ARTICLE pubs.acs.org/est
Nutrient and Trace-Metal Removal by Bauxsol Pellets in Wastewater Treatment Laure M. Despland,*,†,‡,§ Malcolm W. Clark,† Tony Vancov,‡ Dirk Erler,† and Michel Aragno§ †
School of Environmental Science and Management, Southern Cross University, P.O. Box 157, Lismore, New South Wales 2480, Australia ‡ Industry & Investment NSW, 1243 Bruxner Highway, Wollongbar NSW 2477, Australia § Laboratory of Microbiology, Institute of Biology, University of Neuch^atel, P.O. Box 158, 2009 Neuch^atel, Switzerland ABSTRACT: In this study, Bauxsol pellets packed in PVC columns were used to remove nutrients and trace-metals from municipal wastewater during a 6 months field trial. Bauxsol pellet columns showed a high phosphate removal rate via precipitation of PO43- with Ca2þ and Mg2þ ions: at 90% in the 1st month; at 80% from the second to fifth months; and at 60% in the sixth month. Pellet bound total phosphorus and Colwell phosphorus were 7.3 g/kg and 2 g/kg and are about 20 times the concentrations found in most fertile soils. Trace-metals in effluents were bound, probably irreversibly under the columns’ environmental conditions, to the Bauxsol minerals that have high surface area to volume ratios and high charge to mass ratios. Experimental results showed a complex nitrogen cycle operating within the Bauxsol pellet columns including anoxic nitrification, denitrification, and anammox processes. Although a transient pH spike, associated with the release of unreacted CaO from the cement binder used in the pellets, was observed, this may be readily corrected through post-treatment pH adjustment. Hence, the geochemistry of Bauxsol pellets can effectively remove and bind nutrients and trace-metals during wastewater treatment, and further research may show that saturated spent pellets can be used as fertilizer.
’ INTRODUCTION Wastewaters contain high concentrations of phosphorus and nitrogen often leading to eutrophication with potential health hazards to both animal and humans. During the sewage treatment process, nitrogen is primarily removed by biological activities, whereas phosphorus is mostly removed by chemical precipitation.1 In addition to nitrification and denitrification, nitrogen removal may occur by anammox process (NH4þ þ NO2 f N2 þ 2H2O), which is mediated by a group of chemoautotrophic anaerobes. Anammox is noteworthy because, unlike denitrification, it allows nitrogen removal in anoxic environments without requiring a reducing capacity.2 Free phosphorus primarily exists in wastewater as orthophosphates (o-PO43-) and depending on pH may be present as H2PO4, HPO42-, or PO43-. Phosphate anions precipitate readily with soluble metal salts, such as Al2(SO4)3 - alum - and FeCl3, forming metal-phosphate-hydroxo complexes. Lanthanum (La) impregnated bentonite Phoslock is also commonly used as a phosphate removal agent because the resulting LaPO4 is extremely insoluble.3 However, use of chemical precipitation in wastewater treatments is an unsound practice because of the cost, the addition of new contaminants (sulfate, chloride and La ions), the requirement of strict dosing in proportion to the phosphate r 2011 American Chemical Society
concentration (i.e., water may become toxic through excess soluble Al/Fe), and the creation of large quantities of phosphate-rich sludge deposits requiring disposal.1,4,5 In addition, wastewaters contain trace-metals from industrial discharges and urban stormwater runoffs. Trace-metals are generally closely monitored and removed by sorption using metaloxide-based sorbents, because they cannot be broken-down. 6 Consequently, there is a desire to find cost-effective and environmentally beneficial materials to remove nutrients and trace-metals from wastewater. Bauxsol, a modified bauxite refinery residue, is available as powder, slurry, or pellets. Bauxsol reagents have been successfully applied to environmental remediation for acidic contaminated soils and waters.79 Batch studies have also demonstrated the potential of Bauxsol to remove and bind phosphorus from pH-controlled waters;10,11 however, Bauxsol reagents (powders) are limited by their hydraulic conductivity. No studies to date have investigated the geochemical and biological removal of Received: March 21, 2011 Accepted: May 23, 2011 Revised: May 18, 2011 Published: June 03, 2011 5746
dx.doi.org/10.1021/es200934y | Environ. Sci. Technol. 2011, 45, 5746–5753
Environmental Science & Technology nutrients and trace-metals by pellets in circum-neutral waters, such as sewage effluent, although clear cost savings of Bauxsol powders in wastewater treatment have been made.12 Consequently, in a previous study Despland et al.13 developed porous Bauxsol pellets, using an ordinary Portland cement (OPC) binder, that characteristically possess a high surface to volume ratio (up to 100 m2/g); a high binding capacity (metals: >1500 meq/kg; phosphorus: >2% by mass); and moderate acid neutralizing capacity and are insoluble and highly nondispersive.9,11,14 However, unreacted calcium oxide in the cement binder produces a pH/alkalinity spikes when pellets are placed in circum-neutral waters. To mitigate against this, a postproduction curing process using MgCl2 and CaCl2 precipitates most hydroxide alkalinity.13,15 The commercial value of Bauxsol pellets produced this way is about $US700 per ton. The aim of this paper is to investigate the efficiency of cured OPC-bound porous Bauxsol pellets in wastewater treatment to remove nutrients and trace-metals. Specifically, the objectives are to monitor a possible pH spike caused by Bauxsol pellets, to quantify total phosphorus and phosphate loadings removed from the effluent and bound to pellets, to determine the nitrogen removal processes occurring in the system, and to measure trace-metals loadings trapped by pellets over a 6 months period.
