Printer Friendly

Surface charge characteristics and sorption properties of bauxite-processing residue sand.

Introduction

Bauxite mining for alumina production has increased in recent years in response to the strong demand by developing economies, with a worldwide output of 172 Mt of bauxite in 2007. For every tonne of alumina, about 2 t of highly alkaline, highly saline bauxite-processing residue is produced. In Western Australia, Alcoa of Australia Ltd (Alcoa) produces ~15 Mt of residue annually from its refineries (Kwinana, Pinjarra, and Wagerup). In contrast to many other bauxite-processing operations, Alcoa's Western Australian refineries separate residue into 2 distinct size fractions: <150 [micro]m (termed residue mud) and >150 [micro]m (termed residue sand). The residue sand (BRS) fraction represents the primary material used for rehabilitating Alcoa's residue storage areas (RSAs).

The inherently hostile characteristics of BRS (high alkalinity, high salinity, low fertility, and low water holding characteristics) pose severe limitations for establishing sustainable plant cover systems. Furthermore, the mineralogical composition (~58% quartz, 22% hematite, 7% gibbsite, 5% goethite, 1% anatase, 1% desilication product (DSP), 1% muscovite, and 0.3% boehmite; Taylor and Pearson 2001), and the presence of amorphous Fe and Al oxides (0.7%) ,suggest BRS contains colloids that can limit plant nutrient availability due to specific adsorption mechanisms (e.g. phosphorus (P), copper (Cu), zinc (Zn), boron (B)) (Apak et al. 1998). Recent studies by Bell et al. (2008) and Thiyagarajan et al. (2009) found nutrient deficiencies in a range of plant species used in Alcoa's rehabilitation program, and observed a strong correlation between these deficiencies and nutrient availability in BRS.

It is well established that the surface charge of soil colloids plays an important role in nutrient availability, and that surface charge can be classified as either permanent or variable. In fact, most soils contain a mixture of both types of charge (Gallez et al. 1976; Morais et al. 1976; Theng 1980; Phillips and Sheehan 2005). Research on the surface charge characteristics of soils has largely focused on agricultural systems, and the impact of fertilisers and soil amendments on surface charge (e.g. changes in soil solution pH and ionic strength on net colloid surface charge (([[sigma].sub.T]), and the effect of phosphorus on the zero point of charge), with considerable emphasis placed on the variable charge component of highly weathered and tropical soils (van Raij and Peech 1972; Gallez et al. 1976; Morais et al. 1976; Qafoku et al. 2004; Gillman 2007; Wong and Wittwer 2009). Despite the dominance of Fe and A1 oxides and hydrous oxides in BRS, detailed information on the surface charge properties of this material is scarce.

The zero point of net charge (ZPNC), defined as the pH at which there is an equal number of positive and negative surface charges, has been reported to be 8.3 (Apak et al. 1998), 8.5 (Pradhan et al. 1999), and 6.9 (Lopez et al. 1998) for residue mud. Although the surface charge chemistry of residue mud (Wong and Ho 1995) and how fertiliser nutrients interact with it (Meecham and Bell 1977; Snars et al. 2003, 2004) have been studied previously, similar studies involving BRS are scarce (Meecham and Bell 1977; Gherardi and Rengel 2003a, 2003b, 2003c; Courtney and Timpson 2005; Thiyagarajan et al. 2009).

Alcoa's BRS is highly alkaline (pH >11), highly saline (EC>2dS/m in saturated paste extract), and sodium (Na)-dominated in both the solution (primarily as [Na.sub.2]C[O.sub.3]) and exchange (exchangeable sodium percentage (ESP) >60%) phases. Rehabilitation of Alcoa's RSAs involves the use of gypsum to reduce alkalinity, and di-ammonium phosphate (DAP) based fertiliser to supply plant nutrients. The addition of gypsum alters BRS from a system dominated by Na to one dominated by calcium (Ca). Despite the known effects of pH and ionic composition of the pore-water on the sign and magnitude of charge carried by variably charged colloids, there is no information on how these parameters influence the surface charge of BRS. Changes in the sign and magnitude of surface charge will influence the ability of BRS to adsorb (and desorb) applied nutrients, and therefore affect the quantities of nutrients available for plant uptake and/or leaching below the root-zone. Given the importance of nutrient availability for sustainable residue rehabilitation, there is a need to better understand the surface charge characteristics of BRS, and how this parameter affects plant nutrient dynamics.

The objectives of this study were to demonstrate the effect of pH on (1) the surface charge characteristics of BRS, and (2) ammonium (N[H.sub.4]) and phosphorus (P) adsorption by BRS. Outcomes from this research will provide baseline information for future studies aimed at understanding nutrient dynamics in BRS following application of DAP-based fertiliser.

Materials and methods

Sources of residue sand (BRS) and characterisation Samples (n = 2) of BRS were collected from each of Alcoa's 3 refineries in south-west Western Australia: Kwinana (KW), Pinjarra (P J), and Wagemp (WG). Samples of 'fresh' (unweathered) BRS were collected from the sand discharge pipe, and as such had not received any gypsum or fertiliser amendment. Approximately 500 g of field-moist (<2 mm) BRS were wetted-up with deionised water (>18M[ohms]) to achieve a saturated paste (gravimetric water content ([[theta].sub.g] [approximately equal to] 0.3 kg/kg). The paste was allowed to equilibrate for 16 h, after which a sample of the pore-water was extracted under vacuum. After filtering (<0.45[micro]m), the water-extract was analysed for electrical conductivity (EC), pH, cations (Ca, Na, magnesium (Mg), potassium (K), aluminium (Al), iron (Fe), and N[H.sub.4]), anions [P, sulfate (S[O.sub.4]), chloride (CI), nitrate (N[O.sub.3]), bicarbonate (HC[O.sub.3]), and carbonate (C[O.sub.3]), and trace elements (boron (B), copper (Cu), zinc (Zn), and manganese (Mn))]. The pH and EC in water were measured using a combination pH electrode; water-soluble cations (Ca, Na, Mg, K, Al, Fe, and trace elements) were analysed by inductively coupled plasma-atomic emission spectrometry (ICP-AES); water-soluble anions (N[O.sub.3], Cl, and P) and N[H.sub.4] were measured using flow injection analysis (FIA, LaChat Flow Injection Analysis QuickChem 8000) using QuickChem methods 10-117-07-2-B, 10-107-04-1-H, 31-115-01-3-A, and 10-107-06-4-D, respectively; water-soluble S[O.sub.4] by ICP-AES; alkalinity was determined by titration (Method Dla) and reported in terms of HC[O.sub.3] (Method D3a) and C[O.sub.3] (Method D2a) (Rayment and Higginson 1992).

