- Email: [email protected]
Ligand-promoted reductive cleaning of iron-fouled membranes from submerged membrane bioreactors
Journal of Membrane Science 545 (2018) 126–132
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Ligand-promoted reductive cleaning of iron-fouled membranes from submerged membrane bioreactors Zhenghua Zhanga,b, Mark W. Bligha, Xiu Yuana, T. David Waitea,
MARK
⁎
a
Water Research Centre, School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052, Australia Tsinghua-Kangda Research Institute of Environmental Engineering & Nano-Technology, Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055, Guangdong, China b
A R T I C L E I N F O
A B S T R A C T
Keywords: Ascorbic acid Citric acid Membrane cleaning Membrane fouling Ligand-promoted reductive dissolution
Four pathways for cleaning of iron oxide-fouled membranes, namely i) proton (acid)-assisted, ii) ligand-promoted (citric acid), iii) reductive (ascorbic acid), and iv) ligand-promoted reductive (ascorbic acid-citric acidmediated) were examined and their effectiveness were compared in this study. The cleaning effectiveness under oxic conditions followed the order: proton-assisted < ligand-promoted < reductive < ligand-promoted reductive. Iron oxide dissolution rate initially increased with increase of citrate concentration in the ascorbic acidcitric acid solutions but declined at higher citrate concentrations indicating that an intermediate citrate concentration was required for optimal cleaning. The mechanism of ligand-promoted reductive dissolution mediated by ascorbate and citrate was investigated through studies of dissolution behaviour under both oxic and anoxic conditions with citrate shown to have a mitigating effect on the consumption of oxygen, apparently by reducing the iron catalyzed oxidation of ascorbate. Kinetic modelling showed that the dynamics of dissolution could be reasonably well simulated with the inclusion of a surface > Fe(III)-citrate-Fe(II) ternary complex which facilitates the detachment of surface Fe(II). Use of dual reagents (ascorbic acid and citric acid) under oxic conditions is recommended for the cleaning of iron-fouled membranes in view of the extreme cleaning effectiveness though it is critical that cleaning conditions be carefully optimised.
1. Introduction Coagulants such as ferric chloride and aluminium sulphate are widely used for achieving phosphorus (P) removal from effluents of traditional biological unit operations and are now also being used in MBR plants [1–4]. High levels of iron dosing with Fe: P molar ratios of 2–4 are typical of the Fe(III) dosages used in full scale MBR plants in Australia for facilitation of sufficient and consistent P removal (typically 0.01–0.3 mg/L in effluents) in view of the sensitivity of receiving waters to algal growth [5,6]. However, severe membrane fouling resulted from high levels of Fe with amorphous ferric oxyhydroxide (AFO) particles and gelatinous assemblages containing Fe(III) bound to polysaccharide materials particularly responsible for gel layer formation and pore blockage, necessitating the application of an effective membrane chemical cleaning regime to remove iron species from the membrane [7–10]. Chemical reagents such as acids (hydrochloric, sulfuric, citric, oxalic, etc.), bases (caustic soda), oxidants (hypochlorite and hydrogen peroxide) and other chemicals (chelating agents, surfactants, etc.) have been widely used to remove materials from irreversibly fouled membranes [11] though little attention appears to have ⁎
Corresponding author. E-mail address: [email protected] (T.D. Waite).
http://dx.doi.org/10.1016/j.memsci.2017.09.059 Received 22 July 2017; Accepted 16 September 2017 Available online 20 September 2017 0376-7388/ © 2017 Elsevier B.V. All rights reserved.
been given to which of these, if any, are well suited to the cleaning of iron oxide fouled membranes. With regard to the dissolution of iron oxides, there are four general pathways in acidic aqueous suspensions: proton (acid)-assisted, ligandpromoted acid, reductive, and ligand-promoted reductive dissolution [12,13]. The first two pathways involve the adsorption of proton and ligand to the surface of iron oxides which results in weakening of the bonds in the proximity of a surface Fe(III) center followed by slow detachment of the surface Fe(III) to solution [12]. If the surface Fe(III) is reduced to surface Fe(II) by reductants, the detachment of surface Fe (II) into solution is much faster than that of surface Fe(III) due to the higher lability of the Fe(II)–O bond compared to the Fe(III)–O bond [12]. Reductive dissolution of Fe oxide minerals involves many steps, including reductant adsorption, electron transfer, detachment of oxidised reductant, and detachment of Fe(II) (the rate determining step) [14]. Previous workers have reported that the detachment of surface Fe (II) might be accelerated in the presence of a ligand with ligand-promoted reductive dissolution (i.e. ascorbate-EDTA; ascorbate-oxalate; dithionite-citrate) shown to be particularly effective in promoting fast dissolution of iron oxides [13,15,16], however, the mechanism
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
Nomenclature AFO Asc Cit C0 C
EDTA ethylenediaminetetraacetic acid f percentage of undissolved iron on the membrane (%) P phosphorus Fe(II)Cit complex of Fe(II) and citric acid Fe(III)Cit complex of Fe(III) and citric acid >Fe(III)CitFe(II) surface complex of Fe(III), citric acid and Fe(II)
amorphous ferric oxyhydroxide ascorbic acid (mM) citric acid (mM) the initial iron concentration on the membrane (mM) the released iron concentration (mM)
dominating such dissolution processes remains unclear. Citric acid, a tricarboxylic acid that forms strong solution complexes with iron, is not particularly effective when used as a single cleaning agent for inorganics removal [17] while ascorbic acid, an effective Fe (III) reducing agent, has been shown to be especially effective in removing iron species from iron oxide-fouled membrane surface [10,18]. However, ascorbate-mediated reductive dissolution under oxic conditions is much diminished compared to that under anoxic conditions, apparently due to the catalytic oxidation of ascorbate via heterogeneous re-oxidation of surface Fe(II) [18]. It has also been reported that the combination of a reductant and a strong chelating ligand is especially effective in promoting the dissolution of iron oxides due to the possible synergistic effects between these two agents [13]. For example, Banwart et al. [15] reported that in the presence of ascorbate, the adsorption of oxalate (a ligand capable of forming bidentate mononuclear surface complexes) leads to a significant increase in the rate of reductive dissolution of hematite even though the adsorption of oxalate displaces some of the ascorbate from the hematite surface. A similar synergistic effect of reductant and complexant has also been observed in the dissolution of goethite by dithionite and citrate or EDTA [16]. Therefore, it is reasonable to postulate that in a system with both ascorbic acid and citric acid, ligand-promoted reductive dissolution could also occur and be more effective than ascorbate-mediated reductive dissolution for iron oxides-fouled membranes. In this study, we investigated proton-assisted (acid), ligand-promoted (citric acid), reductive (ascorbic acid), and ligand-promoted reductive (ascorbic acid-citric acid-mediated) cleaning of Fe(III)-based foulants from membranes used in submerged membrane bioreactors to which ferric chloride had been added for facilitation of phosphorus removal. Detailed investigations of the ascorbic acid-citric acid-mediated reductive cleaning of the iron-fouled membranes were undertaken under both oxic and anoxic conditions. Based on the findings of these studies, a kinetic model was developed as an aid to elucidating the key factors controlling the effectiveness of the dual reagent cleaning process. This model may also assist in determining the conditions that are optimal for removal of iron foulants in practical cleaning procedures.
pH=4 proton-assisted dissolution
100
Cit 1mM ligand-promoted dissolution
80
Asc 1 mM
f, %
60 reductive dissolution
40
Asc 3 mM
20 0
Asc 10 mM Asc-10 mM + Cit 1 mM ligand-promoted reductive dissolution
0
10
20 30 40 Leaching time, h
50
Fig. 1. Cleaning effectiveness of iron-fouled membranes at pH 4 with i) hydrochloride acid, ii) citric acid (Cit) at concentration of 1 mM, iii) ascorbic acid (Asc) at concentrations of 1, 3 and 10 mM, and iv) Asc-Cit at concentrations of 10 mM of Asc and 1 mM of Cit. Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
2. Materials and methods 2.1. Formation of iron-fouled small membrane modules Accelerated fouling and cleaning studies were conducted using a small membrane module containing four polyvinylidene fluoride (PVDF) hollow fibre membranes (Beijing Origin Water, China) with nominal pore size 0.1–0.3 µm, diameter 0.24 cm, length 15 cm and total surface area 0.0044 m2. The small module was oriented vertically in the membrane bioreactor to which a 37.2 mM ferric iron solution (Fe/P molar ratio of 4) was dosed to facilitate phosphorus removal [9]. The small membrane module was operated at a constant flux of 30 L/ m2 h (with no relaxation) for a period of two weeks before cleaning studies were undertaken.
Fig. 2. Cleaning effectiveness of iron-fouled membranes at pH 4 in ascorbic acid (Asc)citric acid (Cit) solutions with Asc concentration of 10 mM and Cit concentrations of 50, 100, 500, 1000 (a), 3000, 5000 µM (b). Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
immersed in 50 mL of one of the cleaning solutions from the following experimental configurations:
2.2. Cleaning studies of iron-fouled membranes
1. Hydrochloride acid; pH 4 2. Citric acid (Cit); pH 4; 1 and 10 mM
After operation for two weeks, the fouled membranes were rinsed with MQ water and then cut into a number of 5 cm length pieces and 127
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
DO concentration, mg/L
8
pH 4 with membrane Asc 10mM Asc 10mM + Cit 50 µM Asc 10mM + Cit 100 µM Asc 10mM + Cit 500 µM Asc 10mM + Cit 1000 µM
6 4 2 0
0
20
40 60 80 Leaching time, min
100
Fig. 5. DO concentration during the membrane cleaning process by Asc-Cit-mediated solutions with 10 mM Asc and 50–1000 µM Cit at pH 4 under oxic conditions.
