5th .Annual Industrial Wate~.Managrment and Treat~nent Sytiiposiwn. 15.1 6 >la) 2003, Johan~~eshurg, k~th Africa. PAPIX 7: Grrbrn, 14.. Geldcnhup 4.. Maree, J.; Strobos. G. and Ilagger, M. The sustainability of biologically trcatcd nickellcopper mine effluent uitable for irrigation. Proceedings 6th ln~ernational Conference on Acid Rock Drainaye_ 12 - 18 July 2003. Ciarns. Australia. PAPER 4: hlaree. J.P.. Netshidnulu, I.. Strobos. G.; Nengovhela, R. and Geldenhuys, A.J. Iniegrated process for biological sltlphare removal and sulphur recovery. Proceedings WlSA 2001 - Liiennid Conference & Exhibirion. 2-6 ;May 2001: ('spc Town South Africa. J P Maree Busigheid~areabcstuurder Nd - --- EM- / \ i i ward resented to dden huys for the best Envirmrnentek research pubkcation that appeared in press in Ule M03 calendar year. The paper demonstrates the a&ation of a socmd research methoddogy and is an exceHent example of how basic ard appkd research can be cwnbined to identify i~~~tivesdutiortitocomplex~problems. From: Sent: To: Subject: Esther Eloff [EEloff@csir.co.zal 31 May 2004 08:55 Andre.Geldenhuys@afrox.boc.com Fwd: One of our most prestigious awards!!!!!! best pub award citation 2003 L... >>> Rose Clark 31/05/2004 08:51:34 >,> Hi everyone many of you attended the 3 events in each major centre celebrating the success of our top achievers over the past 3 weeks. Given for the strong call for having the citations short and sweet, we were not able to give you all the fuil set of information regarding citations etc for all candidates at each event. An abbreviated version was presented at the functions to ensure that interest was maintained throughout the evening. Attached is the full set of information on one of our most prestigious awards - the best Publication award. Please take the time to read this and acknowledge the fantastic work of your colleagues in this arena. Also many thanks to lnge Kotze for co-ordinating the 22 volunteer reviewers. Bob Scholes for programming the algorithms, and David Le Maitre. Dirk Roux and Alex Weaver for doing the final adjudicatioh. I will also be asking Mandi to ensure that the full list of award winners, with photos and full citations to be made available on our own website so that you can all celebrate their fantastic achievements in style and full detail! Thanks to our awards review panel for adjudicating all your motivations (Brian van Wilaen. Shamilla Pillav. Vasna Ramasar. Steohan Woodborne. I am sure you will all join me in thanking Mandi Titi for organising 3 great events and Khungeka for sponsoring them. Kind rgds Rose Rose Clark 841 4707 (tel) 841 2734 (fax) .. This message has been scanned for viruses and dangerous content by Mailscanner. and is believed to be clean. Mailscanner thanks transtec Computers for their support f* "l*ffff**"**l.*.**~*...*~*.~.***.*~.~.*..*'*.~~~'A..**+*<~.b~*~~A~. Thls footnote conflrrns that this e-mail rnessaoe has been scanned for ENVIRONMENTEK'S BEST PUBLICATION AWARD FOR 2003. The annual award for the best scientific publication by an Ehvironmcntek staff member is arguably thc premier award made to any staffmember on an annual basis. The award is intended to show the importance that thc management team places on quality research, which in turn underpins our effectiveness as a scienze-bzcd organization. We also hope that an award of this type will encourasc the production of more high-quality publications by providing a substantial incentive. This seems to be working because Environmentek produced 33 scicntific publicatiolis in 2002 and 54 in 2003. more than 40% of the total produced by the CSIR. The award for the best publication dthc ycar is in the form of a grant for the researcher or rescai-chers involved to travel to a destination of their choice, worldwidc. to pursue further research or studics, attend a confcrcnce, or to build networks in their research field. subject to the approval of a costed proposal. The option of a single: large award indicates the importance that is attached to this aspect of our business, and although there can only be one winner, even thongh lhcre are oRen many dcsc~ving cases. Environmentek's Kesearch Committee has run nn evalt~ation process that has resulted in the selection of the 