Following vacuum extraction, the moist BRS was analysed for amorphous Fe and Al oxides, organic carbon (Org C), exchangeable cations (Ca, Mg, K, Na, A1, Fe), available nutrients (K, P, N[H.sub.4], N[O.sub.3], S[O.sub.4]), total N, total P, total AI, total Fe, trace elements (Cu, Zn, Mn, Fe), effective cation exchange capacity (ECEC), P buffering index (PBI), particle size distribution, and texture. The concentrations of water-soluble Ca, Mg, K, Na, S[O.sub.4], AI, and Fe were accounted for when calculating the concentrations of exchangeable cations and plant-available nutrients. Exchangeable cations (Ca, Mg, K, Na, Al, Fe) were extracted by shaking BRS with 0.1 M Ba[Cl.sub.2]/0.1 M N[H.sub.4]Cl at a soil to solution ratio of 1 : 10 for 2 h (Method 15E1, Rayment and Higginson 1992) and analysed by ICP-AES. The ECEC refers to the summation of exchangeable Ca, Mg, K, Na, and AI and was used as an estimation of the cation exchange capacity (CEC) (Method 15J1, Rayment and Higginson 1992). Organic C content was determined using the Walkley-Black method (Method 6A1, Rayment and Higginson 1992). The more reactive forms of Fe and Al (hydrous oxides) were extracted using acid (pH 3) ammonium oxalate (Method 13A1, Rayment and Higginson 1992). Available P and K were extracted by shaking for 16h using 0.5M sodium bicarbonate (NaHC[O.sub.3] pH 8.5) at a BRS to solution ratio of 1:100 (Method 9B1, Rayment and Higginson 1992). Phosphorus was determined using FIA (QuickChem method 31-115-01-3-A) and K by atomic absorption spectrometry. Available S was extracted at 40[degrees]C for 3 h with 0.25 M KCl, and the S[O.sub.4]-S measured by ICP (Blair et al. 1991). Ammonium and N[O.sub.3] were extracted by shaking BRS for 1 h with 1 M KCl at a soil to solution ratio of 1 : 5, and the extracts analysed using FIA (QuickChem method 10-107-06-4-D and 10-107-04-1-H, respectively). Total (LECO) N was measured using a TruSpec CN Carbon Nitrogen analyser. Total P was extracted by digesting BRS with sulfuric acid-potassium-copper sulfate, and P measured colourimetrically at a wavelength of 880 nm. Total AI and total Fe were extracted using aqua regia (3 : 1 v/v HCI to HN[O.sub.3], ISO Standard 11466), and analysed by ICP-OES. Trace elements (Cu, Zn, Mn, Fe) were extracted with 0.01 M EDTA at a BRS to solution ratio of 1 : 5 for 1 h. The supernatant was analysed using ICP-AES. The PBI was determined as described in Burkitt et al. (2002). Particle size analysis was undertaken using the modified pipette procedure (Indorante et al. 1990).

Selected BRS samples were also analysed for mineralogy (X-ray diffraction (XRD), Philips PW diffractometer), particle density (Flint and Flint 2002), and surface area using [N.sub.2] adsorption in conjunction with the BET equation (Pennell 2002).

Permanent charge ([[sigma].sub.p])

Permanent surface charge (n-4) was determined using the cesium-adsorption method of Anderson and Sposito (1991) under ambient laboratory conditions (temperature 25 [+ or -] 3[degrees]C).

A known weight of field-moist BRS (equivalent to 0.5 g dry weight) was placed in a 50-mL pre-weighed polypropylene centrifuge tube, and shaken for 0.5 h with 25 mL of 0.5 M CsCl. After shaking, the suspension was centrifuged and the supematant discarded. The BRS was washed twice more with 25 mL of 0.1 M CsCl, discarding the supernatant after each wash. The BRS was washed once with 0.05 M CsCl, and the suspension pH adjusted to 6.5 with 0.05 M CsOH. The suspension was shaken for an additional 1 h when the pH stabilised, after which it was centrifuged and the supernatant discarded. The CsCl-saturated sample was washed with 25 mL of 95% ethanol by shaking for 1 min, after which the suspension was centrifuged and the supernatant discarded. This washing step was repeated until the supematant showed the absence of Cl following addition of 1 drop of 1 M AgN[O.sub.3]. The BRS was dried at 65[degrees]C for 48 h in the tubes to enhance the formation of inner-sphere Cs surface complexes. After drying, the BRS was shaken with 25 mL of 0.01 M LiCl for 0.5 h to displace Cs from variable charge sites. The suspension was centrifuged and filtered (Whatman 42), and the supernatant retained for Cs measurement. The volume of entrained solution was estimated by weighing, after which the BRS was washed with 3 x 30mL washings with 1 M N[H.sub.4]Oac to remove adsorbed Cs. Following each shaking period (0.5h), the suspension was centrifuged and filtered (Whatman 42). The supernatant from each extraction was combined, made to a volume of 100 mL with 1 M N[H.sub.4]OAc, and analysed for Cs. The concentrations of Cs in the LiCl and N[H.sub.4]OAc extracts were determined by atomic emission spectroscopy.

Ion adsorption measurements and point of zero net charge

Potassium and Cl adsorptions as a function of pH (pH 3, 5, 7, 8, 9, 10, 11, 12) were measured as described by Zelazny et al. (1996), using KCl as the background solution (n=2). Four grams of BRS (field-moist <2 mm) were placed into 8 pre-weighed 50-mL centrifuge tubes, and the BRS shaken with 20mL of 1 M KCl for 1 h. A duplicate set of tubes containing no BRS was treated in exactly the same manner as the BRS samples (i.e. blanks). After shaking, the suspension was centrifuged (3000 rpm for 10min) and the clear supernatant discarded. The BRS was shaken with 20mL of 0.2M KCl for 16h, centrifuged, and the clear supernatant discarded. The BRS was further shaken with 2 x 20 mL of 0.2 M KCl for 1 h, after which the clear supernatant was discarded. The BRS was successively washed 3 times with 20mL of 0.01M KCl by shaking for l h, centrifuging, and discarding the clear supernatant. An additional 20 mL of 0.01 M KCl was added to the BRS, and the pH adjusted with HN[O.sub.3] or NaOH to achieve a solution pH of ~3, 5, 7, 8, 9, 10, 11, or 12. The pH of each initial solution was adjusted based on previously measured acid neutralisation curves (data not presented). After shaking for 44 h, the suspension was centrifuged and a 5-mL sample of supernatant removed for pH determination, and the pH adjusted if required. The total volume was adjusted back to 20mL using 0.01 M KCl, after which the suspension was shaken for an additional 4 h, giving a total reaction time of 48 h (preliminary work confirmed that pH stability was achieved within 48h following adding acid to BRS; Carter et al. 2008, 2009). The supernatant was centrifuged and filtered, and subsequently analysed for K and Cl. The volume of entrained solution was estimated by weighing. Adsorbed K and Cl were extracted from the BRS by 5 x 20 mL washings with 0.5 M N[H.sub.4]N[O.sub.3]. For each washing, the BRS was shaken with 0.5 M N[H.sub.4]N[O.sub.3] for 1 h, centrifuged, and the supematant filtered (Whatman 42). The supernatants from each extraction were combined and made to a final volume of 100 mL with 0.5 M N[H.sub.4]N[O.sub.3]. Potassium and Cl in the 0.5 M N[H.sub.4]N[O.sub.3] were measured by atomic absorption spectrometry (AAS) and flow injection analysis (FIA) techniques, respectively.

The point of zero net charge (PZNC) is defined as the pH value at which ion adsorption is zero (Sposito 1992; Gillman 2007). The PZNC was determined from the K and Cl adsorption data using:

([q.sub.+] [q.sub._]) = 0 (pH = PZNC) (1)

where [q.sub.+] is the concentration of adsorbed K at each pH and q is the concentration of adsorbed Cl at each pH. The pH for PZNC was estimated using a second-order polynomial equation fitted to the experimental data.

Ammonium and phosphorus adsorption isotherms

Ammonium and P adsorption isotherms as a function of pH (pH 7, 9, 11) were developed using Ca-saturated BRS from the KW and PJ Refineries. Samples were Ca-saturated to reflect the exchangeable cation status of BRS following gypsum incorporation, and were prepared by successive washings with 1M Ca[Cl.sub.2], then 0.1M Ca[Cl.sub.2], then 0.01M Ca[Cl.sub.2] at a BRS to solution ratio of 350g to 2L. Samples of each BRS were then partially dried by vacuum filtration to remove much of the interstitial Ca[Cl.sub.2] solution. Saturation of the CEC was achieved using Ca[Cl.sub.2] solution rather than a gypsum solution to eliminate competitive effects of S[O.sub.4] on P sorption behaviour.