Fig. 3. Cleaning effectiveness of iron-fouled membranes at pH 4 in ascorbic acid (Asc)citric acid (Cit) solutions with Asc concentration of 3 mM and Cit concentrations of 100, 500, 1000 µM (a) and Asc concentration of 10 mM and Cit concentrations of 100, 1000 µM (b). Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
Fig. 6. Conceptual model of the Asc-Cit-mediated reductive dissolution of iron oxides from fouled membranes at pH 4 under both oxic and anoxic conditions. > Fe(III) – surface Fe(III) species; > Fe(II) – surface Fe(II) species; Fe(II) – solution Fe(II) species; O2 – dissolved oxygen. Numbers represent reactions as numbered in Table 1.
3. Ascorbic acid (Asc); pH 4; 1, 3, 10 mM 4. Ascorbic acid-citric acid (Asc-Cit); pH 4; 3 and 10 mM-Asc and 50 µM to 5 mM-Cit
Fig. 4. Cleaning effectiveness of iron-fouled membranes at pH 4 ascorbic acid (Asc) with concentrations of 10 (a) and 3 mM (b) for i) oxic conditions, ii) oxic conditions with 1000 µM citric acid (Cit), and iii) anoxic conditions with 1000 µM Cit. Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
The effectiveness of the above four solutions in removing iron from the membrane pieces were systematically compared with the most effective cleaning solution investigated in further detail. The pH of the citric acid (purity ≥ 99.5%, Sigma), ascorbic acid (purity ≥ 99%, Sigma), and Asc-Cit solutions was carefully controlled at 4 with an 128
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
Table 1 Modelled reactions and associated rate constants for the Asc-Cit-mediated ligand-promoted reductive dissolution of iron oxides from fouled membranes at pH 4 under both oxic and anoxic conditions. No.
Reaction
Rate constant (M−1 s−1)a
Note
1 2 3 4 5 6 7 8 9 10 11 12 13
> Fe(III) + Asc → > Fe(III)Asc > Fe(III)Asc → > Fe(II) + Asc*b Asc* + Asc* → Asc + Asc**c > Fe(II) → Fe(II) + (1-c) > Fe(III)d > Fe(II) + O2 → > Fe(III)e Fe(II) + Cit → Fe(II)Cit > Fe(II) + Cit → > Fe(II)Cit > Fe(II)Cit → Fe(II)Cit + (1-c) > Fe(III) > Fe(III) + Cit → > Fe(III)Cit > Fe(III)Cit → Fe(III)Cit + (1-c) > Fe(III) > Fe(III) + Fe(II)Cit → > Fe(III)CitFe(II) > Fe(III)CitFe(II) → > Fe(II)CitFe(III) > Fe(II)CitFe(III) → Fe(II) + Fe(III)Cit + (1-c) > Fe(III)
k1 = 4.85 × 10−2 k2 = 4.92 × 1010 (s−1) k3 = 7.12 × 108 k4 = 3.56 × 10−4 (s−1) k5 = 4.98 k6 = 500 k7 = 0.01 k 8 = 2.0 × 10−3 (s−1) k 9 = 0.05 k10 = 1.5 × 10−4 (s−1) k 11 = 35 k12 ≥ 1.0 × 104 (s−1) k13 = 1.0 × 10−3 (s−1)
[18] [18] [18] [18] [18] [20] Fitting parameter Fitting parameter Fitting parameter Fitting parameter Fitting parameter Not sensitive when ≥ 1.0 × 104 Fitting parameter
Note that all the important reactions based on the schematic in Fig. 6 are shown in Table 1 with reactions 1–5 taken from our previous study on Asc-mediated reductive dissolution of iron- fouled membranes [18]. Reactions 6–10 represent solution phase Fe(II)Cit and surface Fe(II)Cit and Fe(III)Cit complexes formation and corresponding dissolution. Reactions 11–13 describe the formation and detachment of surface Fe(III)CitFe(II) complex. a Unless noted otherwise. b Asc* represents ascorbate radical. c Asc** represents dehydroascorbic acid and reaction 3 represents the disproportionation of two molecules of Asc* leading to the regeneration of one molecule of Asc. d c is the ratio of surface iron binding sites to total iron on the membrane and c = 0.6 was used here [18]. e Model input for the dissolved oxygen concentrations are 0.243 mM (saturation concentration under atmosphere pressure) under oxic conditions and zero under anoxic conditions, respectively.
membranes were not stirred during the cleaning operation.
appropriate buffer (sodium acetate (8.2 mM)/acetic acid (1.8 mM)). We confirmed that the sodium acetate/acetic acid buffer negligibly influences the iron oxyhydroxide foulant with less than 2% removal of iron from the iron-fouled membrane surface after 12 h reaction [18]. Fe(III)based materials were the major foulants on these membrane pieces as indicated by SEM and EDX analysis shown in Fig. s1, presumably resulting from the rapid precipitation of amorphous ferric oxyhydroxides (AFO) after Fe addition with AFO depositing on the membrane surface [18]. The effectiveness of the various cleaning solutions in removing Febased foulants was determined qualitatively by examining the membranes using field emission scanning electron microscopy (FESEM, Hitachi S3400-I, Japan) (following the coating of samples with chromium) and energy dispersive x-ray spectroscopy (EDX) (Bruker AXS Microanalysis GmbH Berlin, Germany). A more quantitative measure of rate and extent of removal of Fe-containing foulants from the membranes was obtained by examining the increase in concentrations of Fe in solution as a function of exposure time of the membrane to the cleaning solution (as described below). Aliquots (5 mL) of the membrane-leachate solution were removed after various intervals of immersion of the fouled membrane pieces in cleaning solution for measurement of total Fe by inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Agilent 7300, USA). After 48 h cleaning, each fouled membrane was immersed for a further 24 h in 50 mL of 300 mM sodium dithionite in order to remove any residual iron remaining on the membrane. Based on this technique, a measure of the initial iron concentration on each membrane (C0) was taken to equal the iron concentration in the wash solution after Asc cleaning plus the residual iron concentration in the dithionite wash solution. The cleaning effectiveness of different agents was quantified by comparing the undissolved iron fraction on the membrane (f) (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration).