2003104 winning publication. The first step was to examine all thc publications on the database, or submitted by the authors, and selected a shortlist of 42 publications. The shoe-listcd publications were shared out among 20 reviewers and rated bascd on: 'the quality and innovativeness of thc work; its relevance to Environmentek's mission and impact, the standing of the journal or book and the relative contribution by CSIR authors. This process resulted in a shortlist of six outstanding publications from which a tcam of thrce adjudicators had to select the final one, The final six papers werc as follows: Cave. I.., Beekman, H. and Weaver. J. (2003). Impacts of climate change on yuundwatei recharge estimation. In: C3oundwatcr recharge estimation in southern Akica (eds Y Xu and I1 Beekman). pp 189-197. UNESCO IHP Series No. 64, UNESCO. Paris. Cieldmhuys. A.J., blarer, J.P., De Bcer. hl. and Fflahalela, P. (2003). An integated iimcstone/lime process for partial sulphate renioval. The Joumal of the South African lnslitutc vTMining and Meta1lu1-gy JulyIAugust 2003. pp 1-9. Gelilerblom, C.M.. Van Wilgen, B.W., Nel. J.L.. Sandwith, T., Botha, M. and ilauck. M. (2003). 'Ttunmg strategy into acuun: implementing a conservation action plan in the Capo Floristic Region. Biological C:onservation 1 12: 291 -297. McC'onnachie. h.... De Wit. M.P.. Hill. M.P. and Rymc, M.J. 12003). Economic evaluation of thc successful biological control of Azolir~,/iliculoids in South Africa. B~ologi~al Control. Schosmann, J.J. and Steyn. A. (2003). Nltrate removal with reverse osmosls in a mral arca in South Akica. Iksalmation 155: 15-26. Scholes. K.J., f30nilz W.J. and liclcharill. 11. (2003). Vcgctal~on dynamcs in the Kruger ecosystcm. In: The K~uger experience: ecology and management of sa\anna heterogeneity (eds JT Du Toil. KH Kodgerz and [I(' Higgs), pp 242-262. authors. The significant implications of even a small decline in ramfall for groundwater recharge, and for the communities dependent on groundwater, are explained clearly and simply. The shortcomings and assumptions in the model and in the climate change model information used in the model are described clcarly. 'This chapter is based on a presentation given at a local workshop conference and publication in thrs forn? will give it much wider exposure. The paper by Gt.ldcnhuys imd his co-workers deals with a kcy aspect or the innovative research bcing done by Jannii. Maree and his tcanl to address the severe acid pollution problcms being created by the mining industly. The paper delnollstratcs the applicatton of a sound research methodolo&y and 1s an exccllcnt cxample of how basic and applied research can be combined to identify solohons to genuine problcms. It also demonstrates the full cycle of innovation from idea to application. Thc process it describes reduces the high sulphate concenhations in coal mine water to acceptable levels and simultmeousiy rcduccs magnesluni levels. A key l&slure of the process is that by-products can be rc- used in tlic proccss so that n needs much lower inputs of thc basic resources. Ihe process includes the use of polymers to flocculate out suspended and colloidal material and obtain water that has low turbidity levels. The cost reductions with this new process are about 69% compared with the previously accepted methods which also p~oduced greater quantities of waste and required lime which requires more energy to producc than irmcstone. The paper was produced in a local society journal but one lhat is very relcvant because the society has many members in the mtnmg industry. Carolme and co-worker's paper descrihes the result ofa lot of work which Icd to the successfUl Cape Action Plan for the Environment, a strategic approach to conservation of the unique vegetation Cape Floral Region. The successful complehon of th~s project mvolved workmg with a wide range of partrcipants and stakeholders and getting them to collaborate. This manner is in which this was done and the quality of the final products was recognised by, among others Kathy McKin~~on of the funding agency (GEFJWorld Bank) and promoted as a model for similar projects elsewhere in the world. The strategic plan provided the basis for securing very large amounts of