Eight initial solutions containing 0-50 mg of N (as N[H.sub.4]) and P were obtained by diluting a DAP [((N[H.sub.4]).sub.2]HP[O.sub.4]) stock solution with Milli-Q deionised water (>18M[OMEGA]) to achieve final concentrations representing 0, 2.5, 5, 10, 20, 30, 40, 50, and 100% of the stock solution. Adsorption isotherms were performed at a constant temperature of 25 [+ or -] 3[degrees]C at pH 7, 9, and 11.

Approximately 5.5 g of Ca-saturated BRS (field-moist and <2 mm size fraction; equivalent to an oven-dry mass of 5 g) were placed into pre-weighed 50-mL polypropylene centrifuge tubes. The actual mass of BRS added to each tube was recorded and later convened to an oven-dry weight using measured initial water contents. The initial solutions (25 mL) were added to the centrifuge tubes, and the suspensions shaken end-over-end for 44 h. After shaking, the pH of the suspension was re-checked and adjusted with 0.01 M HCl if required. All samples were then shaken for an additional 4 h, giving a total shaking time of 48 h. After shaking, all samples were centrifuged at 10 000 rpm for 10min, filtered (Whatman 42), and the filtrate stored frozen before analysis for N[H.sub.4] and P (using FIA). The concentrations of sorbed N[H.sub.4] and P were calculated as:

S = (([c.sub.i] - [c.sub.e]) x [V.sub.T] + [m.sub.s]) (2)

where S is sorbed concentration (mmol/kg), [c.sub.i] is initial solution concentration (mmol/L), [c.sub.e] is equilibrium solution concentration (mmol/L), [v.sub.T] is volume of added solution (L), and [m.sub.s] is oven-dry mass of BRS (kg).

Normally, the concentration of sorbed ion is calculated using the difference between the amount of ion added initially and that remaining in solution after shaking. However, given the pH-dependence of N[H.sub.4] behaviour in soil (Cabrera et al. 1991), actual measurement of the quantity of N[H.sub.4] retained by the solids phase of BRS was considered a more realistic estimate of sorbed N[H.sub.4] (hereafter referred to as adsorbed N[H.sub.4]). This cation was extracted from the solid phase as follows. After removing the supernatant, each tube was re-weighed to allow the weight of entrained solution to be estimated. The residue sand was then shaken with 25 mL of 2 M KCl for 2 h, after which the suspension was centrifuged and the supernatant filtered to obtain a clear solution. This washing step was repeated 2 more times. The 3 washings were bulked into a 100-mL volumetric flask then made to volume with 2 M KCl. The solution was analysed for N[H.sub.4] as described above. Thc concentration of adsorbed N[H.sub.4] was calculated as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

where SN[H.sub.4] is adsorbed N[H.sub.4] concentration (mmol/kg), [C.sub.T] = N[H.sub.4] concentration in bulk wash solution (mmol/L), [c.sub.c] = equilibrium N[H.sub.4] solution concentration (mmol/L), vent = volume of entrained solution (L), and [m.sub.s]=oven-dry mass of residue sand (kg).

The N[H.sub.4] and P sorption isotherms were described using the Freundlich equation:

S = [kc.sup.n.sub.e] (4)

where co is the ion concentration in the equilibrium solution (mmol/L), S is the change in ion concentration in the solid phase (mmol/kg), and k and n are empirical constants related to the sorption index and bonding strength respectively.

Statistical methods

Mean and standard deviation of the water-soluble and exchangeable ion concentrations, and surface charge characteristics of the BRS samples, were evaluated using analysis of variance (ANOVA) and comparison of means by least significant difference (l.s.d. at P=0.05) procedures (Analytical Software 1994). All curve fitting was done using the software package Grapher 3 (Golden Software 2000).

Results and discussion

Selected chemical and physical properties

The water-soluble fraction of each BRS was highly alkaline (pH >10), highly saline (EC > 2 dS/m), and dominated by Na and C[O.sub.3] ions, with lesser amounts of Cl and S[O.sub.4] ions (Table 1). All samples exhibited very low concentrations of plant nutrients such as inorganic N (<0.1 mmol/L) and Mg (<0.2mmol/L). These findings are not uncommon for BRS produced using the Bayer process (e.g. Meecham and Bell 1977; Courtney and Timpson 2005), and highlight the need for developing targeted remediation activities before establishing vegetation covers on 'flesh' BRS (Woodard et al. 2008).

Water-extractable P concentrations ranged between 0.09 and 0.30mmol/L, which are considerably higher than those commonly found in unfertilised soil (~0.002 mmol/L; Tisdale et al. 1993). Bauxite used in Alcoa's alumina refining operations contains ~6.5 mmol P/kg (Taylor and Pearson 2001; Carter 2006). Alcoa's alumina refining process also adds P as dihydrogen phosphate (Ca[([H.sub.2]P[O.sub.4]).sub.2] x [H.sub.2]O) to control calcia. This results in BRS deposited at Alcoa's RSAs containing up to 4mmol P/kg (Carter 2006). It is well documented that P solubility increases dramatically under highly alkaline conditions, even when A1 and Fe oxides are the dominant sorbing materials (Barrow 1982; Carter et al. 2009; Huang et al. 2009). It is plausible, therefore, that the highly alkaline conditions of 'flesh' BRS (Table 1) can maintain elevated concentrations of water-soluble P as observed for the 3 samples used in this study. In fact, the geochemical speciation model PHREEQC, in combination with the ionic composition and pH conditions of the water-soluble fraction of each BRS sample (Table 1), simulated that water-soluble P would primarily occur as HP[O.sub.4.sup.2-] (>85%) and NaHP[O.sub.4] ([approximately equal to] 10%) (Table 2). This finding is not unexpected based on P speciation as a function of pH, and the absence of cations capable of forming low solubility phosphates (e.g. Ca and A1).

Soluble A1 concentrations exceeded those typically observed in soil solutions (Adams 1984), with values ranging from ~0.24 mmol/L for WG BRS to 2.03 mmol/L for PJ BRS (Table 1). Elevated water-soluble Al concentrations in alkaline BRS is well documented (Fuller et al. 1982; Kopittke et al. 2004; Carter et al. 2008, 2009), and for highly alkaline BRS, PHREEQC predicted that >99% of the soluble A1 would occur as the monomeric hydroxyl-A1 species aluminate (Al[(OH).sub.4.sup.-]) (Table 2). Aluminium toxicity is considered to be a major limitation to plant growth in 'fresh' bauxite-processing residue (Fuller and Richardson 1986; Woodard et al. 2008). Although Al rhizotoxicity has been demonstrated in nutrient solutions adjusted to pH values >8 (Fuller and Richardson 1986; Kinraide 1990; Eleftheriou et al. 1993; Ma et al. 2003), it is not clear whether Al[(OH).sub.4.sup.-] is the Al species responsible for this effect (Kopittke et al. 2004). This is because Al(OH)4- forms the central core to the [Al.sub.13] species (A1[O.sub.4][Al.sub.12][(OH).sub.24][([H.sub.2]O).sub.12.sup.7+]) which is believed to be more toxic than [Al.sup.3+]. Whether [Al.sub.l3] forms in Alcoa's BRS is unknown. However, given the high valency of 7+ of [Al.sub.l3], and predominantly negative surface charge of BRS at high pH, then much of the Alia may be removed from solution via cation exchange reactions. Kopittke et al. (2004) suggested that at low AI concentrations (<0.025 mmol/L), Al[(OH).sub.4.sup.-] was not toxic to plants; however, BRS from Alcoa's Refineries displayed much higher soluble A1 concentrations (>0.24 mmol/L). Whether A1 concentrations of this magnitude are toxic to Alcoa's rehabilitation is unknown and warrants more attention.

