Bauxite residue issues: I. Current management, disposal and storage practices

Hydrometallurgy 108 (2011) 33–45

Contents lists available at ScienceDirect

Hydrometallurgy j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / h yd r o m e t

Bauxite residue issues: I. Current management, disposal and storage practices G. Power, M. Gräfe, C. Klauber ⁎ CSIRO Process Science and Engineering (Parker CRC), Light Metals National Research Flagship, PO Box 7229, Karawara WA 6152, Australia

a r t i c l e

i n f o

Article history: Received 19 August 2010 Received in revised form 1 February 2011 Accepted 14 February 2011 Available online 24 February 2011 Keywords: Red mud Bauxite residue Bayer process solids Global residue inventory BRDA engineering Dry-stacking Lagooning Residue utilization BRaDD Marine disposal

a b s t r a c t Bauxite residue has been continuously produced since the inception of the alumina/aluminium industry in the late nineteenth century. The global inventory of bauxite residue reached an estimated 2.7 billion tonnes in 2007 increasing at 120 million tonnes per annum. This growth highlights the urgency to develop and implement improved means of storage and remediation, and to pursue large-volume utilization options of residue as an industrial by-product. This review looks at current management practices for disposal and amendment, and how each unit process influences residue properties. Since 1980 the trend has been away from lagoon-type impoundments towards “dry” stacking; this reduces the potential for leakage, reduces the physical footprint and improves recoveries of soda and alumina. Associated technical developments in residue neutralization are considered with possible future practices in residue disposal and how that might best integrate with future utilization. For example, hyperbaric steam filtration is an emerging technology that could discharge residue as a dry, granular material of low soda content. Such properties are beneficial to long term storage and remediation, but importantly also to future utilization. Although residue has a number of characteristics of environmental concern, the most immediate and apparent barrier to remediation and utilization (improved sustainability) is its high alkalinity and sodicity. The sustained alkalinity is the result of complex solid-state and solution phase interactions while its sodicity arises from the use of caustic soda (NaOH) for digestion. This is the first in a series of four related reviews examining bauxite residue issues in detail. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.

1. Introduction Bauxite residue and its appropriate management is of increasing environmental concern. As the global demand for aluminium metal continues in transportation, packaging and construction, from the fabrication of fuselages and car parts to disposable containers and various consumer durables, the legacy of nearly 120 years of alumina extraction from bauxite ore is manifested as an estimated 2.7 billion tonnes (Bt) of hazardous material commonly referred to either as red mud or bauxite residue. This amount is currently increasing by approximately 120 million tonnes per annum (Mtpa). Storage locations vary from purpose built, well-engineered containment areas through to simply utilizing natural landscape depressions and the bottom of the ocean (marine disposal). Despite the long-standing recognition of bauxite residue as a problem, it has never been satisfactorily resolved. In order to understand why this has arisen and to seek possible solutions for better management, it is very useful to consider all aspects of production, chemistry, residue value and barriers to utilization. This is the first in a series of four reviews examining issues pertinent to bauxite residue and its environmental impact (see also Klauber et al., 2011; Gräfe et al., 2011; Gräfe and Klauber, 2011). Ultimately, the aim is to

⁎ Corresponding author. Tel.: + 61 8 9334 8060; fax: + 61 8 9334 8001. E-mail address: [email protected] (C. Klauber).

identify the technologies or practices for either the alternative use of bauxite residues and/or improved storage practices.

2. Overview of bauxite residue management 2.1. History and future of bauxite residue In order to appreciate the trends in both awareness and perceptions of residue, the evolution of technologies, management systems and the regulatory situation, it is worthwhile examining the history of the alumina industry and the inevitable growth in the world residue inventory and how it has been managed. In 1885, Karl Josef Bayer moved from Brno (Czech Republic, then Austro-Hungarian Empire) to St. Petersburg, where he joined the Tentelev Chemical Plant to work on the dying of cotton fabrics using pure aluminium hydroxide, which was a rare and expensive raw material. By 1887 he had discovered that aluminum hydroxide could be precipitated in crystalline form from a cold sodium aluminate (NaAl(OH)4) solution, if a seed of aluminium hydroxide was used. On July 17, 1887, the Imperial Patent Office of Germany issued patent number 43977 “Verfahren zur Darstellung von Thonerdehydrat und Alkalialuminat”.1 Shortly after, he discovered that the required NaAl (OH)4 solution could be prepared by heating bauxite under pressure 1

Process for synthesis of aluminium oxide hydrate (gibbsite) and alkali aluminate.

0304-386X/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.hydromet.2011.02.006

34

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

3000

120

Global residue production and inventory 2500

annual residue production cumulative inventory

100

2000

80

1500

60 1000

40 500

20

0

cumulative residue (Mt)

residue production rate (Mtpa)

140

1900

1920

1940

1960

1980

2000

0

year Fig. 1. Global production rate and cumulative inventory of bauxite residue.

in concentrated caustic soda (NaOH) solution, a process that he patented shortly after (Patent number 65604 issued by the German Imperial Patent Office 31 January 1892). Bayer's two patents formed that basis of the industrial process of extracting and synthesizing gibbsite that now bears his name (Habashi, 1988).2 The potential for the Bayer process to provide alumina as feedstock for the newly invented process for manufacturing aluminium metal was quickly realized. By 1892 Bayer plants had been established in England, France, Italy and Germany for this purpose. Between 1893 and 1910 two further Bayer plants were built in France and the Pittsburgh Reduction Company (later Alcoa) built the first Bayer plant in the USA at East St. Louis (Illinois). In the ensuing 30 years, more plants were built in the USA, Germany, Great Britain, Japan and the Soviet Union. Of these early plants, only the plant at Gardanne (France) is still operating today. From these beginnings the aluminium industry grew rapidly. From 6800 tonnes in 1900 the global production rate of aluminium metal reached 1 Mtpa in 1940. The global inventory of bauxite residue at that time can be estimated to have been approximately 22 Mt, associated with production facilities that were small by today's standards. The available literature up to that time makes little mention of bauxite residue, other than to account for dealing with it as part of the cost of production. It appears that neither the industry, public nor governments viewed bauxite residue as an issue of social or environmental importance. Technology development and economies of scale have driven the development of increasingly larger production facilities. By the 1980s, 1 Mtpa was the size of new plants being built in Australia and Brazil, with plants in Spain and Ireland expanding beyond 1 Mtpa in the early 1990s. Today Alunorte (Barcarena, Brazil) is targeting a production of 6 Mtpa. By 1985 residue was being produced at a rate of approximately 48.5 Mtpa and the global inventory of residue neared the 1 Bt mark. In 2007 the annual production of residue was an estimated 120 Mtpa and the global inventory had risen to over 2.6 Bt. Thus it took approximately 90 years to generate the first billion tonnes of residue, but only approximately 15 years to produce the second. These figures suggest that 3 Bt will be exceeded between 2010 and 2015. Our estimates of the growth in the inventory and rate of production of bauxite residue from 1900 to 2007 are shown in Fig. 1 and are based on alumina and aluminum metal production. Where alumina production data is not directly available, the annual production values have been inferred by applying a factor of 2 to the production values for aluminium metal. This factor is calculated from the stoichiometric ratio of 1.9 plus an additional 0.1 to allow for alumina being produced for non-metallurgical uses (0.08, Roskill, 2

The dates suggested by Habashi (1988) are incorrect as can be evidenced from the original patents. The correct dates are presented here.

2008) and for dust and other losses in the conversion of the oxide to the metal. Bauxite residue tonnage was then estimated by applying an overall ratio of 1.5 to the alumina production (Hamada, 1986). Factors such as bauxite grade, alumina extraction efficiency and caustic losses will vary between refinery operations so a world average ratio value of 1.5 should only be regarded as approximate. 2.2. Need for a cohesive residue database: Development of BRaDD It appears to have been in the early 1980s, when the bauxite residue inventory was approaching the 1 Bt level, that bauxite residue became recognized as an issue of increasing environmental importance to the future of the industry. In 1981, the United Nations Environmental Programme (UNEP) and the United Nations Industrial Development Organization (UNIDO) jointly sponsored a workshop in Paris to examine the environmental aspects of alumina production. Overviews of the workshop, “Environmental Aspects of Alumina Production”, are available as a UNEP pamphlet (United Nations Industrial Development Organisation, 1985), and in a conference paper by (Hamada, 1986). In 1986, The University of the West Indies and The Jamaica Bauxite Institute conducted the first conference and workshop on bauxite tailings in Kingston (Jamaica) co-sponsored by UNIDO and the International Research Development Centre (IRDC) and major producers of alumina and aluminium (Wagh and Desai, 1986). A subsequent bauxite tailings workshop was held in Perth (Australia) 1992 under the sponsorship of the Australian bauxite and alumina producers (Glenister, 1992). Other useful sources of publications relating to bauxite residue storage and disposal have been conference proceedings of the annual Light Metals conferences of The Minerals, Metals and Materials Society (TMS), the Alumina Quality Workshop held every 3 years in Australia, and the International Committee for the Study of Bauxite, Alumina & Aluminium (ICSOBA). In the general academic literature (journal and conference proceedings) only a small number of publications relating to bauxite residues and storage issues have appeared. From the first publication in 1977 to date, there are about 90 papers (excluding pre-1990 conference proceedings and patents) and CSA's Aluminium Industry Abstracts lists 220 publications since 1968 including patents, peer-reviewed journal articles, other articles and conference proceedings. Surveys of the accessible literature on residue revealed that the various data domains placed a severe limitation on the ability to collect, systematize and interrogate information on residue. A number of key features are apparent: 1. Relevant information is scattered over many public domain sources: journals, conference proceedings, patents, company reports, securities and exchange commission reports, newspaper articles, PhD and MSc theses, and other miscellaneous sources. Much of the information is not peer-reviewed, but is essential to construct an understanding of residue management. 2. Each source of information provides only a limited amount of relevant data, and the data type is not consistent across publications. 3. The relevant information is published in a variety of languages, principally Chinese, English, French, German, Portuguese, Russian and Spanish. 4. The nature and scope of the information is owner and/or refinery specific and not consistent in either form or content. 5. Each refinery has unique operating details with respect to bauxite residue technologies, management and engineering practices. 6. The operating details of each refinery may or may not change in any given year. That is, for every refinery, a multifactorial data set propagates over time, which must then be combined over all refineries in order to provide a true reflection of technology, storage practices, engineering,