’ EXPERIMENTAL SECTION PVC columns (130 cm high 10 cm diameter) were filled with 10 L of OPC-bound Bauxsol pellets (510 mm diameter; total weight of 5 kg; experimental columns)13 or ordinary basalt gravel (510 mm diameter; total weight of 13.5 kg; control column). Secondary treated effluent from the South Lismore Sewage Treatment Plant NSW Australia was fed into the columns in an upward direction at 15 mL/min to provide a hydraulic retention time (HRT) of approximately 11 h. Particulate clogging of columns was prevented by placing polyester fiber beneath the caps at both ends of the each column and by a regular cleaning and maintenance of the system. At predetermined bed volumes, liquid samples (1501200 mL) were collected from column inlet, midway column, and outlets during the 6 months trial. At the end of the experiment, the columns were opened vertically, and the solids (Bauxsol pellets and Gravel) were sampled at 0 (bottom; inlet), 25, 50, 75, 105, and 130 (top; outlet) cm points along the column. The pH (APHA 4500 Hþ-b), electrical conductivity (EC) (APHA 2510-b), and temperature (APHA 2550-b) were recorded in situ every six hours at the column inlets and outlets using a Conductivity-TDS-pH-Temperature meter (TPS, WP-81). Individual liquid samples were also tested for dissolved oxygen (DO) (APHA 2810-b) using a Dissolved OxygenTemperature meter (TPS, WP-82Y).16 Alkalinity determinations (APHA 2320-b) were conducted using 20 mL inlet and outlet waters samples using a potentiometric titrator (Metrohm, Titrando 856), while alkalinity speciation was determined by the fixed end point method using the online Alkalinity Calculator.17 Concentrations of total phosphorus (TP) (APHA 4500 P-h), orthophosphate (PO43-) (APHA 4500 P-g), ammonia (NH4þ) (APHA 4500 NH3-h), nitrite (NO2) (APHA 4500 NO2-c), nitrate (NO3) (APHA 4500 NO3-f), total nitrogen (TN) (APHA 4500 N-c), and calcium (Ca) (APHA 3120 ICPOES) were assessed for column inlet, midway point, and outlet liquid
ARTICLE
samples.16 Solids samples (before and after experiment at each distance) were analyzed for the following: TP and Colwell phosphorus,18 NH4þ and NO3 (1:10 KCl extract APHA 4500), trace-metals such as silver (Ag), arsenic (As), lead (Pb), cadmium (Cd), chromium (Cr), copper (Cu), manganese (Mn), nickel (Ni), selenium (Se), zinc (Zn), mercury (Hg), iron (Fe), and aluminum (Al) (Nitric/HCl digest-APHA 3120 ICPMS).16 Nitrogen assays were performed at the end of the trial on solids with a range of stable isotope amendments (mixture of 15N and 14N) in 20 mL glass vials with screw cap lids with Teflon septa. Gravel solids or Bauxsol pellets (triplicate) were incubated with deoxygenated tap water and with 15NO3, 15NH4þ, or a combination of 15NH4þ þ 14NO3. In addition, vials containing Bauxsol pellets and the same three N combinations were mixed with methanol (MeOH) at 15 mg/L. Unamended controls vials were also included. At 0, 1, 3, and 5 days vials were analyzed after stopping microbial activity with 0.5 mL of ZnCl2 (50% w/v). After 5 days incubation, 5 mL of helium headspace was introduced into the remaining vials. The headspace was subsequently analyzed by GC-IRMS (Thermo Trace GC Ultra) interfaced to the isotope ratio mass spectrometer (Thermo Finnigan GC Combustion III). All preparations and incubations were performed in an anoxic glovebag. Production of 29N2 in treatments where only 15NH4þ is added is indicative of nitrification, whereas production of 29N2 in treatments where 15NH4þ þ 14NO3 is added indicates anammox process. The concentration of 29N2 (i.e.15N14N) was calculated using eq 1 - similar equation for 30N2 (i.e.15N15N) - derived by Nielsen19 ! 29 29 A A C29 N 2 ¼ All s All c C ÷ W ð1Þ As Ac where C29N2 = concentration of 29N2 [μmol/g]; 29As = area of 29N2 peak for the sample from GC-IRMS; AllAs = total area of N2 peak for the sample; 29Ac = area of 29N2 peak for the control; All Ac = total area of N2 peak for the control; C = total concentration of N2 in the incubation water; and W = dry weight of Gravel or Bauxsol pellets in each vial. The denitrification rate was calculated using eq 2 - derived by Nielsen19 ! R 29 N 2 Dnt ¼ ð2Þ ÷ ðR 29 N2 þ 2R 30 N 2 Þ 2 R 30 N2 where Dnt = rate of denitrification [μmol/g/d]; R29N2 = production’s rate of 29N2; R30N2 = production’s rate of 30N2; and R29N2 and R30N2 were calculated as the slope of the linear increase in concentration over time.