Sodium dominated the exchange phase of all BRS samples (Table 3), which was not unexpected given that the Bayer process for alumina production is based on the use of concentrated caustic soda (NaOH) to solubilise Al (as Al[(OH).sub.4.sup.-]) from crushed bauxite. The negligible exchangeable A1 concentration (0.01 [cmol.sub.c]/kg) also suggests that A1 not present as Al[(OH).sub.4.sup.-] may exist primarily as Al oxides and hydrous oxides as indicated by BRS mineralogy (Taylor and Pearson 2001).

Concentrations of N, P, K, and trace elements were very low, with the exception of EDTA Fe (67-84 mg/kg). Available (NaHC[O.sub.3]) P concentrations (<3 mg/kg) were well below the median value of 19 mg/kg measured in a dataset of 290 Australian surface soils (Burkitt et al. 2002), while NaHC[O.sub.3] K was only ~1/10 of the critical value required for pasture growth in sandy soils (126 mg/kg; Gourley et al. 2007). In contrast, KCl S concentrations in all BRS samples were found to be greater than that considered adequate for pasture soils (i.e. >8 mg/kg; Gourley et al. 2007). Furthermore, all 3 BRS samples contained negligible organic carbon, which was consistent with the absence of microbial activity in fresh BRS (Banning et al. 2009).

The very high PBI of BRS (Table 3) suggests that applied P may quickly become partitioned into non-plant available fractions (Burkitt et al. 2002). This finding may appear inconsistent with the elevated solution P observed for BRS (Table 1); however, interpretation of P adsorption must consider the pH of the extracting systems. The pH of the saturated paste extract was >10, and its anionic composition was dominated by C[O.sub.3]. The PBI is measured using [H.sub.2]P[O.sub.4] in a 0.01M Ca[Cl.sub.2] solution (1:10 BRS to solution ratio), which would decrease the pH of BRS and its associated C[O.sub.3] concentration. This also provides a reduction in the C[O.sub.3] : HC[O.sub.3] ratio (thereby reducing the concentration of a competing anion for adsorption sites) and decreases the concentration of negative surface charge (see below). These mechanisms enhance P adsorption by colloids dominated by Fe and A1 oxides and hydroxides, and provide a valid explanation for the observed greater P solubility in the saturated paste water extract relative to that which may be expected from the PBI.

All 3 BRS samples were dominated by coarse sand (60-70%), with lower percentages of fine sand (23-28%), and negligible silt (2%) and clay (5-8%) (Table 3). Particle density of each BRS was ~2870 kg/[m.sup.3], which was very similar to the value of 2900 kg/[m.sup.3] reported by Taylor and Pearson (2001). Independent analysis found that BRS has a specific surface area of 6.85 [m.sup.2]/g (measured by [N.sub.2] adsorption using the BET equation; L. Wissmeier, pers. comm.), which was similar to a value of 8 [m.sup.2]/g reported by Taylor and Pearson (2001). Furthermore, this value was similar to those reported for sands (<10[m.sup.2]/g; Sumner 2000), and was considerably smaller than those reported for crystalline (116-184 [m.sup.2]/g) and amorphous (305-412 [m.sup.2]/g) Fe oxides (White 1997; Sumner 2000). The lack of organic matter and clay-sized particles, and small specific surface area, are reflected in the relatively low ECEC of these samples (<5 [cmol.sub.c]/kg, Table 3).

Clearly, in its unamended 'fresh' condition, BRS exhibits limited potential as a growth medium capable of establishing and sustaining long-term plant growth without supplemental fertiliser addition to provide plant nutrients. For Alcoa, this problem is exacerbated because rehabilitation is undertaken during the wet winter months, which provides the greatest potential for nutrient loss through leaching. For these reasons, detailed research is currently under way to identify the most appropriate fertiliser type and form for rehabilitating Alcoa's RSAs.

Permanent ([[sigma].sub.p]) and variable ([[sigma.sub.v]) surface charge density

All samples exhibited ~0.2[cmol.sub.c]/kg of permanent negative surface charge ([[sigma].sub.p]), and between 3.2 (KW) and 3.9 (PJ and WG) [cmol.sub.c]/kg of variable surface charge ([[sigma].sub.v]) (Table 4). The magnitude of negative charge measured by the cesium-adsorption method (Anderson and Sposito 1991) was similar to that estimated by the ECEC (Table 3). Based on a specific surface area of 6.85 [m.sup.2]/g, then the surface charge density of BRS is ~5 x [10.sup.-4] [cmol.sub.c]/[m.sup.2], which is consistent with values reported for various soil components (White 1997).

In terms of nutrient behaviour, BRS appears to be best described as a soil material containing predominantly colloids with variable surface charge. Variable surface charge commonly occurs in soils dominated by Fe and/or Al oxides and hydrous oxides, and/or organic matter (e.g. Theng 1980; Gillman 2007). The mineralogical composition of BRS typically contains ~58% quartz, 22% hematite, 7% gibbsite, 5% goethite, 1% anatase, 1% desilication product (DSP), 1% muscovite, and 0.3% boehmite (Taylor and Pearson 2001), and 0.4-0.7% amorphous Fe and AI oxides (Table 3). This composition is consistent with that obtained from 2 independent analyses using XRD (M. Fey, The University of Western Australia, pers. comm.; R. Haynes, The University of Queensland, pers. comm.). Quartz can act as a diluent of surface negative charge due to its extremely low physiochemical activity (Drees et al. 1989).This, coupled with the lack of clay-sized material and strong red-brown coloration of BRS, suggests that amorphous Fe and A1 materials may coat much of the quartz grains, thereby allowing BRS to exhibit predominantly variable surface charge. More detailed mineralogical research is currently under way to better understand the role of Fe and AI oxides and hydroxides on surface charge properties of BRS (M. Fey, The University of Western Australia, pers. comm.).

Phillips and Sheehan (2005) investigated the surface charge characteristics for a range of soil types from south-east Queensland. Values of [[sigma].sub.p] for the surface (organic-rich) and subsurface (organic-poor) of a Humic Podosol, a Ferrosol, a Vertosol, and a Sodosol were 4.29, 0.54, 3.47, 19.27, and 0.67 [cmol.sub.c]/kg, respectively. Of these soils, the Humic Podosol (organic-poor) and Sodosol exhibited [[sigma].sub.p] values similar to that of BRS. Both of these soils were coarsetextured and dominated by quartz, and contained negligible organic matter. Charlet and Sposito (1987) and Anderson and Sposito (1991) reported [[sigma].sub.p] values approaching zero for a Brazilian Oxisol (0.033 [cmol.sub.c]/kg), and Chorover and Sposito (1995) reported [[sigma].sub.p] values for 4 Brazilian Oxisols ranging from 0.8 to 1.9[cmol.sub.c]/kg. Oxisols (also classified as Ferrosols) typically are the products of intense weathering and as such are dominated by colloids with variable charge. Since BRS contains predominantly Fe and Al oxides hydrous oxides and no organic matter, these samples would be expected to exhibit little to no [[sigma].sub.p].

The finding that BRS exhibits chemical properties similar to soils with predominantly variable surface charge is extremely important with regard to fertiliser efficiency and plant nutrient dynamics. The relatively low cation exchange capacity and high leaching potential (saturated hydraulic conductivity ~20 m/day) of BRS can result in rapid transport of weakly retained cations (K) and anions (Cl and S[O.sub.4]) from the plant root-zone. This is particularly relevant to Alcoa's Western Australian Operations because residue rehabilitation occurs during the wet winter months. On the other hand, the dominance of Fe and AI oxides and hydrous oxides can render nutrients (e.g. P) and many trace elements (e.g. Cu, Zn and Mn) unavailable for plant uptake (Qafoku et al. 2004; Thiyagarajan et al. 2009). A recent assessment of Alcoa's residue rehabilitation has suggested that older rehabilitation (>3 years old) may be deficient in N, K, B, Mg, Zn, and Cu despite receiving fertiliser addition at establishment (Bell et al. 2008). These findings highlight the importance of understanding the chemical properties of BRS and the associated impacts on nutrient dynamics.