2.2.2. Asc-Cit-mediated dissolution under anoxic conditions Asc-Cit-mediated solutions with i) Asc concentration of 10 mM and Cit concentration of 1 mM, and ii) Asc concentration of 3 mM and Cit concentration of 1 mM at controlled pH 4 were applied for periods of up to 4 h. Asc-Cit-mediated cleaning solutions were bubbled with argon gas for 10 min before membrane immersion with bubbling continued during the cleaning experiments in order to ensure the removal of oxygen. 2.3. Kinetic modelling The MS-Excel based kinetic modelling programme Kintecus [19] was used to determine the time varying concentrations of reactants and products for a set of reactions hypothesized to occur during the membrane cleaning process. Rate constants for the various reactions included in the analysis were obtained, for the most part, from the literature with the set of differential equations corresponding to the rate expressions solved numerically. It should be noted that the modelling is not a data fitting exercise but, rather, a means of assessing the relative importance of particular proposed reactions. 3. Results 3.1. Dissolution of iron-fouled membranes by four pathways Results for cleaning effectiveness of iron-fouled membranes at pH 4 using four different pathways are presented in Fig. 1. Negligible effect (~ 5%) in terms of Fe removal by hydrochloride acid (pH 4) was achieved after 48 h soaking (Fig. 1), while excellent cleaning performance (with > 99% of Fe removed) was achieved with 10 mM ascorbic acid (Asc) after 48 h soaking (Fig. 1), most of which occurred in the first 6 – 12 h. In contrast, soaking in 1 mM citric acid for 48 h resulted in a removal of only 26% of Fe, whereas even a solution containing 10 mM (about 0.2% in weight) citric acid is only capable of dissolving 39.9% of Fe after 48 h soaking (Fig. s2), which is much less effective compared to Asc at the same concentration and pH (> 99%) (Fig. 1). As can be seen from the SEM and EDX results in Fig. s3, soaking with 0.1 mM citric acid for 48 h induced negligible removal of the Fe-rich gel layer with a
2.2.1. Asc-Cit-mediated dissolution under oxic conditions The membrane pieces were immersed in Asc-Cit solutions with i) Asc concentration of 10 mM and Cit concentrations of 50, 100, 500, 1000, 3000, 5000 µM, and ii) Asc concentration of 3 mM and Cit concentrations of 100, 500, 1000 µM at controlled pH 4 for periods of up to 4 h. Note that “oxic condition” indicates that the cleaning solutions were open to the atmosphere. The solutions containing the fouled 129
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
Fig. 7. Cleaning effectiveness of iron-fouled membranes at pH 4 in Asc-Cit solutions with Asc concentrations of 3 and 10 mM and Cit concentrations of 50 µM to 5 mM under both oxic and anoxic conditions. Symbols represent the experimental data and solid lines represent the model fit using reactions 1–13 (Table 1) with c = 0.6 representing the ratio of surface iron binding sites to total iron on the membrane. Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
while only gradual dissolution was observed in the presence of Asc alone under oxic conditions (Figs. 2 and 3). At 10 mM Asc, the percentage of iron dissolved from the membrane increased with increase in Cit concentrations from 50 µM to 1000 µM (Fig. 2a), while further increase of Cit concentrations from 1 mM to 3 mM resulted in a decrease in effectiveness of iron oxide dissolution (Fig. 2b). Similarly, at a lower Asc concentration of 3 mM, the dissolution effectiveness increased with increase in Cit concentrations from 100 µM to 1000 µM, though the dissolution performances at Cit concentrations of 500 and 1000 µM are almost identical (Fig. 3a). A similar synergistic effect was observed when ethylenediaminetetraacetic acid (EDTA) was involved in reductive dissolution of magnetite by Asc under acid conditions (pH 3.5), where the dissociation rate reached a maximum at an intermediate concentration of EDTA [13]. This may suggest that at higher concentrations, citrate competes with ascorbate to bind surface sites yielding a surface Fe(III) complex with reduced reactivity and resultant
thick deposit remaining on the membrane surface while the thick Ferich gel layer disappeared after 48 h soaking with 10 mM Asc (pH 4). Extremely efficient iron removal by ligand-promoted reductive dissolution was achieved with 10 mM Asc and 1 mM Cit (pH 4) with > 96% of Fe removed only after 4 h soaking (Fig. 1). In light of these results, an apparent synergistic effect of Asc and Cit appears to be occurring in ligand-promoted reductive dissolution with further investigation of this phenomenon provided below. 3.2. Dissolution of iron-fouled membranes by Asc-Cit-mediated pathway Further experiments have been conducted to investigate the dual reagent Asc-Cit-mediated dissolution of iron-fouled membranes with results shown in Figs. 2 and 3. As can be seen from these results, AscCit-mediated dissolution proceeds in a biphasic manner, with an initial lag period followed by a much faster phase of dissolution over time, 130
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
(II) and/or promoting rapid detachment of Fe(II) thereby reducing the opportunity for heterogeneous oxidation. During the lag period DO levels are high enough to compete with citrate and enable Fe(II) reoxidation to occur, however, when DO levels decline sufficiently, citrate is able to prevent Fe(II) re-oxidation. Further, according to previous studies, the sudden increase in the dissolution rate following the lag period suggests a switch in the system resulting from some process that involves a positive feedback loop. One possibility is that the ligandpromoted reductive dissolution process involves the Fe(II)-citrate complex, which acts as a more effective reductant than Asc, possibly via formation of a > Fe(III)-citrate-Fe(II) ternary complex. Such a process is consistent with the Fe(II)L-mediated process suggested by Afonso et al. [13] who attributed the observed ‘induction period’ to the steady accumulation of Fe(II)L until a point was reached where reduction via Fe(II)L becomes significant. The process is self-reinforcing since the more Fe(II) that is solubilized, the greater the reduction rate and production of solution phase Fe(II).
lower dissolution rate. It is also worth mentioning that though the best dissolution efficiency was reached at Asc-10 mM and Cit-1 mM in this study, we have also found that almost identical dissolution effectiveness were achieved at Asc-10 mM, Cit-100 µM and Asc-3 mM, Cit-1000 µM (Fig. 3b). Asc-Cit-mediated reductive dissolution under anoxic conditions was also investigated with results shown in Fig. 4. Interestingly, under anoxic conditions, the biphasic behaviour observed during dissolution under oxic conditions (Figs. 2 and 3) has been replaced by an initially more rapid dissolution process with no evidence of any lag though similar extents of dissolution have been observed under both oxic and anoxic conditions after approximately 4 h of leaching (Fig. 4). 3.3. DO concentration during membrane cleaning process Dissolved oxygen (DO) concentrations during the membrane cleaning process in Asc-Cit solutions under oxic conditions are shown in Fig. 5. O2 concentrations in the ascorbate solution quickly decreased in the initial 10 min followed by a slower gradual decline over the rest of the experimental period (90 min). Similar trends were evident when low concentrations of citrate co-existed with 10 mM ascorbate though the rate of O2 consumption was markedly less at citrate concentrations of 500 and 1000 µM.
4.2. Modelling of ligand-promoted reductive dissolution A schematic of the hypothesized key processes involved in citrateenhanced ascorbate-mediated reductive dissolution is shown in Fig. 6. We have constructed a kinetic model of these key processes with citrate-induced enhancement of the surface Fe(II) detachment hypothesized to be promoted by the formation of a surface Fe(III)-citrateFe(II) ternary complex. All the important reactions shown in the schematic in Fig. 6 are presented in Table 1 with reactions 1–5 taken from our previous study on Asc-mediated reductive dissolution of iron-fouled membranes [18]. Reactions 6–10 represent solution phase Fe(II)Cit and surface Fe(II)Cit and Fe(III)Cit formation and corresponding dissociation processes. Reactions 11–13 describe the formation and detachment of a ternary surface Fe(III)CitFe(II) complex. Other reactions, such as the oxidation of > Fe(II)Cit species by O2, and the reduction of solution Fe(III)Cit species by ascorbic acid, were shown to be unimportant and thus neglected with variation of the rate constants of these reactions over a reasonable range leading to minimal impact on model output. Model results for all experimental data including 3 and 10 mM Asc and 50 µM to 5 mM Cit under both oxic and anoxic conditions are shown in Fig. 7 with these results reasonably describing all the experimental data obtained. At the reaction rate constants proposed here, the formation of the ternary surface complex Fe(III)CitFe(II) occurs at a relatively fast rate (reaction 11) and, as such, is central to the proposed dissolution process. Similar ternary species have been proposed in previous studies. For example, Sulzberger et al. reported that the occurrence of > Fe(III)oxalate-Fe(II) species accounted for the catalytic effect of Fe(II) in the dissolution of hematite and proposed that besides oxalate, malonate and citrate are also suitable bridging ligands for ternary complex formation with Fe(II) and Fe(III) (hydr)oxides [12]. Borghi et al., in studies of the dissolution of magnetite in solutions containing EDTA and Fe(II), suggested that the adsorption of Fe(II)EDTA2– onto > Fe(III) to form > Fe(III)-EDTA-Fe(II) surface species is highly favorable and should proceed at a high rate (on the order of 1250 mol–1 L) [21]. In view of these earlier studies, it would seem reasonable to hypothesize that the > Fe(III)CitFe(II) ternary surface complex will be formed at a rate considerably faster than that of the binary surface complexes (Table 1). It should be noted however that we provide no direct evidence for formation of the ternary species other than that its formation is consistent with the findings of other studies and accounts adequately for the observed iron oxide dissolution behaviour. It can also be deduced from the model that the lag period in dissolution under oxic conditions (especially at higher concentrations of Cit) was due to the formation of the ternary surface complex > Fe (III)CitFe(II) (reaction 11) which, in the first instance, hinders iron oxide dissolution by 1) returning solution Fe(II) species to the surface via complex formation between Cit and dissolved Fe(II) (reaction 6)
4. Discussion 4.1. Effect of citrate and DO on ligand-promoted reductive dissolution In our previous study, ascorbate-mediated reductive dissolution under oxic conditions has been found to be much reduced compared to that under anoxic conditions, apparently due to the catalytic oxidation of ascorbate via heterogeneous re-oxidation of surface Fe(II) [18]. In this study, citrate has been shown to almost completely alleviate the retarding effects of DO on dissolution for both low and high concentrations of Asc under oxic conditions (Fig. 2a and Fig. 3a). The mechanism by which citrate enhances the reductive dissolution of iron oxides under oxic conditions is not clear, though based on previous studies, it could be attributed to either the enhancement of Fe(II) detachment [15] or enhanced Fe(III) reduction as a result of formation of the Fe(II)Cit complex [13]. Meanwhile, for a given concentration of Asc, there is an optimal concentration of Cit above which the dissolution rate starts to decrease (Fig. 2b and Fig. 3a). Additionally, a notable difference in the effect of citrate between oxic and anoxic conditions is the presence of an initial lag period under oxic conditions during which the dissolution rate is similar to that in the absence of citrate (Fig. 4). At the end of the lag period, the rate increases dramatically until dissolution is nearly equivalent to the anoxic case (Fig. 4). This lag effect prior to enhanced dissolution has also been observed under oxic conditions with addition of phenanthroline, a strong ligand for Fe(II) [18]. As such, there is clearly some interplay between O2 and citrate in the ligand-promoted reductive dissolution process with the presence of citrate somehow mitigating the effects of O2 on reductive dissolution, although only after an initial lag period. One possibility is that citrate complexation of surface Fe(II) may be reducing access of O2 to surfacelocated Fe(II) thereby slowing heterogeneous oxidation of Fe(II) and the associated catalytic oxidation of ascorbate. Measurements of DO concentrations also provide some insight into possible mechanisms of citrate-mediated reductive dissolution and reasons for the occurrence of the lag period. In the absence of citrate, DO is rapidly consumed as a result of the heterogeneous re-oxidation of Fe(II) resulting in low steady state DO concentrations (Fig. 5) [18]. However, in the presence of citrate, O2 consumption is much slower and only reaches DO levels equivalent to that where citrate is absent after a period that is, interestingly, equivalent to the lag period (Fig. 5). Under oxic conditions, the most likely action of citrate is the prevention of Fe (II) oxidation either by competing with DO for adsorption to surface Fe 131
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
surface. 4. Use of dual reagents ascorbic acid and citric acid under oxic conditions is recommended for the cleaning of iron-fouled membranes in view of the extreme cleaning effectiveness. Further analysis of cleaning effectiveness and cost of full-scale implementation of ascorbate-citrate-mediated cleaning is required before definitive recommendations in terms of optimal reagent dose and cleaning durations for removal of iron-containing foulants can be made.