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

science, utilization options and chemistry relevant to bauxite residue. The absence of a means of collecting and managing information on bauxite residue is a key impediment to elucidate knowledge gaps about bauxite residue. This prompted the creation of an on-line database that would contain details on refinery practices on bauxite residue storage, disposal, and technologies gathered from a wide range of sources with an aim to provide a comprehensive resource through which data could be compared and analyzed in a variety of ways. The database, known as the Bauxite, Residue and Disposal Database (BRaDD), is publicly available to interested parties at no cost (see Gräfe et al., 2009, 2010). BRaDD has been an essential enabling tool for the collation and interpretation of information required for this and the subsequent reviews in this series. It is hoped that later versions of BRaDD, with data input from the industry, will provide invaluable support to the development of a global strategy for the future of bauxite residue management. 2.3. Bauxite residue production process: red side The success of the Bayer process is demonstrated by the fact that, despite enormous technological advances and the need to process a wide range of ore types, the basic chemistry and operational steps of a modern Bayer plant are fundamentally the same as originally described in Bayer's patents. In combination with the Hall–Héroult process for smelting alumina to metal, it remains the only viable process for the production of the metal, a situation that is highly unusual, if not unique, in the history of metallurgical processes. The chemical and physical nature of bauxite residue is determined by the nature of the bauxite and the effect that the Bayer process has on it. Hence it is relevant to consider the main aspects of a modern Bayer plant that affect residue properties. The half of the Bayer plant that affects the properties of the residue produced includes a series of unit processes that start with bauxite and finish with bauxite residue; this is often referred to as the “red side”. A contemporary Bayer plant's red side generally includes the unit processes of bauxite milling, pre-desilication, digestion, clarification, and counter-current decantation (CCD) washing (thickener series). A further thickening and/or filtration step prior to residue discharge (usually to a bauxite residue disposal area or BRDA) follows these unit processes. A schematic of these basic steps is shown in Fig. 2. Each unit process influences the chemical composition of the residues, as well as their physical and mineralogical attributes and the way they behave subsequently in the BRDA. Bauxite residue management begins at the point of separation of the green liquor from the solids remaining after digestion. This is usually achieved in a thickener (also known as a settler) which separates residues from saturated sodium aluminate (NaAl(OH)4) solutions. After the green liquor has been separated, the residue underflow is transferred to the counter-current decantation (CCD) washer trains where the residues are washed in several stages in raked thickeners (settlerwashers) to recover NaOH and NaAl(OH)4 from the remaining entrained liquor and from the solids that slowly release these constituents into solution. After washing in the CCD train, the residue is usually filtered or treated in a final thickener to increase the solids content prior to being transported to the BRDA. Filtration is used in some plants currently, but the majority of plants use thickeners for this purpose. The thickener may typically be either a deep cone thickener (Alcan) or a superthickener (Alcoa). The most important influences of each process in relation to bauxite residue are summarized in Tables 1 and 2.

35

simplest method, the residue slurry from the washing circuit is disposed directly into the sea, usually via a pipeline that takes the slurry well offshore for discharge into the deep ocean. For lagooning the residue slurry from the washing circuit is pumped into land-based ponds. Such ponds may be formed within natural depressions using dams and other earthworks to ensure secure containment. More recently the emphasis has been on dry disposal, namely “dry” stacking and dry cake disposal. “Dry” stacking is a misnomer as the residue is actually not dry at time of discharge. Each method has advantages and disadvantages that are summarized in Table 3. The main factors determining the choice of disposal method for a given refinery are local rainfall and topography and the land availability in the context of the grade and mineralogy of the bauxite and the size of the refinery (which together determine the rate of residue production) and of course legislative requirements. Secondary factors that can influence choice are the economics of trucking versus pumping to the BRDA. French and Japanese practices have historically favoured disposal at sea as the best option on economic and environmental grounds. North American practice has favoured landbased disposal in lagoons, even for refineries on the seaboard. The 1970s saw a major expansion of the alumina industry in response to growth in primary aluminium production, resulting in a rapid growth in the production rate and global inventory of bauxite residue (Fig. 1) that in turn led to the construction of new refineries throughout the world. In particular, increasingly large refineries were built in the major laterite regions of first Australia and then South America. The distribution and timing of refineries are shown in Fig. 3. Apart from a number of refineries that we have been unable to locate

3. Storage practices, engineering and science 3.1. Development of disposal methods Prior to the 1970s, there were only two disposal methods in general use, marine discharge and lagooning. Marine discharge is the

Fig. 2. Schematic of a general Bayer process “red side”, highlighting those processes affecting the disposal and storage of bauxite residues. The right arrows for the countercurrent wash train (e.g. 1–5) indicate the flow of wash water opposite to the flow of residue. The final residue preparation stage may be a deep cone type thickener (5), superthickener (e.g. 6) or a filter. Some refineries practice a variety of pre-disposal neutralization amendments.

36

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

Table 1 Bayer plant unit processes affecting bauxite residue characteristics relevant to disposal and storage. Plant process

Effects relevant to disposal and storage

Bauxite milling

• Sets the effective particle size distribution and specific surface area of bauxite particles for digestion • Particle size and surface area are key factors in determining physical properties (e.g. Settling rates, mud rheology) and chemical properties (e.g. Adsorption of flocculants, organics and metal ions, dissolution rates of minerals and precipitates) • Introduces sodium alumino-silicates (desilication products or DSP) to the residue • High temperature, high pressure environment which alters the elemental and mineralogical make-up of the original bauxite particles into bauxite residues • Nature of the bauxite dictates temperature, pressure and NaOH concentrations (see Table 2) • Introduces large amounts NaOH to the solids increasing their ph, alkali and Na+ content • Al is solubilized and removed from the solid phase • Mineral transformation reactions occur: less stable minerals dissolve and re-precipitate; other minerals undergo solid-state dehydration reactions which convert them to more stable mineral phases • Conducted in high volume raked settlers • Slurry is mixed with chemical additives to promote settling and to clarify the overflow liquor • Bauxite residue is separated from the green liquor • Conducted in high volume raked washers • Flocculants added to achieve required underflow densities • Recovery of NaOH and NaAl(OH)4 in the remaining liquor of the residues to be returned to the main Bayer process liquor • Underflow of the last washer is optimized according to the disposal method • Increases solids density to meet storage or disposal requirements • Further flocculant or filtration aid added • Additional washing

Pre-desilication Digestion

Settling

Washing

Final thickening or filtration

in China, we believe this distribution map accurately accounts for the majority of global alumina production. This rapid expansion coincided with increasing environmental awareness. Regulation in the developed world from the 1970s onwards led to the development of new approaches to residue management. Table 4 shows residue disposal methods used for 17 refineries for which information could be found for each of the years nominated. These 17 refineries accounted for 44% of the 76 million tonnes of alumina produced globally in 2007 (U.S. Geologic Survey, 2008). Noteworthy is the consistent trend away from marine disposal to land-based disposal, and from wet to “dry” disposal methods. In 1965, one third of the refineries considered were disposing of residue by marine dumping, and the other two thirds were impounding it in lagoons. In the 1970s research into alternatives led to the development of “dry” stacking methods. The subsequent increase in the number of plants using “dry” stacking, and the conversion of a number of refineries from wet to dry disposal, reflects its emergence as the favoured option for residue disposal (Banerjee, 2003; Cooling, 1989) and certainly the preferred practice in the construction of new refineries. By 1985 43% of the refineries were using “dry” stacking, growing to 70% by 2007. As noted in Table 3 it has the particular advantages of reducing the land area required for residue storage, minimizing the potential for liquor release to the surrounding environment and also maximizing the recovery of liquor (and entrained value) to the refinery. One serious drawback however is the increased risk of fugitive dust requiring mitigation methods. 3.2. Marine disposal The 1981 UNEP/UNIDO workshop concluded that “disposal [of bauxite residue] into rivers has been strongly discouraged and