Ion adsorption and PZNC

Each BRS displayed increasing K adsorption (CEC) and decreasing C1 (AEC) adsorption with increasing pH, confirming that all samples contained colloids with predominantly variable surface charge (Fig. 1). Potassium adsorption, hence negative charge, was negligible at pH < 7, after which adsorption increased significantly. Under acidic conditions, all BRS samples exhibited significant positive surface charge (~5 [cmol.sub.c]c/kg at pH 1), which decreased with increasing pH and approached zero at about pH 7. Similar behaviour by soils with variable charge has been reported by many workers, particularly for Oxisols (e.g. Gallez et al. 1976; Morais et al. 1976; Theng 1980; Phillips and Sheehan 2005).

All BRS samples exhibited negative K adsorption at low pH (Fig. 1). Under acidic conditions, considerable AEC was developed (2-5[cmol.sub.c]c/kg between pH 4 and pH 1). Under these conditions, a proportion of the positive charge (AEC) may have been balanced not by adsorption of negatively charged diffuse layer ions (e.g. CI), but by repulsion of positively charged ions (e.g. K) from the diffuse layer (Qafoku et al. 2004; Gillman 2007). The opposite effect at high pH values may explain the presence of adsorbed C1 and the subsequent measurement of positive charge in the BRS samples (Qafoku et al. 2004; Gillman 2007).

[FIGURE 1 OMITTED]

The net surface charge ([q.sub.K] - [q.sub.c1]) as a function of pH for each BRS is presented in Fig. 2. The PZNC was observed at pH 6.96, 6.89, and 5.98 for the KW, PJ, and WG samples, respectively. As stated above, the mineral composition of residue sand is dominated by quartz, with lesser amounts of hematite, gibbsite, goethite, and amorphous Fe and A1 oxides, and these materials exhibit a PZNC of ~3, 8.5, 9.8, 8.8, 8.0, and 9.3, respectively (Theng 1980; Goldberg et al. 2000). Although quartz comprises a major proportion of the minerals in BRS, it resides primarily in the sand fraction and is generally considered chemically inert (Drees et al. 1989). Thus, quartz can be regarded as a diluent to the more reactive materials (i.e. hematite, gibbsite, goethite, and amorphous Fe and A1 oxides). Since the overall PZNC reflects the weighted average of all colloids with surface charge (permanent and variable), the PZNC for the BRS samples appears to reflect the relative contributions of the more reactive materials, hematite, gibbsite, goethite, and amorphous Fe and A1 oxides.

Soils dominated by Fe and A1 oxides and hydrous oxides typically exhibit a PZNC within the range 2.7-6.5, depending on the organic matter content (Theng 1980). The BRS samples tended to exhibit PZNC values at the higher end of this range, which may be due to the lack of organic matter in all samples (Table 3). It is well established that the presence of organic matter will decrease the PZNC towards lower pH values, while the presence of hydrous Fe and A1 oxides will increase the PZNC towards higher pH values (van Raij and Peech 1972; Theng 1980).

Ammonium and phosphorus sorption isotherms N[H.sub.4]

Retention of N[H.sub.4] by BRS varied markedly depending on whether isotherms were developed using Eqn 2 (assumes any loss of solution N[H.sub.4] occurred solely through interactions with the surface charge) or Eqn 3 (based on measured changes in exchangeable N[H.sub.4]). The concentration of exchangeable N[H.sub.4] calculated using Eqn 2 always exceeded the measured concentration, particularly at pH 11. In fact, measured N[H.sub.4] adsorption decreased dramatically as pH increased from 9 to 11, despite an increase in negative surface charge (Figs 1 and 3).

[FIGURE 2 OMITTED]

The isotherms also show that while a small (-20%) proportion of added N[H.sub.4] was retained by cation exchange (pH 7 and 9 only), the remaining N[H.sub.4] was lost from solution (particularly at pH 11). Due to the lack of obvious N[H.sub.4] sinks in fresh BRS (e.g. nitrifying bacteria and organic carbon substrates), ammonia (NH3) volatilisation was considered to be the primary mechanism responsible for loss of solution N[H.sub.4]. Ammonia volatilisation is well known to occur following application of N[H.sub.4]-based fertilisers to alkaline soils, following urea application and subsequent hydrolysis, and from animal urine patches (e.g. Cabrera et al. 199l; Vaio et al. 2008). Reported NH3 losses from urea applied to grasslands range from 5 to 48% of the applied urea (Vaio et al. 2008). The importance of N[H.sub.3] volatilisation as an N loss mechanism in alkaline BRS is currently under investigation within Alcoa's residue rehabilitation research program.

[FIGURE 3 OMITTED]

The measured N[H.sub.4] isotherms were reasonably well described by the Freundlich equation (Eqn 4), with [r.sup.2] values exceeding 0.94 (Fig. 3). The sorption index at pH 7 and 9 was 0.73 and 0.85 for KW, and 0.56 and 0.58 for P J, respectively, but decreased to <0.2 at pH 11 for both KW and PJ samples. The magnitude of the N[H.sub.4] sorption parameters determined for BRS at pH 7 and 9 are similar to other sandy-textured soils which contain negligible organic matter (Phillips 2002; Phillips and Sheehan 2005).

Phosphorus

Phosphorus sorption isotherms showed that at pH 7 and 9, essentially all of the added P was lost from solution (Fig. 4). The loss of solution P is attributed to the presence of significant amounts of Fe and A1 oxides and hydrous oxides (Table 3), and the presence of Ca which can result in the formation of low solubility Ca-phosphates (Sanyal and De Datta 1991).

[FIGURE 4 OMITTED]

Phosphorus sorption was found to decrease with increasing pH (Fig. 4). For example, at pH 7, P sorption at the highest addition was ~9 mmol/kg, while at pH 11, P sorption was ~6.5 mmol/kg. This was associated with a significant (P<0.05) decrease in the sorption index (Eqn 4) from ~100 L/kg at pH 7 to 10 L/kg at pH 11 (Fig. 4). Fresh BRS contains >500 mmol C[O.sub.3]/L (Table 1) and exhibits increasing negative surface charge with increasing pH (Figs 1 and 2). Both of these factors can reduce P sorption with increasing pH due to the presence of competing anions (Parfitt 1978), and anion repulsion (Bowden et al. 1980). A similar effect of pH on P sorption has been reported by Parfitt (1980) on goethite, and Barrow (1982) on residue mud.

Conclusions

Bauxite-processing residue sand exhibits many properties which can restrict optimum plant establishment and growth. These include excessive salinity and alkalinity, extreme sodicity, elevated soluble AI, and extremely low organic matter and microbial activity. Although some of these properties can be managed using amendments such as gypsum and inorganic fertilisers, many challenges exist to produce a growth medium that can support a sustainable vegetation cover.

Much of the surface charge of BRS is variable in nature due to the dominance of Fe and A1 oxides and hydrous oxides. The PZNC of BRS was within the range of pH 6-7, which is consistent with values reported for the individual reactive mineral components of BRS. Since BRS exhibits pH values exceeding 7 for many years after placement at a RSA, this material would be expected to carry a net negative surface charge. This has significant impact on fertiliser selection for BRS because rehabilitation at Alcoa's RSAs is undertaken during the winter wet season, and plant nutrient anions (e.g. C1 and N[O.sub.3]) and weakly retained cations (e.g. K) would be highly susceptible to loss through leaching.