money that have revitallscd ecologcal research. It also set the stage for a number of other conservation planning initiatives and for further involvement of CSR staff in these initiatives, Much of the attention has focused on but thls would be ofl~ttle used without the model for its implementation provided by the CAPE project. Martin De Wit is the second author of 1111s paw but a key contributor because the reason for the research was to provide a sc~cntilically acceptable assessment of the economic benelits of biocontrol of a major weed species. Azollrr filic~rloidr~ is a very successful invader of aquahc sysicms. particularly of slow flowing rivers and dam.. with nutrientknriched water. It rcduccs water quality and can develop such thick mats on the surface that sheep havc mistaken it for grass, walked into dams and drowned (all we like shcep ....). Thc analysis confirnms that the quantifiable financial benefits which could be quantified in hard cash outweighed the costs with ratios of 2.5: 1 or more. The savlngs (Net Present Value) camc to about R7300 pcr ha or aboul R1.3 billion nationally. An miportant issue raised by this work is who should pay for this type of rcscarch? Currently the government pays and other sectors benefit hut the government conld potentially put this money to other uses. Maybe a fund is needed that could he sustained by the beneficiaries? An interesting issue that drservcs furthor research. Schocnman and Steyn's paper is on work done to rcduce thc high nitrate and salmty lcvels which are oftcn found in rural warer supplies based on groundwater. These high levels a-cate severe health problenls for people who often do not have the resources needed to deal w~th tho problems that arrse. High nitrates can stunt the developn~ent 01' yoimg babies, especially if they are on a pour diet. 'fhc trevcrsc osmosls (KO1 technolo&y rhat was used 1s not easy to mamtain so the project was armed at dcmonsln~ting that il could work and dclivcr measurable benefits when operated by a typical rural cnn~munity. Thc KO system was s~iccesifully operated and mamtamed by thc community and produced vevy good quality water. The concentrated by-products could bc used lor stock watering. I'lic capltal and opetational cos~s 01' lhc 1<0 systcn? wmr to about R3.00 pcr m', ahaut KO.08 per pcrson per day ibr the 25 1 basic human rcqu~rcn~ent. Ihc costs can he reduced by blending it with unporiticd wJter wii~lc mainwmlng ;icceptable nitratc and salim~y levels. An asscssmcnt of thc value of the benefits to human hcalrh and wall-brinp would bc a valusblr cxtms~on of this work. The wxk was innovativr. iargets a kvy issue Tor many rural comrnuntties and prondes a woikahle solorion. Hob Scl>olcs ;and co-aiithors haw coniptled n rhrough and ~ssefui i-ewrw oi lhu lindings oi local and ii~tem;uionill rcscvch Inlo the dynamics of savanna vcgetation. They docunicnt some intcresimg contiask lxtween savanna in dil.feferenr pats of thc world and ho~ thcse tropical rcosystcrns dliii~ Fiom the temp.rate that have lcd to most of the general~sat~ons about plant ecology. Thc vmou drwng factors in savanna are rwiwcd, focusing on water avilii~bility. the direct and ~ndircet impas of large plant eaters - mcgahcrhtvorcs - and fires. Fires have dramatic impacts on vcgetation ti seems that wtlhout fires the wetter part? ot.Soulh Aha would not have the highvcld grasslands and uvannas that am the source olmost of our tourism income. The chaptcr con~ins a good example of r synthes~s of a lot dinfornation to show how savanna arc shaped by multxplr. ~mtetucting Pactors. The selection of thc linal winners was not easy, it was a very. very close contest. All the papers are eoutstandiny: and some hard thousht and iterations were required before the final . . . winners were chosen. Wc would like to congratulate all of the authors of the ahove papers for making the final shortlist from which we were ultinmely ford to choose just one. We urse and encotrage our young scientists to make the time to read these papers and leam from them. Our choice for the winners goes to Andre Cieldenhuys and his co authors - congratulations and well done! PO Rox 3808; Garsfantein East 0060 Tel012 993 2787 Mobile 072 182 3592 wrlssoniam&hsa~nai!.~.~:.