with this species further complexed by surface Fe(III) to generate the ternary surface complex at a relatively fast rate (reaction 11); and 2) capturing surface Fe(III) and forming the ternary surface complex (reaction 11) with O2 acting in a facilitating role by providing more surface Fe(III) species via heterogeneous re-oxidation of surface Fe(II) (reaction 5). Though slow formation of > Fe(II)Cit (reaction 7) could also contribute to the lag period by partially circumventing the direct detachment via reaction 4, this pathway is less important due to the fast oxidation of surface Fe(II) species by oxygen (reaction 5). On the other hand, once the concentration of the ternary surface complex reaches a certain level, it will quickly undergo internal electron transfer followed by dissociation into solution Fe(II) and Fe(III) species with this process accounting for the fast dissolution after the induction period. Further, the adverse impact on dissolution effectiveness exerted by increasing the Cit concentrations to very high levels (Fig. 2b and Fig. 3a) could also be explained by the model. More specifically, since reaction 1 and reaction 9 proceed with similar rates, increasing the concentrations of Cit (to levels comparable to Asc) will increase the importance of reaction 9 thereby enabling it to compete with reaction 1, yielding more surface Fe(III)Cit complexes which further dissociate at a lower rate (reaction 10) compared to that of surface Fe(III)Asc complexes. A similar conclusion was reached when EDTA and Asc were employed for reductive dissolution of magnetite at pH 3.5, where the dissociation rate initially increased then decreased on continuing increase in EDTA concentration with this behaviour recognised to occur because EDTA outcompetes Asc for surface Fe(III) sites yielding surface-located Fe(III) EDTA with reduced reactivity and resultant lower dissolution rate at high EDTA concentrations.
Acknowledgements Funding provided by the Australian Research Council, Beijing Origin Water, Sydney Water Corporation and Water Research Australia through ARC Linkage Grant (LP100100056) is greatly acknowledged. The authors would also like to acknowledge Beijing Origin Water Technology for provision of hollow fibre membranes. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2017.09.059. References [1] R.I. Sedlak, Phosphorus and Nitrogen Removal from Municipal Wastewater: Principles and Practice, Lewis Publishers, Florida, USA, 1991. [2] T.D. Waite, Challenges and opportunities in the use of of iron in water and wastewater treatment, Environ. Sci. Bio/Technol. 1 (2002) 9–15. [3] J. Kim, Q. Deng, M.M. Benjamin, Simultaneous removal of phosphorus and foulants in a hybrid coagulation/membrane filtration system, Water Res. 42 (2008) 2017–2024. [4] A.H. Caravelli, E.M. Contreras, N.E. Zaritzky, Phosphorous removal in batch systems using ferric chloride in the presence of activated sludges, J. Hazard. Mater. 177 (2010) 199–208. [5] Metcalf, Eddy, G. Tchobanoglous, F. Burton, H.D. Stensel (Eds.), Wastewater Engineering: Treatment and Reuse, 4th ed., McGraw-Hill, New York, 2003. [6] M. Manzouri, H.K. Shon, Rectification methods for the fouling of ultrafiltration hollow fibre membranes as a result of excessive soluble iron, Desalin. Water Treat. 32 (2011) 437–444. [7] Y. Wang, G.L. Leslie, T.D. Waite, Impact of iron dosing of membrane bioreactors on membrane fouling, Chem. Eng. J. 252 (2014) 239–248. [8] Y. Wang, K.H. Tng, H. Wu, G.L. Leslie, T.D. Waite, Removal of phosphorus from wastewaters using ferrous salts – a pilot scale membrane bioreactor study, Water Res. 57 (2014) 140–150. [9] Z.H. Zhang, Y. Wang, G.L. Leslie, T.D. Waite, Effect of ferric and ferrous iron addition on phosphorus removal and fouling in submerged membrane bioreactors, Water Res. 69 (2015) 210–222. [10] Z.H. Zhang, M.W. Bligh, Y. Wang, G.L. Leslie, H. Bustamante, T.D. Waite, Cleaning strategies for iron-fouled membranes from submerged membrane bioreactor treatment of wastewaters, J. Membr. Sci. 475 (2015) 9–21. [11] Z. Wang, J. Ma, C.Y. Tang, K. Kimura, Q. Wang, X. Han, Membrane cleaning in membrane bioreactors: a review, J. Membr. Sci. 468 (2014) 276–307. [12] B. Sulzberger, D. Suter, C. Siffert, S. Banwart, W. Stumm, Dissolution of Fe(III) (hydr)oxides in natural waters: laboratory assessment on the kinetics controlled by surface coordination, Mar. Chem. 28 (1989) 127–144. [13] M.D. Afonso, P.J. Morando, M.A. Blesa, S. Banwart, W. Stumm, The reductive dissolution of iron-oxides by ascorbate – the role of carboxylate anions in accelerating reductive dissolution, J. Colloid Interface Sci. 138 (1990) 74–82. [14] D. Suter, S. Banwart, W. Stumm, Dissolution of hydrous iron (III) oxides by reductive mechanisms, Langmuir 7 (1991) 809–813. [15] S. Banwart, S. Davies, W. Stumm, The role of oxalate in accelerating the reductive dissolution of hematite (α-Fe2O3) by ascorbate, Colloids Surf. 39 (1989) 303–309. [16] E.H. Rueda, M.C. Ballesteros, R.L. Grassi, M.A. Blessa, Dithionite as a dissolving reagent for goethite in the presence of EDTA and citrate: application to soil analysis, Clays Clay Miner. 40 (1992) 575–585. [17] T.D. Waite, F.M.M. Morel, Photoreductive dissolution of colloidal iron oxide: effect of citrate, J. Colloid Interface Sci. 102 (1984) 121–137. [18] Z.H. Zhang, M.W. Bligh, T.D. Waite, Ascorbic acid-mediated reductive cleaning of iron-fouled membranes from submerged membrane bioreactors, J. Membr. Sci. 477 (2015) 194–202. [19] James C. Ianni, Kintecus, Windows Version 5.20, 2014. 〈www.kintecus.com〉. [20] A.L. Rose, T.D. Waite, Kinetics of iron complexation by dissolved natural organic matter in coastal waters, Mar. Chem. 84 (1–2) (2003) 85–103. [21] E.B. Borghi, A.E. Regazzoni, A.J.G. Maroto, M.A. Blesa, Reductive dissolution of magnetite by solutions containing EDTA and FeII, J. Colloid Interface Sci. 130 (1989) 299–310.
5. Conclusion In this study, four iron oxide dissolution pathways: i) proton (acid)assisted, ii) ligand-promoted (citric acid), iii) reductive (ascorbic acid), and iv) ligand-promoted reductive (citric acid-ascorbic acid) for removal of iron foulants from iron oxide-fouled membranes from submerged membrane bioreactors under both oxic and anoxic conditions were investigated and their dissolution effectiveness were compared. A kinetic model for the ascorbate-citrate-mediated reductive dissolution of ferric iron from iron oxide-fouled membranes has been developed as an aid to determining the major pathways involved in the dissolution process. The following conclusions can be drawn: 1. Among the four general pathways for release of iron from iron oxide-fouled membranes, the cleaning effectiveness under oxic conditions followed the order: proton-assisted < ligand-promoted < reductive < ligand-promoted reductive dissolution. After 48 h soaking, > 99% of iron foulant was removed by 10 mM Asc (pH 4) while > 96% of the foulant was removed after only 4 h soaking by dual reagents of 10 mM Asc and 1 mM citric acid (pH 4). 2. The dissolution effectiveness increased with increase in citrate concentration in the ascorbate-citrate solutions though there was an optimal intermediate citrate concentration after which the dissolution performance decreased with further increase in citrate concentrations. Under oxic conditions, an initial lag period followed by a sharp increase in dissolution rate was observed while no lag in dissolution was evident under anoxic conditions for the ascorbatecitrate-mediated reductive dissolution process. The dissolution performance at the latter stage under oxic conditions was comparable to that under anoxic conditions. 3. The presence of O2 reduced ascorbic acid cleaning effectiveness of iron-fouled membranes due to the Fe(III)-catalyzed oxygen oxidation of ascorbate. It appears likely that in spite of the initial induction period, citric acid could completely alleviate the retarding effects of O2 on the dissolution because of the formation of a ternary surface complex > Fe(III)CitFe(II) which readily detaches from the
132
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Ligand-promoted reductive cleaning of iron-fouled membranes from submerged membrane bioreactors Zhenghua Zhanga,b, Mark W. Bligha, Xiu Yuana, T. David Waitea,
MARK
⁎
a
Water Research Centre, School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052, Australia Tsinghua-Kangda Research Institute of Environmental Engineering & Nano-Technology, Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055, Guangdong, China b
A R T I C L E I N F O
A B S T R A C T
Keywords: Ascorbic acid Citric acid Membrane cleaning Membrane fouling Ligand-promoted reductive dissolution
Four pathways for cleaning of iron oxide-fouled membranes, namely i) proton (acid)-assisted, ii) ligand-promoted (citric acid), iii) reductive (ascorbic acid), and iv) ligand-promoted reductive (ascorbic acid-citric acidmediated) were examined and their effectiveness were compared in this study. The cleaning effectiveness under oxic conditions followed the order: proton-assisted < ligand-promoted < reductive < ligand-promoted reductive. Iron oxide dissolution rate initially increased with increase of citrate concentration in the ascorbic acidcitric acid solutions but declined at higher citrate concentrations indicating that an intermediate citrate concentration was required for optimal cleaning. The mechanism of ligand-promoted reductive dissolution mediated by ascorbate and citrate was investigated through studies of dissolution behaviour under both oxic and anoxic conditions with citrate shown to have a mitigating effect on the consumption of oxygen, apparently by reducing the iron catalyzed oxidation of ascorbate. Kinetic modelling showed that the dynamics of dissolution could be reasonably well simulated with the inclusion of a surface > Fe(III)-citrate-Fe(II) ternary complex which facilitates the detachment of surface Fe(II). Use of dual reagents (ascorbic acid and citric acid) under oxic conditions is recommended for the cleaning of iron-fouled membranes in view of the extreme cleaning effectiveness though it is critical that cleaning conditions be carefully optimised.