Ammonium adsorption by BRS decreased with increasing pH despite an increase in net negative surface charge. Although cation exchange was considered responsible for some of the N[H.sub.4] adsorption, much of the loss of soluble N[H.sub.4] may have occurred through volatilisation. This finding has significant implications for selecting the most appropriate form of inorganic N (N[H.sub.4] or N[O.sub.3]) fertiliser for use in BRS rehabilitation as much of the applied N[H.sub.4] and/or N[O.sub.3] could be lost by volatilisation and/or leaching, respectively. The observed absence of inorganic N in BRS and plants from established rehabilitation (Bell et al. 2008) clearly shows more research is required to quantify N dynamics in BRS, particularly with respect to volatilisation and leaching.

Reduced P sorption with increasing pH suggests that fertiliser P is most available for plant uptake and most susceptible to leaching when BRS is fresh. However, despite the strong sorption properties of BRS, P deficiencies in BRS rehabilitation are not obvious. This could be because the use of phosphogypsum in Alcoa's residue rehabilitation program provides ~30 kg P/ha (I. R. Phillips, unpubl, data). Thus, losses of inorganic P may be compensated for by release of P via gypsum dissolution. More research is required to quantify P (and other key nutrients) dynamics in BRS, particularly in terms of plant availability in both the short and long term to ensure development of a sustainable ecosystem.

10.1071/SR09056

Acknowledgments

The authors thank Alcoa's Western Australia Residue Operations group for funding this study, CSBP Ltd and the Western Australia Chemistry Centre for chemical analyses, Mr Laurin Wissmeier, Federal Technical University of Lausanne for specific surface area measurement, and Professor Martin Fey, The University of Western Australia, and Professor Richard Haynes, The University of Queensland, for mineralogical analysis and helpful comments during preparation of this manuscript.

Manuscript received 1 April 2009, accepted 9 September 2009

References

Adams F (1984) 'Soil acidity and liming.' Agronomy Series No. 12. (Soil Science Society of America: Madison, WI) Analytical Software (1994) 'Statistix.' (Analytical Software: Tallahassee, FL)

Anderson SJ, Sposito G (1991) Cesium-adsorption method for measuring accessible structural surface charge. Soil Science Society of America Journal 55, 1569-1576.

Apak R, Guclu K, Turgut MH (1998) Modeling of copper(II), cadmium(II), and lead(II) adsorption on red mud. Journal of Colloid and Interface Science 203, 122-130. doi:10.1006/jcis.1998.5457

Banning NC, Phillips IR, Jones DL, Murphy DV (2009) Rapid development of microbial diversity and functional potential in alkaline, nutrient-limited bauxite residue sand under rehabilitation. Restoration Ecology, (in press).

Barrow NJ (1982) Possibility of using caustic residue from bauxite for improving the chemical and physical properties of sandy soils. Australian Journal of Agricultural Research 33, 275-285. doi: 10.1071/AR9820275

Bell RW, Dell B, Thiyagarajan C (2008) A preliminary analysis of the nutrient status of plants and residue at the Pinjarra and Kwinana residue disposal areas of Alcoa World Alumina Australia. Murdoch University, Perth, Australia.

Blair GJ, Chinoim N, Lefroy RDB, Anderson GC, Crocker GJ (1991) A soil sulphur test for pastures and crops. Australian Journal of Soil Research 29, 619-626. doi:10.1071/SR9910619

Bowden JW, Posner AM, Quirk JP (1980) Adsorption and charging phenomena in variable charge soils. In 'Soils with variable charge'. (Ed. BKG Theng) pp. 147-166. (New Zealand Society of Soil Science: Lower Hutt, New Zealand)

Burkitt LL, Moody PW, Gourley CJP (2002) A simple phosphorus buffering index of Australian soils. Australian Journal of Soil Research 40, 497-513. doi:10.1071/SR01050

Cabrera ML, Kissel DE, Brock BR (1991) Urea hydrolysis in soils: effects of urea concentration and soil pH. Soil Biology & Biochemistry 23, 1121-1124. doi:10.1016/0038-0717(91)90023-D

Carter C (2006) Composition of AWA bauxite and bauxite residue results of a 12 month survey. Technology Delivery Group, Alcoa World Alumina Australia, Perth, Australia.

Carter C, van der Sloot HA, Cooling D (2009) pH-dependant extraction of soil and soil amendments to understand the factors controlling element mobility. European Journal of Soil Science 60, 622 637. doi: 10.1111/ j.1365-2389.2009.01139.x

Carter C, van der Sloot HA, Cooling D, van Zomeren A, Matheson T (2008) Characterisation of untreated and neutralised bauxite residue for improved waste management. Environmental Engineering Science 25, 475-488. doi: 10.1089/ees.2006.0234

Charlet L, Sposito G (1987) Monovalent ion adsorption by an Oxisol. Soil Science Society of America Journal 51, 1155-1160.

Chorover J, Sposito G (1995) Surface charge characteristics of kaolinitic tropical soils. Geochimica et Cosmochimica Acta 59, 875-884. doi:l0.1016/0016-7037(94)00357-2

Courtney RG, Timpson JP (2005) Reclamation of fine fraction bauxite processing residue (red mud) amended with coarse fraction residue and gypsum. Water, Air, and Soil Pollution 164, 91-102. doi:10.1007/ s11270-005-2251-0

Drees LR, Wilding LP, Smeck NE, Senkayi AL (1989) Silica in soils: quartz and disordered silica polymorphs. In 'Minerals in soil environments'. (Eds JB Dixon, SB Weed) pp. 913-974. (Soil Science Society of America: Madison, WI)

Eleftheriou EP, Moustakas M, Fragiskos N (1993) Aluminate-induced changes in morphology and ultrastructure of Thinopyrum roots. Journal of Experimental Botany 44, 427-436. doi:10.1093/jxb/44.2.427

Flint LE, Flint AL (2002) Particle density. In 'Methods of soil analysis. Part 4. Physical methods'. (Eds JH Dane, G Clarke Topp) pp. 241-254. (Soil Science Society of America: Madison, WI)

Fuller RD, Nelson EDP, Richardson CJ (1982) Reclamation of red mud (bauxite residues) using alkaline-tolerant grasses with organic amendments. Journal of Environmental Quality 11, 533-539.

Fuller RD, Richardson CJ (1986) Aluminate toxicity as a factor controlling plant-growth in bauxite residue. Environmental Toxicology and Chemistry 5, 905-915. doi:l0.1897/1552-8618(1986)5[905:ATAAFC] 2.0.CO;2

Gallez A, Juo ASR, Herbillon AJ (1976) Surface charge characteristics of selected soils in the tropics. Soil Science Society of America Journal 40, 601-608.

Gherardi MJ, Rengel Z (2003a) Genotypes of lucerne (Medicago sativa L.) show differential tolerance to manganese deficiency and toxicity when grown in bauxite residue sand. Plant and Soil 249, 287-296. doi:l0.1023/A:1022872524844

Gherardi MJ, Rengel Z (2003b) Deep banding improves residual effectiveness of manganese fertiliser for bauxite residue revegetation.

Australian Journal of Soil Research 41, 1273-1282. doi:10.1071/ SR03030

Gherardi MJ, Rengel Z (2003c) Deep placement of manganese fertiliser improves sustainability of lucerne growing on bauxite residue: a glasshouse study. Plant and Soil 257, 85-95. doi:10.1023/ A:1026252114933

Gillman GP (2007) An analytical tool for understanding the properties and behaviour of variable charge soils. Australian Journal of Soil Research 45, 83-90. doi:10.1071/SR06117

Goldberg S, Lebron I, Suarez DL (2000) Soil colloidal behavior. In 'Handbook of soil science'. (Ed. ME Sumner)pp. B195-B240. (CRC Press: New York)

Golden Software (2000) 'Grapher[TM] User's Guide.' (Golden Software: Golden, CO)

Gourley CJP, Melland AR, Waller RA, Awry IM, Smith AP, Peverill KI, Hannah MC (2007) Making better fertiliser decisions for grazed pastures in Australia. Victorian Government, Department of Primary Industries, Victoria.