’ RESULTS AND DISCUSSION pH Spike and Stabilization. Although the pH of the gravel column’s outlet was similar to the inlet (pH 6.37.2), a pH spike (10.7) was recorded in the Bauxsol columns’ outlets during the first 15 bed volumes (≈1 week). The pH fell to 9.5 after another 15 bed volumes (≈2 weeks) and slowly decreased to 8.5 over the next 100 bed volumes (≈2 months). A further 260 bed volumes (≈4 months) saw the pH decline to 7.8 - Figure 1. This pH spike resulted from the release of unreacted calcium oxide in the cement binder. However, the spike was short-lived because most of the cement derived hydroxide was converted to low solubility 5747
dx.doi.org/10.1021/es200934y |Environ. Sci. Technol. 2011, 45, 5746–5753
Environmental Science & Technology
5748
0.38 0.31 0.50 0.43 0.22 0.15 0.94 ( 0.17 0.84 ( 0.14 0.67 ( 0.25 1.10 ( 0.35 0.80 ( 0.10 1.71 ( 0.22 aluminuma
a
Bp = Bauxsol pellets, Gc = Gravel; 0 cm = inlet, and 130 cm = outlet (intermediate distances are also provided). b All data are mg/kg except iron and aluminium, which are in weight %; ( standard error for Bp.
0.00
1.97 1.69
0.00 0.00
1.71 1.88
0.00 0.00
1.03 1.22
0.00 0.08 ( 0.02
2.20 ( 0.36 2.02 ( 0.53
0.06 ( 0.01 0.05 ( 0.01
1.41 ( 0.82 1.40 ( 0.30 4.05 ( 0.22
2.62 ( 1.08
0.05 ( 0.00 0.09 ( 0.02 mercury
ironb
0.07 ( 0.01
0.09
16.91 22.25 28.70 26.83 8.39 2.99 37.06 ( 8.35 49.89 ( 4.13 70.47 ( 13.36 52.71 ( 6.05 79.19 ( 14.48 zinc
64.67 ( 11.61
6.32 7.57
0.01 0.02
8.69 7.81
0.21 0.18 0.27
4.04 2.55 3.97 ( 1.65
0.13 ( 0.04
4.40 ( 0.69
0.21 ( 0.04 0.15 ( 0.04
6.71 ( 1.48
0.40 ( 0.03
7.59 ( 0.72 6.06 ( 0.76
0.24 ( 0.05
13.25 ( 0.66 nickel
selenium
0.30 ( 0.08
3.17
7.43 259.35 5.82 207.38 9.18 297.20 6.56 271.54 5.46 190.73 0.45 107.61 5.45 ( 3.64 49.71 ( 14.45 12.76 ( 5.99 50.41 ( 21.37 10.82 ( 3.84 56.69 ( 13.94 8.99 ( 3.03 89.04 ( 15.86 28.01 ( 1.05 244.09 ( 11.06 copper manganese
14.18 ( 3.30 84.48 ( 12.01
0.01 0.03
2.74 2.88
0.12 0.02
3.54 2.66 1.64
0.04 0.01 0.02 ( 0.18
44.79 ( 4.76
0.02 ( 0.03
41.99 ( 12.26 33.90 ( 18.71
0.27 ( 0.16 0.12 ( 0.17 0.01 ( 0.01
33.92 ( 5.11
0.20 ( 0.12
85.72 ( 15.59
cadmium
chromium
53.96 ( 22.14
0.00
0.72 0.48 0.62 0.48 0.12 0.03 6.03 ( 4.37 5.50 ( 1.78 12.82 ( 3.06 6.52 ( 1.14 14.53 ( 1.32 lead
10.01 ( 3.09
0.40 0.40
0.18 0.25
0.39 0.39
0.26 0.32
0.37 0.30
0.04 3.60 ( 1.52
0.26 ( 0.13 0.15 ( 0.05
4.41 ( 1.45 4.32 ( 1.36
0.12 ( 0.02 0.19 ( 0.02
1.93 ( 0.60
5.75 ( 1.15
0.16 ( 0.02 0.88 ( 0.10
5.37 ( 0.54
silver
arsenic
Gc 25 cm Gc 0 cm Bp 130 cm Bp 105 cm Bp 75 cm Bp 50 cm Bp 25 cm Bp 0 cm
hydroxyl-carbonates during pellet curing;13 carbonate/hydroxide buffer limit of pH 10.2. Similar results (including pH spike causes), reported by Despland et al.15 using microcosm studies, show the pH spike was slightly greater and takes longer to decline in the current study, presumably because of column scale-up, i.e. larger column and pellet volumes (10 L vs 300 mL in the small-scale). A freshwater rinse of the pellets pretreatment and/ or a pH correction post-treatment may prevent pH spike in circum-neutral waters. Similarly an EC spike recorded after the first bed volume (≈11 h) in the Bauxsol columns’ outlets (14000 μS) decreased to 965 μS after 15 bed volumes (≈1 week) and then mirrored the Gravel column outlets and inlets EC (≈700 μS) (data not shown). The observed EC spike correlated with the calcium spike in Bauxsol columns’ outlet samples (1730 mg/L) is attributed to the release of unreacted CaO (cement binder) in the effluent, i.e. a rise of calcium greatly increases EC.20 However, a