Article pubs.acs.org/est
Regrowth of Broken Hydroxide Flocs: Effect of Added Fluoride Wen-Zheng Yu,† John Gregory,*,‡ and Nigel Graham† †
Department of Civil and Environmental Engineering, Imperial College London, South Kensington Campus, London SW7 2AZ, U.K. Department of Civil, Environmental, and Geomatic Engineering, University College London, Gower Street, London WC1E 6BT, U.K.
‡
S Supporting Information *
ABSTRACT: Hydrous oxides of Al(III) and Fe(III) play a large part in environmental processes and in the action of coagulants used in water and wastewater treatment. Aggregates (flocs) of hydroxide precipitates can be rather weak and are easily broken by applied shear. It is usually found that broken flocs do not fully regrow under low-shear conditions, and this could be a serious disadvantage in practical applications. The irreversible nature of floc breakage suggests that some form of specific, chemical interaction between precipitate particles must be at least partly responsible. On the basis of experiments reported here and elsewhere, we propose that hydroxyl bridges between particles play a part. When these are broken, there is a reduction in the number of “active” surface groups that are able to form new bridges. When small amounts of fluoride are added during breakage of Al flocs, there can be significant improvement in floc regrowth, although this depends on a number of factors, especially pH. With Fe flocs, fluoride has no noticeable effect. These results can be explained by the formation of soluble Al−F complexes and some dissolution of the Al(OH)3 precipitate. This creates a new surface with more “active” groups that can form new hydroxyl bridges.
■
INTRODUCTION Hydrous oxides of iron and aluminum are of great environmental importance and play a large part in coagulation processes in water and wastewater treatment. Hydroxide precipitates have relatively large specific surface areas and can adsorb a wide range of pollutants, a fact which is exploited in water treatment for the removal of dissolved organic material, toxic metals, and anions such as fluoride and phosphate. When Al and Fe(III) coagulants are added to water at typical dosages and around neutral pH, the solution is highly supersaturated with respect to the amorphous hydroxide, and there is rapid precipitation of nanosized particles.1 These primary nanoparticles then aggregate, eventually forming large flocs with sizes up to several hundred μm in many cases. Soluble and particulate impurities in water are incorporated in the growing flocs by adsorption and enmeshment, a process known as “sweep coagulation”.2 The amorphous hydroxide precipitates may undergo a slow transition to crystalline forms such as hematite and gibbsite in the case of Fe and Al salts, respectively. However, because of the short time scales involved, these crystalline forms are not generally relevant in water-treatment applications. Hydroxide flocs are of only limited strength and can be readily broken by an increased shear rate. When the shear rate is reduced, in most cases, the flocs do not fully regrow.3 This implies that the binding of particles within flocs is not entirely due to physical (e.g., van der Waals) interactions but that some form of specific, chemical binding is at least partly responsible. Irreversible floc breakage can be a significant practical problem in © 2016 American Chemical Society
water treatment, and the aim of this paper is to seek a better understanding of the mechanisms involved. From previous work,4 it has been shown that fresh hydroxide precipitate, formed during floc breakage as a result of a small additional dosage of coagulant, can greatly improve the regrowth. Anions such as phosphate and fluoride can also influence floc regrowth, and here we investigate the effect of fluoride. A wide range of soluble impurities can be removed by adsorption on the hydroxide precipitate, which is of great practical value in water treatment. Anions such as phosphate and fluoride may be removed from water in this way. In the natural environment, adsorption on hydrous oxides can affect the transport and ultimate fate of pollutants, and amorphous Al oxides are thought to be the most important environmental sink for fluoride.5 However, another important effect with fluoride and Al oxides is that large amounts of soluble aluminum may be released into aquifers as a result of formation of soluble Al−F complexes.6 These complexes are the predominant aqueous Al species in some forest soils.7 The release of Al is especially significant with amorphous precipitates. At high F/Al ratios and pH < 6, it has been found that most of an amorphous Al(OH)3 gel can be dissolved.8 Harrington et al.9 found that fluoride could cause a significant release of Al from iron-oxide-rich soils, but no Received: Revised: Accepted: Published: 1828
October 29, 2015 January 2, 2016 January 12, 2016 January 12, 2016 DOI: 10.1021/acs.est.5b05334 Environ. Sci. Technol. 2016, 50, 1828−1833
Article
Environmental Science & Technology
Figure 1. Flocculation of kaolin suspensions with AlCl3. Fluoride added during floc breakage (at about 750 s). Fluoride concentrations shown on curves.
■
dissolved Fe was detected. Nordin et al.10 studied the kinetics of fluoride-enhanced dissolution of bayerite (β-Al(OH)3) and boehmite (γ-AlOOH) using 19F-NMR spectroscopy. The dissolution process was found to be quite complex, involving a number of coupled pathways, with the dissolution rate depending greatly on the concentration of adsorbed fluoride. Fluoride-substituted bridges and sites where aluminum atoms are bonded to several fluorides are especially important. High levels of fluoride are found in groundwater in many parts of the world, including India, China, and large areas of Africa,11 which can lead to serious health problems, including skeletal fluorosis and a number of other conditions.12 There are several possible methods for removal of fluoride in drinking-water treatment,13 the most promising of which is adsorption on substrates such as alumina.14 At high fluoride concentrations, formation of soluble Al−F complexes may reduce the effectiveness of defluoridation.7 In this paper, we consider the effect of quite low levels of added fluoride (up to 1 mM) on the regrowth of Al and Fe flocs after breakage. There is no suggestion that fluoride addition would be a feasible method for promoting floc regrowth in practice, but the results should improve our understanding of an important practical problem. The growth, breakage, and regrowth of flocs at different pH values were monitored by a continuous optical technique, and fluoride was added during floc breakage. Corresponding ζ potential measurements have been carried out to provide further information. On the basis of the results, a possible explanation for irreversible floc breakage will be considered.
MATERIALS AND METHODS Materials. Ferric chloride and aluminum chloride were used as coagulants. Stock solutions were prepared at 0.1 M and kept in a refrigerator at 4 °C. Sodium fluoride was prepared as a 0.1 M solution. All other reagents used were of analytical grade, and stock solutions were prepared in deionized (DI) water. Kaolin clay (Kaolin light, Fisher Chemicals, U.K.) was used as a model suspension in the tests. A total of 200 g of kaolin was dispersed in 500 mL of DI water in a high-speed blender. To achieve full dispersion, we raised the pH of the suspension to about 7.5 by adding NaOH solution. After blending at 800 rpm for about 30 min, the suspension was diluted to 1 L in DI water. Flocculation Tests. Flocculation was carried out in stirred vessels, following a procedure similar to that of Yu et al.15 Coagulant was added to a 50 mg/L kaolin suspension in a solution of 5 mM NaHCO3 with a predetermined amount of 0.1 M HCl to give the required final pH (6, 7, or 8). In all cases, the coagulant dosage was 0.1 mM Al or Fe. The suspension was stirred at 200 rpm for 1 min after dosing to give good mixing and then at 50 rpm for 10 min to allow flocs to form. The stirring speed was then increased to 200 rpm for 1 min, causing floc breakage, followed by stirring at 50 rpm so that regrowth of flocs could occur. Fluoride, at different concentrations, was added halfway through the 1 min breakage period. In one case, a longer breakage period of 10 min was used and fluoride added at different times during this period. Floc size was monitored throughout the process by the “turbidity fluctuation” technique (PDA 3000, Rank Brothers Ltd., U.K.). The sample was circulated from the stirred vessel at about 20 mL per minute by a peristaltic pump that was placed 1829
DOI: 10.1021/acs.est.5b05334 Environ. Sci. Technol. 2016, 50, 1828−1833
Article
Environmental Science & Technology
smoother trace, but fluoride addition of up to 1 mM gives no noticeable improvement in regrowth. All of the previous experiments were conducted with floc breakage periods of 1 min. Figure 3 shows the effect of adding
downstream of the monitor to minimize floc breakage problems. Readings were taken every second. The output of the PDA 3000 was converted to a “flocculation index” (FI), which is strongly correlated with floc size. The FI is the dimensionless ratio of the rms value of the fluctuations in intensity of light transmitted through the flowing sample to the average intensity16 and is expressed here as a percentage. This technique is not sufficiently sensitive for some hydroxide flocs, which have rather low lightscattering coefficients. However, with included particles, such as kaolin, flocs can easily be monitored.17 ζ potentials of samples, taken mostly just after floc breakage, were measured with a Zetasizer Nano ZS90 (Malvern, U.K.).
■
RESULTS Floc Growth, Breakage, and Regrowth. Kaolin suspensions (50 mg/L) were flocculated by AlCl3 and FeCl3 (both at a concentration of 0.1 mM) at final pH values of 6, 7, and 8. The results in Figures 1 and 2 show the changing FI value during the formation, breakage, and regrowth of hydroxide flocs. Figure 3. Flocculation of kaolin suspensions with AlCl3 at pH 7 with fluoride (1 mM) added at 0.5 and 9.5 min during short (1 min) and long (10 min) breakage periods, respectively. The corresponding results without fluoride addition are also shown.
fluoride during a 10 min breakage period. In this case Al flocs were formed at pH 7, and 1 mM F was added at 0.5 and 9.5 min into the long breakage period. For comparison, the effect of a 1 min breakage period, without fluoride, is also shown. Without fluoride, the longer breakage period gives reduced floc regrowth, as has been previously found.3 With fluoride additions (1 mM) at 0.5 and 9.5 min into the long breakage period, very different floc regrowth results are found. When fluoride was added near the beginning of the breakage period, there was only limited floc regrowth; slightly less than for the 10 min breakage without fluoride. However, fluoride addition near the end of the breakage period gave substantial regrowth, so that the final FI value was slightly greater than that before breakage. This result shows that the effect of fluoride addition is quite short-lived and is lost during a lengthy breakage period. A similar phenomenon has been observed when a small additional dosage of coagulant is added during floc breakage. This can greatly improve floc regrowth but not when added at the start of a long breakage period.18 A possible explanation of these fluoride effects in terms of soluble Al−F complexes will be considered in the Discussion of Floc Formation, Breakage, and Regrowth section. ζ Potentials. Figure 4 shows the effect of fluoride on the ζ potential of broken Al and Fe flocs at the same three pH values as for the previous experiments. Also included are results for the flocs after rapid mixing (i.e., without fluoride). Floc breakage causes the ζ potentials of Al and Fe flocs to become slightly more negative. Generally, the ζ potentials become more negative with increasing pH, as is expected for hydroxide precipitates. Fluoride gives more negative ζ potentials, but this effect is more pronounced with Al flocs than Fe flocs. The largest effect is with 1 mM F at pH 6 when the ζ potential of Al flocs changes from about +10 mV to −50 mV. Under the same conditions, the ζ potential of Fe flocs becomes only about 6 mV more negative as a result of fluoride addition. This is most likely due to the stronger adsorption of fluoride on Al flocs.
Figure 2. Flocculation of kaolin suspensions with FeCl3 at pH 7. Fluoride (0−1 mM) added during floc breakage.
During the growth phase, the FI value increases rapidly at first and then reaches a plateau, representing a maximum or steadystate floc size. This depends on the floc strength and the effective shear rate (determined by the stirring speed) and can be thought of in terms of a balance between floc growth and breakup. After 10 min of stirring at 50 rpm, the stirrer speed was increased to 200 rpm, which caused an immediate and rapid reduction in FI because of floc breakage. On restoring the stirring speed to 50 rpm, flocs are allowed to regrow and the FI rises, although not usually back to the value before breakage. The results for Al flocs are in Figure 1 and show very different behavior for the three pH values. At all pH values, without fluoride, floc regrowth is quite limited, but at pH 6, the addition of only 0.2 mM fluoride during floc breakage gives a marked improvement. However, with increasing fluoride concentration, the regrowth deteriorates. At the highest fluoride levels, the FI value increases initially after floc breakage but then starts to decline. At pH 7, increasing levels of fluoride cause more floc regrowth, and at 1 mM F, the FI value for the regrown flocs is greater than that before breakage. The results at pH 8 show that fluoride has almost no effect on floc regrowth. Similar experiments with Fe flocs show no significant effect on floc regrowth at any of the pH values. As an example, results at pH 7 are shown in Figure 2. In this case, flocs are considerably larger than for Al, as shown by the higher FI value. Because the flocs are larger, they are fewer in number, and this gives more scatter in the FI values. The regrown flocs show a lower FI and a 1830
DOI: 10.1021/acs.est.5b05334 Environ. Sci. Technol. 2016, 50, 1828−1833
Article
Environmental Science & Technology
Figure 4. Effect of fluoride on the ζ potentials of Fe and Al hydroxide precipitates (with included kaolin particles).
■
DISCUSSION OF FLOC FORMATION, BREAKAGE, AND REGROWTH It is clear from Figures 1 and 2 that floc formation readily occurs under conditions where the measured ζ potentials (before floc breakage) have absolute values up to 30 mV or more (Figure 4). For the low ionic strength solutions used in this work, such high ζ potentials would be expected to give strong repulsion between colloid particles. However, it should be noted that all of the flocs have included kaolin particles, which are negatively charged in the pH range 6−8. These particles are embedded within the flocs and cause the measured ζ potentials of the flocs to become more negative than those for the precipitate particles alone. Thus, the overall ζ potential may not be a good guide to interactions between flocs, which involve contacts between individual precipitate particles on the floc surfaces. When hydroxide flocs are broken in the absence of fluoride, they do not fully regrow (Figures 1 and 2). This is now a wellknown phenomenon,19 but so far, there has not been a satisfactory explanation of the effect. The flocs consist of aggregates of very small precipitate particles, and the binding between these particles determines the strength of the flocs. (The included kaolin particles have no significant effect on floc strength17). If the binding were solely due to physical interactions, such as van der Waals forces, then it is difficult to see why floc breakage should not be reversible. This suggests that some form of chemical interaction must be at least partly responsible for binding within the floc structure and that floc breakage causes some loss of “active” binding sites. When hydrolyzed metals in solution form precipitates, the initial step should be binding between individual metal ions, and it is thought that bonds are formed between OH2 and OH ligands in the inner coordination spheres of the metal ions. These groups form −OH·OH2− bridges between M(III) sites.20 These bridges (written as H3O2) can then release water to give OH bridges:
MOH(−1/2) + H+ ↔ MOH 2(+1/2)
The fractional charges arise from Pauling’s concept of bond valence, and it is assumed that the charge on a metal ion such as Al3+ is equally distributed between all of its six ligands. According to this model, only two types of surface group contribute to the surface charge, and the net charge depends on the relative proportions of these groups. Around neutral pH there are significant numbers of both OH and OH2 groups on Al and Fe (hydr)oxide surfaces, and, because these groups are oppositely charged, the formation of hydroxyl bridges between different particles, by a mechanism like that in eq 1, would be likely. This form of specific, chemical binding between precipitate particles would give enhanced floc strength but may also be responsible for the irreversible nature of floc breakage. Floc breakage involves rupture of interparticle bonds and, in the case of hydroxyl bridges, this would leave one surface OH group and a metal site on the other surface that is not fully coordinated. The latter would very rapidly react with water to give an OH2 surface group. Thus, the breakage process could be written: M−OH−M + H 2O → M−OH(−1/2) + MOH 2(+1/2) (3)
In this case, the surface groups remaining after breakage would be just the same as those present before formation of the hydroxyl bridge, and there would no reason why another bridge could not be formed. However, both of the surface groups created by breakage are subject to pH-dependent acid−base reactions according to eq 2. For instance, the MOH2 group could lose a proton: ≡MOH 2(+1/2) + OH− → ≡MOH(−1/2) + H 2O
(4)
The overall effect of bond breakage would then be to leave two surface OH groups, which could not form a hydroxyl bridge according to eq 1. Depending on pH, protonation of the M−OH group might occur giving M−OH2. In this case, the effect of bond breakage would be to leave two M−OH2 groups and again no hydroxyl bridge formation would be possible. So, floc breakage could give a change in the relative numbers of OH and OH2 groups, making hydroxyl bridge formation less likely. This is a possible explanation for irreversible floc breakage. Fewer hydroxyl bridges between broken floc fragments would mean weaker and hence smaller regrown flocs, as is usually observed. At higher pH, the deprotonation in eq 4 would be enhanced, and there should be still less opportunity for hydroxyl bridges to be formed between broken floc fragments; thus, less
≡M−OH + ≡M−OH 2 → ≡M−(H3O2 )−M≡ → ≡M−(OH)−M≡+H 2O
(2)
(1)
Because precipitated hydroxide particles have surface −OH and −OH2 groups, it is likely that these could form hydroxyl bridges between particles. According to the Charge Distribution - Multisite Complexation (CD-MUSIC) model, charging of oxide surfaces can be represented by the following equilibrium reaction,21 assuming singly coordinated oxygens: 1831
DOI: 10.1021/acs.est.5b05334 Environ. Sci. Technol. 2016, 50, 1828−1833
Article
Environmental Science & Technology
changes, but there is another possible explanation. As mentioned in the introduction, it is known that fluoride can promote the dissolution of Al (hydr)oxides and that this effect is more apparent with amorphous, rather than crystalline, materials. Al release by fluoride is important in soils and can result in significant increase in soluble Al in groundwater. Aluminum forms very strong complexes with fluoride, stronger than with any other halide. Unlike all other common ligands, F− binds about 10 times more strongly to Al(III) than to Fe(III).22 Complexes of the form AlFn are known, with n = 1−6. There are also ternary complexes with hydroxyl, Al(OH)mFn,, with m + n ≤ 4. With appropriate stability constants, it is possible to calculate the concentrations of the various complexes in equilibrium with the amorphous hydroxide precipitate over a range of pH values and total fluoride concentrations. Details of the calculations are given in the Supporting Information, and the results are shown in Figure 6a over a range of pH values and for a total fluoride concentration of 1 mM. The results show that, around neutral pH, there is a very significant increase in Al solubility, up to nearly 3 orders of magnitude. At low and high pH, Al solubility is due mainly to Al3+ and Al(OH)4−, respectively, and the fluoride complexes contribute very little to the total. The effect of total fluoride concentration on Al solubility at pH 7 is shown in Figure 6b. It is clear that, beyond about 0.5 mM F, Al solubility increases almost linearly with fluoride concentration. These calculations assume that all of the fluoride is in solution, either free or complexed, and take no account of adsorbed fluoride. They apply to solutions where Al solubility is just exceeded, so that there is very little precipitate present and hence very little adsorption. In the experiments reported here, fluoride was added after precipitate had formed, and fluoride adsorption would occur initially and eventually promote Al dissolution by mechanisms such as those proposed by Nordin et al.10 In some cases, the dissolution could be quite slow, but nevertheless, even if only a small amount of precipitate dissolved, new surface would be exposed and have more −OH2 groups; this could promote regrowth of broken flocs in a manner similar to that when small amounts of coagulant are added during the breakage process.4 This would explain the enhanced regrowth with fluoride addition at pH 7 (Figure 1). At pH 6, quite low fluoride concentrations can enhance floc regrowth to some extent, but with 1 mM of F,
regrowth should occur at higher pH. Analysis of the results in Figure 2 for Al flocs without fluoride addition supports this conclusion. From the FI values we can calculate a regrowth factor (RF), given by
RF =
F3 − F2 F1 − F2
(5)
where F1, F2, and F3 are the FI values for the flocs before breakage, after breakage, and after regrowth, respectively. Figure 5 shows the regrowth factors for Al flocs at pH 6, 7, and 8, without fluoride. There is a steady decrease of RF from 0.53 to
Figure 5. Regrowth factors for Al−kaolin flocs at different pH values (without fluoride).
0.38 over this pH range, which is consistent with the above hypothesis. It has been found4 that quite small amounts of additional coagulant added during floc breakage can greatly enhance regrowth of broken flocs. The most likely explanation is that fresh hydroxide precipitate is formed with −OH and −OH2 surface groups, which can form hydroxyl bridges according to eq 1. If floc breakage is continued for several minutes after the fresh coagulant is added, then there is no improvement in floc regrowth, probably due to the loss of active binding sites, as suggested above.
■
EFFECT OF FLUORIDE It is difficult to explain the effects of added fluoride on floc regrowth when fluoride is added during the breakage of Al flocs (Figure 1 and 3), in terms of fluoride adsorption and ζ potential
Figure 6. (a) Effect of 1 mM fluoride on the solubility of Al in equilibrium with amorphous Al(OH)3 precipitate. (b) Effect of fluoride concentration on Al solubility at pH 7. 1832
DOI: 10.1021/acs.est.5b05334 Environ. Sci. Technol. 2016, 50, 1828−1833
Article
Environmental Science & Technology there is a short period of floc regrowth and then a decrease in FI value. This might be a result of the higher solubility and more rapid dissolution of the amorphous precipitate at pH 6. Farrah et al.8 showed that at pH < 6 and a F/Al ratio of >2.5, most of an amorphous Al(OH)3 gel dissolved due to the formation of soluble Al−F complexes. In our experiments, with 1 mM F, there is a 10-fold excess over Al, so that significant dissolution of the precipitate would be expected at pH 6. This would give a decreasing floc size and FI value, as observed. The results in Figure 3 show that, at pH 7, the addition of 1 mM fluoride during floc breakage only enhances floc regrowth if it occurs near the end of the breakage period. If breakage is continued for several minutes after fluoride addition then there is no significant effect on regrowth. This suggests that the new surface exposed by dissolution becomes changed by floc breakage in such a way that the number of active groups is reduced. Because of the weaker interaction of fluoride with Fe(III), the solubility of amorphous ferric hydroxide is affected very little by low fluoride concentrations, which could explain the results in Figure 2 that show no significant effect on Fe floc regrowth. This work has provided further evidence that regrowth of broken hydroxide flocs can be improved by the generation of new hydroxide surface. In earlier work,18 this was achieved by the addition of small additional amounts of coagulant during the breakage process, producing fresh precipitate. Essentially, the same effect can be produced by dissolving some of the precipitated hydroxide from broken flocs to give fresh surface. In the present case, the dissolution occurred as a result of the formation of soluble Al−F complexes, but other methods might also be effective. Through a better understanding of the phenomenon of irreversible floc breakage, it may be possible to find practical ways of overcoming this problem.
■
(5) Omueti, J. A. I.; Jones, R. L. Fluoride adsorption by Illinois soils. J. Soil Sci. 1977, 28 (4), 564−572. (6) Fletcher, H. R.; Smith, D. W.; Pivonka, P. Modeling the sorption of fluoride onto alumina. J. Environ. Eng. 2006, 132 (2), 229−246. (7) Mitrovic, B.; Milacic, R. Speciation of aluminium in forest soil extracts by size exclusion chromatography with UV and ICP-AES detection and cation exchange fast protein liquid chromatography with ETAAS detection. Sci. Total Environ. 2000, 258 (3), 183−194. (8) Farrah, H.; Slavek, J.; Pickering, W. F. Fluoride interactions with hydrous aluminum-oxides and alumina. Aust. J. Soil Res. 1987, 25 (1), 55−69. (9) Harrington, L. F.; Cooper, E. M.; Vasudevan, D. Fluoride sorption and associated aluminum release in variable charge soils. J. Colloid Interface Sci. 2003, 267 (2), 302−313. (10) Nordin, J. P.; Sullivan, D. J.; Phillips, B. L.; Casey, W. H. Mechanisms for fluoride-promoted dissolution of bayerite beta-Al(OH) (3)(s) and boehmite gamma-AlOOH: F-19-NMR spectroscopy and aqueous surface chemistry. Geochim. Cosmochim. Acta 1999, 63 (21), 3513−3524. (11) George, S.; Pandit, P.; Gupta, A. B. Residual aluminium in water defluoridated using activated alumina adsorption - Modeling and simulation studies. Water Res. 2010, 44 (10), 3055−3064. (12) Mondal, P.; George, S. A review on adsorbents used for defluoridation of drinking water. Rev. Environ. Sci. Bio/Technol. 2015, 14 (2), 195−210. (13) Velazquez-Jimenez, L. H.; Vences-Alvarez, E.; Flores-Arciniega, J. L.; Flores-Zuniga, H.; Rangel-Mendez, J. R. Water defluoridation with special emphasis on adsorbents-containing metal oxides and/or hydroxides: A review. Sep. Purif. Technol. 2015, 150, 292−307. (14) Gong, W. X.; Qu, J. H.; Liu, R. P.; Lan, H. C. Adsorption of fluoride onto different types of aluminas. Chem. Eng. J. 2012, 189, 126− 133. (15) Yu, W.; Gregory, J.; Campos, L. C. Breakage and re-growth of flocs formed by charge neutralization using alum and polyDADMAC. Water Res. 2010, 44 (13), 3959−3965. (16) Gregory, J. Monitoring particle aggregation processes. Adv. Colloid Interface Sci. 2009, 147, 109−123. (17) Yu, W. Z.; Gregory, J.; Campos, L. C.; Graham, N. Dependence of floc properties on coagulant type, dosing mode and nature of particles. Water Res. 2015, 68, 119−126. (18) Yu, W. Z.; Gregory, J.; Campos, L. The effect of additional coagulant on the re-growth of alum-kaolin flocs. Sep. Purif. Technol. 2010, 74 (3), 305−309. (19) Jarvis, P.; Jefferson, B.; Parsons, S. A. Floc structural characteristics using conventional coagulation for a high doc, low alkalinity surface water source. Water Res. 2006, 40 (14), 2727−2737. (20) Rustad, J. R.; Casey, W. H. A molecular dynamics investigation of hydrolytic polymerization in a metal-hydroxide gel. J. Phys. Chem. B 2006, 110 (14), 7107−7112. (21) Hiemstra, T.; Van Riemsdijk, W. H. A surface structural model for ferrihydrite I: Sites related to primary charge, molar mass, and mass density. Geochim. Cosmochim. Acta 2009, 73 (15), 4423−4436. (22) Martin, R. B. Ternary complexes of Al3+ and F- with a third ligand. Coord. Chem. Rev. 1996, 149, 23−32.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b05334. Details of calculations and a table showing formation constants for complexes and other constants needed. (PDF)
■ ■
AUTHOR INFORMATION
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS This research was supported by a Marie Curie International Incoming Fellowship (FP7-PEOPLE-2012-IIF-328867) within the 7th European Community Framework Program.
■
REFERENCES
(1) Yu, W. Z.; Liu, T.; Gregory, J.; Li, G. B.; Liu, H. J.; Qu, J. H. Aggregation of nano-sized alum-humic primary particles. Sep. Purif. Technol. 2012, 99, 44−49. (2) Duan, J. M.; Gregory, J. Coagulation by hydrolysing metal salts. Adv. Colloid Interface Sci. 2003, 100, 475−502. (3) Yukselen, M. A.; Gregory, J. The reversibility of floc breakage. Int. J. Miner. Process. 2004, 73 (2), 251−259. (4) Yu, W. Z.; Gregory, J.; Campos, L. Breakage and Regrowth of AlHumic Flocs - Effect of Additional Coagulant Dosage. Environ. Sci. Technol. 2010, 44 (16), 6371−6376. 1833
DOI: 10.1021/acs.est.5b05334 Environ. Sci. Technol. 2016, 50, 1828−1833