Table 2 Bayer digest variables. Hudson (1987). Main Al mineral

Gibbsite Al(OH)3

Boehmite γ-AlOOH

Temperature (°C) Pressure (atm) NaOH (M)

104–145 1.0–3.0 8.9–3.6

200–232 6.0 5.0–3.6

marine disposal regarded as a last resort where suitable land disposal is not available” (United Nations Industrial Development Organisation, 1985). This appears to reflect the opinion of the majority of companies involved in alumina refining. There has been a shift away from direct marine or river disposal as noted in Table 4. To our knowledge no new refineries were built after 1970 that employ marine disposal. Presently only 2–3% of global alumina production results in marine disposal of residue. Residue from the Gramercy plant in Louisiana was disposed into the Mississippi River until 1974, at which time a transition was made to lagooning. The operating company (Kaiser) voluntarily removed residue from the river and relocated it to land-based lagoons. Effort was also directed towards dewatering and treating the residue to enable the creation of useful products in order to limit the requirements for additional area for residue disposal (Kirkpatrick, 1996). Plants in Japan are restricted in relation to land area available for disposal of residues, and so have historically marine discharged. Responses to find beneficial uses for residue (e.g. as a supplement to cement manufacture) have proved insufficient. Other approaches to minimize residue have been to maximize the imported bauxite grade and to directly import smelter feed “hydrate” (International Maritime Organization, 2005). Japan made a commitment in 2005 to the International Maritime Organization that disposal of bauxite residue to the sea would be discontinued by 2015. The country reported that in the previous 5 years residue had been disposed to the sea at two locations at a total rate of 1.0 Mtpa. Environmental impact assessments at both sites concluded no impact on water quality and marine organisms in the water column in the region of sedimentation but some differences in the nature of the seabed between the dumping sites and control sites, i.e. only limited impact was reportedly observed. It was intended that further monitoring campaigns would be carried out every 2 to 3 years (International Maritime Organization, 2005). The plants at Gardanne in France and Viotia in Greece that still use marine dumping are actively pursuing alternatives (Martinet-Catalot et al., 2002). Dauvin (2010) summarizes the most recent impact assessments for the Cassidaigne canyon. It is noteworthy that as of 2010 only 0.18 Mt (dry weight) will be discharged at Gardanne and they (along with Japan) will cease all marine discharge by 2015. A study of the discharge from the Viotia plant into the Gulf of Corinth showed that an extensive deposit has been formed close to the point of discharge from a pipeline at a depth of 100 m. The residue also

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

37

Table 3 Summary of the advantages and disadvantages of the four residue disposal methods. Advantages

Disadvantages

Marine disposal No land required for storage No closure and rehabilitation of storage areas No impacts associated with land-based storage (dusting, contamination of surface and ground waters, and leaching of toxic metals) Lower capital and operating costs than for land-based disposal methods No additional thickening or filtration of the slurry prior to disposal Does not return impurities to the Bayer plant liquor

Covering of sea-bed and destruction of associated ecosystem Potential for release of toxic metals to marine environment Increased turbidity of seawater due to dispersion of original residue and the formation of colloidal magnesium and aluminium compounds Unknown effects on ecosystems and food chains Increasing community/regulatory pressure to discontinue the practice Loss of soda and alumina values with entrained liquor

Lagooning (low wt% solids) Lowest capital cost land-based disposal method No additional thickening or filtration of the slurry prior to disposal Dust potential minimized by maintaining a supernatant liquor layer Does not return impurities to the Bayer plant liquor

“Dry” stacking (48–55 wt.% solids) Land area required for storage is minimized Structures required for secure containment are minimized Soda and alumina values in the liquor are returned to the Bayer plant Potential for leakage to groundwater is minimized Risk of surface water contamination is lowered by smaller area of open contaminated water and returning runoff to the Bayer plant Rapid de-liquoring assists area management, including rehabilitation Hazards associated with open caustic lakes are minimized Recovery of solids for alternative uses is facilitated Dry cake disposal (N 65 wt.% solids) Land area required for storage is minimized Collapse risk of a raised mound or impoundment is minimized Soda and alumina values in the liquor are returned to the Bayer plant which reduces the need for washing stages Low leakage risk to groundwater due to the low pore volume Environmental and safety hazards associated with the presence of open caustic lakes are avoided Residue can be recovered by standard earthmoving equipment Rehabilitation and closure are facilitated by the ability to landscape the deposit and by its low alkali content

extends over the sloping shelf of the sea floor and into an abyssal plain to a depth of 860 m at distances of up to 17 km from the point of discharge. The main mechanism of transport of the residue from the shelf to the abyss was found to be by gravitational flow triggered by seismic events, which are common in the region (Varnavas et al., 1986). 3.3. Lagooning The residue slurry from the washing circuit is pumped into landbased impoundments. In the absence of suitable natural topography, ponds may exist as fully fabricated structures, generally excavated into the existing land. In either case, best practice is to line the ponds with sealants to minimize liquor leakage to the underlying ground and ground water. The linings vary in relation to both the materials used and the complexity of the construction. The simplest is a single layer of compacted clay which separates the residue from the original soil or rock of the storage area, but the integrity of such layers cannot generally be guaranteed (Thomas et al., 2002). Additional security can be achieved by the use of multiple layers featuring impermeable plastic or geo-membrane materials to form a seal between the residue and the supporting clay or other layer beneath. Minimizing the hydrostatic pressure by removing liquor by decanting and in some cases by also under-draining the deposit further contributes to the reliability of the seal (Cooling, 1989). Storage problems can be further mitigated by neutralizing the slurry prior to discharge, for example by

Provision of substantial areas of land Long term planning and funds for closure and rehabilitation Loss of soda and alumina values with entrained liquor If not neutralized: Poorly compacted highly alkaline residue and highly alkaline lake Range of hazards (risk of contact of humans and wildlife with caustic liquor and residue, and contamination of surface and ground waters) Costly engineering to prevent catastrophic failure of impoundments Difficult to close and rehabilitate because of large amounts of caustic liquor — possibility of an indefinite legacy

Sufficient areas of land for storage Planning and funds allocation for long-term closure Additional stage of thickening or filtration prior to discharge. Liquor impurities are returned to the Bayer plant (process impacts) “Dry” stack surfaces are subject to fugitive dust lift-off and mitigation processes may be required (e.g. sprinklers) Difficult to achieve in areas of high rainfall and low net evaporation Significant compaction density hinders the establishment of vegetation

Provision of sufficient areas of land for storage Long term planning and funds for closure and rehabilitation Installation and operation of a filtration plant capable of handling entire residue production Liquor impurities are returned to the Bayer plant (process impacts) Surfaces are subject to dust lift-off so that mitigation processes may be required

mixing with seawater or by the addition of mineral acids (e.g. sulphuric acid). Lagooning is the simplest land-based disposal method, but due to the low solids content is also the most dependent on good engineering practice for the residue containment. The increased risk with low solids containment is tragically demonstrated by the impoundment wall collapse at the Ajka Timföldgyár alumina refinery in October 2010 (Enserink, 2010). Effective low solids in an impoundment area can also arise from high rainfall events. 3.4. “Dry” stacking “Dry” stacking is a significantly different disposal concept and could more generically be termed thickened tailings disposal (TTD). Unlike lagooning residue at a low solids level, the residue slurry from the washing circuit is thickened to a paste (in the initial range of approximately 48 to 55% solids) prior to discharge. The paste is thixotropic, and so is amenable to pumping and pipeline transport. Upon discharge from the pipe it will flow down a slope without segregating or settling to form layers of uniform thickness at an angle of repose in a range of 2° to 6° (in nonfreezing conditions). The residue deposit is built into a stack by discharging the paste progressively in thin layers that are allowed to dewater and air-dry before being overlain with the next layer. Upon standing, the paste consolidates (with liquor decant) to a density of 62 to 65% solids, with a final dry density around 70% and an accompanying

38

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

Fig. 3. World map showing the distribution of alumina refineries based on establishment date (before or after 1970), closed or unknown.

increase in shear strength. Shear strengths of 20 kPa or more may be reached (Nikraz et al., 2007). The consolidation occurs as the solids settle and liquor is removed by a combination of decantation and surface evaporation. Modern designs also include under-draining to facilitate the consolidation process. The process is designed such that the dried paste becomes substantially self-supporting so that it can be safely stacked to considerable heights within containment bunds, minimizing the physical footprint of the disposal area (Cooling, 1989). As with lagooning the hazards associated with alkalinity may be reduced by neutralization but any changes to paste rheology need to be considered. Methods for disposing of residue with prior de-liquoring were pioneered in Germany with a view to minimizing the land required for disposal and maximizing the return of soda and alumina to the refinery. The term “dry stacking” appears to have been first used to describe the process developed at the Ludwigshafen plant in the 1970s (Anon, 1982; Pohland et al., 1981; Pohland and Tielens, 1983, 1986). Residue slurry from the washing stage was filtered on drum filters that discharged a cake at about 55% solids by weight. The cake was fluidized by mechanical agitation for pumping to the drying area where it formed a mound with a shallow angle of repose (approx-

imately 5°). The deposition schedule was managed to optimize surface drying while minimizing dusting potential. It was found that: • The residue consolidates to about 62–67% solids, enabling it to be traversed by machinery minimizing the requirements of the containment dykes • Degree of fluid decant was minimal, and rainwater could be collected and either discharged to the environment or returned to the Bayer plant according to purity • Material compacts on standing to become highly impermeable (permeability in the order of 10−9 cm/s), removing the necessity to line the disposal area to prevent seepage of impurities into the underlying soil and ground water A similar system was developed and implemented by Alcan at the Arvida plant in Canada (Paradis, 1992, 1993; Robinsky, 1982). In this case special thickeners were developed to produce a residue of sufficient solids density to behave as a non-segregating slurry. This led to the development of the Alcan Deep Cone Thickener, a new technology for producing densities of over 50% solids suitable for