small spike of alkalinity as OH (9 mg/L) and as CO32 (80 mg/L) found in the Bauxsol column outlets during the first 30 bed volumes (≈2 weeks) suggests that inlet waters are poorly buffered, and small changes in concentration have large effects in physicochemical measure such as pH and EC. During this study, water temperature ranged from 24 to 29 °C, while dissolved oxygen (DO) ranged from 0.5 to 3 mg/L at the inlets and from 1.5 to 3 mg/L at column outlets. This disparity in DO is partially attributed to the presence of trapped air in sampling taps during the initial stages, although for Bauxsol pellet columns DO concentration differences may reflect the high porosity of pellets. Moreover, greater porosity and surface area to volume ratio of Bauxsol pellets may have provided additional anoxic microzones within the columns; similar effects were also found in a small scale column experiment.15 Trace-Metals Bound to Solids. Although trace-metal concentrations in sewage effluent are typically low, an assessment of their accumulation on the substrates is important. Bauxsol pellets and Gravel naturally contain trace-metals. The metal concentrations in the Bauxsol pellets reflect the average composition of Bauxsol powder for As, Pb, Cd, Al and Ni.9 However, Cu, Mn, and Zn are higher, whereas Cr and Fe are lower in pellet formulations. These differences are from the use of OPC-cement in pellet production, where Cu, Mn, and Zn are enhanced by OPC either as contaminants or changing surface chemistries brought about by a rise in pH, whereas Cr and Fe concentrations are diluted by the cement mass in the pellets.2123 Trace-metal scans from Bauxsol and Gravel columns solids at 0 (inlet), 25, 50, 75, 105, and 130 (outlet) cm after 6 months of
metal [mg/kg]
Table 1. Bound Trace-Metal Content of Mesocosm Substrates after 6 Months Accumulation As Measured by Aqua Regia Digesta
Figure 1. pH of column inlet and outlet waters samples (Bauxsol pellets and Gravel).
Gc 50 cm
Gc 75 cm
Gc 105 cm
Gc 130 cm
ARTICLE
dx.doi.org/10.1021/es200934y |Environ. Sci. Technol. 2011, 45, 5746–5753
Environmental Science & Technology operation generally showed increases in metal concentrations on Bauxsol pellets but a decline for Gravel samples (Table 1). Gravels lost approximately 76% of the initial concentrations of Ag; 2023% of initial Cu, Mn, Fe, and Al; and 1317% of the initial Pb, Cd, Cr, Ni, Se, and Zn. These losses are most likely caused by chemical weathering of minerals surfaces and loss of weakly bound trace-metals, especially from reactive olivine surfaces. Increased metal binding capacity of Bauxsol pellets reflects the high surface to volume ratio and high charge to mass ratio of Bauxsol.21 Compared to findings by Clark et al.22 and by Genc-Fuhrman et al.6,24 trace-metal binding capacity of the Bauxsol pellets is quite high even after 6 months of field trials. The results indicate that Ag, Cr, Cu, Mn, Ni, Se, Fe, and Al are preferentially bound to Bauxsol in the first quarter of the columns, whereas As, Pb, Cd, Zn, and Hg were found more evenly distributed along the column length. This may be explained by the geochemical structure of Bauxsol pellets and the competition of metals for available sites. Metals are initially bound to the Bauxsol by surface charge attracting dissolved metals and then by crystal growth via new mineral precipitation and intracrystalline diffusion.8,11,21,22 Recent synchrotron data
Figure 2. Percentage removal of orthophosphate in the columns’ outlets (Bauxsol pellets and Gravel).
