- Email: [email protected]
Laboratory investigation of the phosphorus removal (SRP and TP) from eutrophic lake water treated with aluminium
ARTICLE IN PRESS WAT E R R E S E A R C H
40 (2006) 2713 – 2719
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/watres
Laboratory investigation of the phosphorus removal (SRP and TP) from eutrophic lake water treated with aluminium Franck Auvraya, Eric D. van Hullebuscha,b, Veronique Deluchata, Michel Baudua, a
Laboratoire des Sciences de l’Eau et de l’Environnement, Faculte´ des Sciences et Techniques, 123, Avenue Albert Thomas, F-87060 Limoges, Cedex, France b Laboratoire des Ge´omate´riaux et Ge´ologie de l’Inge´nieur, Universite´ de Marne la Valle´e, 5, Boulevard Descartes – Champs sur Marne, 77454 Marne La Valle´e, Cedex 2, France
art i cle info
A B S T R A C T
Article history:
Mechanisms involved in phosphorus (P) removal from eutrophic lake water with
Received 31 March 2004
aluminium (Al) were assessed by jar tests. For this purpose, eutrophic lake water enriched
Received in revised form
with soluble reactive phosphorus (SRP), algae or sediments in order to mimic the various
26 April 2006
conditions found in shallow eutrophic lakes was studied. Total phosphorus (TP) removal
Accepted 27 April 2006
was reached after floc settling, the maximal TP removal efficiency (90–95%) was obtained
Available online 30 June 2006
for an Al concentration ranging from 2 to 5 mg L1, depending on the organic matter (OM)
Keywords:
origin (algae or sediments). Algae appeared to limit macro-floc formation (those able to
Eutrophic lake water
settle). In contrast, in the presence of sediments, macro-floc formation was favoured at low
Aluminium
Al dose (2–3 mg L1 Al). High SRP removal was obtained with the lowest Al dose (1 mg L1 Al)
Phosphorus removal
and remained greater than 60% for an SRP concentration up to 350 mg L1. SRP removal was
Flocculation
not influenced by the OM origin. The experimental data and literature were used to suggest a hypothetical model for floc formation and P removal with Al under the conditions observed in treated lakes. & 2006 Elsevier Ltd. All rights reserved.
1.
Introduction
The over enrichment of surface water with nutrients (phosphorus (P) and nitrogen) is the major cause of eutrophication (Correll, 1998). As high P concentrations are often associated with cyanobacterial blooms, this nutrient appears to play an important role in the phytoplankton dynamics. The dominance of cyanobacteria in summer results in high turbidity as well as possible cyanotoxins production, limiting uses of freshwater lakes (i.e., recreational activities, reservoir for drinking water production). There is no universal link between the P concentration and cyanobacterial dominance, but high spring bioavailability of P and high summer P release by sediments are the main factors responsible for the occurrence of cyanobacterial blooms in shallow lakes (Reeders et al., 1998; Krivtsov et al., 2001). Corresponding author. Tel.: +33 5 55 45 73 67; fax: +33 5 55 45 72 03.
E-mail address: [email protected] (M. Baudu). 0043-1354/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2006.04.042
Consequently, P removal from the water column and immobilisation in the sediment with aluminium (Al) has been used to a large extent. Al is chosen due to its ability to complex P and its insensitivity to redox potential fluctuations in sediments (Burley et al., 2001). The conventional aim of Al treatment is to increase the sediment P retention capacity. Generally, the recommended Al dose ranges from 5 to 30 mg L1 (Rydin and Welch, 1998 ; Reitzel et al., 2003). However, recent studies (van Hullebusch et al., 2002a; van Hullebusch et al., 2003) have shown that lower doses (i.e., 1.5 mg L1 as Al) could both decrease the soluble reactive phosphorus (SRP) concentration in the water column and introduce freshly formed Al hydroxides onto superficial sediments (with newly settled particles) which enhance P retention, thus reducing the summer algae biomass. Moreover, Kopacek et al. (2001) have shown that a high Al content
ARTICLE IN PRESS 2714
WAT E R R E S E A R C H
40 (2006) 2713– 2719
in the in-flow, following acid rain, results in Al hydroxide formation in the lake water column when the pH is close to neutral. This phenomenon was shown to enhance the phosphorus sorption capacity of sediments subsequent to Al hydroxide floc settling. However, all of these studies only take into consideration the results of phosphorus depletion with no attention paid to mechanisms involved in floc formation and/or SRP sorption. Moreover, many studies on drinking water treatment (Galarneau and Gehr, 1997; Gregor et al. 1997; Kim et al., 2001), have underlined the major role played by organic matter (OM) flocs. Floc structure is one of the dominant factors that influences its behaviour, physical (i.e. sedimentation) and/or chemical (i.e. sorption) properties in natural systems. However, physicochemical conditions in water treatment plants are very different in contrast to those occurring in eutrophic lakes. In these aquatic systems, the organic P pool is mainly constituted by dissolved organic species, algae and sediment organic particles present in the water column following re-suspension events (Pettersson, 2001; Selig et al., 2002). The relative contribution of sediment and algae to the P pool may vary a lot depending on the season and weather (wind, storms, etc.) which, in the end, can enhance the growth of algae or favour sediment resuspension. Hejzlar et al. (1998) have shown that algae greatly influence water treatability by coagulation, but the impact of sediments on coagulation, flocculation and sedimentation processes are less well known. The aim of this study is to improve understanding of P removal from eutrophic lake water treated with Al salts. As this should be closely related to floc formation and structure, which are greatly influenced by OM content and origin, these factors must also be considered. For this purpose, laboratory experiments were carried out using jar tests and natural water from a eutrophic lake. Firstly, mechanisms involved in total phosphorus (TP) removal from water containing mostly particulate OM from two origins (sediments or algae) were established. Secondly, SRP removal at different OM matter concentrations was monitored.
2.
Materials and experimental methods
2.1.
Water and organic matter sources
Water and superficial sediments were taken from a shallow polymictic eutrophic lake (Lake Courtille, France). This lake was previously treated with Al to avoid cyanobacterial blooms (van Hullebusch et al., 2002a,b). Lake water and sediments (only the upper 2 cm of sediment) were collected in May 2002 under oxic conditions and the samples were stocked in acidcleaned polyethylene flasks, stored at 41C and used within 48h after collection. Sediments were wet sieved (63 mm), stirred vigorously with lake water for 1 h followed by 2 h of undisturbed settling. Supernatants were collected to form the sediment suspension. The sediment Al and iron content were respectively 10.5 and 27 mg g1 of dry weight (more details regarding the sediment analytical data and methods can be found in van Hullebusch et al., 2003). van Hullebusch et al. (2003) showed
that treatment with Al didn’t significantly increase the total Al concentration in sediment. Scenedesmus subspicatus (276/22 CCAP strain purchased from Collection and Culture of Algae and Protozoa collection, UK) were grown axenically in Chu-10 growth medium (Chu, 1942) at 25 1C with continuous stirring. Cells in the exponential phase were harvested by centrifugation (5 min, 1500g) and rinsed twice with distilled water. Optical microscopy was used to verify cell integrity and a mean size, around 872 mm (n ¼ 100), was measured. Algae were then mixed with lake water to form algae-rich water samples. As sediments or algae are able to adsorb or release SRP, adjustments were required to obtain the desired value of SRP in water samples enriched with suspensions. Preliminary kinetic studies showed that after 1 h of equilibration (under slow stirring) no significant release or adsorption occurred within 10 h (data not shown). Therefore, before each experiment, all samples enriched with algal or sediment suspension were allowed to equilibrate for 1 h after SRP adjustment with SRP stock solution (KH2PO4, 5 mg L1 P); the initial SRP concentration was measured after equilibration. In this way, SRP removal measured during experiments would only be the result of Al additions and not a contribution of natural Al or ferrous hydroxides initially present in the suspension.
2.2.
Jar-test procedure
Coagulation and flocculation experiments were carried out with a multiple paddle stirrer at 25 1C. The samples were mixed for 5 min at 100 round per minute (rpm) followed by 30 min at 20 rpm. Aliquots of Al stock solutions (Al2(SO4)3, 16 H2O; 1 g L1 Al) were added to water samples during the first mixing phase to give the required concentrations. The pH was adjusted to 6.5 7 0.1 by addition of NaOH or H2SO4. Samples were then allowed to settle for 45 min, as no change was observed over longer periods. pH values were checked and supernatants were sampled for TP, SRP, OM and Al content determination. The sediment suspension contained natural Al and Fe which increased their dissolved concentrations (respectively up to 1 and 3 mg L1) following sediment suspension enrichment. As these metals are known to have coagulant and sorption properties, control jar tests (0 mg L1 Al) were always conducted to assess the impact of the stirring procedures on sedimentation and SRP removal. In each case, without Al treatment, no sedimentation (of algae or sediment) was observed. Moreover, no SRP removal was observed in supernatants indicating that SRP sorption equilibrium onto natural Fe and Al oxides was probably reached. The increase in dissolved Al and Fe concentrations following sediment enrichment did not therefore contribute to the SRP removal observed in jar tests treated with Al.
2.3.
Total phosphorus removal
TP removal was studied in lake water enriched with sediment or algal suspensions; additions were made to mimic realistic environmental OM concentrations present in shallow eutrophic lakes (van Hullebusch et al., 2002 a, b, 2003). The water characteristics resulting from enrichments are summarised in Table 1. A lot of studies (Huang and Shiu, 1996; Ratnaweera
ARTICLE IN PRESS
et al., 1999) have pointed out the predominant role of the OM level in the coagulation–flocculation process, conditions with similar OM content were therefore chosen. As algae have higher P content than sediments, P concentrations in water enriched with algae are higher. Jar-test procedures were carried out with Al in the range of 1–6 mg L1 Al.
2.4.
Soluble reactive phosphorus removal
SRP removal was studied using lake water enriched with P from a stock solution (KH2PO4, 5 mg L1 P). Final SRP concentrations were in the range of 20–320 mg L1 which largely cover SRP concentrations occurring in Lake Courtille. In another experimental set-up, lake water enriched with different amounts of sediment or algal suspensions to give a final range of OM from 6.2 to 9.5 mg L1 total organic carbon (TOC) was used. SRP concentrations were adjusted to 100720 mg L1 P. Jar-test procedures were carried out with 1, 2 and 6 mg L1 Al.
2.5.
2715
4 0 (200 6) 271 3 – 271 9
Analytical methods
SRP was determined in filtered samples by the phosphomolybdate method (AFNOR procedure NFT 90-023). TP was converted to SRP in crude samples after digestion at 105 1C with sulphuric acid and potassium persulphate. SRP is often assumed to represent the bioavailable P fraction for algae in the water column. The detection limit of this method was 5 mg L1. All P analyses were performed in duplicate with a standard deviation of 5%. OM concentration was determined by wet oxidation with persulphate and UV radiation using a TOC Phoenix 8000 analyser (Dohrman). The standard deviation and detection limit were, respectively, 8% and 0.5 mg L1 TOC. Al determination was done by atomic absorption spectrophotometry using a graphite furnace with a Zeeman effect (Varian GTA 100). Total Al (AlTOT) was determined after digestion with nitric acid. The detection limit and standard deviation were, respectively, 10 mg L1 and 7%.
the range of 1–6 mg L1, and SRP concentrations were adjusted to 40720 mg L1 P. Water compositions are reported in Table 1 and results obtained are plotted in Fig. 1. Al addition induces floc formation and sedimentation that results in a TP decrease in the supernatants. Removal curves show conventional trends for jar-test curves: a first lag phase where removal increases with coagulant addition, then a second one where removal was maximal. However, in the first lag phase (below 4 mg L1 Al), two different trends are obtained with the different OM origins: the increase in OM originating from the sediment suspension resulted in an improvement in TP removal, whereas the addition of algae greatly decreased the Al efficiency (particularly at doses lower than 3 mg L1 Al). Previous studies (Galarneau and Gehr, 1997) have shown that TP removal depends basically on the ability of Al to settle OM associated with P. Al associated with OM induces the formation of some flocs, that have size, charge, density and structure inconsistencies which determine their ability to settle. A simple way to characterise them was proposed by Hereit et al. (1980) and used by Gregor et al. (1997). Macro-flocs are operationally defined as flocs settled within a given time (45 min) and micro-flocs as flocs remaining in suspension after this time. In our experiments, no measurable sedimentation was noticed beyond this time. If this nomenclature is used due to its simplicity, it should not be forgotten that the denomination of macro- and micro-flocs is not directly associated with their size but with their ability to settle under given conditions. The results presented in Fig. 1 indicate that the sediment enhanced macro-floc formation, whereas algae stabilised micro-flocs.
100 TP removal (%)
WAT E R R E S E A R C H
80 Control
60
Water 1S 40
Water 2S Water 1A
20
3.
Results and discussion
3.1.
Total phosphorus removal
Jar-test procedures were applied to lake water enriched with two organic P sources (algae and sediments), Al was dosed in
Water 2A
0 0
1
2 3 4 Aluminium (mg.L-1 Al)
5
6
Fig. 1 – TP removal from lake water (control), enriched with sediment (1S and 2S) or algae (1A and 2A). Al concentration ranging from 0 to 6 mg L1 Al.
Table 1 – Compositions of waters used to assess TP removal (SRP concentrations were adjusted to 40720 mg L1)
Minimum and maximum values reported from van Hullebusch et al. (2002a,b) Without phosphorus enrichment Low organic P enrichment with sediment suspension High organic P enrichment with sediment suspension Low organic P enrichment with algae suspension High organic P enrichment with algae suspension
Water samples
TP (mg L1)
TOC (mg L1)
Control Water 1S Water 2S Water 1A Water 2A
80–150 100 110 140 150 230
6.3–8 6.2 7.3 9.4 7.3 9.2
ARTICLE IN PRESS 2716
WAT E R R E S E A R C H
40 (2006) 2713– 2719
Moreover previous studies have shown that a competitive sorption of phosphate with OM can occur (Kastelan-Macan and Petrovic 1995, 1996). As the OM sources appeared to have a significant impact on TP removal, their influence on SRP removal needed to be studied.
SRP removal (%)
3.2.
100
Soluble reactive phosphorus removal
80 60 1 mg.L-1 Al 2 mg.L-1 Al 6 mg.L-1 Al
40 20 0
SRP removal (%)
The SRP removal curve (Fig. 2) shows the same trend for the three Al concentrations studied (1, 2 and 6 mg L1 Al). Even at the lowest Al concentration the SRP removal efficiency remains above 55% whatever the initial SRP concentration, in the range of 20–320 mg L1 P. For an SRP concentration up to 150 mg L1 P an improvement in the SRP removal with increasing Al was noticed. The maximum percentage removal occurred with initial SRP concentration higher than 100 mg L1 P. Maximum removal percentages were 80% for a 1 mg L1 Al treatment and 95% for Al treatment at 2 and 6 mg L1. When the SRP exceeded 150 mg L1, the removal efficiency remained at the maximum for 2 and 6 mg L1 Al, whereas a slight decrease in the removal efficiency was observed at the lowest Al dose (1 mg L1 Al) probably due to the saturation of the Al hydroxides accumulation capacity. These results, displayed on Fig. 2, show that 2 mg L1 is sufficient to remove more than 80% of the SRP at initial SRP concentrations over the range of 50–320 mg L1 P. Fig. 3 shows SRP removal from lake water with an initial SRP concentration of 100720 mg L1 P. OM concentrations (enrichment with sediment) from 6.2 to 9.7 mg L1 TOC and three different Al concentrations (1, 2 and 6 mg L1 Al) were used. Enrichment with algae gave similar results (data not shown). Al hydroxides display a high affinity with OM, since the SRP removal efficiency remained constant with increasing OM concentration (Fig. 3). These results show a high Al effectiveness in SRP removal. According to Galarneau and Gehr (1997) and Boisvert et al. (1997) two mechanisms can be involved in SRP removal at pH 6.5: Al-PO4(s) precipitation and SRP adsorption with Al hydrolysed species (Al involved in the interaction can be part of a polymeric species or on the surface of solid particles). Galarneau and Gehr (1997) showed that SRP concentrations well above the concentrations used in this study (and usually found in natural waters) are needed to allow Al-PO4(s) formation at pH 7.
100 80 60 1 mg.L-1 Al 2 mg.L-1 Al 6 mg.L-1 Al
40 20 0 0
50
100
150
200
250
300
350
Initial concentration of SRP (µg.L-1 P) Fig. 2 – Removal of SRP from solution with different concentrations of Al (1, 2 and 6 mg L1 Al). Different initial SRP concentrations were obtained using lake water enriched with K2HPO4.
6.0
7.0
8.0
OM concentration
9.0 (mg.L-1
10.0
TOC)
Fig. 3 – Influence of OM concentration on SRP removal with different concentrations of Al (1, 2 and 6 mg L1 Al). Lake water was enriched with sediment suspension; SRP concentrations were adjusted at 100720 lg L1.
At pH near 6.5, Boisvert et al. (1997) showed that Al hydroxide surfaces were negatively charged inducing repulsive electrostatic interactions with negative ions. At this pH, 2 the dominant SRP species are H2PO 4 and HPO4 , which are also negatively charged. Thus, Al hydrolysed species could exchange OH ligands with SRP resulting in an inner sphere complex. Even if the influence of pH and the SRP concentration on adsorption have been studied (Lijklema, 1980; Zhou et al., 2005), little attention has been paid to competition with OM even though Al is supposed to form specific complexes with SRP and OM. Figs. 2 and 3 display Al effectiveness for SRP removal (even at low doses), which suggest high affinity interactions between Al and SRP. The data reported in Fig. 3 does, however, not display any evidence of a competitive effect of particulate OM on SRP sorption under our operational conditions. In contrast, Kastelan-Macan and Petrovic (1996) have shown that dissolved OM (i.e., fulvic acid) can reduce SRP sorption on mineral particles when introduced at concentration levels similar to those observed in sediment pore water (around 50 mg L1 C). Even if increasing the particulate OM concentration in the water column was not found to compete with SRP sorption onto alum flocs, the OM influence when alum flocs are incorporated into lake sediments still needs to be investigated.
3.3.
Flocs properties in relation with structure
As described in the first section, the alum floc settleability towards the sediment phase is shown to play a major role in the TP removal efficiency. The knowledge of the floc structure should give a more global view of Al reactivity and behaviour in aquatic systems. Floc structures have been studied with limited resources in our experiments, but the literature provides experimental features for a better description of our results. Within the neutral pH range, Al forms many species with variable reactivity. At pH 6.5, the concentrations of Al cationic monomer species, which are toxic for microorganisms (Gensemer and Playle, 1999), are negligible. The Al hydroxides species are essentially present in polymerised (dimer, trimer, Al13, etc.) or solid (Al(OH)3) form. The occurrence of these species is consistent with coagulation related to the complexation or entrapment of OM (Gregor et al., 1997; Kim et al.,
ARTICLE IN PRESS
Residual aluminium after sedimentation (mg.L-1 Al)
Water 2S Water 1A
3 2 2 1
100 Removal (%)
2001). However, sweep-floc formation involves rapid settling of solid Al hydroxides, which requires a high Al concentration. Kim et al. (2001) showed that more than 6 mg L1 Al is required in most cases. This agrees with Fig. 4, where Al remaining in solution after settling (associated with microflocs) is reported versus Al dose. The higher Al concentration observed in water 2S than in 1A resulted from sediment suspension enrichment, which contained a high natural Al concentration in this lake. Our experimental data (Fig. 4) are well correlated with the occurrence of complexation as described by Gregor et al. (1997). Residual Al increases when the Al dose is increased up to 2 mg L1, this indicates complexation of Al hydroxides with OM to form Al–OM aggregates that are unable to settle (micro-flocs). Increasing the Al dose results in larger Al–OM aggregates (able to settle) where Al acts as a bridging agent between organic matter. Few studies have focused on the structure of these bridges. Masion et al. (2000) identified, at the molecular level, the Al species involved in flocs formed by coagulation of OM by Al salts in lake water. Al species in flocs appeared to be poorly polymerised, the biggest polynuclear Al species seemed to be Al13. According to these authors, this was the result of depolymerisation of Al polynuclear species by organic ligands. The role of the size of the particles (algae or sediments) involved in floc formation is shown in Fig. 5 where SRP, TP and OM removal are reported for lake water enriched with sediment or algae. On Fig. 5, with water 2S, TP removal with the lowest Al dose occurs, indicating macro-floc formation. This occurred from 1 to 4 mg L1 Al. This can be attributed to the wide range of sediment particle size. Large particles (up to 63-mm) could form macro-flocs with only few particles and bridge together at low Al concentrations, whereas smaller particles contained in sediment suspensions need more Al. Algae have, in contrast (water 1A), a uniform size (5–10 mm) which induces macro-floc formation with higher Al concentrations over a reduced Al range (3–4 mg L1). Petrovic and Kastelan-Macan (1996) have previously described the uptake of inorganic P by insoluble metal humic substances, where Al acted as a bridging agent between OM and orthophosphate. However, Fig. 5 shows that SRP were removed with low Al concentrations (1–3 mg L1 Al) only when micro-flocs were present. This indicated that phosphate anions must be bound to hydrolysed Al species. The results of this study lead to the suggestion of a mechanism of
3
2717
4 0 (200 6) 271 3 – 271 9
80 60 40
TP OM SRP
20
Water 2S
0 0
1
2
4
3
5
6
Aluminium (mg.L-1 Al)
(a)
Removal (%)
WAT E R R E S E A R C H
100 90 80 70 60 50 40 30 20 10 0
TP OM SRP Water 1A
0
1
2
3
4
5
6
-1
Aluminium (mg.L Al)
(b)
Fig. 5 – SRP, TP and organic matter removal from two waters (2S and 1A) enriched, respectively, with sediment or algae suspension.
specific phosphate anion sorption by poorly polymerised Al contained in OM flocs. This mechanism is shown in Fig. 6. A recent study from Berkowitz et al. (2005) showed that in case of the addition of higher alum doses (20–25 mg L1 Al), flocs are essentially composed of Al hydroxides in the amorphous phase. These authors studied the effect of ageing (i.e., incubation in laboratory conditions) on the flocs’ structure, which became more crystalline compared to the amorphous dominant Al hydroxides species obtained a few hours after the alum treatment. These alum flocs were also shown to lose part of their P sorption capacity (Berkowitz et al., 2006). The maximum phosphate sorption of the alum floc aged for 6 months was about 50% lower than freshly precipitated floc (Berkowitz et al., 2006). On the other hand, Rydin et al. (2000), Lewandowski et al. (2003) and Reitzel et al. (2005) have shown that flocs formed during lake treatment and buried in the sediments do not lose their P sorption capacity. Lewandowski et al. (2003) performed P sorption batch experiments, illustrating that the sediment Al(OH)3 layer’s sorptive capacity is still not exhausted with further P sorption. Moreover, Rydin et al. (2000) found that Al sampled from sediment previously treated with Al several years ago (i.e., from 2 to 8 years after Al treatment) appeared to be loosely crystallised compared to natural Al. These results could support our hypothetical mechanism.
1 0 0
1
2
3
4
5
6
4.
Conclusion
-1
Aluminium dose (mg.L Al) Fig. 4 – Residual Al concentration in supernatant of water 1S and 1A (TOC ¼ 7.3 mg L1) after coagulation versus Al dose.
This study underlines two mechanisms in phosphorus (P) removal: soluble reactive phosphorus (SRP) complexation with Al and organic P sedimentation. It shows also that the
ARTICLE IN PRESS 2718
WAT E R R E S E A R C H
40 (2006) 2713– 2719
OM
Aluminium poorly polymerised
Orthophosphates
Fig. 6 – Mechanism suggested for orthophosphate adsorption by aluminium–organic matter flocs.
maximal TP removal efficiency was obtained with an Al concentration ranging from 2 to 5 mg L1 depending on organic matter (OM) origins (algae or sediments). However, OM sources and content were shown not to influence the SRP removal and 2 mg L1 Al was sufficient to obtain 85% SRP removal under natural conditions occurring in eutrophic lakes. According to the present study, the formation of an SRP–Al–OM aggregate structure is proposed to describe the mechanisms involved in P removal. Even if the obtained experimental results are well correlated, further studies at a molecular level are needed to detail our hypothesis. The structure of flocs formed in the water column after treatment differ greatly from that of Al(OH)3(s) alone. Thus laboratory studies of Al SRP sorption properties in sediments should not be performed with pure Al(OH)3(s) as is usually done. Further research on the Al affinity with phosphate should be carrried out with preformed flocs as proposed in this study. The behaviour of flocs when buried in the sediment could be described. This would be also helpful to investigate the evolution of the floc binding capacity due to ageing effects.
Acknowledgements This study was supported financially by Agence de l’Eau Loire Bretagne; the Conseil Regional du Limousin, and Aquagestion-Limousin Aquaculture society. The authors would like to thank Virginie Pallier for her technical assistance. The two reviewers are also thanked for their helpful comments on a previous version of the paper. R E F E R E N C E S
Berkowitz, J., Anderson, M.A., Graham, R.C., 2005. Laboratory investigation of aluminum solubility and solid-phase properties following alum treatment of lake waters. Water Res. 39 (16), 3918–3928. Berkowitz, J., Anderson, M.A., Amrhein, C., 2006. Influence of aging on phosphorus sorption to alum floc in lake water. Water Res. 40 (5), 911–916.
Boisvert, J.P., To, T.C., Berrak, A., Jolicoeur, C., 1997. Phosphate adsorption in flocculation processes of aluminium sulphate and poly aluminium-silicate-sulphate. Water Res. 31 (8), 1939–1946. Burley, K.L., Prepas, E.E., Chambers, P.A., 2001. Phosphorus release from sediments in hardwater eutrophic lakes: the effects of redox-sensitive and -insensitive chemical treatments. Freshwater Biology 46 (8), 1061–1074. Chu, S.P., 1942. The influence of mineral composition of the medium on the growth of planktonic 1. Methods and culture media. Journal of Ecol. 30, 284–325. Correll, D.L., 1988. The role of phosphorus in the eutrophication of receiving waters: a review. J. Environ. Qual. 27, 261–266. Galarneau, E., Gehr, R., 1997. Phosphorus removal from wastewaters: experimental and theoretical support for alternative mechanisms. Water Res. 31 (2), 328–338. Gensemer, R.W., Playle, R.C., 1999. The bioavailability and toxicity of aluminium in aquatic environments. Crit. Rev. Environ. Sci. Technol. 29 (4), 315–450. Gregor, J.E., Nokes, C.J., Fenton, E., 1997. Optimising natural organic matter removal from low turbidity waters by controlled pH adjustment of aluminium coagulation. Water Res. 31 (12), 2949–2958. Hejzlar, J., Dolejs, P., Komarkova, J., Sed’a, J., Simek, K., Vyhnalek, V., 1998. Effect of biomanipulation on the structuring of the planktonic food web and water treatability by coagulation. Water Sci. Technol. 37 (2), 105–119. Hereit, F., Mutl, S., Vagner, V., 1980. The formation of separable suspensions and the methods of its assessment. Aqua 5, 895–899. Huang, C., Shiu, H., 1996. Interactions between alum and organics in coagulation. Colloids and Surfaces A 113, 155–163. Kastelan-Macan, M., Petrovic, M., 1995. Competitive sorption of phosphate and marine humic substances on suspended particulate matter. Water Sci. Technol. 32 (9–10), 349–355. Kastelan-Macan, M., Petrovic, M., 1996. The role of fulvic acids in phosphorus sorption and release from mineral particles. Water Sci. Technol. 34 (7-8), 259–265. Kim, S.-H., Moon, B.H., Lee, H.I., 2001. Effects of pH and dosage on pollutant removal and floc structure during coagulation. Microchem. J. 68, 197–203. Kopacek, J., Ulrich, K.U., Hejzlar, J., Borovec, J., Stuchlik, E., 2001. Natural inactivation of phosphorus by aluminium atmospherically acidified water bodies. Water Res. 35 (16), 3783–3790. Krivtsov, V., Goldspink, C., Sigee, D.C., Bellinger, E.G., 2001. Expansion of the model ‘Rostherne’ for fish and zooplankton: role of top-down effects in modifying the prevailing pattern of ecosystem functioning. Ecol. Model. 138 (1–3), 153–171. Lewandowski, J., Schauser, I., Hupfer, M., 2003. Long term effects of phosphorus precipitations with alum in hypereutrophic Lake Su¨sser See (Germany). Water Res. 37 (13), 3194–3204. Lijklema, L., 1980. Interaction of orthophosphate with iron (III) and aluminum hydroxides. Environ. Sci. Technol. 14 (5), 537–541. Masion, A., Vilge´-Ritter, A., Rose, J., Stone, W.E.E., Teepen, B.J., Rybacki, D., Bottero, J.-Y., 2000. Coagulation-flocculation of natural organic matter with salts: Speciation and structure of the aggregates. Environ. Sci. and Technol. 34, 3242–3246. Petrovic, M., Kastelan-Macan, M., 1996. The uptake of insoluble phosphorus by insoluble metal–humic complexes. Water Sci. Technol. 34 (7–8), 253–258. Pettersson, K., 2001. Phosphorus characteristics of settling and suspended particles in lake Erken. Sci. Tot. Environ. 266, 79–86. Ratnaweera, H., Gjessing, E., Oug, E., 1999. Influence of physical– chemical characteristics of natural organic matter (NOM) on
ARTICLE IN PRESS WAT E R R E S E A R C H
4 0 (200 6) 271 3 – 271 9
coagulation properties: an analysis of eight Norwegian water sources. Water Sci. Technol. 40, 89–95. Reeders, H.H., Boers, P.C.M., van der Molen, D.T., Helmerhorst, T.H., 1998. Cyanobacterial dominance in the lakes Veluwemeer and Wolderwijd, The Netherlands. Water Sci. Technol. 37 (3), 85–92. Reitzel, K., Hansen, J., Jensen, H.S., Andersen, F., Hansen, K., 2003. Testing aluminium addition as a tool for lake restoration in shallow, eutrophic lake Sønderby, Denmark. Hydrobiologia 506–509, 781–787. Reitzel, K., Hansen, J., Andersen, F.O., Hansen, K.S., Jensen, H.S., 2005. Lake restoration by dosing aluminium relative to mobile phosphorus in the sediment. Environ. Sci. Technol. 39 (11), 4134–4140. Rydin, E., Welch, E.B., 1998. Aluminium dose required to inactivate phosphate in lake sediments. Water Res. 32 (10), 2969–2976. Rydin, E., Huser, B., Welch, E.B., 2000. Amount of phosphorus inactivated by alum treatments in Washington lakes. Limnol. Oceanogr. 45 (1), 226–230.
2719
Selig, U., Hubener, T., Michalik, M., 2002. Dissolved and particulate phosphorus forms in a eutrophic shallow lake. Aquat. Sci. 64, 97–105. van Hullebusch, E., Deluchat, V., Chazal, P.M., Baudu, M., 2002a. Environmental impact of two chemical treatments in a small shallow eutrophied lake: (I) case of aluminium sulphate. Environ. Pollut. 120 (3), 617–626. van Hullebusch, E., Deluchat, V., Chazal, P.M., Baudu, M., 2002b. Environmental Impact of two chemical treatments in a small shallow eutrophied lake: (II) case of copper sulphate. Environ. Pollut. 120 (3), 627–634. van Hullebusch, E., Auvray, F., Deluchat, V., Chazal, P.M., Baudu, M., 2003. Phosphorus fractionation and short term mobility in the surface sediment of a shallow polymictic lake treated with low dose of alum (Courtille lake, France). Water, Air, Soil Pollut. 146 (1–4), 75–91. Zhou, A., Tang, H., Wang, D., 2005. Phosphorus adsorption on natural sediments: modeling and effects of pH and sediment composition. Water Res. 39 (7), 1245–1254.
40 (2006) 2713 – 2719
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/watres
Laboratory investigation of the phosphorus removal (SRP and TP) from eutrophic lake water treated with aluminium Franck Auvraya, Eric D. van Hullebuscha,b, Veronique Deluchata, Michel Baudua, a
Laboratoire des Sciences de l’Eau et de l’Environnement, Faculte´ des Sciences et Techniques, 123, Avenue Albert Thomas, F-87060 Limoges, Cedex, France b Laboratoire des Ge´omate´riaux et Ge´ologie de l’Inge´nieur, Universite´ de Marne la Valle´e, 5, Boulevard Descartes – Champs sur Marne, 77454 Marne La Valle´e, Cedex 2, France
art i cle info
A B S T R A C T
Article history:
Mechanisms involved in phosphorus (P) removal from eutrophic lake water with
Received 31 March 2004
aluminium (Al) were assessed by jar tests. For this purpose, eutrophic lake water enriched
Received in revised form
with soluble reactive phosphorus (SRP), algae or sediments in order to mimic the various
26 April 2006
conditions found in shallow eutrophic lakes was studied. Total phosphorus (TP) removal
Accepted 27 April 2006
was reached after floc settling, the maximal TP removal efficiency (90–95%) was obtained
Available online 30 June 2006
for an Al concentration ranging from 2 to 5 mg L1, depending on the organic matter (OM)
Keywords:
origin (algae or sediments). Algae appeared to limit macro-floc formation (those able to
Eutrophic lake water
settle). In contrast, in the presence of sediments, macro-floc formation was favoured at low
Aluminium
Al dose (2–3 mg L1 Al). High SRP removal was obtained with the lowest Al dose (1 mg L1 Al)
Phosphorus removal
and remained greater than 60% for an SRP concentration up to 350 mg L1. SRP removal was
Flocculation
not influenced by the OM origin. The experimental data and literature were used to suggest a hypothetical model for floc formation and P removal with Al under the conditions observed in treated lakes. & 2006 Elsevier Ltd. All rights reserved.
1.
Introduction
The over enrichment of surface water with nutrients (phosphorus (P) and nitrogen) is the major cause of eutrophication (Correll, 1998). As high P concentrations are often associated with cyanobacterial blooms, this nutrient appears to play an important role in the phytoplankton dynamics. The dominance of cyanobacteria in summer results in high turbidity as well as possible cyanotoxins production, limiting uses of freshwater lakes (i.e., recreational activities, reservoir for drinking water production). There is no universal link between the P concentration and cyanobacterial dominance, but high spring bioavailability of P and high summer P release by sediments are the main factors responsible for the occurrence of cyanobacterial blooms in shallow lakes (Reeders et al., 1998; Krivtsov et al., 2001). Corresponding author. Tel.: +33 5 55 45 73 67; fax: +33 5 55 45 72 03.
E-mail address: [email protected] (M. Baudu). 0043-1354/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2006.04.042
Consequently, P removal from the water column and immobilisation in the sediment with aluminium (Al) has been used to a large extent. Al is chosen due to its ability to complex P and its insensitivity to redox potential fluctuations in sediments (Burley et al., 2001). The conventional aim of Al treatment is to increase the sediment P retention capacity. Generally, the recommended Al dose ranges from 5 to 30 mg L1 (Rydin and Welch, 1998 ; Reitzel et al., 2003). However, recent studies (van Hullebusch et al., 2002a; van Hullebusch et al., 2003) have shown that lower doses (i.e., 1.5 mg L1 as Al) could both decrease the soluble reactive phosphorus (SRP) concentration in the water column and introduce freshly formed Al hydroxides onto superficial sediments (with newly settled particles) which enhance P retention, thus reducing the summer algae biomass. Moreover, Kopacek et al. (2001) have shown that a high Al content
ARTICLE IN PRESS 2714
WAT E R R E S E A R C H
40 (2006) 2713– 2719
in the in-flow, following acid rain, results in Al hydroxide formation in the lake water column when the pH is close to neutral. This phenomenon was shown to enhance the phosphorus sorption capacity of sediments subsequent to Al hydroxide floc settling. However, all of these studies only take into consideration the results of phosphorus depletion with no attention paid to mechanisms involved in floc formation and/or SRP sorption. Moreover, many studies on drinking water treatment (Galarneau and Gehr, 1997; Gregor et al. 1997; Kim et al., 2001), have underlined the major role played by organic matter (OM) flocs. Floc structure is one of the dominant factors that influences its behaviour, physical (i.e. sedimentation) and/or chemical (i.e. sorption) properties in natural systems. However, physicochemical conditions in water treatment plants are very different in contrast to those occurring in eutrophic lakes. In these aquatic systems, the organic P pool is mainly constituted by dissolved organic species, algae and sediment organic particles present in the water column following re-suspension events (Pettersson, 2001; Selig et al., 2002). The relative contribution of sediment and algae to the P pool may vary a lot depending on the season and weather (wind, storms, etc.) which, in the end, can enhance the growth of algae or favour sediment resuspension. Hejzlar et al. (1998) have shown that algae greatly influence water treatability by coagulation, but the impact of sediments on coagulation, flocculation and sedimentation processes are less well known. The aim of this study is to improve understanding of P removal from eutrophic lake water treated with Al salts. As this should be closely related to floc formation and structure, which are greatly influenced by OM content and origin, these factors must also be considered. For this purpose, laboratory experiments were carried out using jar tests and natural water from a eutrophic lake. Firstly, mechanisms involved in total phosphorus (TP) removal from water containing mostly particulate OM from two origins (sediments or algae) were established. Secondly, SRP removal at different OM matter concentrations was monitored.
2.
Materials and experimental methods
2.1.
Water and organic matter sources
Water and superficial sediments were taken from a shallow polymictic eutrophic lake (Lake Courtille, France). This lake was previously treated with Al to avoid cyanobacterial blooms (van Hullebusch et al., 2002a,b). Lake water and sediments (only the upper 2 cm of sediment) were collected in May 2002 under oxic conditions and the samples were stocked in acidcleaned polyethylene flasks, stored at 41C and used within 48h after collection. Sediments were wet sieved (63 mm), stirred vigorously with lake water for 1 h followed by 2 h of undisturbed settling. Supernatants were collected to form the sediment suspension. The sediment Al and iron content were respectively 10.5 and 27 mg g1 of dry weight (more details regarding the sediment analytical data and methods can be found in van Hullebusch et al., 2003). van Hullebusch et al. (2003) showed
that treatment with Al didn’t significantly increase the total Al concentration in sediment. Scenedesmus subspicatus (276/22 CCAP strain purchased from Collection and Culture of Algae and Protozoa collection, UK) were grown axenically in Chu-10 growth medium (Chu, 1942) at 25 1C with continuous stirring. Cells in the exponential phase were harvested by centrifugation (5 min, 1500g) and rinsed twice with distilled water. Optical microscopy was used to verify cell integrity and a mean size, around 872 mm (n ¼ 100), was measured. Algae were then mixed with lake water to form algae-rich water samples. As sediments or algae are able to adsorb or release SRP, adjustments were required to obtain the desired value of SRP in water samples enriched with suspensions. Preliminary kinetic studies showed that after 1 h of equilibration (under slow stirring) no significant release or adsorption occurred within 10 h (data not shown). Therefore, before each experiment, all samples enriched with algal or sediment suspension were allowed to equilibrate for 1 h after SRP adjustment with SRP stock solution (KH2PO4, 5 mg L1 P); the initial SRP concentration was measured after equilibration. In this way, SRP removal measured during experiments would only be the result of Al additions and not a contribution of natural Al or ferrous hydroxides initially present in the suspension.
2.2.
Jar-test procedure
Coagulation and flocculation experiments were carried out with a multiple paddle stirrer at 25 1C. The samples were mixed for 5 min at 100 round per minute (rpm) followed by 30 min at 20 rpm. Aliquots of Al stock solutions (Al2(SO4)3, 16 H2O; 1 g L1 Al) were added to water samples during the first mixing phase to give the required concentrations. The pH was adjusted to 6.5 7 0.1 by addition of NaOH or H2SO4. Samples were then allowed to settle for 45 min, as no change was observed over longer periods. pH values were checked and supernatants were sampled for TP, SRP, OM and Al content determination. The sediment suspension contained natural Al and Fe which increased their dissolved concentrations (respectively up to 1 and 3 mg L1) following sediment suspension enrichment. As these metals are known to have coagulant and sorption properties, control jar tests (0 mg L1 Al) were always conducted to assess the impact of the stirring procedures on sedimentation and SRP removal. In each case, without Al treatment, no sedimentation (of algae or sediment) was observed. Moreover, no SRP removal was observed in supernatants indicating that SRP sorption equilibrium onto natural Fe and Al oxides was probably reached. The increase in dissolved Al and Fe concentrations following sediment enrichment did not therefore contribute to the SRP removal observed in jar tests treated with Al.
2.3.
Total phosphorus removal
TP removal was studied in lake water enriched with sediment or algal suspensions; additions were made to mimic realistic environmental OM concentrations present in shallow eutrophic lakes (van Hullebusch et al., 2002 a, b, 2003). The water characteristics resulting from enrichments are summarised in Table 1. A lot of studies (Huang and Shiu, 1996; Ratnaweera
ARTICLE IN PRESS
et al., 1999) have pointed out the predominant role of the OM level in the coagulation–flocculation process, conditions with similar OM content were therefore chosen. As algae have higher P content than sediments, P concentrations in water enriched with algae are higher. Jar-test procedures were carried out with Al in the range of 1–6 mg L1 Al.
2.4.
Soluble reactive phosphorus removal
SRP removal was studied using lake water enriched with P from a stock solution (KH2PO4, 5 mg L1 P). Final SRP concentrations were in the range of 20–320 mg L1 which largely cover SRP concentrations occurring in Lake Courtille. In another experimental set-up, lake water enriched with different amounts of sediment or algal suspensions to give a final range of OM from 6.2 to 9.5 mg L1 total organic carbon (TOC) was used. SRP concentrations were adjusted to 100720 mg L1 P. Jar-test procedures were carried out with 1, 2 and 6 mg L1 Al.
2.5.
2715
4 0 (200 6) 271 3 – 271 9
Analytical methods
SRP was determined in filtered samples by the phosphomolybdate method (AFNOR procedure NFT 90-023). TP was converted to SRP in crude samples after digestion at 105 1C with sulphuric acid and potassium persulphate. SRP is often assumed to represent the bioavailable P fraction for algae in the water column. The detection limit of this method was 5 mg L1. All P analyses were performed in duplicate with a standard deviation of 5%. OM concentration was determined by wet oxidation with persulphate and UV radiation using a TOC Phoenix 8000 analyser (Dohrman). The standard deviation and detection limit were, respectively, 8% and 0.5 mg L1 TOC. Al determination was done by atomic absorption spectrophotometry using a graphite furnace with a Zeeman effect (Varian GTA 100). Total Al (AlTOT) was determined after digestion with nitric acid. The detection limit and standard deviation were, respectively, 10 mg L1 and 7%.
the range of 1–6 mg L1, and SRP concentrations were adjusted to 40720 mg L1 P. Water compositions are reported in Table 1 and results obtained are plotted in Fig. 1. Al addition induces floc formation and sedimentation that results in a TP decrease in the supernatants. Removal curves show conventional trends for jar-test curves: a first lag phase where removal increases with coagulant addition, then a second one where removal was maximal. However, in the first lag phase (below 4 mg L1 Al), two different trends are obtained with the different OM origins: the increase in OM originating from the sediment suspension resulted in an improvement in TP removal, whereas the addition of algae greatly decreased the Al efficiency (particularly at doses lower than 3 mg L1 Al). Previous studies (Galarneau and Gehr, 1997) have shown that TP removal depends basically on the ability of Al to settle OM associated with P. Al associated with OM induces the formation of some flocs, that have size, charge, density and structure inconsistencies which determine their ability to settle. A simple way to characterise them was proposed by Hereit et al. (1980) and used by Gregor et al. (1997). Macro-flocs are operationally defined as flocs settled within a given time (45 min) and micro-flocs as flocs remaining in suspension after this time. In our experiments, no measurable sedimentation was noticed beyond this time. If this nomenclature is used due to its simplicity, it should not be forgotten that the denomination of macro- and micro-flocs is not directly associated with their size but with their ability to settle under given conditions. The results presented in Fig. 1 indicate that the sediment enhanced macro-floc formation, whereas algae stabilised micro-flocs.
100 TP removal (%)
WAT E R R E S E A R C H
80 Control
60
Water 1S 40
Water 2S Water 1A
20
3.
Results and discussion
3.1.
Total phosphorus removal
Jar-test procedures were applied to lake water enriched with two organic P sources (algae and sediments), Al was dosed in
Water 2A
0 0
1
2 3 4 Aluminium (mg.L-1 Al)
5
6
Fig. 1 – TP removal from lake water (control), enriched with sediment (1S and 2S) or algae (1A and 2A). Al concentration ranging from 0 to 6 mg L1 Al.
Table 1 – Compositions of waters used to assess TP removal (SRP concentrations were adjusted to 40720 mg L1)
Minimum and maximum values reported from van Hullebusch et al. (2002a,b) Without phosphorus enrichment Low organic P enrichment with sediment suspension High organic P enrichment with sediment suspension Low organic P enrichment with algae suspension High organic P enrichment with algae suspension
Water samples
TP (mg L1)
TOC (mg L1)
Control Water 1S Water 2S Water 1A Water 2A
80–150 100 110 140 150 230
6.3–8 6.2 7.3 9.4 7.3 9.2
ARTICLE IN PRESS 2716
WAT E R R E S E A R C H
40 (2006) 2713– 2719
Moreover previous studies have shown that a competitive sorption of phosphate with OM can occur (Kastelan-Macan and Petrovic 1995, 1996). As the OM sources appeared to have a significant impact on TP removal, their influence on SRP removal needed to be studied.
SRP removal (%)
3.2.
100
Soluble reactive phosphorus removal
80 60 1 mg.L-1 Al 2 mg.L-1 Al 6 mg.L-1 Al
40 20 0
SRP removal (%)
The SRP removal curve (Fig. 2) shows the same trend for the three Al concentrations studied (1, 2 and 6 mg L1 Al). Even at the lowest Al concentration the SRP removal efficiency remains above 55% whatever the initial SRP concentration, in the range of 20–320 mg L1 P. For an SRP concentration up to 150 mg L1 P an improvement in the SRP removal with increasing Al was noticed. The maximum percentage removal occurred with initial SRP concentration higher than 100 mg L1 P. Maximum removal percentages were 80% for a 1 mg L1 Al treatment and 95% for Al treatment at 2 and 6 mg L1. When the SRP exceeded 150 mg L1, the removal efficiency remained at the maximum for 2 and 6 mg L1 Al, whereas a slight decrease in the removal efficiency was observed at the lowest Al dose (1 mg L1 Al) probably due to the saturation of the Al hydroxides accumulation capacity. These results, displayed on Fig. 2, show that 2 mg L1 is sufficient to remove more than 80% of the SRP at initial SRP concentrations over the range of 50–320 mg L1 P. Fig. 3 shows SRP removal from lake water with an initial SRP concentration of 100720 mg L1 P. OM concentrations (enrichment with sediment) from 6.2 to 9.7 mg L1 TOC and three different Al concentrations (1, 2 and 6 mg L1 Al) were used. Enrichment with algae gave similar results (data not shown). Al hydroxides display a high affinity with OM, since the SRP removal efficiency remained constant with increasing OM concentration (Fig. 3). These results show a high Al effectiveness in SRP removal. According to Galarneau and Gehr (1997) and Boisvert et al. (1997) two mechanisms can be involved in SRP removal at pH 6.5: Al-PO4(s) precipitation and SRP adsorption with Al hydrolysed species (Al involved in the interaction can be part of a polymeric species or on the surface of solid particles). Galarneau and Gehr (1997) showed that SRP concentrations well above the concentrations used in this study (and usually found in natural waters) are needed to allow Al-PO4(s) formation at pH 7.
100 80 60 1 mg.L-1 Al 2 mg.L-1 Al 6 mg.L-1 Al
40 20 0 0
50
100
150
200
250
300
350
Initial concentration of SRP (µg.L-1 P) Fig. 2 – Removal of SRP from solution with different concentrations of Al (1, 2 and 6 mg L1 Al). Different initial SRP concentrations were obtained using lake water enriched with K2HPO4.
6.0
7.0
8.0
OM concentration
9.0 (mg.L-1
10.0
TOC)
Fig. 3 – Influence of OM concentration on SRP removal with different concentrations of Al (1, 2 and 6 mg L1 Al). Lake water was enriched with sediment suspension; SRP concentrations were adjusted at 100720 lg L1.
At pH near 6.5, Boisvert et al. (1997) showed that Al hydroxide surfaces were negatively charged inducing repulsive electrostatic interactions with negative ions. At this pH, 2 the dominant SRP species are H2PO 4 and HPO4 , which are also negatively charged. Thus, Al hydrolysed species could exchange OH ligands with SRP resulting in an inner sphere complex. Even if the influence of pH and the SRP concentration on adsorption have been studied (Lijklema, 1980; Zhou et al., 2005), little attention has been paid to competition with OM even though Al is supposed to form specific complexes with SRP and OM. Figs. 2 and 3 display Al effectiveness for SRP removal (even at low doses), which suggest high affinity interactions between Al and SRP. The data reported in Fig. 3 does, however, not display any evidence of a competitive effect of particulate OM on SRP sorption under our operational conditions. In contrast, Kastelan-Macan and Petrovic (1996) have shown that dissolved OM (i.e., fulvic acid) can reduce SRP sorption on mineral particles when introduced at concentration levels similar to those observed in sediment pore water (around 50 mg L1 C). Even if increasing the particulate OM concentration in the water column was not found to compete with SRP sorption onto alum flocs, the OM influence when alum flocs are incorporated into lake sediments still needs to be investigated.
3.3.
Flocs properties in relation with structure
As described in the first section, the alum floc settleability towards the sediment phase is shown to play a major role in the TP removal efficiency. The knowledge of the floc structure should give a more global view of Al reactivity and behaviour in aquatic systems. Floc structures have been studied with limited resources in our experiments, but the literature provides experimental features for a better description of our results. Within the neutral pH range, Al forms many species with variable reactivity. At pH 6.5, the concentrations of Al cationic monomer species, which are toxic for microorganisms (Gensemer and Playle, 1999), are negligible. The Al hydroxides species are essentially present in polymerised (dimer, trimer, Al13, etc.) or solid (Al(OH)3) form. The occurrence of these species is consistent with coagulation related to the complexation or entrapment of OM (Gregor et al., 1997; Kim et al.,
ARTICLE IN PRESS
Residual aluminium after sedimentation (mg.L-1 Al)
Water 2S Water 1A
3 2 2 1
100 Removal (%)
2001). However, sweep-floc formation involves rapid settling of solid Al hydroxides, which requires a high Al concentration. Kim et al. (2001) showed that more than 6 mg L1 Al is required in most cases. This agrees with Fig. 4, where Al remaining in solution after settling (associated with microflocs) is reported versus Al dose. The higher Al concentration observed in water 2S than in 1A resulted from sediment suspension enrichment, which contained a high natural Al concentration in this lake. Our experimental data (Fig. 4) are well correlated with the occurrence of complexation as described by Gregor et al. (1997). Residual Al increases when the Al dose is increased up to 2 mg L1, this indicates complexation of Al hydroxides with OM to form Al–OM aggregates that are unable to settle (micro-flocs). Increasing the Al dose results in larger Al–OM aggregates (able to settle) where Al acts as a bridging agent between organic matter. Few studies have focused on the structure of these bridges. Masion et al. (2000) identified, at the molecular level, the Al species involved in flocs formed by coagulation of OM by Al salts in lake water. Al species in flocs appeared to be poorly polymerised, the biggest polynuclear Al species seemed to be Al13. According to these authors, this was the result of depolymerisation of Al polynuclear species by organic ligands. The role of the size of the particles (algae or sediments) involved in floc formation is shown in Fig. 5 where SRP, TP and OM removal are reported for lake water enriched with sediment or algae. On Fig. 5, with water 2S, TP removal with the lowest Al dose occurs, indicating macro-floc formation. This occurred from 1 to 4 mg L1 Al. This can be attributed to the wide range of sediment particle size. Large particles (up to 63-mm) could form macro-flocs with only few particles and bridge together at low Al concentrations, whereas smaller particles contained in sediment suspensions need more Al. Algae have, in contrast (water 1A), a uniform size (5–10 mm) which induces macro-floc formation with higher Al concentrations over a reduced Al range (3–4 mg L1). Petrovic and Kastelan-Macan (1996) have previously described the uptake of inorganic P by insoluble metal humic substances, where Al acted as a bridging agent between OM and orthophosphate. However, Fig. 5 shows that SRP were removed with low Al concentrations (1–3 mg L1 Al) only when micro-flocs were present. This indicated that phosphate anions must be bound to hydrolysed Al species. The results of this study lead to the suggestion of a mechanism of
3
2717
4 0 (200 6) 271 3 – 271 9
80 60 40
TP OM SRP
20
Water 2S
0 0
1
2
4
3
5
6
Aluminium (mg.L-1 Al)
(a)
Removal (%)
WAT E R R E S E A R C H
100 90 80 70 60 50 40 30 20 10 0
TP OM SRP Water 1A
0
1
2
3
4
5
6
-1
Aluminium (mg.L Al)
(b)
Fig. 5 – SRP, TP and organic matter removal from two waters (2S and 1A) enriched, respectively, with sediment or algae suspension.
specific phosphate anion sorption by poorly polymerised Al contained in OM flocs. This mechanism is shown in Fig. 6. A recent study from Berkowitz et al. (2005) showed that in case of the addition of higher alum doses (20–25 mg L1 Al), flocs are essentially composed of Al hydroxides in the amorphous phase. These authors studied the effect of ageing (i.e., incubation in laboratory conditions) on the flocs’ structure, which became more crystalline compared to the amorphous dominant Al hydroxides species obtained a few hours after the alum treatment. These alum flocs were also shown to lose part of their P sorption capacity (Berkowitz et al., 2006). The maximum phosphate sorption of the alum floc aged for 6 months was about 50% lower than freshly precipitated floc (Berkowitz et al., 2006). On the other hand, Rydin et al. (2000), Lewandowski et al. (2003) and Reitzel et al. (2005) have shown that flocs formed during lake treatment and buried in the sediments do not lose their P sorption capacity. Lewandowski et al. (2003) performed P sorption batch experiments, illustrating that the sediment Al(OH)3 layer’s sorptive capacity is still not exhausted with further P sorption. Moreover, Rydin et al. (2000) found that Al sampled from sediment previously treated with Al several years ago (i.e., from 2 to 8 years after Al treatment) appeared to be loosely crystallised compared to natural Al. These results could support our hypothetical mechanism.
1 0 0
1
2
3
4
5
6
4.
Conclusion
-1
Aluminium dose (mg.L Al) Fig. 4 – Residual Al concentration in supernatant of water 1S and 1A (TOC ¼ 7.3 mg L1) after coagulation versus Al dose.
This study underlines two mechanisms in phosphorus (P) removal: soluble reactive phosphorus (SRP) complexation with Al and organic P sedimentation. It shows also that the
ARTICLE IN PRESS 2718
WAT E R R E S E A R C H
40 (2006) 2713– 2719
OM
Aluminium poorly polymerised
Orthophosphates
Fig. 6 – Mechanism suggested for orthophosphate adsorption by aluminium–organic matter flocs.
maximal TP removal efficiency was obtained with an Al concentration ranging from 2 to 5 mg L1 depending on organic matter (OM) origins (algae or sediments). However, OM sources and content were shown not to influence the SRP removal and 2 mg L1 Al was sufficient to obtain 85% SRP removal under natural conditions occurring in eutrophic lakes. According to the present study, the formation of an SRP–Al–OM aggregate structure is proposed to describe the mechanisms involved in P removal. Even if the obtained experimental results are well correlated, further studies at a molecular level are needed to detail our hypothesis. The structure of flocs formed in the water column after treatment differ greatly from that of Al(OH)3(s) alone. Thus laboratory studies of Al SRP sorption properties in sediments should not be performed with pure Al(OH)3(s) as is usually done. Further research on the Al affinity with phosphate should be carrried out with preformed flocs as proposed in this study. The behaviour of flocs when buried in the sediment could be described. This would be also helpful to investigate the evolution of the floc binding capacity due to ageing effects.
Acknowledgements This study was supported financially by Agence de l’Eau Loire Bretagne; the Conseil Regional du Limousin, and Aquagestion-Limousin Aquaculture society. The authors would like to thank Virginie Pallier for her technical assistance. The two reviewers are also thanked for their helpful comments on a previous version of the paper. R E F E R E N C E S
Berkowitz, J., Anderson, M.A., Graham, R.C., 2005. Laboratory investigation of aluminum solubility and solid-phase properties following alum treatment of lake waters. Water Res. 39 (16), 3918–3928. Berkowitz, J., Anderson, M.A., Amrhein, C., 2006. Influence of aging on phosphorus sorption to alum floc in lake water. Water Res. 40 (5), 911–916.
Boisvert, J.P., To, T.C., Berrak, A., Jolicoeur, C., 1997. Phosphate adsorption in flocculation processes of aluminium sulphate and poly aluminium-silicate-sulphate. Water Res. 31 (8), 1939–1946. Burley, K.L., Prepas, E.E., Chambers, P.A., 2001. Phosphorus release from sediments in hardwater eutrophic lakes: the effects of redox-sensitive and -insensitive chemical treatments. Freshwater Biology 46 (8), 1061–1074. Chu, S.P., 1942. The influence of mineral composition of the medium on the growth of planktonic 1. Methods and culture media. Journal of Ecol. 30, 284–325. Correll, D.L., 1988. The role of phosphorus in the eutrophication of receiving waters: a review. J. Environ. Qual. 27, 261–266. Galarneau, E., Gehr, R., 1997. Phosphorus removal from wastewaters: experimental and theoretical support for alternative mechanisms. Water Res. 31 (2), 328–338. Gensemer, R.W., Playle, R.C., 1999. The bioavailability and toxicity of aluminium in aquatic environments. Crit. Rev. Environ. Sci. Technol. 29 (4), 315–450. Gregor, J.E., Nokes, C.J., Fenton, E., 1997. Optimising natural organic matter removal from low turbidity waters by controlled pH adjustment of aluminium coagulation. Water Res. 31 (12), 2949–2958. Hejzlar, J., Dolejs, P., Komarkova, J., Sed’a, J., Simek, K., Vyhnalek, V., 1998. Effect of biomanipulation on the structuring of the planktonic food web and water treatability by coagulation. Water Sci. Technol. 37 (2), 105–119. Hereit, F., Mutl, S., Vagner, V., 1980. The formation of separable suspensions and the methods of its assessment. Aqua 5, 895–899. Huang, C., Shiu, H., 1996. Interactions between alum and organics in coagulation. Colloids and Surfaces A 113, 155–163. Kastelan-Macan, M., Petrovic, M., 1995. Competitive sorption of phosphate and marine humic substances on suspended particulate matter. Water Sci. Technol. 32 (9–10), 349–355. Kastelan-Macan, M., Petrovic, M., 1996. The role of fulvic acids in phosphorus sorption and release from mineral particles. Water Sci. Technol. 34 (7-8), 259–265. Kim, S.-H., Moon, B.H., Lee, H.I., 2001. Effects of pH and dosage on pollutant removal and floc structure during coagulation. Microchem. J. 68, 197–203. Kopacek, J., Ulrich, K.U., Hejzlar, J., Borovec, J., Stuchlik, E., 2001. Natural inactivation of phosphorus by aluminium atmospherically acidified water bodies. Water Res. 35 (16), 3783–3790. Krivtsov, V., Goldspink, C., Sigee, D.C., Bellinger, E.G., 2001. Expansion of the model ‘Rostherne’ for fish and zooplankton: role of top-down effects in modifying the prevailing pattern of ecosystem functioning. Ecol. Model. 138 (1–3), 153–171. Lewandowski, J., Schauser, I., Hupfer, M., 2003. Long term effects of phosphorus precipitations with alum in hypereutrophic Lake Su¨sser See (Germany). Water Res. 37 (13), 3194–3204. Lijklema, L., 1980. Interaction of orthophosphate with iron (III) and aluminum hydroxides. Environ. Sci. Technol. 14 (5), 537–541. Masion, A., Vilge´-Ritter, A., Rose, J., Stone, W.E.E., Teepen, B.J., Rybacki, D., Bottero, J.-Y., 2000. Coagulation-flocculation of natural organic matter with salts: Speciation and structure of the aggregates. Environ. Sci. and Technol. 34, 3242–3246. Petrovic, M., Kastelan-Macan, M., 1996. The uptake of insoluble phosphorus by insoluble metal–humic complexes. Water Sci. Technol. 34 (7–8), 253–258. Pettersson, K., 2001. Phosphorus characteristics of settling and suspended particles in lake Erken. Sci. Tot. Environ. 266, 79–86. Ratnaweera, H., Gjessing, E., Oug, E., 1999. Influence of physical– chemical characteristics of natural organic matter (NOM) on
ARTICLE IN PRESS WAT E R R E S E A R C H
4 0 (200 6) 271 3 – 271 9
coagulation properties: an analysis of eight Norwegian water sources. Water Sci. Technol. 40, 89–95. Reeders, H.H., Boers, P.C.M., van der Molen, D.T., Helmerhorst, T.H., 1998. Cyanobacterial dominance in the lakes Veluwemeer and Wolderwijd, The Netherlands. Water Sci. Technol. 37 (3), 85–92. Reitzel, K., Hansen, J., Jensen, H.S., Andersen, F., Hansen, K., 2003. Testing aluminium addition as a tool for lake restoration in shallow, eutrophic lake Sønderby, Denmark. Hydrobiologia 506–509, 781–787. Reitzel, K., Hansen, J., Andersen, F.O., Hansen, K.S., Jensen, H.S., 2005. Lake restoration by dosing aluminium relative to mobile phosphorus in the sediment. Environ. Sci. Technol. 39 (11), 4134–4140. Rydin, E., Welch, E.B., 1998. Aluminium dose required to inactivate phosphate in lake sediments. Water Res. 32 (10), 2969–2976. Rydin, E., Huser, B., Welch, E.B., 2000. Amount of phosphorus inactivated by alum treatments in Washington lakes. Limnol. Oceanogr. 45 (1), 226–230.
2719
Selig, U., Hubener, T., Michalik, M., 2002. Dissolved and particulate phosphorus forms in a eutrophic shallow lake. Aquat. Sci. 64, 97–105. van Hullebusch, E., Deluchat, V., Chazal, P.M., Baudu, M., 2002a. Environmental impact of two chemical treatments in a small shallow eutrophied lake: (I) case of aluminium sulphate. Environ. Pollut. 120 (3), 617–626. van Hullebusch, E., Deluchat, V., Chazal, P.M., Baudu, M., 2002b. Environmental Impact of two chemical treatments in a small shallow eutrophied lake: (II) case of copper sulphate. Environ. Pollut. 120 (3), 627–634. van Hullebusch, E., Auvray, F., Deluchat, V., Chazal, P.M., Baudu, M., 2003. Phosphorus fractionation and short term mobility in the surface sediment of a shallow polymictic lake treated with low dose of alum (Courtille lake, France). Water, Air, Soil Pollut. 146 (1–4), 75–91. Zhou, A., Tang, H., Wang, D., 2005. Phosphorus adsorption on natural sediments: modeling and effects of pH and sediment composition. Water Res. 39 (7), 1245–1254.