1. Introduction Coagulants such as ferric chloride and aluminium sulphate are widely used for achieving phosphorus (P) removal from effluents of traditional biological unit operations and are now also being used in MBR plants [1–4]. High levels of iron dosing with Fe: P molar ratios of 2–4 are typical of the Fe(III) dosages used in full scale MBR plants in Australia for facilitation of sufficient and consistent P removal (typically 0.01–0.3 mg/L in effluents) in view of the sensitivity of receiving waters to algal growth [5,6]. However, severe membrane fouling resulted from high levels of Fe with amorphous ferric oxyhydroxide (AFO) particles and gelatinous assemblages containing Fe(III) bound to polysaccharide materials particularly responsible for gel layer formation and pore blockage, necessitating the application of an effective membrane chemical cleaning regime to remove iron species from the membrane [7–10]. Chemical reagents such as acids (hydrochloric, sulfuric, citric, oxalic, etc.), bases (caustic soda), oxidants (hypochlorite and hydrogen peroxide) and other chemicals (chelating agents, surfactants, etc.) have been widely used to remove materials from irreversibly fouled membranes [11] though little attention appears to have ⁎
Corresponding author. E-mail address: [email protected] (T.D. Waite).
http://dx.doi.org/10.1016/j.memsci.2017.09.059 Received 22 July 2017; Accepted 16 September 2017 Available online 20 September 2017 0376-7388/ © 2017 Elsevier B.V. All rights reserved.
been given to which of these, if any, are well suited to the cleaning of iron oxide fouled membranes. With regard to the dissolution of iron oxides, there are four general pathways in acidic aqueous suspensions: proton (acid)-assisted, ligandpromoted acid, reductive, and ligand-promoted reductive dissolution [12,13]. The first two pathways involve the adsorption of proton and ligand to the surface of iron oxides which results in weakening of the bonds in the proximity of a surface Fe(III) center followed by slow detachment of the surface Fe(III) to solution [12]. If the surface Fe(III) is reduced to surface Fe(II) by reductants, the detachment of surface Fe (II) into solution is much faster than that of surface Fe(III) due to the higher lability of the Fe(II)–O bond compared to the Fe(III)–O bond [12]. Reductive dissolution of Fe oxide minerals involves many steps, including reductant adsorption, electron transfer, detachment of oxidised reductant, and detachment of Fe(II) (the rate determining step) [14]. Previous workers have reported that the detachment of surface Fe (II) might be accelerated in the presence of a ligand with ligand-promoted reductive dissolution (i.e. ascorbate-EDTA; ascorbate-oxalate; dithionite-citrate) shown to be particularly effective in promoting fast dissolution of iron oxides [13,15,16], however, the mechanism
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
Nomenclature AFO Asc Cit C0 C
EDTA ethylenediaminetetraacetic acid f percentage of undissolved iron on the membrane (%) P phosphorus Fe(II)Cit complex of Fe(II) and citric acid Fe(III)Cit complex of Fe(III) and citric acid >Fe(III)CitFe(II) surface complex of Fe(III), citric acid and Fe(II)
amorphous ferric oxyhydroxide ascorbic acid (mM) citric acid (mM) the initial iron concentration on the membrane (mM) the released iron concentration (mM)
dominating such dissolution processes remains unclear. Citric acid, a tricarboxylic acid that forms strong solution complexes with iron, is not particularly effective when used as a single cleaning agent for inorganics removal [17] while ascorbic acid, an effective Fe (III) reducing agent, has been shown to be especially effective in removing iron species from iron oxide-fouled membrane surface [10,18]. However, ascorbate-mediated reductive dissolution under oxic conditions is much diminished compared to that under anoxic conditions, apparently due to the catalytic oxidation of ascorbate via heterogeneous re-oxidation of surface Fe(II) [18]. It has also been reported that the combination of a reductant and a strong chelating ligand is especially effective in promoting the dissolution of iron oxides due to the possible synergistic effects between these two agents [13]. For example, Banwart et al. [15] reported that in the presence of ascorbate, the adsorption of oxalate (a ligand capable of forming bidentate mononuclear surface complexes) leads to a significant increase in the rate of reductive dissolution of hematite even though the adsorption of oxalate displaces some of the ascorbate from the hematite surface. A similar synergistic effect of reductant and complexant has also been observed in the dissolution of goethite by dithionite and citrate or EDTA [16]. Therefore, it is reasonable to postulate that in a system with both ascorbic acid and citric acid, ligand-promoted reductive dissolution could also occur and be more effective than ascorbate-mediated reductive dissolution for iron oxides-fouled membranes. In this study, we investigated proton-assisted (acid), ligand-promoted (citric acid), reductive (ascorbic acid), and ligand-promoted reductive (ascorbic acid-citric acid-mediated) cleaning of Fe(III)-based foulants from membranes used in submerged membrane bioreactors to which ferric chloride had been added for facilitation of phosphorus removal. Detailed investigations of the ascorbic acid-citric acid-mediated reductive cleaning of the iron-fouled membranes were undertaken under both oxic and anoxic conditions. Based on the findings of these studies, a kinetic model was developed as an aid to elucidating the key factors controlling the effectiveness of the dual reagent cleaning process. This model may also assist in determining the conditions that are optimal for removal of iron foulants in practical cleaning procedures.
pH=4 proton-assisted dissolution
100
Cit 1mM ligand-promoted dissolution
80
Asc 1 mM
f, %
60 reductive dissolution
40
Asc 3 mM
20 0
Asc 10 mM Asc-10 mM + Cit 1 mM ligand-promoted reductive dissolution
0
10
20 30 40 Leaching time, h
50
Fig. 1. Cleaning effectiveness of iron-fouled membranes at pH 4 with i) hydrochloride acid, ii) citric acid (Cit) at concentration of 1 mM, iii) ascorbic acid (Asc) at concentrations of 1, 3 and 10 mM, and iv) Asc-Cit at concentrations of 10 mM of Asc and 1 mM of Cit. Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
2. Materials and methods 2.1. Formation of iron-fouled small membrane modules Accelerated fouling and cleaning studies were conducted using a small membrane module containing four polyvinylidene fluoride (PVDF) hollow fibre membranes (Beijing Origin Water, China) with nominal pore size 0.1–0.3 µm, diameter 0.24 cm, length 15 cm and total surface area 0.0044 m2. The small module was oriented vertically in the membrane bioreactor to which a 37.2 mM ferric iron solution (Fe/P molar ratio of 4) was dosed to facilitate phosphorus removal [9]. The small membrane module was operated at a constant flux of 30 L/ m2 h (with no relaxation) for a period of two weeks before cleaning studies were undertaken.
Fig. 2. Cleaning effectiveness of iron-fouled membranes at pH 4 in ascorbic acid (Asc)citric acid (Cit) solutions with Asc concentration of 10 mM and Cit concentrations of 50, 100, 500, 1000 (a), 3000, 5000 µM (b). Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
immersed in 50 mL of one of the cleaning solutions from the following experimental configurations:
2.2. Cleaning studies of iron-fouled membranes
1. Hydrochloride acid; pH 4 2. Citric acid (Cit); pH 4; 1 and 10 mM
After operation for two weeks, the fouled membranes were rinsed with MQ water and then cut into a number of 5 cm length pieces and 127
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
DO concentration, mg/L
8
pH 4 with membrane Asc 10mM Asc 10mM + Cit 50 µM Asc 10mM + Cit 100 µM Asc 10mM + Cit 500 µM Asc 10mM + Cit 1000 µM
6 4 2 0
0
20
40 60 80 Leaching time, min
100
Fig. 5. DO concentration during the membrane cleaning process by Asc-Cit-mediated solutions with 10 mM Asc and 50–1000 µM Cit at pH 4 under oxic conditions.
Fig. 3. Cleaning effectiveness of iron-fouled membranes at pH 4 in ascorbic acid (Asc)citric acid (Cit) solutions with Asc concentration of 3 mM and Cit concentrations of 100, 500, 1000 µM (a) and Asc concentration of 10 mM and Cit concentrations of 100, 1000 µM (b). Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
Fig. 6. Conceptual model of the Asc-Cit-mediated reductive dissolution of iron oxides from fouled membranes at pH 4 under both oxic and anoxic conditions. > Fe(III) – surface Fe(III) species; > Fe(II) – surface Fe(II) species; Fe(II) – solution Fe(II) species; O2 – dissolved oxygen. Numbers represent reactions as numbered in Table 1.
3. Ascorbic acid (Asc); pH 4; 1, 3, 10 mM 4. Ascorbic acid-citric acid (Asc-Cit); pH 4; 3 and 10 mM-Asc and 50 µM to 5 mM-Cit
Fig. 4. Cleaning effectiveness of iron-fouled membranes at pH 4 ascorbic acid (Asc) with concentrations of 10 (a) and 3 mM (b) for i) oxic conditions, ii) oxic conditions with 1000 µM citric acid (Cit), and iii) anoxic conditions with 1000 µM Cit. Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
The effectiveness of the above four solutions in removing iron from the membrane pieces were systematically compared with the most effective cleaning solution investigated in further detail. The pH of the citric acid (purity ≥ 99.5%, Sigma), ascorbic acid (purity ≥ 99%, Sigma), and Asc-Cit solutions was carefully controlled at 4 with an 128
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
Table 1 Modelled reactions and associated rate constants for the Asc-Cit-mediated ligand-promoted reductive dissolution of iron oxides from fouled membranes at pH 4 under both oxic and anoxic conditions. No.
Reaction
Rate constant (M−1 s−1)a
Note
1 2 3 4 5 6 7 8 9 10 11 12 13
> Fe(III) + Asc → > Fe(III)Asc > Fe(III)Asc → > Fe(II) + Asc*b Asc* + Asc* → Asc + Asc**c > Fe(II) → Fe(II) + (1-c) > Fe(III)d > Fe(II) + O2 → > Fe(III)e Fe(II) + Cit → Fe(II)Cit > Fe(II) + Cit → > Fe(II)Cit > Fe(II)Cit → Fe(II)Cit + (1-c) > Fe(III) > Fe(III) + Cit → > Fe(III)Cit > Fe(III)Cit → Fe(III)Cit + (1-c) > Fe(III) > Fe(III) + Fe(II)Cit → > Fe(III)CitFe(II) > Fe(III)CitFe(II) → > Fe(II)CitFe(III) > Fe(II)CitFe(III) → Fe(II) + Fe(III)Cit + (1-c) > Fe(III)
k1 = 4.85 × 10−2 k2 = 4.92 × 1010 (s−1) k3 = 7.12 × 108 k4 = 3.56 × 10−4 (s−1) k5 = 4.98 k6 = 500 k7 = 0.01 k 8 = 2.0 × 10−3 (s−1) k 9 = 0.05 k10 = 1.5 × 10−4 (s−1) k 11 = 35 k12 ≥ 1.0 × 104 (s−1) k13 = 1.0 × 10−3 (s−1)
[18] [18] [18] [18] [18] [20] Fitting parameter Fitting parameter Fitting parameter Fitting parameter Fitting parameter Not sensitive when ≥ 1.0 × 104 Fitting parameter
Note that all the important reactions based on the schematic in Fig. 6 are shown in Table 1 with reactions 1–5 taken from our previous study on Asc-mediated reductive dissolution of iron- fouled membranes [18]. Reactions 6–10 represent solution phase Fe(II)Cit and surface Fe(II)Cit and Fe(III)Cit complexes formation and corresponding dissolution. Reactions 11–13 describe the formation and detachment of surface Fe(III)CitFe(II) complex. a Unless noted otherwise. b Asc* represents ascorbate radical. c Asc** represents dehydroascorbic acid and reaction 3 represents the disproportionation of two molecules of Asc* leading to the regeneration of one molecule of Asc. d c is the ratio of surface iron binding sites to total iron on the membrane and c = 0.6 was used here [18]. e Model input for the dissolved oxygen concentrations are 0.243 mM (saturation concentration under atmosphere pressure) under oxic conditions and zero under anoxic conditions, respectively.
membranes were not stirred during the cleaning operation.
appropriate buffer (sodium acetate (8.2 mM)/acetic acid (1.8 mM)). We confirmed that the sodium acetate/acetic acid buffer negligibly influences the iron oxyhydroxide foulant with less than 2% removal of iron from the iron-fouled membrane surface after 12 h reaction [18]. Fe(III)based materials were the major foulants on these membrane pieces as indicated by SEM and EDX analysis shown in Fig. s1, presumably resulting from the rapid precipitation of amorphous ferric oxyhydroxides (AFO) after Fe addition with AFO depositing on the membrane surface [18]. The effectiveness of the various cleaning solutions in removing Febased foulants was determined qualitatively by examining the membranes using field emission scanning electron microscopy (FESEM, Hitachi S3400-I, Japan) (following the coating of samples with chromium) and energy dispersive x-ray spectroscopy (EDX) (Bruker AXS Microanalysis GmbH Berlin, Germany). A more quantitative measure of rate and extent of removal of Fe-containing foulants from the membranes was obtained by examining the increase in concentrations of Fe in solution as a function of exposure time of the membrane to the cleaning solution (as described below). Aliquots (5 mL) of the membrane-leachate solution were removed after various intervals of immersion of the fouled membrane pieces in cleaning solution for measurement of total Fe by inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Agilent 7300, USA). After 48 h cleaning, each fouled membrane was immersed for a further 24 h in 50 mL of 300 mM sodium dithionite in order to remove any residual iron remaining on the membrane. Based on this technique, a measure of the initial iron concentration on each membrane (C0) was taken to equal the iron concentration in the wash solution after Asc cleaning plus the residual iron concentration in the dithionite wash solution. The cleaning effectiveness of different agents was quantified by comparing the undissolved iron fraction on the membrane (f) (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration).
2.2.2. Asc-Cit-mediated dissolution under anoxic conditions Asc-Cit-mediated solutions with i) Asc concentration of 10 mM and Cit concentration of 1 mM, and ii) Asc concentration of 3 mM and Cit concentration of 1 mM at controlled pH 4 were applied for periods of up to 4 h. Asc-Cit-mediated cleaning solutions were bubbled with argon gas for 10 min before membrane immersion with bubbling continued during the cleaning experiments in order to ensure the removal of oxygen. 2.3. Kinetic modelling The MS-Excel based kinetic modelling programme Kintecus [19] was used to determine the time varying concentrations of reactants and products for a set of reactions hypothesized to occur during the membrane cleaning process. Rate constants for the various reactions included in the analysis were obtained, for the most part, from the literature with the set of differential equations corresponding to the rate expressions solved numerically. It should be noted that the modelling is not a data fitting exercise but, rather, a means of assessing the relative importance of particular proposed reactions. 3. Results 3.1. Dissolution of iron-fouled membranes by four pathways Results for cleaning effectiveness of iron-fouled membranes at pH 4 using four different pathways are presented in Fig. 1. Negligible effect (~ 5%) in terms of Fe removal by hydrochloride acid (pH 4) was achieved after 48 h soaking (Fig. 1), while excellent cleaning performance (with > 99% of Fe removed) was achieved with 10 mM ascorbic acid (Asc) after 48 h soaking (Fig. 1), most of which occurred in the first 6 – 12 h. In contrast, soaking in 1 mM citric acid for 48 h resulted in a removal of only 26% of Fe, whereas even a solution containing 10 mM (about 0.2% in weight) citric acid is only capable of dissolving 39.9% of Fe after 48 h soaking (Fig. s2), which is much less effective compared to Asc at the same concentration and pH (> 99%) (Fig. 1). As can be seen from the SEM and EDX results in Fig. s3, soaking with 0.1 mM citric acid for 48 h induced negligible removal of the Fe-rich gel layer with a
2.2.1. Asc-Cit-mediated dissolution under oxic conditions The membrane pieces were immersed in Asc-Cit solutions with i) Asc concentration of 10 mM and Cit concentrations of 50, 100, 500, 1000, 3000, 5000 µM, and ii) Asc concentration of 3 mM and Cit concentrations of 100, 500, 1000 µM at controlled pH 4 for periods of up to 4 h. Note that “oxic condition” indicates that the cleaning solutions were open to the atmosphere. The solutions containing the fouled 129
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
Fig. 7. Cleaning effectiveness of iron-fouled membranes at pH 4 in Asc-Cit solutions with Asc concentrations of 3 and 10 mM and Cit concentrations of 50 µM to 5 mM under both oxic and anoxic conditions. Symbols represent the experimental data and solid lines represent the model fit using reactions 1–13 (Table 1) with c = 0.6 representing the ratio of surface iron binding sites to total iron on the membrane. Fraction (f) is representative of the percentage of undissolved iron on the membrane (f = (C0 – C)/C0, where C is the released iron concentration and C0 is the initial iron concentration) as a function of leaching time.
while only gradual dissolution was observed in the presence of Asc alone under oxic conditions (Figs. 2 and 3). At 10 mM Asc, the percentage of iron dissolved from the membrane increased with increase in Cit concentrations from 50 µM to 1000 µM (Fig. 2a), while further increase of Cit concentrations from 1 mM to 3 mM resulted in a decrease in effectiveness of iron oxide dissolution (Fig. 2b). Similarly, at a lower Asc concentration of 3 mM, the dissolution effectiveness increased with increase in Cit concentrations from 100 µM to 1000 µM, though the dissolution performances at Cit concentrations of 500 and 1000 µM are almost identical (Fig. 3a). A similar synergistic effect was observed when ethylenediaminetetraacetic acid (EDTA) was involved in reductive dissolution of magnetite by Asc under acid conditions (pH 3.5), where the dissociation rate reached a maximum at an intermediate concentration of EDTA [13]. This may suggest that at higher concentrations, citrate competes with ascorbate to bind surface sites yielding a surface Fe(III) complex with reduced reactivity and resultant
thick deposit remaining on the membrane surface while the thick Ferich gel layer disappeared after 48 h soaking with 10 mM Asc (pH 4). Extremely efficient iron removal by ligand-promoted reductive dissolution was achieved with 10 mM Asc and 1 mM Cit (pH 4) with > 96% of Fe removed only after 4 h soaking (Fig. 1). In light of these results, an apparent synergistic effect of Asc and Cit appears to be occurring in ligand-promoted reductive dissolution with further investigation of this phenomenon provided below. 3.2. Dissolution of iron-fouled membranes by Asc-Cit-mediated pathway Further experiments have been conducted to investigate the dual reagent Asc-Cit-mediated dissolution of iron-fouled membranes with results shown in Figs. 2 and 3. As can be seen from these results, AscCit-mediated dissolution proceeds in a biphasic manner, with an initial lag period followed by a much faster phase of dissolution over time, 130
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
(II) and/or promoting rapid detachment of Fe(II) thereby reducing the opportunity for heterogeneous oxidation. During the lag period DO levels are high enough to compete with citrate and enable Fe(II) reoxidation to occur, however, when DO levels decline sufficiently, citrate is able to prevent Fe(II) re-oxidation. Further, according to previous studies, the sudden increase in the dissolution rate following the lag period suggests a switch in the system resulting from some process that involves a positive feedback loop. One possibility is that the ligandpromoted reductive dissolution process involves the Fe(II)-citrate complex, which acts as a more effective reductant than Asc, possibly via formation of a > Fe(III)-citrate-Fe(II) ternary complex. Such a process is consistent with the Fe(II)L-mediated process suggested by Afonso et al. [13] who attributed the observed ‘induction period’ to the steady accumulation of Fe(II)L until a point was reached where reduction via Fe(II)L becomes significant. The process is self-reinforcing since the more Fe(II) that is solubilized, the greater the reduction rate and production of solution phase Fe(II).
lower dissolution rate. It is also worth mentioning that though the best dissolution efficiency was reached at Asc-10 mM and Cit-1 mM in this study, we have also found that almost identical dissolution effectiveness were achieved at Asc-10 mM, Cit-100 µM and Asc-3 mM, Cit-1000 µM (Fig. 3b). Asc-Cit-mediated reductive dissolution under anoxic conditions was also investigated with results shown in Fig. 4. Interestingly, under anoxic conditions, the biphasic behaviour observed during dissolution under oxic conditions (Figs. 2 and 3) has been replaced by an initially more rapid dissolution process with no evidence of any lag though similar extents of dissolution have been observed under both oxic and anoxic conditions after approximately 4 h of leaching (Fig. 4). 3.3. DO concentration during membrane cleaning process Dissolved oxygen (DO) concentrations during the membrane cleaning process in Asc-Cit solutions under oxic conditions are shown in Fig. 5. O2 concentrations in the ascorbate solution quickly decreased in the initial 10 min followed by a slower gradual decline over the rest of the experimental period (90 min). Similar trends were evident when low concentrations of citrate co-existed with 10 mM ascorbate though the rate of O2 consumption was markedly less at citrate concentrations of 500 and 1000 µM.
4.2. Modelling of ligand-promoted reductive dissolution A schematic of the hypothesized key processes involved in citrateenhanced ascorbate-mediated reductive dissolution is shown in Fig. 6. We have constructed a kinetic model of these key processes with citrate-induced enhancement of the surface Fe(II) detachment hypothesized to be promoted by the formation of a surface Fe(III)-citrateFe(II) ternary complex. All the important reactions shown in the schematic in Fig. 6 are presented in Table 1 with reactions 1–5 taken from our previous study on Asc-mediated reductive dissolution of iron-fouled membranes [18]. Reactions 6–10 represent solution phase Fe(II)Cit and surface Fe(II)Cit and Fe(III)Cit formation and corresponding dissociation processes. Reactions 11–13 describe the formation and detachment of a ternary surface Fe(III)CitFe(II) complex. Other reactions, such as the oxidation of > Fe(II)Cit species by O2, and the reduction of solution Fe(III)Cit species by ascorbic acid, were shown to be unimportant and thus neglected with variation of the rate constants of these reactions over a reasonable range leading to minimal impact on model output. Model results for all experimental data including 3 and 10 mM Asc and 50 µM to 5 mM Cit under both oxic and anoxic conditions are shown in Fig. 7 with these results reasonably describing all the experimental data obtained. At the reaction rate constants proposed here, the formation of the ternary surface complex Fe(III)CitFe(II) occurs at a relatively fast rate (reaction 11) and, as such, is central to the proposed dissolution process. Similar ternary species have been proposed in previous studies. For example, Sulzberger et al. reported that the occurrence of > Fe(III)oxalate-Fe(II) species accounted for the catalytic effect of Fe(II) in the dissolution of hematite and proposed that besides oxalate, malonate and citrate are also suitable bridging ligands for ternary complex formation with Fe(II) and Fe(III) (hydr)oxides [12]. Borghi et al., in studies of the dissolution of magnetite in solutions containing EDTA and Fe(II), suggested that the adsorption of Fe(II)EDTA2– onto > Fe(III) to form > Fe(III)-EDTA-Fe(II) surface species is highly favorable and should proceed at a high rate (on the order of 1250 mol–1 L) [21]. In view of these earlier studies, it would seem reasonable to hypothesize that the > Fe(III)CitFe(II) ternary surface complex will be formed at a rate considerably faster than that of the binary surface complexes (Table 1). It should be noted however that we provide no direct evidence for formation of the ternary species other than that its formation is consistent with the findings of other studies and accounts adequately for the observed iron oxide dissolution behaviour. It can also be deduced from the model that the lag period in dissolution under oxic conditions (especially at higher concentrations of Cit) was due to the formation of the ternary surface complex > Fe (III)CitFe(II) (reaction 11) which, in the first instance, hinders iron oxide dissolution by 1) returning solution Fe(II) species to the surface via complex formation between Cit and dissolved Fe(II) (reaction 6)
4. Discussion 4.1. Effect of citrate and DO on ligand-promoted reductive dissolution In our previous study, ascorbate-mediated reductive dissolution under oxic conditions has been found to be much reduced compared to that under anoxic conditions, apparently due to the catalytic oxidation of ascorbate via heterogeneous re-oxidation of surface Fe(II) [18]. In this study, citrate has been shown to almost completely alleviate the retarding effects of DO on dissolution for both low and high concentrations of Asc under oxic conditions (Fig. 2a and Fig. 3a). The mechanism by which citrate enhances the reductive dissolution of iron oxides under oxic conditions is not clear, though based on previous studies, it could be attributed to either the enhancement of Fe(II) detachment [15] or enhanced Fe(III) reduction as a result of formation of the Fe(II)Cit complex [13]. Meanwhile, for a given concentration of Asc, there is an optimal concentration of Cit above which the dissolution rate starts to decrease (Fig. 2b and Fig. 3a). Additionally, a notable difference in the effect of citrate between oxic and anoxic conditions is the presence of an initial lag period under oxic conditions during which the dissolution rate is similar to that in the absence of citrate (Fig. 4). At the end of the lag period, the rate increases dramatically until dissolution is nearly equivalent to the anoxic case (Fig. 4). This lag effect prior to enhanced dissolution has also been observed under oxic conditions with addition of phenanthroline, a strong ligand for Fe(II) [18]. As such, there is clearly some interplay between O2 and citrate in the ligand-promoted reductive dissolution process with the presence of citrate somehow mitigating the effects of O2 on reductive dissolution, although only after an initial lag period. One possibility is that citrate complexation of surface Fe(II) may be reducing access of O2 to surfacelocated Fe(II) thereby slowing heterogeneous oxidation of Fe(II) and the associated catalytic oxidation of ascorbate. Measurements of DO concentrations also provide some insight into possible mechanisms of citrate-mediated reductive dissolution and reasons for the occurrence of the lag period. In the absence of citrate, DO is rapidly consumed as a result of the heterogeneous re-oxidation of Fe(II) resulting in low steady state DO concentrations (Fig. 5) [18]. However, in the presence of citrate, O2 consumption is much slower and only reaches DO levels equivalent to that where citrate is absent after a period that is, interestingly, equivalent to the lag period (Fig. 5). Under oxic conditions, the most likely action of citrate is the prevention of Fe (II) oxidation either by competing with DO for adsorption to surface Fe 131
Journal of Membrane Science 545 (2018) 126–132
Z. Zhang et al.
surface. 4. Use of dual reagents ascorbic acid and citric acid under oxic conditions is recommended for the cleaning of iron-fouled membranes in view of the extreme cleaning effectiveness. Further analysis of cleaning effectiveness and cost of full-scale implementation of ascorbate-citrate-mediated cleaning is required before definitive recommendations in terms of optimal reagent dose and cleaning durations for removal of iron-containing foulants can be made.
with this species further complexed by surface Fe(III) to generate the ternary surface complex at a relatively fast rate (reaction 11); and 2) capturing surface Fe(III) and forming the ternary surface complex (reaction 11) with O2 acting in a facilitating role by providing more surface Fe(III) species via heterogeneous re-oxidation of surface Fe(II) (reaction 5). Though slow formation of > Fe(II)Cit (reaction 7) could also contribute to the lag period by partially circumventing the direct detachment via reaction 4, this pathway is less important due to the fast oxidation of surface Fe(II) species by oxygen (reaction 5). On the other hand, once the concentration of the ternary surface complex reaches a certain level, it will quickly undergo internal electron transfer followed by dissociation into solution Fe(II) and Fe(III) species with this process accounting for the fast dissolution after the induction period. Further, the adverse impact on dissolution effectiveness exerted by increasing the Cit concentrations to very high levels (Fig. 2b and Fig. 3a) could also be explained by the model. More specifically, since reaction 1 and reaction 9 proceed with similar rates, increasing the concentrations of Cit (to levels comparable to Asc) will increase the importance of reaction 9 thereby enabling it to compete with reaction 1, yielding more surface Fe(III)Cit complexes which further dissociate at a lower rate (reaction 10) compared to that of surface Fe(III)Asc complexes. A similar conclusion was reached when EDTA and Asc were employed for reductive dissolution of magnetite at pH 3.5, where the dissociation rate initially increased then decreased on continuing increase in EDTA concentration with this behaviour recognised to occur because EDTA outcompetes Asc for surface Fe(III) sites yielding surface-located Fe(III) EDTA with reduced reactivity and resultant lower dissolution rate at high EDTA concentrations.
Acknowledgements Funding provided by the Australian Research Council, Beijing Origin Water, Sydney Water Corporation and Water Research Australia through ARC Linkage Grant (LP100100056) is greatly acknowledged. The authors would also like to acknowledge Beijing Origin Water Technology for provision of hollow fibre membranes. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2017.09.059. References [1] R.I. Sedlak, Phosphorus and Nitrogen Removal from Municipal Wastewater: Principles and Practice, Lewis Publishers, Florida, USA, 1991. [2] T.D. Waite, Challenges and opportunities in the use of of iron in water and wastewater treatment, Environ. Sci. Bio/Technol. 1 (2002) 9–15. [3] J. Kim, Q. Deng, M.M. Benjamin, Simultaneous removal of phosphorus and foulants in a hybrid coagulation/membrane filtration system, Water Res. 42 (2008) 2017–2024. [4] A.H. Caravelli, E.M. Contreras, N.E. Zaritzky, Phosphorous removal in batch systems using ferric chloride in the presence of activated sludges, J. Hazard. Mater. 177 (2010) 199–208. [5] Metcalf, Eddy, G. Tchobanoglous, F. Burton, H.D. Stensel (Eds.), Wastewater Engineering: Treatment and Reuse, 4th ed., McGraw-Hill, New York, 2003. [6] M. Manzouri, H.K. Shon, Rectification methods for the fouling of ultrafiltration hollow fibre membranes as a result of excessive soluble iron, Desalin. Water Treat. 32 (2011) 437–444. [7] Y. Wang, G.L. Leslie, T.D. Waite, Impact of iron dosing of membrane bioreactors on membrane fouling, Chem. Eng. J. 252 (2014) 239–248. [8] Y. Wang, K.H. Tng, H. Wu, G.L. Leslie, T.D. Waite, Removal of phosphorus from wastewaters using ferrous salts – a pilot scale membrane bioreactor study, Water Res. 57 (2014) 140–150. [9] Z.H. Zhang, Y. Wang, G.L. Leslie, T.D. Waite, Effect of ferric and ferrous iron addition on phosphorus removal and fouling in submerged membrane bioreactors, Water Res. 69 (2015) 210–222. [10] Z.H. Zhang, M.W. Bligh, Y. Wang, G.L. Leslie, H. Bustamante, T.D. Waite, Cleaning strategies for iron-fouled membranes from submerged membrane bioreactor treatment of wastewaters, J. Membr. Sci. 475 (2015) 9–21. [11] Z. Wang, J. Ma, C.Y. Tang, K. Kimura, Q. Wang, X. Han, Membrane cleaning in membrane bioreactors: a review, J. Membr. Sci. 468 (2014) 276–307. [12] B. Sulzberger, D. Suter, C. Siffert, S. Banwart, W. Stumm, Dissolution of Fe(III) (hydr)oxides in natural waters: laboratory assessment on the kinetics controlled by surface coordination, Mar. Chem. 28 (1989) 127–144. [13] M.D. Afonso, P.J. Morando, M.A. Blesa, S. Banwart, W. Stumm, The reductive dissolution of iron-oxides by ascorbate – the role of carboxylate anions in accelerating reductive dissolution, J. Colloid Interface Sci. 138 (1990) 74–82. [14] D. Suter, S. Banwart, W. Stumm, Dissolution of hydrous iron (III) oxides by reductive mechanisms, Langmuir 7 (1991) 809–813. [15] S. Banwart, S. Davies, W. Stumm, The role of oxalate in accelerating the reductive dissolution of hematite (α-Fe2O3) by ascorbate, Colloids Surf. 39 (1989) 303–309. [16] E.H. Rueda, M.C. Ballesteros, R.L. Grassi, M.A. Blessa, Dithionite as a dissolving reagent for goethite in the presence of EDTA and citrate: application to soil analysis, Clays Clay Miner. 40 (1992) 575–585. [17] T.D. Waite, F.M.M. Morel, Photoreductive dissolution of colloidal iron oxide: effect of citrate, J. Colloid Interface Sci. 102 (1984) 121–137. [18] Z.H. Zhang, M.W. Bligh, T.D. Waite, Ascorbic acid-mediated reductive cleaning of iron-fouled membranes from submerged membrane bioreactors, J. Membr. Sci. 477 (2015) 194–202. [19] James C. Ianni, Kintecus, Windows Version 5.20, 2014. 〈www.kintecus.com〉. [20] A.L. Rose, T.D. Waite, Kinetics of iron complexation by dissolved natural organic matter in coastal waters, Mar. Chem. 84 (1–2) (2003) 85–103. [21] E.B. Borghi, A.E. Regazzoni, A.J.G. Maroto, M.A. Blesa, Reductive dissolution of magnetite by solutions containing EDTA and FeII, J. Colloid Interface Sci. 130 (1989) 299–310.
5. Conclusion In this study, four iron oxide dissolution pathways: i) proton (acid)assisted, ii) ligand-promoted (citric acid), iii) reductive (ascorbic acid), and iv) ligand-promoted reductive (citric acid-ascorbic acid) for removal of iron foulants from iron oxide-fouled membranes from submerged membrane bioreactors under both oxic and anoxic conditions were investigated and their dissolution effectiveness were compared. A kinetic model for the ascorbate-citrate-mediated reductive dissolution of ferric iron from iron oxide-fouled membranes has been developed as an aid to determining the major pathways involved in the dissolution process. The following conclusions can be drawn: 1. Among the four general pathways for release of iron from iron oxide-fouled membranes, the cleaning effectiveness under oxic conditions followed the order: proton-assisted < ligand-promoted < reductive < ligand-promoted reductive dissolution. After 48 h soaking, > 99% of iron foulant was removed by 10 mM Asc (pH 4) while > 96% of the foulant was removed after only 4 h soaking by dual reagents of 10 mM Asc and 1 mM citric acid (pH 4). 2. The dissolution effectiveness increased with increase in citrate concentration in the ascorbate-citrate solutions though there was an optimal intermediate citrate concentration after which the dissolution performance decreased with further increase in citrate concentrations. Under oxic conditions, an initial lag period followed by a sharp increase in dissolution rate was observed while no lag in dissolution was evident under anoxic conditions for the ascorbatecitrate-mediated reductive dissolution process. The dissolution performance at the latter stage under oxic conditions was comparable to that under anoxic conditions. 3. The presence of O2 reduced ascorbic acid cleaning effectiveness of iron-fouled membranes due to the Fe(III)-catalyzed oxygen oxidation of ascorbate. It appears likely that in spite of the initial induction period, citric acid could completely alleviate the retarding effects of O2 on the dissolution because of the formation of a ternary surface complex > Fe(III)CitFe(II) which readily detaches from the
132