Huang X, Foster GD, Honeychuck RV, Schreifels JA (2009) The maximum of phosphate adsorption at pH 4.0: why it appears on aluminium oxides but not on iron oxides. Langmuir 25, 4450-4461. doi:10.1021/ la803302m

Indorante SJ, Hammer RD, Koenig PG (1990) Particle-size analysis by a modified pipette procedure. Soil Science Society of America Journal 54, 560-563.

Kinraide TB (1990) Assessing the rhizotoxicity of the aluminate ion, [Al(OH).sub.4]. Plant Physiology 93, 1620-1625. doi: 10.1104/pp.93.4.1620

Kopittke DG, Menzies NW, Blamey FPC (2004) Rhizotoxicity of aluminate and polycationic aluminium at high pH. Plant and Soil 266, 177-186. doi: 10.1007/s11104-005-2229-0

Lopez E, Soto B, Arias M, Nunez A, Rubinos D, Barral MT (1998) Adsorbent properties of red mud and its use for wastewater treatment. Water Research 32, 1314-1322. doi: 10.1016/S0043-1354(97)00326-6

Ma G, Rengasamy P, Rathjen AJ (2003) Phytotoxicity of aluminium to wheat plants in high-pH solutions. Australian Journal of Experimental Agriculture 43, 497-501. doi:10.1071/EA01153

Meecham JR, Bell LC (1977) Revegetation of alumina refinery wastes. I. Properties and amelioration of the materials. Australian Journal of Experimental Agriculture 17, 679-588. doi:10.1071/EA9770679

Morais FI, Page AL, Lund LJ (1976) The effect of pH, salt concentration, and nature of electrolytes on the charge characteristics of Brazilian tropical soils. Soil Science Society of America Journal 40, 521-527.

Parfitt RL (1978) Anion adsorption by soils and soil materials. Advances in Agronomy 30, 1-50. doi:10.1016/S0065-2113(08)60702-6

Parfitt RL (1980) Chemical properties of variable charge soils. In 'Soils with variable charge'. (Ed. BKG Theng) pp. 167-194. (New Zealand Society of Soil Science: Lower Hutt, New Zealand)

Parkhurst DL, Appelo AAJ (1999) 'User's guide to PHREEQC (version 2) A computer program for speciation, batch-reaction, one dimensional transport, and inverse geochemical modeling.' (U.S. Geological Survey: USA)

Pennell KD (2002) Specific surface area. In 'Methods of soil analysis. Part 4. Physical methods'. (Eds JH Dane, G Clarke Topp) pp. 295-315. (Soil Science Society of America: Madison, WI)

Phillips IR (2002) Phosphorus sorption and nitrogen transformations in soils treated with piggery wastewater. Australian Journal of Soil Research 40, 335-349. doi:10.1071/SR01040

Phillips IR , Sheehan KJ (2005) Importance of surface charge characteristics when selecting soils for wastewater re-use. Australian Journal of Soil Research 43, 915-927. doi: 10.1071/SR05045

Pradhan J, Das SN, Singh Thakur R (1999) Adsorption of hexavalent chromium from aqueous solution by using activated red mud. Journal of Colloid and Interface Science 217, 137-141. doi:10.1006/jcis. 1999.6288

Qafoku NP, van Ranst E, Noble A, Baert G (2004) Variable charge soils: their mineralogy, chemistry and management. Advances in Agronomy 84, 159-215. doi: 10.1016/S0065-2113(04)84004-5

Rayment GE, Higginson FR (1992) 'Australian laboratory handbook of soil and water chemical methods.' (Inkata Press: Melbourne, Vic.)

Sanyal SK, De Datta SK (1991) Chemistry of phosphorus transformations in soil. Advances in Soil Science 16, 1-120.

Snars K, Gilkes R, Hughes J (2003) Effect of soil amendment with bauxite Bayer process residue (red mud) on the availability of phosphorus in very sandy soils. Australian Journal of Soil Research 41, 1229-1241. doi:10.1071/SR02140

Snars KE, Gilkes RJ, Wong MTF (2004) The liming effect of bauxite processing residue (red mud) on sandy soil. Australian Journal of Soil Research 42, 321-328. doi:10.1071/SR02140

Sposito G (1992) Characterization of particle surface charge. In 'Environmental particles. Vol. 1'. (Eds J Buffle, H van Leeuwen) pp. 291-314. (International Union of Pure and Applied Chemistry, Lewis Publishers: London)

Sumner ME (2000) 'Handbook of soil science.' (CRC Press: Boca Raton, FL)

Taylor S, Pearson N (2001) Properties of Bayer process solids from Alcoa WA Refineries and their component minerals. Technology Delivery Group, Alcoa World Alumina Australia, Perth, Australia.

Theng BKG (1980) 'Soils with variable charge.' (New Zealand Society of Soil Science: Lower Hutt, New Zealand)

Thiyagarajan C, Phillips IR, Dell B, Bell RW (2009) Micronutrient fractionation and plant availability in bauxite-processing residue sand. Australian Journal of Soil Research 47, 518-528.

Tisdale SL, Nelson WL, Beaton JD, Havlin JL (1993) 'Soil fertility and fertilizers.' 5th edn, pp. 176-229. (Macmillan Publishing Company: New York)

Vaio PN, Cabrera ML, Kissel DE, Rema JA, Newsome JF, Calvert VH (2008) Ammonia volatilisation from urea-based fertilisers applied to tall fescue pastures in Georgia, USA. Soil Science Society of America Journal 72, 1665-1671. doi:10.2136/sssaj2007.0300

van Raij B, Peech M (1972) Electrochemical properties of some Oxisols and Alfisols of the tropics. Soil Science Society of America Proceedings 36, 587-593.

White RE (1997) 'Principles and practice of soil science.' 3rd edn (Blackwell Sciences Pty Ltd: Melbourne, Vic.)

Wong JWC, Ho GE (1995) Cation-exchange behavior of bauxite refining residues from Western-Australia. Journal of Environmental Quality 24, 461-466.

Wong MTF, Wittwer K (2009) Positive charge discovered across Western Australian wheatbelt soils challenges key soil and nitrogen management assumptions. Australian Journal of Soil Research 47, 127-135. doi: 10.1071/SR08098

Woodard HJ, Hossner L, Bush J (2008) Ameliorating caustic properties of aluminum extraction residue to establish a vegetative cover. Journal of Environmental Science and Health. Part A 43, 1157-1166. doi: 10.1080/ 10934520802171659

Zelazny LW, He L, van Wormhoudt AN (1996) Charge analysis of soils and anion exchange. In 'Methods of soil analysis. Part 3. Chemical methods'. (Ed. DL Sparks) (Soil Science Society of America: Madison, WI)

I. R. Phillips (A,C) and C. Chen (B)

(A) Alcoa of Australia Ltd, PO Box 172, Pinjarra, WA 6208, Australia.

(B) Griffith University, Nathan, Qld 4111, Australia.

(C) Corresponding author. Email: Iaan.Phillips@alcoa.com.a
Table 1. Mean and standard deviation (s.d.) for major ions in
saturated paste water extract for Kwinana (KW), Pinjarra (PJ), and
Wagerup (WG) bauxite residue sand

All concentrations are expressed in units of mmol/L unless indicated
otherwise. bdl, Below detection limit; SAR, sodium adsorption ratio

                   KW                  PJ                WG

              Mean      s.d.      Mean    s.d.      Mean    s.d.