Table 4 Storage practices at 17 refineries (representing 44% of global alumina production in 2007) for which information is available over the period 1965–2007. Startup date

Refinery

1965

1975

1985

2007

1959 1963 1965 1967 1972 1972 1972 1973 1973 1980 1983 1983 1984 1984 1987 1995 2005

Kirkvine Kwinana Aluminium de Grece Queensland Alumina Ltd Alcoa Alumino S.A. Gove Alumina Pinjarra VAW Stade Eurallumina San Ciprian Aughinish CVG Bauxilum Worlsey Alumina Wagerup Dhamanjodi Alunorte Barcarena Yarwun

Lagoon Lagoon Sea disposal

Lagoon Lagoon Sea disposal Lagoon Lagoon Lagoon Lagoon Unknown Lagoon

Lagoon “Dry” stack Sea disposal Lagoon Lagoon Lagoon “Dry” stack “Dry” stack Lagoon “Dry” stack “Dry” stack Lagoon “Dry” stack Lagoon

Lagoon “Dry” stack Sea disposal Lagoon “Dry” stack “Dry” stack “Dry” stack “Dry” stack Lagoon “Dry” stack “Dry” stack Lagoon “Dry” stack “Dry” stack “Dry” stack “Dry” stack “Dry” stack

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

the production of thickened tailings without the need for filtration (Peloquin and Simard, 2002). In 1996 Alcan entered into a license agreement with EIMCO for the further development and marketing of Deep Cone thickeners. According to EIMCO (now FLSmidth Minerals) there are over 50 Alcan Deep Cone thickeners now operating in alumina plants, and the technology has been deployed to other industries including copper and lead/zinc processing (Emmett and Klepper, 1991; Emmett et al., 1992). Thickener performance issues are reviewed by Farrow et al. (2000). The National Aluminium Company (NALCO) plant at Orissa, India, applied the Alcan technology (Banerjee, 2003; Das et al., 2003) to replace the original wet disposal method and they simply refer to it as TTD. The system was installed as part of a major plant expansion from 0.8 to 1.575 Mtpa. The existing wet disposal method would not have been capable of supporting the projected residue production over the ensuing 20 plus years, because of limitations on the availability of suitable land area and environmental concerns associated with wet disposal. The alternative of dry cake disposal was also considered, but was rejected, primarily on grounds of capital and operating costs. Concerns related to the safety and environmental aspects of trucking the material in hilly terrain were also cited (Banerjee, 2003). In addition to the “dry” stacking advantages outlined in Table 3, the TTD system was seen to have additional advantages in relation to wet storage of reduced earthquake and erosion risks and lower energy consumption and pipeline requirements. The rheological properties of bauxite residues vary widely between plants and successful “dry” stacking requires an understanding of the rheology (Farrow et al., 2000; Sofra and Boger, 2002) as slurries are thixotropic and their flowability increases with the amount and manner of mixing energy applied (Pashias and Boger, 1996). For example, slurry that has been allowed to settle to 50% solids will be firm and seemingly un-pumpable, but if sufficient mixing energy is introduced, the viscosity is decreased as the pore liquids create a network enabling the slurry to be pumped. Without sustained energy input it will slowly revert back to a firm condition. These flow behaviour changes occur without the solids density of the material changing. The key rheological parameters for slurry thickening

39

and transport design, and for spreading behaviour and slope formation prediction are yield stress and viscosity as a function of solids content and the shear and compression history of the material (Green and Boger, 1997; Sofra and Boger, 2002). The particle size distribution is also important (Doucet, 1999) as is the nature and rates of the chemical additives (in particular flocculants and flowability aids). These principles have also been successfully applied at the MOTIM plant in Hungary, where the residue is filtered on vacuum drum filters to 50% solids, fluidized by high-intensity stirring, and then pumped to the residue storage area where it is stacked in layers to dry. After drying times between 80 days (summer) and 160 days (winter), the solids content rises to around 65%, at which point the material can be mechanically spread using conventional earth-moving machines (Solymar et al., 2002). Alcoa determined that underflow densities in the region of 50 wt% solids were sufficient as the starting point for a “dry” stacking system provided with underdrainage for efficient de-liquoring (Cooling, 1989) of the deposited residue and return of the recovered liquor to the Bayer plant. The required solids density was achieved in a specially designed, large diameter (90 m) thickener (a “superthickener”) and associated rake system developed in collaboration with EIMCO (Marunczyn and Laros, 1992, 1993). As the raking loads would exceed the capability of any centre-drive design, a peripheral traction drive was employed (Emmett and Klepper, 1991). Between 1985 and 1992, Alcoa implemented at its Western Australian refineries a stacking system in which residue is deposited into rectangular cells on a rotational basis to assist drying and reduce the land area. It was found that final densities of ~ 70 wt.% solids are achievable, although the time required was not reported (Cooling and Glenister, 1992). A schematic of this system is shown in Fig. 4. At 70 wt.% the shear strength of the stack is N 30 kPa (i.e. high structural integrity) and not only is the storage volume requirement reported to be about 60% of that required for lagooning, but greater impoundment heights are more easily constructed. Further under-drainage may be introduced progressively as the stack is built. Some mechanical ploughing of the surface is required to optimize drying rates, and water sprinklers are used on an as needs basis to mitigate fugitive dust lift-off. Fig. 5 shows a satellite image of the Pinjarra residue

Fig. 4. Schematic of Alcoa dry-stacking method for bauxite residue disposal. Redrawn from Alcoa (2005a).

40

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

Fig. 5. Google Earth™ image of Pinjarra refinery and residue area in 2006. Captions inferred from Pinjarra long term residue management plan (Alcoa, 2005b). GOOGLE is a trademark of Google Inc.

area (the labeling of the areas is according to the Pinjarra long term residue plan (Alcoa, 2005b). This residue management has since been deployed to other operations including those in Suriname, Spain, and Pt Comfort Texas (Alcoa, 2001). It should be noted that the storage volume advantage over lagooning may be operationally relevant for a time, but the final compaction of the solids in a lagooned impoundment would be similar to dry disposal. Issues such as yield stress values and permeabilities at a given weight% solids must also be considered in the context of residue particle size distributions, shape and mineralogy. For example, Jamaican residues settle to considerably lower% solids than other residues (Roach et al., 2001) but exhibit similar viscosities, leading to the conclusion that comparisons of residue behaviour should be made on a volume/volume basis and not% solids. 3.5. Dry cake disposal Dry cake disposal refers to the practice of mechanically removing as much water as possible from the residue to produce a dry cake with a solid's content of N 65 wt% prior to disposal. The special case of dry cake disposal is not always specified as such in available information and “dry” stacking tends to be used to encompass both high wt% solids methods. The dry cake is not pumpable, so is generally moved to the disposal area by conveyors or trucks. There is no significant further de-liquoring once the residue has been delivered to the storage area and these factors distinguish it from “dry” stacking. Similarly to the other methods, any hazards associated with entrained alkalinity can be further reduced by washing or neutralizing the cake on the filters. Filtration on vacuum drum filters followed by dry cake disposal has been applied successfully in practice, for example at the Hindalco plant at Renukoot (Shah and Gararia, 1995). It is not possible to achieve a dry cake by means of thickening alone, and so a filtration stage is required. Vacuum filtration is not generally satisfactory for this purpose, as it is limited in the available pressure drop across the filter cake to b 0.75 bar (Oeberg and Steinlechner, 1996). This is insufficient to overcome the capillary resistance inside the pores of the

residue to fully de-liquor the cake. By using pressure filtration it is possible to achieve differential pressures in the order of 6 bar which is sufficient to fully de-liquor. The non-thixotropic nature of this cake is fundamental to the success of the process. It means that the cake can be readily discharged from the filter by air blow-back or mechanical means, and can then be transported to the disposal area by truck or conveyor where it can be stacked with a minimum of additional treatment in relatively simple containment facilities (Kainuma et al., 1979; Oeberg and Steinlechner, 1996; Steinlechner et al., 1996). Recent trials of BOKELA hyperbaric filtration technology have been successfully demonstrated using a Bokela Hi-Bar® at the pilot plant scale at the Aluminiumoxid Stade (Germany) plant (Bott et al., 2002). This trial demonstrates the possibility of a marked improvement in the removal of fluids and salts from bauxite residue in which solids contents of N 75% can be achieved with the application of steam in the filter. This filtration method produces soda contents as low as 3 g/kg as compared to 6 to 12 g/kg by vacuum filtration. The resulting cake is described as “extremely dry and well washed … with a crumbly, sandy-like and nonsticky consistency and a low soda content” (Bott et al., 2002). Similar results were also achieved by Maschinenfabrik Andritz at the CVG plant in Venezuela (Oeberg and Steinlechner, 1996). Producing a residue with these characteristics greatly simplifies handling and storage requirements, reduces the potential for environmental impact, and broadens the options for rehabilitation and/or reuse of the material. A number of these advantages were previously demonstrated by Nippon Light Metals at their Tomakomai works (Kainuma et al., 1979). Given that the technology has been demonstrated at a plant scale, the challenge is to develop it as an economically attractive alternative to current methods, and then to deploy it to existing and future plants. 3.6. Bauxite residue amendment Except for dry cake disposal bauxite residue is discharged as a slurry or paste whose high alkalinity (pH N 11) and high sodium content

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

favours particle dispersion. In addition, the solids in the slurry contain a high proportion of fine, silt- to clay-sized particles. The combination of these chemical and physical factors creates significant difficulties in the handling, storage and remediation or utilization of bauxite residue. For long-term storage or utilization of the residue, it may be necessary to release liquor to the surrounding environment. This will generally require it to be neutralized in some way and diluted prior to discharge. The pore fluids remaining in the stack will generally also need to be diluted and/or neutralized. Solid alkalinity can also have a profound effect on attempts to neutralize the pore water and is a major factor to be considered in any evaluation of possible future storage or utilization scenarios. The need to amend bauxite residue grew with the rapid growth of industry output and increasing levels of environmental awareness and regulation. Although significant general progress has been made with the advent of “dry” stacking, it is reasonable to conclude that pressure filtration, washing and at least partial pH neutralization would remain essential to any long term disposal solution. However, pressure filtration in particular is costly and its implementation without regulatory incentive is unlikely. The key to successful amendment is the complex chemistry, this is treated in detail in part three of this review series (Gräfe et al., 2011). 3.6.1. Alkalinity of bauxite residue The object of the digestion step of the Bayer process is to dissolve aluminium-containing minerals while minimizing the dissolution of other mineral phases. This is achieved in a strongly alkaline solution at temperatures above 100 °C (mostly in the range 145 to 250 °C), so that minerals such as gibbsite (Al(OH)3) and boehmite (γ-AlO(OH)) are dissolved to produce [Al(OH)4]− (aluminate) ions in solution, according to the following reactions: −

−

AlðOHÞ3ðsÞ þ OH ⇔½AlðOHÞ4 ðaqÞ −

−

γ  AlOðOHÞðsÞ þ OH þ H2 O⇔½AlðOHÞ4 ðaqÞ The bauxite residue is then separated from the main liquor stream and washed to recover soda and alumina. The resulting residue contains pore water that is considerably more dilute (by a factor of at least 10) than the main Bayer plant liquor, but is nevertheless highly alkaline. If this residual solution is then neutralized, aluminium hydroxides will re-precipitate. The nature of the precipitates depends on the method and conditions of precipitation, but will generally include Al(OH)3 and/or AlOOH. The chemistry of the residue slurry is thus determined by the interactions between the bauxite and the plant liquor during digestion, followed by the changes that occur after the bulk solids are separated from the digest liquor and progressively washed (usually CCD) to remove sodium hydroxide, aluminate and carbonate. During the washing process various chemicals are added, most notably flocculants to settle the residue and clarify the solution for return to the Bayer liquor circuit, and lime to remove carbonate ions from the liquor by precipitation of CaCO3 and replacement of carbonate by hydroxide ions in solution, in a process called “causticization”, nominally: 2−

CaO þ CO3 þ H2 O→CaCO3ðsÞ þ 2OH

−

This produces pH buffers in the form of calcium carbonate, tricalcium aluminate (TCA, stoichiometric formula Ca3Al2(OH)12) and calciumaluminate carbonate (hydrocalumite, stoichiometric formula Ca4Al2 (OH)12.CO3.6H2O). Apart from the intended solubilization of aluminium species, there are a number of other reactions that occur during and prior to digestion. One of the most important is the solubilization of phyllosilicate minerals (clays), e.g. kaolinite Al4Si4O10(OH)8. The dissolution of these minerals results in aluminate and silicate ions in solution. The digester discharge

41

liquor, which is the solution produced by the digestion process, also contains Na+ ions in high concentration. As a result, the liquor becomes unstable with respect to a range of sodium alumino-silicates, in particular sodalite3 and/or cancrinite.4 These sodium alumino-silicate solids that form during pre-desilication and pre-digestion are collectively referred to as “desilication products” or DSPs because they remove silica, sodium and hydroxide ions from solution by incorporation. DSP is a form of solid alkalinity, stable while the solution is alkaline, but decomposing in neutral to acid solutions, releasing sodium, hydroxide, aluminate and silicate ions into solution. In this way the DSPs act as significant solid alkaline pH buffers. A number of side reactions can occur between the added lime and the aluminate ions in the Bayer solution. The exact reactions and the nature of the products depend on conditions, but the general reactions may be illustrated by the formation of hydrated calcium aluminate and aluminate carbonate as follows (Rosenberg et al., 2004): −

−

3CaðOHÞ2 þ 2½AlðOHÞ4  þ nH2 O→Ca3 Al2 ðOHÞ12 :nH2 OðsÞ þ 4OH −

2−

−

4CaðOHÞ2 þ 2½AlðOHÞ4  þ CO3 þ nH2 O→½Ca2 AlðOHÞ6 2 :CO3 :nH2 OðsÞ þ 4OH

Alkaline solids can also be formed with other anions such as phosphate, zincate and titanate, and with other cations such as magnesium and potassium. Both the quantity and forms of alkalinity must be taken into account in any process for neutralizing the residue. For example, the amount of carbonate in solution determines the buffering capacity of the solution above pH 10. In the region of neutral pH additional buffering is provided by (solid) aluminium hydroxides, and if significant carbonate is present, also by sodium alumino-carbonates, in particular dawsonite (NaAl(OH)2CO3). 3.6.2. Seawater remediation Apart from direct marine disposal, such as has been practiced at Gardanne since the very early days of the industry (Martinet-Catalot et al., 2002), the first example of onshore remediation of residue by mixing with seawater appears to be at the Queensland Alumina refinery. In this case seawater was used as a transport fluid for pumping it to disposal lagoons in order to conserve fresh water (Hanahan, 2004). Remediation by neutralization was a consequence, but not the primary aim of seawater use. By adding seawater, hydroxide, carbonate and aluminate ions are removed from solution by reaction with Mg2+ and Ca2+ ions to form alkaline solids, in particular hydrotalcite (Mg6Al2(CO3)(OH)16·4(H2O)) and calcite (CaCO3). This buffers the solution in the range of pH 8 to 9. It has been shown that the phosphate adsorption capacity of the solids is substantially increased by seawater neutralization (Hanahan, 2004), and that the solids have a high trapping capacity for trace metals (McConchie et al., 2002). These features are useful in relation to revegetation of bauxite residue areas, and for creating various reuse products (Tillotson, 2006). Seawater is also used to neutralize the decant liquor from ponding operations prior to discharge. Neutralization by simply mixing the decant liquor with seawater results in the formation of white colloidal precipitates which have the potential to produce contamination downstream of the discharge point (Baseden and Grey, 1976). At the Gove refinery, hydrotalcite, calcium carbonate and calcium oxalate were identified as the main components of such precipitates along with trace amounts of other salts (Anderson et al., 2008). Further research led Alcan Gove to develop a seawater treatment method for decant water to meet stringent standards of purity for 3 4

Na6[Al6Si6O24]·[2NaOH,Na2SO4]·nH2O. Na6[Al6Si6O24]·2[CaCO3]·nH2O.

42

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

discharge to the environment (Haggard and Smith, 2002). The purity parameters of particular interest are pH, turbidity and trace metals. In addition to meeting the designated water quality criteria, ecological toxicity testing was done to provide additional confirmation that the treated water was safe for discharge to the marine environment. It was found that a two-stage treatment was required in order to meet the design performance requirements, particularly in relation to vanadium levels. The first stage of the process is the mixing of the decant water with a six-fold excess of heated seawater (waste cooling water from the plant evaporators) to form the hydroxide precipitates. The resulting suspension is then mixed with residue slurry and treated with flocculants in a thickener to produce a clarified overflow stream for discharge. Continuous flow (bench scale) trials confirmed that the required performance parameters can be met with this design, which has been proposed for implementation at Gove (Anderson et al., 2008). Successful implementation of this technology would be a significant milestone in the development of seawater neutralization technology. It would be the first use of a Deep Cone thickener as a mixer to produce overflow water suitable for direct discharge and would eliminate the need of a labyrinthine discharge channel required to provide residence time necessary for formation and sedimentation of precipitates. 3.6.3. Mineral acid remediation There are few examples in the literature of neutralization of bauxite residue by mineral acids and whilst addressed in general terms by McConchie et al. (2002), no examples are given. The authors point out that large quantities of acid would need to be available at low cost, opining that this is unlikely to be the case for alumina refineries generally. It is also noted that residue neutralized by mineral acid contains no residual (solid) alkalinity, and so would not be useful as a starting material for the range of applications envisaged by those authors. In addition it does not improve the physical properties of the material, both of which occur if either seawater or carbon dioxide is used for neutralization. It has been reported that sulphuric acid is used as a supplementary neutralizing agent at the Yarwun refinery (Queensland, Australia) due to (pipeline) limitations in seawater supply (Sturt, 2008). The performance of sulphuric acid was assessed in terms of its ability to reduce the level of aluminate in the discharge liquor. It was found that the mineral acid was significantly less effective in removing aluminate from solution because it relies on the precipitation of aluminium hydroxide, whereas if seawater is used the solid phase formed is hydrotalcite which is less soluble at the target pH of 8.5. No references were found in the open literature concerning the neutralization of bauxite residue decant waters with mineral acids in a specific refinery context. Such a practice would be expected for some refineries in order to meet discharge license requirements. This would result in the formation of colloidal and/or gelatinous precipitates (primarily aluminium hydroxides and hydroxy-carbonates) requiring a clarification process prior to discharge. 3.6.4. Carbon dioxide remediation A process for reacting bauxite residue with gaseous carbon dioxide was piloted by Alcan for the Saramenha Ouro Preto refinery in Brazil in 1983 (Versiani, 1983). The concept has since been developed by Alcoa as a means of reducing the pH of the slurry prior to “dry” stacking (Cooling et al., 2002). Extensive laboratory and pilot-scale testing was carried out by Alcoa between 1991 and 1996. This was followed by small-scale field trials in which several mixing devices were tested. Carbonated residue was deposited in layers so that its longer term behaviour could be studied in a “dry” stacking operation. The optimum rate of CO2 addition was found to be 25 kg CO2 per m3 of residue slurry, at a slurry density of 48% solids. It was demonstrated that a pH of 9 could be achieved in the carbonated residue. The pH of the leachate collected from the under-drainage system remained steady at around 10.5 over an extended period, while the pH of the

leachate from a non-carbonated drying bed was consistently greater than 13. On the basis of these trials, full scale residue carbonation capability was implemented at the Kwinana refinery in 2000 (Cooling et al., 2002). Prior to neutralization, residue slurries are highly alkaline. The pH of the slurry is generally in the region of 13, equivalent to an OH− concentration of 0.1 M. Introducing CO2 to the solution causes rapid neutralization according to the following reaction: −

−

CO2ðgÞ þ OH →HCO3 ðaqÞ This reaction causes the pH to drop rapidly until buffering by the carbonate/bicarbonate equilibrium becomes apparent below a pH of 10. At that pH solid hydroxides begin to form, creating additional buffering as they equilibrate with the ions in solution. The Alcoa trials demonstrated that it was possible to achieve a pH as low as 8.5 by reacting residue slurry with CO2 under pressure (Cooling et al., 2002). However, after the CO2 overpressure was removed, the pH would slowly rise until a final pH of about 10.5 was reached. This rise was explained by those authors and by Smith et al. (2003) as being due to the dissolution of alkaline solid phases that become soluble in this pH range, for example TCA. As noted, TCA is present in bauxite residue as a result of the use of lime in the Bayer process for control of carbonate levels in the liquor (Rosenberg et al., 2004) and as an aid to settling and filtration (Whittington, 1996). Dissolution of TCA results in the release of OH− ions to the solution, thus raising the pH. This is accompanied by the formation of new solid phases, including calcium carbonate (calcite), sodium aluminium hydroxyl-carbonate (dawsonite) and aluminum hydroxide (gibbsite), possibly according to the following equations: −

Ca3 Al2 ðOHÞ12 þ 3HCO3 þ

ðaqÞ

þ

þ 4Na →3CaCO3 þ 2AlðOHÞ3 þ 3H2 O

−

þ4Na þ 3OH

−

2Ca3 Al2 ðOHÞ12 þ 10HCO3

ðaqÞ

þ

þ 4Na →6CaCO3 þ 4NaAlðOHÞ2 CO3

−

þ10H2 O þ 6OH

Due to the buffering effects of these reactions, the pH can be expected to stabilize in the region of 10.5. These observations are consistent with recent work by Khaitan et al. (2009a), who also observed that the amount of TCA present is a major factor in determining the carbon sequestration capacity of a bauxite residue (Khaitan et al., 2009b). Alcoa found that carbonation of residue affects the physical properties of the material as well as its chemical composition. Carbonated residue dries more rapidly than uncarbonated residue, possibly because of the elimination of surface crusting and improved cracking behaviour that provides more effective surface area for evaporation. It was also found that the mechanical strength of the residue was enhanced by carbonation. The combination of these effects results in a significant reduction in the drying time required to reach a target minimum strength to ensure long-term stability, which in turn reduces the required cycle time for drying and improves the utilization of the residue drying area. Some of the perceived benefits of carbonation of residue have been summarized (Cooling et al., 2002) as: • Reduces risk to containment seal material (clay or synthetic) and therefore reduced risk of groundwater contamination • Improves quality of drainage water • Provides a sink for greenhouse gasses • Reduces risk of future classification of the residue as a hazardous waste • Facilitates development of productive uses for the residue in the future CO2 supply is a key issue to be considered in relation to carbonation of residue. The carbonation technology applied at Kwinana requires concentrated CO2 as the raw material to be mixed with the residue

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

slurry. The Kwinana alumina plant is located within an industrial complex that includes an ammonia plant which produces pure CO2 as a by-product, this is supplied to the residue area by a dedicated pipeline (Cooling et al., 2002). Such examples of industrial synergy are rare, so alternatives need to be considered if carbonation is to be applied more generally. One alternative is to extract CO2 from flue gasses associated with the alumina refinery. The level of CO2 in such flue gasses is generally in the order of 10%, so a concentration process in which it is separated from the other gasses (mainly nitrogen) is needed. The current best technologies for capturing and concentrating CO2 from flue gasses are based on amine scrubbing (Simmonds et al., 2003). This requires the installation of a significant additional operation to an already complex alumina plant, and carries the potential for amine emissions. A promising amine-free alternative is based on carbonation of residue liquor by contacting it with flue gasses in a high-efficiency scrubber. The carbonated solution is then recirculated with residue solids to carbonate the overall residue (Guilfoyle et al., 2005). 3.6.5. Sulphur dioxide remediation The reaction of SO2 with residue has been developed as a means of scrubbing SO2 from flue gasses. The technology utilizes the reaction of SO2 with the sodium alumino-silicates in the residue to produce sulphites that can be removed by oxidation by air (Yamada et al., 1979a,b). The process was patented by the Sumitomo Aluminium Smelting Co (Yamada et al., 1979b), and has been used at a number of sites, including the Eurallumina plant (Sardinia, Italy; Fois et al., 2007). The process is aimed at the removal of SO2 from gas streams and the amount of residue neutralized is minor in relation to the total of amount of residue produced. Reaction of bauxite residue with SO2 is therefore an example of a beneficial use for bauxite residue, but it does not appear to be a potential treatment for large quantities of residue. Although not believed to be in use it is worth noting that a process using SO2 to recover soda and alumina values form residue has been patented (Cresswell et al., 1987). 4. Conclusions Bauxite residue has been continuously produced since the inception of the alumina/aluminium industry in the late nineteenth century. We have estimated that the global inventory of bauxite residue reached 1 billion tonnes in 1985, 93 years since the establishment of the first Bayer process plant. This inventory grew to 2 Bt by 2000, a doubling time of only 15 years. The inventory was almost 2.7 Bt by 2007, growing at 120 million tonnes per annum. The prospect of an ever decreasing doubling time, and an inventory of 4 Bt possibly before 2015, highlights the urgency of the need to develop and implement improved means of storage and remediation, and to pursue large-volume utilization options. Despite the size of the industry and the residue problem, it is surprisingly difficult to

43

accurately gauge the volume and nature of current residue output, let alone the very substantial legacy material in impoundment. Very little useful information is in the public domain. We have established an online database framework to try and address this problem — the Bauxite, Residue and Disposal Database (BRaDD). BRaDD has been an essential enabling tool for the collation and interpretation of information required for this and the associated three reviews in the series. With further programming refinements, and with data input from the industry, BRaDD should be an invaluable tool in the development of a global strategy for the future of bauxite residue management. Prior to 1980, most of the inventory of bauxite residue was held in lagoon-type impoundments. This storage method has a number of disadvantages, in particular the large areas of land required and the potential for leakage of caustic liquors to the environment. Since 1980 the trend has been towards “dry” stacking in bunded areas that are lined with a variety of natural and synthetic barrier materials. Improved methods for thickening and washing of the residues prior to storage, and recovery of decant water during storage, have been developed to increase the recovery of valuable soda and alumina to the Bayer process plants and to also minimize the potential for leakage to the surrounding environment. The current trend in residue storage practice is towards increasing use of “dry” stacking as the preferred technology, and further research to optimize this technology is appropriate. Hyperbaric steam filtration is an emerging technology that produces a dry, granular residue of low soda content. Such properties are beneficial to long term storage, remediation, and utilization. Support for the development and implementation of hyperbaric filtration as a potential breakthrough technology is warranted. A range of methods has been employed to neutralize both residue and decant waters. The most established methods to date involve either mixing with seawater to precipitate hydroxide, carbonate and aluminate ions with magnesium and calcium, or with carbon dioxide to produce calcium carbonate and calcium alumino-carbonates. Neutralization with mineral acids is less successful due to the influence of solid hydroxides in the residue and the deleterious effects that this method of neutralization has on the physical properties of the residue. The most important barrier to remediation, utilization and long term sustainability of bauxite residue management is its high alkalinity and sodicity. The alkalinity of bauxite residue is a result of complex solid-state and solution phase interactions. A better understanding of these interactions is required to support progress in the development of sustainable bauxite residue management and re-use options, so ongoing research on the chemistry and technology of residue neutralization is required. Based on this review of bauxite management practices a set of knowledge gaps and research priorities are summarized in Table 5.

Table 5 Summary of the key knowledge gaps concerning management, disposal and storage practices for bauxite residue and areas of suggested research. Knowledge gap

Research project

Economic technology to reduce the alkali and soda levels of bauxite residue. Understanding the residue fundamentals of solid alkalinity and surface equilibria; ion exchange, leaching and neutralization. Detailed worldwide bauxite residue inventory knowledge (historical and current); production, characteristics and disposal/storage practice. Is dry disposal the best technical answer to residue disposal?

Development of alternate processes, e.g. economically viable hyperbaric steam filtration. Fundamental residue chemistry investigations of ion exchange, leaching and neutralization. Release of key residue information from industry practitioners into public domain forums, e.g. BRaDD database. Investigation of the relative long-term environmental impact risks of “dry” stacking, dry cake disposal, seawater and carbonation methods. Technology to reduce alkali and soda levels within existing BRDA locations. Fully understanding long term CO2 equilibria as a function of residue composition and the role of direct carbonation compared to natural absorption processes with time. Development of methods for direct carbonation of residue by stack gas carbon dioxide.

BRDA closure technology that improves on simple capping and re-vegetation. Quantification and validation of carbon sequestration potential of residue carbonation Methods for direct carbonation of residue by stack gas carbon dioxide.

44

G. Power et al. / Hydrometallurgy 108 (2011) 33–45

Acknowledgements This project received funding from the Australian Government as part of the Asia-Pacific Partnership on Clean Development and Climate. The views expressed herein are not necessarily the views of the Commonwealth, and the Commonwealth does not accept responsibility for any information or advice contained herein. The support of the CSIRO Light Metals National Research Flagship and the Parker CRC for Integrated Hydrometallurgy Solutions (established and supported under the Australian Government's Cooperative Research Centres Program) is gratefully acknowledged. The authors gratefully acknowledge Dave Pang for the development of the BRaDD interface. References Alcoa, 2001. Lessening the Impact of Bauxite Residue. Alcoa Inc.. http://www.alcoa. com/global/en/environment/initiatives/dry_stacking.asp?initCountry=0. Alcoa, 2005b. Pinjarra Refinery Long Term Residue Management Strategy. Sinclair Knight Merz, Perth. pp 1–1 and 13–14. Alcoa, 2005a. Kwinana Refinery Long Term Residue Management Strategy. Sinclair Knight Merz, Perth 4, 4. Anderson, J., Smith, M., Zeiba, M., Clegg, R., Peloquin, G., Vellacott, S., 2008. Implementing seawater neutralisation at Gove. In: Armstrong, L. (Ed.), 8th International Alumina Quality Workshop. AQW Inc., Darwin, pp. 301–306. Anon, 1982. Thixotropic Sludge Deposition esp. Red Mud: Giulini Chem GmbH. Banerjee, S.K., 2003. Conversion of conventional wet disposal of red mud into thickened tailing disposal (TTD) at NALCO Alumina Refinery, Damanjodi. In: Crepeau, P.N. (Ed.), Light Metals. TMS, Vancouver, pp. 125–132. Baseden, S., Grey, D., 1976. Environmental study of the disposal of red mud waste. Marine Pollut. Bull. 7 (1), 4–7. Bott, R., Langeloh, T., Hahn, J., 2002. Dry bauxite residue by hi-barR steam pressure filtration. In: Chadrashekar, S. (Ed.), Proceedings of the 6th International Alumina Quality Workshop. AQW Inc., Brisbane, pp. 24–32. Cooling, D.J., 1989. Developments in the disposal of residue from the alumina refining industry. In: Campbell, P.G. (Ed.), Light Metals. TMS, Halifax, pp. 49–54. Cooling, D.J., Glenister, D.J., 1992. Practical aspects of dry residue disposal. Light Metals. TMS, pp. 25–31. Cooling, D.J., Hay, P.S., Guilfoyle, L., 2002. Carbonation of bauxite residue. In: Chandrashekar, S. (Ed.), Proceedings of the 6th International Alumina Quality Workshop. AQW Inc., Brisbane, pp. 185–190. Cresswell, P.J., Grayson, I.L., Smith, A.H., 1987. Recovery of Sodium Aluminate from Bayer Process Red Mud. . US-4668485. Das, P., Roy, D., Brahma, R., 2003. Experience with thickened disposal of red mud. In: Crepeau, P.N. (Ed.), Light Metals. TMS, Vancouver, pp. 133–135. Dauvin, J.-C., 2010. Towards an impact assessment of bauxite red mud waste on the knowledge of the structure and functions of bathyal ecosystems: the example of the Cassidaigne canyon (north-western Mediterranean Sea). Mar. Pollut. Bull. 60 (2), 197–206. Doucet, J., 1999. Effect of sand addition on the rheological properties of red mud slurries. Proceedings of the 5th International Alumina Quality Workshop. AQW Inc., Bunbury, pp. 478–488. Emmett, R.C., Klepper, R.P., 1991. High density red mud thickeners. In: Rooy, E.L. (Ed.), Light Metals. TMS, New Orleans, pp. 229–233. Emmett, R.C., Laros, T.J., Paulson, K.A., 1992. Recent developments in solid liquid separation technology in the alumina industry. In: Cutshall, E.R. (Ed.), Light Metals. TMS, San Diego, pp. 87–90. Enserink, M., 2010. After red mud flood, scientists try to halt wave of fear and rumors. Science 330, 432–433 (22 October). Farrow, J.B., Fawell, P.D., Johnston, R.R.M., Nguyen, T.B., Rudman, M., Simic, K., Swift, J.D., 2000. Recent developments in techniques and methodologies for improving thickener performance. Chem. Eng. J. 80 (Sp. Iss. 1–3), 149–155. Fois, E., Lallai, A., Mura, G., 2007. Sulfur dioxide absorption in a bubbling reactor with suspensions of Bayer red mud. Industr. Eng. Chem. Res. 46 (21), 6770–6776. Glenister, D.J. (Ed.), 1992. Conference Proceedings of an International Bauxite Tailings Workshop. Allied Colloids Pty Ltd., Perth. Gräfe, M. and Klauber, C., to be published. Bauxite residue issues: IV. Old obstacles and new pathways for in situ residue bioremediation. Hydrometallurgy (2011), doi:10.1016/j.hydromet.2011.02.005. Gräfe, M., Power, G., Klauber, C., 2009. The Asia-Pacific partnership: an important new initiative for a sustainable alumina industry. In: Bearne, G. (Ed.), Light Metals. TMS, San Francisco, pp. 5–10. Gräfe, M., Klauber, C., Pang, D., 2010. BRaDD: Bauxite Residue and Disposal Database. . http://www.csiro.au/products/Bauxite-residue-database.html. Gräfe, M., Power, G. and Klauber, C., to be published. Bauxite residue issues: III. Alkalinity and associated chemistry. Hydrometallurgy (2011), doi:10.1016/j. hydromet.2011.02.004. Green, M., Boger, D., 1997. Yielding of suspensions in compression. Industr. Eng. Chem. Res. 36 (11), 4984–4992. Guilfoyle, L., Hay, P., Cooling, D., 2005. Use of flue gas for carbonation of bauxite residue. In: McKinnon, A. (Ed.), Proceedings of the 7th International Alumina Quality Workshop. AQW Inc., Perth, pp. 218–220.

Habashi, F., 1988. A hundred years of the Bayer Process for alumina production. In: Boxall, L.G. (Ed.), Light Metals. TMS, Phoenix, pp. 3–11. Haggard, N., Smith, H.D., 2002. Supernatant liquor neutralisation with seawater and Bayer tailings mud via deep cone thickener. In: Chandrashekar, S. (Ed.), Proceedings of the 6th International Alumina Quality Workshop. AQW Inc., Brisbane, pp. 174–179. Hamada, T., 1986. Environmental management of bauxite residue — A review. In: Wagh, A.S., Desai, P. (Eds.), Proceedings of the International Conference on Bauxite Tailings. The Jamaica Bauxite Institute, The University of the West Indies, Kingston, Jamaica, pp. 109–117. Hanahan, C., 2004. Dissolution of hydroxide minerals in the 1 M sodium acetate, pH 5, extracting solution in sequential extraction schemes. Environ. Geol. 45 (6), 864–868. Hudson, L.K., 1987. Alumina production. Crit. Rep. Appl. Chem. 20, 11–46 (Prod. Alum. Alumina). International Maritime Organization, 2005. Report of the twenty-eighth meeting of the scientific group. Scientific Group Meeting — 28th Meeting. IMO, London, UK, p. 54. Kainuma, A., Takahashi, M., Higuchi, T., Itoh, D., 1979. Utilization of filter pressed bauxite residue. In: Peterson, W.S. (Ed.), Light Metals. TMS, New Orleans, pp. 107–127. Khaitan, S., Dzombak, D.A., Lowry, G.V., 2009. Chemistry of the acid neutralization capacity of bauxite residues. Environ. Eng. Sci. 26 (00), 1–9. Khaitan, S., Dzombak, D.A. and Lowry, G.V., (2009). Mechanisms of neutralization of bauxite residue by carbon dioxide. J. Environ. Eng.-ASCE. 135 (6):433-438. Kirkpatrick, D.B., 1996. Red mud product development. In: Hale, W. (Ed.), Light Metals. TMS, Anaheim, pp. 75–80. Klauber, C., Gräfe, M. and Power, G., to be published. Bauxite residue issues: II. Options for residue utilization. Hydrometallurgy (2011), doi:10.1016/j.hydromet.2011.02.007. Martinet-Catalot, V., Lamerant, J.-M., Tilmant, G., Bacou, M.-S., Ambrost, J.P., 2002. Bauxaline: a new product for various applications of Bayer process red mud. In: Schneider, W. (Ed.), Light Metals. TMS, Columbus, pp. 125–131. Marunczyn, M., Laros, T.J., 1992. Bauxite residue disposal at Alcoa of Australia using the Eimco hi-density thickener. International Bauxite Tailings Workshop. Allied Colloids (Australia), Perth, Western Australia, p. 32. Marunczyn, M., Laros, T.J., 1993. Bauxite residue disposal at Alcoa of Australia using hidensity thickener. In: Leung, W.W.-F. (Ed.), Advances in Filtration and Separation Technology: System Approach to Separation and Filtration Process Equipment. American Filtrations & Separations Society, Chicago, pp. 180–188. McConchie, D., Clark, M., Davies-McConchie, F., 2002. New strategies for the management of bauxite refinery residues (red mud). In: Chandrashekar, S. (Ed.), Proceedings of the 6th International Alumina Quality Workshop. AQW Inc., Brisbane, pp. 327–332. Nikraz, H.R., Bodley, A.J., Cooling, D.J., Kong, P.Y.L., Soomro, M., 2007. Comparison of physical properties between treated and untreated bauxite residue mud. J. Mat. Civ. Eng. 19 (1), 2–9. Oeberg, N.C.R., Steinlechner, E.H., 1996. Red mud and sands handling: new thoughts on an old problem. In: Hale, W. (Ed.), Light Metals. TMS, Anaheim, pp. 67–73. Paradis, R.D., 1992. Disposal of Red Mud Using Wet Stacking Technology. International Bauxite Tailings Workshop, Perth, Western Australia. pp. 179. Paradis, R.D., 1993. Application of Alcan's deep thickener technology for thickening and clarifying. In: Leung, W.W.-F. (Ed.), Advances in Filtration and Separation Technology: System Approach to Separation and Filtration Process Equipment. American Filtration & Separations Society, Chicago, pp. 176–179. Pashias, N., Boger, D.V., 1996. The importance of rheology in optimising a disposal strategy for bauxite residue. Proceedings of the 4th International Alumina Quality Workshop. AQW Inc., Darwin, pp. 507–515. Peloquin, G., Simard, G., 2002. Improved thickener for red mud. In: Chandrashekar, S. (Ed.), Proceedings of the 6th International Alumina Quality Workshop. AQW Inc., Brisbane, pp. 40–45. Pohland, H.H., Tielens, A.J., 1983. A new bayer liquor purification process. In: Adkins, E.M. (Ed.), Light Metals. TMS, Atlanta, pp. 211–222. Pohland, H.H., Tielens, A.J., 1986. Design and operation of non-decanting red mud ponds in Ludwigshafen, West Germany. In: Wagh, A.S., Desai, P. (Eds.), Bauxite Tailings International Conference. The Jamaica Bauxite Institute, The University of the West Indies, Kingston, Jamaica, p. 87. Pohland, H., Haerter, M., Bayer, G., 1981. Thixotropic Mud, esp. Red Mud, Disposal — By Mechanically Liquefying and Transferring to Dump with Embankment Contg. Cation-Exchange Material: Giulini Chem Gmbh (Giul). Roach, G.I.D., Jamieson, E., Pearson, N., Yu, A.B., 2001. Effect of particle characteristics on the solids density of Bayer mud slurries. In: Anjier, J.L. (Ed.), Light Metals. TMS, New Orleans, pp. 51–58. Robinsky, E.I., 1982. Thickened tailings disposal in any topography. In: Andersen, J.E. (Ed.), Light Metals. TMS, Dallas, pp. 239–248. Rosenberg, S.P., Wilson, D.J., Heath, C.A., 2004. Processes for the Causticisation of Bayer Liquors in an Alumina Refinery. Worsley Alumina Pty Ltd.. US 6676910 B1. Roskill, 2008. The Economics of Bauxite & Alumina, 7th Ed. Roskill Information Services Ltd, London, UK, p. 48. Shah, R.P., Gararia, S.N., 1995. Upgradation of alumina refinery at Hindalco, Renukoot (India). In: Evans, J. (Ed.), Light Metals. TMS, Las Vegas, pp. 25–29. Simmonds, M., Hurst, P., Wilkinson, M.B., Reddy, S., Khambaty, S., 2003. Amine based CO2 capture from gas turbine. In: Klara, S. (Ed.), Second Annual Conference on Carbon Sequestration. USDOE, Alexandria, VA, pp. 1–10. Smith, P.G., Pennifold, R.M., Davies, M.G., Jamieson, E.J., 2003. Reactions of carbon dioxide with tri-calcium aluminate. In: Young, C., et al. (Ed.), Fifth International Symposium on Hydrometallurgy. TMS, Vancouver, pp. 1705–1715. Sofra, F., Boger, D., 2002. Environmental rheology for waste minimisation in the minerals industry. Chem. Eng. J. 86 (3), 319–330.

G. Power et al. / Hydrometallurgy 108 (2011) 33–45 Solymar, K., Ferenczi, T., Papanastassiou, D., 2002. Digestion alternatives of the Greek diasporic bauxite. In: Schneider, W. (Ed.), Light Metals. TMS, Seattle, pp. 75–81. Steinlechner, E., Oeberg, N., Kern, R., 1996. Red mud, a challenge for filtration. Proceedings of the 4th International Alumina Quality Workshop. AQW Inc., Darwin, pp. 251–261. Sturt, A., 2008. Effect on aluminium solubility when acid is used to neutralise residue. In: Armstrong, L. (Ed.), Proceedings of the 8th International Alumina Quality Workshop. AQW Inc., Darwin, pp. 93–95. Thomas, G.A., Allen, D.G., Wyrwoll, K.-H., Cooling, D., Glenister, D., 2002. Capacity of clay seals to retain residue leachate. In: Chandrashekar, S. (Ed.), Proceedings of the 6th International Alumina Quality Workshop. AQW Inc., Brisbane, pp. 233–239. Tillotson, S., 2006. Phosphate removal: an alternative to chemical dosing. Filtr. Sep. 43 (5), 10–12. U.S. Geological Survey, 2008. Bauxite and alumina statistics and information. In: Kelly, D.T., Matos, G.R. (Eds.), Historical Statistics for Mineral and Material Commodities in the United States. United Nations Industrial Development Organization, 1985. Environmental Aspects of Alumina Production. UNEP, Nairobi.

45

Varnavas, S., Ferentinos, G., Collins, M., 1986. Dispersion of bauxitic red mud in the Gulfof-Corinth, Greece. Mar. Geol. 70 (3–4), 211–222. Versiani, F., 1983. Neutralization of red mud by boiler stack gases. In: Adkins, E.M. (Ed.), Light Metals. TMS, Atlanta, pp. 337–343. Wagh, A.S., Desai, P. (Eds.), 1986. Proceedings of an International Conference on Bauxite Tailings. The Jamaica Bauxite Institute, The University of the West Indies, Kingston, Jamaica. Whittington, B., 1996. The chemistry of CaO and Ca(OH)2 relating to the Bayer process. Hydrometallurgy 43 (1–3), 13–35. Yamada, K., Fukunaga, T., Harato, T., 1979a. Process of SO2 removal from waste gas by red mud. In: Peterson, W.S. (Ed.), 108th AIME Annual Meeting. : Light Metals. The Metallurgical Society of AIME, New Orleans, Louisiana, pp. 67–79. Yamada, K., Harato, T., Shiozaki, Y., 1979b. Removing Sulphur Oxide(s) from Exhaust Gas with Red Mud Slurry — Giving Efficient Removal and Use of Soda Values without Scaling. Sumitomo Aluminium Smelting: Sumitomo Chem Ind Kk. FR2418019-A1.