ARTICLE
(Dr. Richard Collins, personal communication, 2011) and isotopic exchange experiments 25 supports these observations. Moreover, different metal removal mechanisms may be at play as the composition of the effluent changes overtime. Furthermore, Genc- -Fuhrman et al. 6 reported that trace-metal speciation is pH-dependent and concluded that the metal removal processes changes with time. Clark et al. 22 and McConchie et al. 9 also showed that bound metal leaching from Bauxsol decreases as loaded samples are allowed to age. Consequently, it is unlikely that many of the trace-metals accumulated on Bauxsol pellets will leach under natural conditions if applied to soils. Several studies (e.g. refs 22, 23, and 26) demonstrate the long-term fate of bound trace-metals on the Bauxsol matrix. Phosphate Removal. Phosphate concentrations in the inlet ranged from 3.0 to 9.2 mg/L. Bauxsol pellet columns effectively eliminated phosphate from the effluent. During the first 60 bed volumes (≈1 month), >95% of the orthophosphate was removed and subsequently stabilized at around 80% for the next 240 bed volumes (≈5 months), before falling to 60% for the remaining 60 bed volumes (≈6 months; Figure 2). These results are below the Australian concentrations for tertiary treated effluent (postchemical precipitation).27 No significant phosphate removal was observed in the Gravel column; phosphate binding was transitory and eventually leached away. Analyses of Bauxsol pellets following 6 months exposure showed substantially higher TP binding and Colwell P (available phosphate) than the Gravel. Most of the bound TP was observed on the first quarter of the Bauxsol columns (7.3 g/kg) with the remaining three-quarters binding 3.3 g/kg, gradually declining to 1.5 g/kg at the outlet (Figure 3). Phosphate was bound evenly along the first half of the columns (≈2 g/kg) and then dropped to 1.6 g/kg in the second half (Figure 3). Contrary to this, Gravel samples showed little phosphate binding and even signs of TP loss (Figure 3). Observed differences in phosphate binding capacity between Gravel and Bauxsol pellets reflects that Bauxsol has a recorded phosphate binding capacity of ≈2% by mass.14 According to Akhurst et al.10 the adsorption of PO43- by Bauxsol increases with decreasing pH (i.e., change in dominant phosphate speciation,
Figure 3. Bound total phosphorus and Colwell phosphorus on Gravel and Bauxsol pellets after 6 months experiment along the length of the columns (0 cm = inlet, 130 cm = outlet); ( standard error for Bauxsol pellet. 5749
dx.doi.org/10.1021/es200934y |Environ. Sci. Technol. 2011, 45, 5746–5753
Environmental Science & Technology
ARTICLE
Table 2. a) NH4þ, NO2, NO3, and TN Concentration [mg/L] in Inlets, Bauxsol Pellets (Bp) Columns Outlets, Gravel (Gc) Column’s Outlets at 1, 60, 180, and 360 Bed Volumes and b) NH4þ and NO3 Concentration [mg/kg] on Solids before Experiment (Raw) and Bound on Substrates (Bp = Bauxsol Pellets; Gc = Gravel) after 6 Months Experiment at Different Length along the Columns (0 cm = inlet; 130 cm = outlet); ( Standard Error for Bp A) Liquids sample time 1 bed volume (11 h) samples port
inlet
Bp outlet
ammonia
13.04 0.24
nitrite
60 bed volume (1 mth)
Gc outlet
inlet
14.88 ( 0.97
2.16
1.02 ( 0.09
0.64
150 bed volumes (2.5 mths)
Bp outlet
Gc outlet
inlet
10.93
9.32 ( 1.11
13.05
0.44
1.13 ( 0.10
0.80
360 bed volume (6 mths)
Bp outlet
Gc outlet
inlet
Bp outlet
Gc outlet
13.70
6.77 ( 0.13
5.00
1.17
0.04 ( 0.00
0.58
0.25
1.21 ( 0.11
0.06
0.12
1.20 ( 0.11
0.58
nitrate
17.56
9.66 ( 1.98
13.85
12.47
7.45 ( 0.75
6.10
20.90
20.23 ( 0.37
10.80
12.03
9.44 ( 0.44
8.71
total nitrogen
35.10
30.60 ( 0.10
21.60
27.30
18.60 ( 0.17
20.10
37.80
30.33 ( 0.73
17.70
16.56
11.22 ( 0.49
9.95
b) Solids sample time before experiment sample type ammonia nitrate
raw Bp
raw Gc
bound after experiment Bp 0 cm
Bp 75 cm
Bp 130 cm
Gc 0 cm
Gc 75 cm
Gc 130 cm
1.38
0.19
6.29 ( 0.55
6.00 ( 0.77
4.58 ( 0.71
3.89
2.50
2.16
14.38
0.79
17.27 ( 2.21
43.92 ( 9.27
61.79 ( 10.21
3.29
2.06
2.22
which facilitate the electrostatic and chemical attraction of phosphate ions). In the present study, phosphate speciation for inlet water is predominantly H2PO4 at pH 6.37, but as the effluent passes through the Bauxsol column the speciation changes to HPO42- with minor amounts of H2PO4 at pH of 810.20 Accordingly, this should result in reduced phosphate removals at the beginning of this experiment. However, excess Ca2þ and Mg2þ ions (i.e., Bauxsol pellets curing process) present in our system precipitates the HPO42- to form MgHPO4 and CaHPO4. Additionally, it is thought that the HPO42- and NH4þ may also be precipitated as struvite (MgNH4PO4);28 however, confirmation by X-ray diffraction (XRD) is difficult because of the dominance of the Bauxsol mineralogy. Based on present and past geochemical studies,11,14,22 it appears that the main phosphate removal mechanism proceeds via precipitation of PO43- with Ca and Mg ions; therefore, back flushing of the material is not possible. Bauxsol pellets also demonstrated equivalent or greater P-removal rates compared to other materials under similar experimental conditions (i.e., bed volumes, pH, and initial phosphate concentration). Studies on alum showed between 60 and 80% P-removal.4,5 Wollastonite used by Brooks et al.29 and zeolite used by Sakadevan and Bavor30 demonstrated P-removal around 50%. Laterite used by Wood and McAtamney31 in a batch system and a constructed wetland showed a P-removal at 85% and 95%, respectively. Gray et al.32 found in their pilot-scale constructedwetland using Maerl material (dead deposits of calcareous red algae) 98% P-removal. Blast furnace slag mixed with sand in the Johansson study33 revealed similar results to Bauxsol pellets. Comparing to Akhurst et al. study,10 Bauxsol pellets are superior to Bauxsol powder (10 mg/L) interfere with struvite formation28 and phosphate was most likely bound as MgHPO4 and CaHPO4. N-assay experiments revealed that denitrification was active in both Gravel and Bauxsol pellets incubations but higher in Bauxsol pellets incubations; after 24 h the production of 15 NN2 was negligible, so all rate calculations were made for the first 24 h (Figure 4). This difference may be explained by the greater porosity and surface to volume ratio of the Bauxsol pellets that could have provided additional anoxic microzones within the column. Denitrification of the added NO3 in the Bauxsol pellets incubation accounted for up to 1.5% without and 5.5% with MeOH in the incubations. The addition of MeOH most likely provided a source of labile dissolved organic carbon for denitrifiers and also led to the consumption of any available O2, further increasing the availability of anoxic microzones. The increased denitrification from added MeOH may provide an improvement to wastewater treatment using Bauxsol pellets. The presence of 29N2 detected in all the 15NH4 incubations points to the presence of nitrification producing 15NO3 that has mixed with 14NO3 from the nitrification of internally mineralized of 14NH4þ. In the incubations amended with 15NH4þ þ 14 NO3, nitrification will also have produced 15NO3. However, this would have mixed with a pool of available 14NO3 and diluted the 29N2 signal produced via denitrification. Hence, the presence of nitrification indicates a small oxygen contamination in the incubations. Although the incubations were carried out in an anoxic glovebox, only a very small concentration (5 μM O2) is required to facilitate nitrification. Interestingly, the presence of 29N2 increased in the Bauxsol pellets incubations amended with MeOH. Consequently, the MeOH consumed available O2
leading to nitrification under anoxic conditions. Anoxic nitrification can occur in the presence of electron acceptors other than O2 such as manganese oxides37 and iron oxides.38 Metal scans (Table 1) show high levels of Mn bound to the Bauxsol pellets. Moreover recent synchrotron work showed that Mn binds to the hematite and sodalite minerals in the Bauxsol and a partial oxidation consistent with Mn(IV) occurs (Dr. Richard Collins, personal communication, 2011). Consequently, this Mn(IV) would be a suitable electron acceptor as well as the iron-rich Bauxsol pellets.9 In addition, high nitrates loadings on Bauxsol pellets in the last quarter of the columns (i.e., nitrification is proceeding along the length of the column) also seems to support this hypothesis, although further investigations are necessary. The presence of anammox was detected only in Bauxsol pellet incubations with MeOH (Figure 4). This is counterintuitive given that MeOH is a known inhibitor of anammox.39 However, if it is considered that MeOH resulted in the consumption of available O2 at the surface of the Bauxsol pellets, then it is possible that anoxia may have been prevalent in microzones within the pellets, and, consequently, these conditions may have favored anammox. The cycling of N is clearly complex within the Bauxsol pellet columns and further investigations are also warranted, in particular the use of molecular biology to isolate different bacterial processes and pathways. This study demonstrates that OPC-bound Bauxsol pellets used to filter municipal secondary treated wastewater successfully removed effluent nutrients and trace-metals. The short-lived pH spike resulting from the release of unreacted CaO from the cement binder used in pellet production should be monitored and corrected post-treatment during the initial stages of column establishment. Bauxsol pellets showed a large capacity to bind phosphate during the course of the 6 months trial and were yet to reach saturation; however, further studies are required to determine pellet saturations. Furthermore, experimental results indicated that a complex nitrogen cycle may be functioning within the Bauxsol pellets columns, i.e. anoxic nitrification, denitrification, and anammox processes associated with nutrient binding. Further chemical and biological investigations (i.e., molecular biology) are required to elucidate and confirm these processes. Spent Bauxsol pellets (waste from columns), providing supporting data from further investigations, may be suitable for use as fertilizer because they contain high concentrations of plant available phosphate (≈2 g/kg) but only small loadings of 5751
dx.doi.org/10.1021/es200934y |Environ. Sci. Technol. 2011, 45, 5746–5753
Environmental Science & Technology trace-metals, which under normal environmental conditions (pH 58) appear irreversibly bound.
’ AUTHOR INFORMATION Corresponding Author
*Phone: þ612 6620 3650. Fax: þ612 6621 2669. E-mail: laure.
[email protected].
’ ACKNOWLEDGMENT The paper draws inspiration and builds on the previous work of the late Professor David McConchie, Southern Cross University, who saw potential where others saw none. This work was also supported by an Australian Research Council Grant LP0056012 and the Stipend provided to Dr. Clark from CRCCARE. ’ REFERENCES (1) Hammer, M. J.; Hammer, M. J., Jr. Water and Wastewater Technology, 5th ed.; Prentice Hall: Singapore, 2005. (2) Mulder, A.; Vandegraaf, A. A.; Robertson, L. A.; Kuenen, J. G. Anaerobic ammonium oxidation discovered in a denitrifying fluidizedbed reactor. FEMS Microbiol. Ecol. 1995, 16 (3), 177–183. (3) Haghseresht, F.; Wang, S. B.; Do, D. D. A novel lanthanummodified bentonite, Phoslock, for phosphate removal from wastewaters. Appl. Clay Sci. 2009, 46 (4), 369–375. (4) Galarneau, E.; Gehr, R. Phosphorus removal from wastewaters: Experimental and theoretical support for alternative mechanisms. Water Res. 1997, 31 (2), 328–338. (5) Omoike, A. I.; Vanloon, G. W. Removal of phosphorus and organic matter removal by alum during wastewater treatment. Water Res. 1999, 33 (17), 3617–3627. (6) Genc-Fuhrman, H.; Mikkelsen, P. S.; Ledin, A. Simultaneous removal of As, Cd, Cr, Cu, Ni and Zn from stormwater: Experimental comparison of 11 different sorbents. Water Res. 2007, 41 (3), 591–602. (7) Lapointe, F.; Fytas, K.; McConchie, D. Efficiency of Bauxsol in permeable reactive barriers to treat acid rock drainage. Mine Water Environ. 2006, 25 (1), 37–44. (8) Clark, M. W.; McConchie, D.; Munro, L.; Faux, D.; Walsh, S.; Blair, D.; Fergusson, L. The development of porous pellets of seawaterneutralised bauxite refinery residues(red mud) and bauxsol for use in water treatment systems, In Proceedings of the 7th International Alumina Quality Workshop, Alumina Quality Workshop: Perth, Western Australia, 2005; pp 5658. (9) McConchie, D.; Clark, M.; Hanahan, C.; Fawkes, R. The use of seawater-neutralised bauxite refinery residues(red mud) in environmental remediation programs., In Proceedings of the Rewas’99 Global Symposium on Recycling, Waste Treatment,and Clean Technology, San Sebastion, Spain, 1999; Gaballah, I., Hager, J., Solozabal, R., Eds. ;The Minerals, Metals, and Materials Society: Warrendale, PA; pp 391400. (10) Akhurst, D. J.; Jones, G. B.; Clark, M.; McConchie, D. Phosphate removal from aqueous solutions using neutralised bauxite refinery residues(Bauxsol(TM)). Environ. Chem. 2006, 3 (1), 65–74. (11) Hanahan, C.; McConchie, D.; Pohl, J.; Creelman, R.; Clark, M.; Stocksiek, C. Chemistry of seawater neutralization of bauxite refinery residues(red mud). Environ. Eng. Sci. 2004, 21 (2), 125–138. (12) Clark, M. W.; McConchie, D. M.; Johnston, M., Bauxite refinery residues, potential for waste utilisation. In Waste Management: Research, Technology and Developments; Lavelle, J. R., Ed.; Nova Science Publishers: New York, 2008. (13) Despland, L. M.; Clark, M. W.; Aragno, M.; Vancov, T. Minimising alkalinity and pH spikes from Portland cement-bound Bauxsol(seawater-neutralized red mud) pellets for pH circum-neutral waters. Environ. Sci. Technol. 2010, 44 (6), 2119–2125.
ARTICLE
(14) Clark, M.; Gnanam, F.; McConchie, D.; Fergusson, L. The treatment of wastewaters with high phosphorous and high ammonia loadings using modified bauxite refinery residues. Chin. J. Geochem. 2006, 25 (Suppl), 141. (15) Despland, L. M.; Vancov, T.; Clark, M. W.; Aragno, M., Bauxsol pellets: bacterial communities and phosphate removal in wastewater treatment, In Proceedings of the IWA World Water Congress & Exhibition, 1924 September 2010, Montreal, Canada, 2010; International Water Association: Montreal, Canada, 2010. (16) American Public Health Association. Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association, American Water Works Association and Water Environment Federation: Washington DC, 1998; p 356. (17) USGS Alkalinity Calculator Web site. http://or.water.usgs. gov/alk/ (accessed April 23, 2009). (18) Rayment, G. E.; Higginson, F. R. Australian laboratory handbook of soil and water chemical methods; Inkata Press: Port Melbourne, 1992; p 330. (19) Nielsen, L. P. Denitrification in sediment determined from nitrogen isotope pairing. FEMS Microbiol. Ecol. 1992, 86 (4), 357–362. (20) Stumm, W.; Morgan, J. J. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed.; John Wiley & Sons: New York, 1996. (21) McConchie, D.; Clark, M.; Hanahan, C.; Davies-McConchie, F. The use of seawater-neutralised bauxite refinery residues in the management of acid sulphate soils, sulphidic mine tailings and acid mine drainage, In Sustainable environmental solutions for industry and government: proceedings of the 3rd Queensland Environmental Conference; Gaul, K., Ed.; Environmental Engineering Society (Queensland Chapter): Brisbane, Australia; pp 201208. (22) Clark, M.; Berry, J.; McConchie, D. The long term stability of metal-laden Bauxsol reagent under different geochemical conditions. Geochem. Explor. Environ. Anal. 2009, 9 (1), 101–112. (23) Munro, L.; Clark, M.; McConchie, D.; Bauxsol, A (TM)-based permeable reactive barrier for the treatment of acid rock drainage. Mine Water Environ. 2004, 23 (4), 183. (24) Genc-Fuhrman, H.; Tjell, J. C.; McConchie, D. Increasing the arsenate adsorption capacity of neutralized red mud(Bauxsol). J. Colloid Interface Sci. 2004, 271 (2), 313–320. (25) Clark, M. W.; Harrison, J. J.; Payne, T. E. The reversibility of Uranium and Thorium binding to modified Bauxite refinery residue. J. Colloid Interface Sci. 2011, 356 (2), 699–705. (26) Clark, M. W.; McConchie, D. M.; Munro, L.; Us Samed, A. J. F.; Best, D., Bauxsol based barriers for the treatment of metal contaminated ground waters. In Environmental geotechnics: opportunities, challenges and responsiblities for environmental geotechnics: Proceedings of the ISSMGE’s Fifth International Congress on Environmental Geotechnics, Cardiff, Wales, 2630th June, 2006. Vol. 1; Thomas, H. R., Ed.; Thomas Telford Publishers: London; 2006; pp 110117. (27) Australian Guidelines for Sewerage Systems, Effluent Management; National Water Quality Management Strategy 11; Agriculture and Resource Management Council of Australia and New Zealand; Australian and New Zealand Environment and Conservation Council: Canberra, Australia, 1997. (28) Babic-Ivancic, V.; Kontrec, J.; Kralj, D.; Brecevic, L. Precipitation diagrams of struvite and dissolution kinetics of different struvite morphologies. Croat. Chem. Acta 2002, 75 (1), 89–106. (29) Brooks, A. S.; Rozenwald, M. N.; Geohring, L. D.; Lion, L. W.; Steenhuis, T. S. Phosphorus removal by wollastonite: A constructed wetland substrate. Ecol. Eng. 2000, 15 (12), 121–132. (30) Sakadevan, K.; Bavor, H. J. Phosphate adsorption characteristics of soils, slags and zeolite to be used as substrates in constructed wetland systems. Water Res. 1998, 32 (2), 393–399. (31) Wood, R. B.; McAtamney, C. F. Constructed wetlands for waste water treatment: The use of laterite in the bed medium in phosphorus and heavy metal removal. Hydrobiologia 1996, 340 (13), 323–331. (32) Gray, S.; Kinross, J.; Read, P.; Marland, A. The nutrient assimilative capacity of maerl as a substrate in constructed wetland systems for waste treatment. Water Res. 2000, 34 (8), 2183–2190. 5752
dx.doi.org/10.1021/es200934y |Environ. Sci. Technol. 2011, 45, 5746–5753
Environmental Science & Technology
ARTICLE
(33) Johansson, L. Blast furnace slag as phosphorus sorbents - column studies. Sci. Total Environ. 1999, 229 (12), 89–97. (34) Peverill, K.; Sparrow, L.; Reuter, D. Soil analysis: an interpretation manual; CSIRO: Melbourne, Australia, 1999. (35) Yan, Y.; Renping, W.; Clark, M. W. Phosphate removal by hydrothermal modified fumed silica and pulverised oyster shell. J. Colloid Interface Sci. 2010, 350 (2), 538–543. (36) Barbhuiya, S. A.; Basheer, P. A. M.; Clark, M. W.; Rankin, G. I. B.; Frost, P. An investigation into the behaviour of concrete containing seawater-neutralised bauxite refinery residues in silage effluent. Biosystems Eng. 2010, 106 (4), 433–439. (37) Hulth, S.; Aller, R. C.; Gilbert, F. Coupled anoxic nitrification manganese reduction in marine sediments. Geochim. Cosmochim. Acta 1999, 63 (1), 49–66. (38) Straub, K. L.; Benz, M.; Schink, B. Iron metabolism in anoxic environments at near neutral pH. FEMS Microbiol. Ecol. 2001, 34 (3), 181–186. (39) Jensen, M. M.; Thamdrup, B.; Dalsgaard, T. Effects of specific inhibitors on anarnmox and denitrification in marine sedirnents. Appl. Environ. Microbiol. 2007, 73 (10), 3151–3158.
5753
dx.doi.org/10.1021/es200934y |Environ. Sci. Technol. 2011, 45, 5746–5753