Na            123.6      2.77     261.3    4.00     166.2    2.15
Mg             0.08     0.029      0.10   0.038      0.17   0.021
Al             0.68     0.026      2.03   0.034      0.24   0.018
P              0.09     0.023      0.30   0.030      0.19   0.016
S[0.sub.4]     1.27     0.125      2.61   0.157      1.60   0.108

Cl             4.45     0.219      7.42   0.345      3.03   0.133
K              0.35     0.027      0.52   0.024      0.46   0.049
Ca             0.17     0.034      0.39   0.058      0.54   0.048
Fe             0.00     0.000      0.00   0.000      0.00   0.000
NH4            0.06     0.000      0.06   0.000      0.06   0.000
N03            0.02     0.000      0.05   0.003      0.02   0.000
EC (dS/m)      1.91     0.148      3.56   0.122      2.16   0.161
pH            10.58     0.212     10.67   0.152     10.44   0.191
[[theta]
  .sub.g],
  (kg/kg)      0.28     0.014      0.31   0.014      0.30   0.014
HC[O.sub.3]     bdl                 bdl               bdl
C[O.sub.3]      508         3       968       2       548       3
SAR             351        36       532      45       280      10

Table 2. Using PHREEQC (pH redox equilibrium; Parkhurst and
Appelo 1999) to speciate the major ions in the saturated paste
water extract for Kwinana (KW), Pinjarra (PJ), and Wagerup
(WC) bauxite residue sand

The proportion of each dominant species is expressed as a
percentage of its total solution ion concentration

Ion   Dominant species          KW           PJ           WJ

Na    NaC[0.sub.3],           70, 29       88, 11       69, 30
      [Na.sup.+]

Mg    MgC[0.sub.3]              96           98           95

Al    Al[(OH).sub              100          100          100
      .4.sup.-]

P     HP[0.sub.4.sup.2],     88, 8, 3     91, 7, 1    85, 11, 2
      NaHP[0.sub.4],
      P[0.sub.4.sup.3]

S04   S[0.sub.4.sup.2] ,      93, 7        89, 11       90, 10
      NaS[0.sub.4]

CI    Cl                       100          100          100

K     [K.sub.+]                100           99          100

Ca    CaC[0.sub.3]              97           98           96

Fe    Fe[(OH).sub.4]            98           99           97

C     C[0.sub.3.sup.2-],    64, 18, 17   54, 22, 24   57, 22, 21
      HC[0.sub.3],
      NaC[0.sub.3.sup-]

Table 3. Mean and standard deviation (n=2) for selected
physical and chemical properties for Kwinana (KW), Pinjarra (PJ),
and Wagerup (WG) bauxite residue sand

                                   KW               PJ

                              Mean    s.d.     Mean    s.d.

KCI N[0.sub.3] (mg/kg)          1.0    0.00      1.0    0.00
KCI N[H.sub.4] (mg/kg)          1.0    0.00      2.0    0.21
NaHC[0.sub.3] P (mg/kg)         3.2    0.61      2.3    0.21
NaHC[0.sub.3]-K (mg/kg)        11.7    0.21     13.2    0.21
KCl S (mg/kg)                  14.1    1.70     32.3    2.12
Organic C (%)                  0.13   0.082     0.07   0.014
Oxalate Fe (mg/kg)             4690    56.6     7740    91.9
Oxalate Al (mg/kg)               90     7.1       97    21.2
EDTA Cu (mg/kg)                0.27   0.042     0.37   0.099
EDTA Zn (mg/kg)                0.03   0.003     0.13   0.008
EDTA Mn (mg/kg)                0.77   0.042     0.78   0.021
EDTA Fe (mg/kg)               78.83   2.744    84.28   3.041
Exch. Ca ([cmol.sub.c]/kg)     1.28   0.226     1.91   0.212
Exch. Mg ([cmol.sub.c]/kg)     0.12   0.014     0.06   0.021
Exch. Na ([cmol.sub.c]/kg)     3.57   0.325     2.92   0.212
Exch. K ([cmol.sub.c]/kg)      0.01   0.014     0.01   0.021
Exch. Al ([cmol.sub.c]/kg)     0.01   0.007     0.01   0.007
Exch. Fe ([cmol.sub.c]/kg)     0.56   0.013     0.46   0.021
ECEC ([cmol.sub.c]/kg)         4.99   0.451     4.91   0.414
PBI                             565     4.2      551     9.9
Total N (%)                    0.03   0.014     0.01   0.014
Total Al (mg/kg)              44660   169.7    45110    72.1
Total Fe (mg/kg)             243700   260.2   251200   220.6
Total P (mg/kg)                26.3    1.98     24.6    2.12
Coarse sand (200-2000
  [micro]m) (%)                  65       1       67       2
Fine sand (20-200
  [micro]m) (%)                  28       2       23       2
Silt (>2-201 [micro]m) (%)        2       0        2       0
Clay (<20 [micro]m) (%)           5       0        8       0

                                   WG

                              Mean    s.d.

KCI N[0.sub.3] (mg/kg)          1.0    0.00
KCI N[H.sub.4] (mg/kg)          1.0    0.00
NaHC[0.sub.3] P (mg/kg)         0.9    0.18
NaHC[0.sub.3]-K (mg/kg)         6.9    0.91
KCl S (mg/kg)                  22.3    3.25
Organic C (%)                  0.05   0.051
Oxalate Fe (mg/kg)             4370    48.1
Oxalate Al (mg/kg)              122    15.6
EDTA Cu (mg/kg)                0.24   0.057
EDTA Zn (mg/kg)                0.01   0.001
EDTA Mn (mg/kg)                0.63   0.113
EDTA Fe (mg/kg)               67.22   2.093
Exch. Ca ([cmol.sub.c]/kg)     1.25   0.240
Exch. Mg ([cmol.sub.c]/kg)     0.01   0.014
Exch. Na ([cmol.sub.c]/kg)     2.40   0.170
Exch. K ([cmol.sub.c]/kg)      0.01   0.007
Exch. Al ([cmol.sub.c]/kg)     0.01   0.007
Exch. Fe ([cmol.sub.c]/kg)     0.13   0.033
ECEC ([cmol.sub.c]/kg)         3.68   0.423
PBI                             521    15.6
Total N (%)                    0.02   0.014
Total Al (mg/kg)              39310   154.1
Total Fe (mg/kg)             183800   155.6
Total P (mg/kg)                28.1    2.26
Coarse sand (200-2000
  [micro]m) (%)                  69       3
Fine sand (20-200
  [micro]m) (%)                  24       3
Silt (>2-201 [micro]m) (%)        1       0
Clay (<20 [micro]m) (%)           6       0

Table 4. Surface charge characteristics
of fresh BRS from each refinery

Values designated by the same letter are
not significantly (P> 0.05) different
from each other (l.s.d.=0.334)

Site      Type of            Conc.
           charge       ([cmol.sub.c]/kg)

KW     [sigma].sub.p]        0.203a
Pi     [sigma].sub.p]        0.263a
WG     [sigma].sub.p]        0.193a
KW     [sigma].sub.v]        3.238b
Pi     [sigma].sub.v]        3.905c
WG     [sigma].sub.v]        3.913c
COPYRIGHT 2010 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Phillips, I.R.; Chen, C.
Publication:Australian Journal of Soil Research
Article Type:Report
Geographic Code:8AUST
Date:Feb 1, 2010
Words:9773
Previous Article:Soil spatial variability of drainage properties in relation to phosphate retention and mineralogy on a river terrace of northern Manawatu, New...
Next Article:Inheritance and formation of smectite in a texture contrast soil in the Pilliga State Forests, New South Wales.
Topics:

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters