Cationic Pullulan Derivatives Based Flocculants for Removal of Some Metal Oxides from Simulated Wastewater

Modified polysaccharides have been increasingly used as flocculants in wastewater treatment due to their non-toxicity, low price, biodegradability, etc. However, the pullulan derivatives are less used in wastewater purification processes. Therefore, this article presents some data regarding FeO and TiO2 particle removal from model suspensions by some pullulan derivatives with pendant quaternary ammonium salt groups, trimethylammonium propyl carbamate chloride (TMAPx–P). The influence of the polymer ionic content, dose, and initial solution concentration as well as of the dispersion pH and composition (metal oxide content, salts, and kaolin) on the separation efficacy were considered. UV-Vis spectroscopy measurements have shown a very good removal efficacy of TMAPx–P for the FeO particles (around 95% and more), irrespective of the polymer and suspension characteristics; a lower clarification of the TiO2 particles suspension (removal efficiency between 68% and 75%) was noticed. Both the zeta potential and the particle aggregates size measurements revealed the charge patch as the main mechanism which governs the metal oxide removal process. The surface morphology analysis/EDX data provided supplementary evidence regarding the separation process. A good removal efficiency (90%) of the pullulan derivatives/FeO flocs for the Bordeaux mixture particles from simulated wastewater was found.


Introduction
Pullulan, a non-ionic polysaccharide produced by some microorganisms (Aureobasidium pullulans, Teloschistes flavicans, Cytariaharioti, etc.) and different carbon sources (jackfruit seed, cassava bagasse, ricehull, etc.) [1], has attracted expanding interest due to the advantages encountered, generally, of the polymeric materials belonging to polysaccharides class (non-toxicity, low price, easy availability, biodegradability), but also related to its chemical structure (high flexibility of the chain, high solubility in water, ability to be modified by different chemical reactions, etc) [1 -3]. The wide range of chemical reactions used to modify this polysaccharide (esterification, sulfation, copolymerization, and oxidation) [4][5][6][7][8][9][10] lead to pullulan derivatives that, apart from the properties of the unmodified sample, may exhibit new ones, suitable for certain applications. For example, the introduction of ionizable groups such as carboxylic, quaternary ammonium, or amino ones, which dissociate in polar solvents and lead to charged polymer chains, can give the ionic pullulan derivatives the possibility to interact with oppositely charged particles by means of the electrostatic attraction forces, thus becoming suitable polymers for using in the wastewater purification field. Microspheres of pullulan-graft-poly(3-acrylamidopropyl trimethylammonium chloride) and some soluble pullulan derivatives containing either tertiary amine groups (aminopropyl dimethylamine) or quaternary ammonium salt ones (grafted chains, poly[(3-acrylamidopropyl)-trimethylammonium chloride]) have been already tested and successfully removed some organic particles, such as dyes (Methyl Orange, Acid Orange 7, Azocarmine B, Methylene Blue, Reactive Blue 2) [11] and commercial pesticide No flocculation in the case of the suspension prepared in distilled, but only in the tap water [18] Dextran derivatives TiO 2 1.2 ÷ 1.4 mg L −1 50 ÷ 60 Suspension prepared in water [17] Our research considered the influence of some tunable parameters, such as the pullulan derivatives ionic content, polymer dose, and initial concentration (c ip ), dispersion pH and composition (contaminant content (c FeO ), salts and kaolin presence), as the main parameters influencing the separation efficacy.
UV-Vis spectroscopy, zeta potential, the particle aggregates size measurements together with surface morphology, and EDX analyses were used as tools to evaluate the flocculation efficacy of the pullulan derivatives investigated herein and the possible mechanisms controlling the metal oxides removal process.

Results and Discussion
The optimum dose (dose op ) (the polymer dose (PD) corresponding to the maximum removal efficiency of contaminants) is one of the terms, routinely used to express the efficiency of the flocculation process. It can be influenced by parameters related either to the chemical structure of polymers (molecular mass, ionic group content and nature, the alkyl chain length, etc.) or dispersion medium characteristics (composition, pH, temperature, the polymer and contaminant concentration, etc.). In the following, some data regarding the influence of some of the aforementioned parameters on the dose op , and hence the separation efficacy of the pullulan derivatives investigated in dispersions of FeO and TiO 2 particles, are presented.

Pullulan Derivatives Ionic Group Content
a.
FeO suspension The ionic group content of a polyelectrolyte chain is, probably, the parameter with the greatest impact on the chain conformation in solution, which in turn, influences the flocculation ability of an ionic polymer. The impact of a series of TMAP x -P having different ionic content (see Table 3, Section 3.1. Materials) on the removal of FeO (c Feo = 0.4 g L −1 ) and TiO 2 (c TiO2 = 0.05 g L −1 ) particles has been examined and presented below. The residual FeO absorbance vs. TMAP x −P dose plots ( Figure 1) show a very good removal of the FeO particles (removal efficiency around 95% and more) in the polymer dose interval between 0.14 mg L −1 (TMAP 0.7 -P) and 1 mg L −1 (TMAP 0.2 -P); according to the calibration curve, the residual FeO concentration is 0.014 g L −1 (TMAP 0.7 -P) and 0.015 g L −1 (TMAP 0.2 -P).
The optimum dose (doseop) (the polymer dose (PD) corresponding to the maximu removal efficiency of contaminants) is one of the terms, routinely used to express efficiency of the flocculation process. It can be influenced by parameters related either the chemical structure of polymers (molecular mass, ionic group content and nature, alkyl chain length, etc.) or dispersion medium characteristics (composition, p temperature, the polymer and contaminant concentration, etc.). In the following, so data regarding the influence of some of the aforementioned parameters on the doseop, a hence the separation efficacy of the pullulan derivatives investigated in dispersions of F and TiO2 particles, are presented.

a. FeO suspension
The ionic group content of a polyelectrolyte chain is, probably, the parameter w the greatest impact on the chain conformation in solution, which in turn, influences flocculation ability of an ionic polymer. The impact of a series of TMAPx-P havi different ionic content (see Table 3, Section 3.1. Materials) on the removal of FeO (cFeo = g L −1 ) and TiO2 (cTiO2 = 0.05 g L −1 ) particles has been examined and presented below. T residual FeO absorbance vs. TMAPx−P dose plots ( Figure 1) show a very good removal the FeO particles (removal efficiency around 95% and more) in the polymer dose inter between 0.14 mg L −1 (TMAP0.7-P) and 1 mg L −1 (TMAP0.2-P); according to the calibration cur the residual FeO concentration is 0.014 g L −1 (TMAP0.7-P) and 0.015 g L −1 (TMAP0.2-P). The investigations regarding the polymer-particle interactions have brought out diversity of forces (electrostatic attractions, hydrogen bonding, ion binding, hydropho The investigations regarding the polymer-particle interactions have brought out the diversity of forces (electrostatic attractions, hydrogen bonding, ion binding, hydrophobic ones) implied in these systems, depending on the characteristic of the partners investigated [24]. Taking into account (i) the presence of the quaternary ammonium salt groups on the TMAP x -P chains and the negatively charged FeO particles (ζ = −26.7 mV-see the Experimental Section 3) at the working pH (7.1) and (ii) the poor performance of the unmodified pullulan sample in the separation of this type of particles (maximum removal efficiency around 55%) (Figure 1, inset), one may assert that the electrostatic attraction forces between the oppositely charged sites on the TMAP x -P and FeO particles prevail in the separation process. This assumption is supported by another finding, worthy to underline, namely the decreasing of dose op values with augmentation of the ionic group content (Table 2).
A higher amount of charged groups on the polymer chains means an increased ability to electrostatically attract the negative charges on the FeO particle s surface, and hence the separation process. The neutralization process is also helped by the more extended chains of the highest charged sample (TMAP 0.7 -P), conformation resulting from the increased electrostatic repulsion between the positively charged segments of the chain. Moreover, the enhancing efficacy in FeO/TMAP x -P interactions with augmentation pullulan derivatives charged content is reflected in flocs size (median values D50, µm) recorded at dose op for each pullulan derivative sample, which follows the next order: D50 TMAP0.2-P = 5.23 ± 0.16 µm < D50 TMAP0.4-P = 5.99 ± 0.33 µm < D50 TMAP0.7-P = 8.32 ± 0.39 µm (Figure 2). groups on the TMAPx-P chains and the negatively charged FeO particles (ζ = −26.7 mVsee the Experimental Section 3) at the working pH (7.1) and (ii) the poor performance of the unmodified pullulan sample in the separation of this type of particles (maximum removal efficiency around 55%) (Figure 1, inset), one may assert that the electrostatic attraction forces between the oppositely charged sites on the TMAPx-P and FeO particles prevail in the separation process. This assumption is supported by another finding, worthy to underline, namely the decreasing of doseop values with augmentation of the ionic group content (Table 2). A higher amount of charged groups on the polymer chains means an increased ability to electrostatically attract the negative charges on the FeO particleʹs surface, and hence the separation process. The neutralization process is also helped by the more extended chains of the highest charged sample (TMAP0.7-P), conformation resulting from the increased electrostatic repulsion between the positively charged segments of the chain. Moreover, the enhancing efficacy in FeO/TMAPx-P interactions with augmentation pullulan derivatives charged content is reflected in flocs size (median values D50, μm) recorded at doseop for each pullulan derivative sample, which follows the next order: D50TMAP0.2-P = 5.23 ± 0.16 μm < D50TMAP0.4-P = 5.99 ± 0.33 μm < D50TMAP0.7-P = 8.32 ± 0.39 μm ( Figure 2). Anyway, the narrow size distribution and small size of the FeO/TMAPx-P aggregates, irrespective of the pullulan derivatives ionic group content, compared with that of the untreated FeO particles (D50FeO = 2.19 ± 0.02 μm) enforces the mechanisms based on the electrostatic attraction between the oppositely charged particles (charge Figure 2. Particle size distribution for initial FeO particles (square) and aggregates obtained at dose op as a function of the ionic group content, TMAP 0.7 -P (circle); TMAP 0.4 -P (inverted triangle); TMAP 0.2 -P (star); and settling time 120 min. The particle size distribution curves represent the mean of three independent measurements. Anyway, the narrow size distribution and small size of the FeO/TMAP x -P aggregates, irrespective of the pullulan derivatives ionic group content, compared with that of the untreated FeO particles (D50 FeO = 2.19 ± 0.02 µm) enforces the mechanisms based on the electrostatic attraction between the oppositely charged particles (charge neutralization or charge patch), as the main ones in the separation of the FeO particles. According to the finding of Claesson et al. [25], the flocs size increase with the ionic group content of polymer samples indicating that the charge patch mechanism prevails in the FeO particle removal. We would like to recall the reader that in this type of mechanism, aggregation occurs as a consequence of the electrostatic attraction forces between oppositely charged patches on different particles [24]; a higher number of charged groups at the adsorption site entail an enhanced electrostatic attraction between the oppositely charged patches on the surfaces of different particles. This result is in agreement with those found in the flocculation of some clay particles by two series of polyelectrolytes with different charge densities, namely (poly(diallyldimethylammonium chloride)-PDADMAC [26] and cationic dextran derivatives (D40-EtX) [27]. Another argument in favor of this mechanism was obtained by corroborating the UV-Vis spectroscopy results with those provided by zeta potential measurements (ζ), performed, also, in dependence of the TMAP x -P dose. As seen in Figure 3, ζ values corresponding to dose op are negative for all of the polymer samples (−14.9 mV (TMAP 0.7 -P), −15.46 mV (TMAP 0.4 -P), and −16.86 mV (TMAP 0.2 -P)) that pleads for the charge patch mechanism, where zeta potential needs not to be zero [28], as for the neutralization mechanism [29].
oppositely charged patches on different particles [24]; a higher number of charged groups at the adsorption site entail an enhanced electrostatic attraction between the oppositely charged patches on the surfaces of different particles. This result is in agreement with those found in the flocculation of some clay particles by two series of polyelectrolytes with different charge densities, namely (poly(diallyldimethylammonium chloride)-PDADMAC [26] and cationic dextran derivatives (D40-EtX) [27]. Another argument in favor of this mechanism was obtained by corroborating the UV-Vis spectroscopy results with those provided by zeta potential measurements (ζ), performed, also, in dependence of the TMAPx-P dose. As seen in Figure 3, ζ values corresponding to doseop are negative for all of the polymer samples (−14.9 mV (TMAP0.7-P), −15.46 mV (TMAP0.4-P), and −16.86 mV (TMAP0.2-P)) that pleads for the charge patch mechanism, where zeta potential needs not to be zero [28], as for the neutralization mechanism [29].  One also observes that ζ increases for polymer doses higher than dose op up to positive values, namely 22.4 mV in the presence of TMAP 0.7 -P, 16.8 mV for TMAP 0.4 -P, and 9.81 mV for TMAP 0.2 -P. The charge inversion of the particles and hence, their electrostatic repulsion could explain the restabilization of the FeO suspension, noticed in the UV-Vis spectroscopy measurement results (Figure 1). The inefficiency of the flocculation process at polymer doses higher than dose op was also confirmed in the case of TMAP 0.7 -P by particle size distribution measurements ( Figure 4).  The smaller FeO/TMAP0.7-P aggregates size, D50 = 5.36 ± 0.17 μm, was noticed for the higher pullulan derivative dose (1.4 mg L −1 ) than that recorded at doseop, D50 = 8.32 ± 0.39 μm (doseop = 0.14 mg L −1 ) might be a consequence of a surplus of positively charged polymer chains adsorbed onto the FeO particles surface; repulsion between the positive charges on the microflocs prevents their growth in larger ones.

b. TiO2 suspension
In the case of the TiO2 particles, the ionic content influence on the particle's flocculation resembles that noticed for the FeO particles. Therefore, only the separation performances of TMAP0.2-P and TMAP0.7-P are depicted in Figure 5a. The minimum The smaller FeO/TMAP 0.7 -P aggregates size, D50 = 5.36 ± 0.17 µm, was noticed for the higher pullulan derivative dose (1.4 mg L −1 ) than that recorded at dose op , D50 = 8.32 ± 0.39 µm (dose op = 0.14 mg L −1 ) might be a consequence of a surplus of positively charged polymer chains adsorbed onto the FeO particles surface; repulsion between the positive charges on the microflocs prevents their growth in larger ones.

b. TiO 2 suspension
In the case of the TiO 2 particles, the ionic content influence on the particle's flocculation resembles that noticed for the FeO particles. Therefore, only the separation performances of TMAP 0.2 -P and TMAP 0.7 -P are depicted in Figure 5a. The minimum residual TiO 2 absorbance values were located only around 32% for the lowest charged sample (TMAP 0.2 -P) and 25% for the highest charged one (TMAP 0.7 -P) at dose op values which increased with the ionic group content (see Table 3, Section 3.1. Materials). . Particle size distribution for FeO/TMAP0.7-P aggregates at doseop: 0.14 mg L -1 (red circle) and polymer dose: 1.4 mg L -1 (blue circle), settling time 120 min. The particle size distribution curves represent the mean value of the three independent measurements.
The smaller FeO/TMAP0.7-P aggregates size, D50 = 5.36 ± 0.17 μm, was noticed for the higher pullulan derivative dose (1.4 mg L −1 ) than that recorded at doseop, D50 = 8.32 ± 0.39 μm (doseop = 0.14 mg L −1 ) might be a consequence of a surplus of positively charged polymer chains adsorbed onto the FeO particles surface; repulsion between the positive charges on the microflocs prevents their growth in larger ones.

b. TiO2 suspension
In the case of the TiO2 particles, the ionic content influence on the particle's flocculation resembles that noticed for the FeO particles. Therefore, only the separation performances of TMAP0.2-P and TMAP0.7-P are depicted in Figure 5a. The minimum residual TiO2 absorbance values were located only around 32% for the lowest charged sample (TMAP0.2-P) and 25% for the highest charged one (TMAP0.7-P) at doseop values which increased with the ionic group content (see Table 3, Section 3.1. Materials).  The poor separation performance could be an outcome of the weak non-electrostatic and electrostatic attractive forces between the pullulan derivatives and TiO 2 particles. The former kind of force was checked in an experiment where the unmodified pullulan sample has been used as a flocculant (Figure 5a, inset). The minimum residual TiO 2 absorbance was 73%. This means that the attractive electrostatic interactions between the negative charges on the TiO 2 particle surface (TiO − ) (ζ = −38.1 mV) and the cationic groups on the polymeric chains play the most important role in the TiO 2 particle separation. However, they are quite weak, as mentioned above, a finding also revealed by zeta potential measurements (Figure 5b). For both of the TMAP x -P samples, the increase in ζ values was small compared to the initial value (ζ TiO2 = −38.1 mV), over the entire range of the pullulan derivative doses investigated (up to ζ = −20 mV for the TMAP 0.2 -P and ζ = −16.8 mV for the TMAP 0.7 -P). The electrostatic repulsion interactions between the positively charged pullulan derivative segments adsorbed on different TiO 2 particles and/or the steric ones (between the uncharged segments on the polyelectrolyte chains extended into solution, as tails, loops) could, probably, obstruct, to some measure, the polyelectrolytes/TiO 2 particles interactions. However, one has also to remark that the pullulan derivatives efficacy in the removal of this oxide from dispersion prepared in water was higher than those of some cationic dextran derivatives (hydrophilic and amphiphilic)-removal efficiency values between 50% and 60% [17]-and PEG and the Pluronic-type polymers-removal efficiency values 56% and 20%, respectively [19] (see Table 1). Previously, Divakaran and Pillai [18] found that chitosan did not clarify a suspension of TiO 2 particles prepared in distilled water but only in tap water; the authors surmised that some species present in the tap water played an important role in the flocculation of TiO 2 particles.
From the data presented above, the pullulan derivatives proved to be better flocculants for the FeO particles than for the TiO 2 ones; therefore, the impact of some other parameters, namely c ip , dispersion pH, and composition (c FeO , salts, and kaolin presence) on their removal, was further investigated and presented below. As the impact of these parameters on the flocculation efficacy was similar for all of the three pullulan derivatives (data not shown), only one polymer sample was chosen as an example: TMAP 0.4 -P for the c ip influence and TMAP 0.7 -P for the influence of the other parameters.

Initial Solution Concentration of the Pullulan Derivatives (c ip ) and FeO Dispersion (c FeO )
The initial solution concentration of the pullulan derivatives, c ip , refers to the concentration of the polyelectrolyte stock solution added to the clay suspension. The polymer chains conformation may vary as a function of concentration regimes, from the extended (rigid) (dilute regime) to the overlapped or even entangled one (semidilute regime) which could lead to modification of some properties, including the separation one. To find out the impact of this parameter on the FeO particle removal, flocculation tests have been undertaken with solutions of TMAP 0.4 -P with three different initial polymer concentrations, namely c ip = 0.1 g L −1 , 0.3 g L −1 , and 1 g L −1 . They are located below, close to, and above the critical overlap concentration, c* = 0.35 g L −1 , which delimits the diluted and semi-diluted concentration regimes. c* was estimated by means of Equation (1) [30]: (1) where However, the doseop values were different for the three cip investigated. They are higher for cip located either in the dilute regime (cip =0.1 g L −1 ) or in the semidilute one (cip = 1 g L −1 ), doseop= 0.5 mg L −1 and doseop = 0.6 mg L −1 , respectively, than that recorded for cip close to c* (doseop = 0.3 mg L −1 ). Probably, at this concentration, the polymer chains have a better arrangement on the FeO particleʹs surface than in the other two cases, making easier the electrostatic attractions between the oppositely charged groups of the polymer chains and FeO particles, improving, thus, the flocculation process. One has to underline here, that this result, namely the dependence of doseop on cip is different from that obtained when hydrophilic ionic polymers based on dextran were used for the flocculation of some clay particles [27]-the doseop values were almost the same irrespective of cip; only the residual turbidity corresponding to doseop varied.
These investigations revealed clearly the cip influence on the FeO particles removal process showing that a suitable concentration of the initial solution of TMAP0.4-P is required to get the best contaminant removal: cip = 0.3 g L −1 in the flocculation processes where low polymer doses are required while cip = 1 g L −1 is more suitable when a higher removal efficiency is preferred.
As regards the impact of the FeO particles concentration (cFeO) on the flocculation However, the dose op values were different for the three c ip investigated. They are higher for c ip located either in the dilute regime (c ip = 0.1 g L −1 ) or in the semidilute one (c ip = 1 g L −1 ), dose op = 0.5 mg L −1 and dose op = 0.6 mg L −1 , respectively, than that recorded for c ip close to c* (dose op = 0.3 mg L −1 ). Probably, at this concentration, the polymer chains have a better arrangement on the FeO particle s surface than in the other two cases, making easier the electrostatic attractions between the oppositely charged groups of the polymer chains and FeO particles, improving, thus, the flocculation process. One has to underline here, that this result, namely the dependence of dose op on c ip is different from that obtained when hydrophilic ionic polymers based on dextran were used for the flocculation of some clay particles [27]-the dose op values were almost the same irrespective of c ip ; only the residual turbidity corresponding to dose op varied.
These investigations revealed clearly the c ip influence on the FeO particles removal process showing that a suitable concentration of the initial solution of TMAP 0.4 -P is required to get the best contaminant removal: c ip = 0.3 g L −1 in the flocculation processes where low polymer doses are required while c ip = 1 g L −1 is more suitable when a higher removal efficiency is preferred.
As regards the impact of the FeO particles concentration (c FeO ) on the flocculation efficiency of the pullulan derivatives, in addition to c FeO = 0.4 g L −1 (Figure 1), experiments on FeO particles suspensions with c FeO = 0.2 g L −1 and 0.6 g L −1 at the natural suspensions pH were performed; for this test, the TMAP 0.7 -P sample was used as a flocculant (Figure 6b).
These figures point out a negligible influence of c FeO on the dose op when suspensions with c FeO = 0.2 g L −1 and 0.4 g L −1 were used, namely 0.1 mg L −1 in case of the former concentration and 0.14 mg L −1 for the latter one. A notable influence of c FeO on the dose op was observed for the highest c FeO when a removal efficiency of around 95% and more was obtained for an interval of dose op between 0.14 mg L −1 and 0.8 mg L −1 . This, probably, occurs because a greater content of FeO particles in dispersion requires a higher TMAP 0.7 -P chains amount for neutralization and aggregation, hence the increase in dose op .

Suspension pH
It is well known that in real wastewater, the suspension pH could change and influence, thus, the flocculation process as it can affect the charge content on the polymer chains (hence, their conformation) or particle surface. To find out the influence of this parameter on the flocculation efficacy, the experimental tests were carried out using the TMAP 0.7 -P sample on suspension of FeO particles with pH 9.5, and the results (Figure 7) were compared to those obtained on the initial suspension of FeO particles with pH 7.1 (Figure 1). The suspension pH was adjusted using 0.1M NaOH solution. The investigation of the pullulan derivative in the removal of the FeO particles at lower pH values than the natural one has not led to consistent results due to the instability of the FeO particles.  TMAP0.7-P (like all the TMAPx-P samples) is a strong polyelectrolyte which means it is completely ionized irrespective of the pH value and consequently, its conformation does not depend on this parameter. As regards the FeO particles, a small increase in the medium pH from the natural one, pH 8.34 to pH 9.5, determined an increase in ζ values from −26.7 mV to −47.6 mV, which indicated an increase in the (FeO − ) groups concentration on the oxide particles surface against that at natural pH. As seen in Figure  8, a slight enhancement of the removal efficacy (98%) and a pronounced one of the doseop (6 mg L −1 ) were recorded; this could be explained by the stronger electrostatic attraction interactions between the FeO particles and the positive groups of TMAP0.7-P. This result shows that the pullulan derivative sample could be successfully used for the removal of TMAP 0.7 -P (like all the TMAP x -P samples) is a strong polyelectrolyte which means it is completely ionized irrespective of the pH value and consequently, its conformation does not depend on this parameter. As regards the FeO particles, a small increase in the medium pH from the natural one, pH 8.34 to pH 9.5, determined an increase in ζ values from −26.7 mV to −47.6 mV, which indicated an increase in the (FeO − ) groups concentration on the oxide particles surface against that at natural pH. As seen in Figure 8, a slight enhancement of the removal efficacy (98%) and a pronounced one of the dose op (6 mg L −1 ) were recorded; this could be explained by the stronger electrostatic attraction interactions between the FeO particles and the positive groups of TMAP 0.7 -P. This result shows that the pullulan derivative sample could be successfully used for the removal of FeO particles at a higher pH value than the natural one. Figure 7. The residual FeO absorbance (%) dependence on the polymer dose (TMAP0.7-P) at pH 9.5, cFeO = 0.4 g L −1 , cip = 1 g L −1 , settling time 120 min. TMAP0.7-P (like all the TMAPx-P samples) is a strong polyelectrolyte which means it is completely ionized irrespective of the pH value and consequently, its conformation does not depend on this parameter. As regards the FeO particles, a small increase in the medium pH from the natural one, pH 8.34 to pH 9.5, determined an increase in ζ values from −26.7 mV to −47.6 mV, which indicated an increase in the (FeO − ) groups concentration on the oxide particles surface against that at natural pH. As seen in Figure  8, a slight enhancement of the removal efficacy (98%) and a pronounced one of the doseop (6 mg L −1 ) were recorded; this could be explained by the stronger electrostatic attraction interactions between the FeO particles and the positive groups of TMAP0.7-P. This result shows that the pullulan derivative sample could be successfully used for the removal of FeO particles at a higher pH value than the natural one.

FeO Particles in Suspensions Containing Salts, Kaolin, or TiO 2 Particles
The real wastewater may contain low molecular salts, clay particles, mixtures of oxides, dyes, etc., which could influence, in a certain extent, the flocculation efficacy of a polymer. In order to evaluate the influence of other contaminants on the FeO removal percent by the pullulan derivatives investigated here, experiments on suspensions containing along with FeO particles (0.2 g L −1 ), different salts (NaCl, Na 2 SO 4 , NaNO 3 , CaCl 2 , MgCl 2 ) (each salt with c s = 1 × 10 −3 M) and kaolin (0.2 g L −1 ) or TiO 2 particles (FeO/TiO 2 (w/w, 1/1)) were carried out.
As can be seen in Figure 8a, the presence of salts and kaolin particles in the FeO suspension has not decreased the separation performance of TMAP 0.7 -P, the residual FeO absorbance being very low, between 1.31% and 3.33%, at polymer doses located in the interval 0.14 mg L −1 and 2 mg L −1 , after a settling time of 30 min.
In order to see the effects of salts or kaolin particles on the FeO particle suspensions stability, two kinds of suspensions have been prepared: one containing FeO particles and salts and the other one, FeO and kaolin particles. The residual FeO absorbance declined to about 52.66% in the former case and 93% in the latter one, after the same settling time. This means that even though the residual FeO absorbance decrease is much lower than in the presence of TMAP 0.7 -P, the salts in suspension play a certain role in the separation process.
As both metal oxides could be found in wastewater (for example in the ceramic ones), the flocculation performance of TMAP 0.7 -P (the sample with the smallest dose op used for both metal oxides) on suspensions containing mixtures of FeO/TiO 2 (w/w, 1/1) has been checked (Figure 8b). It turned out that the polymer sample was a very good separation agent for the suspension containing the mixture of FeO and TiO 2 , a removal efficiency of about 90% of both metal oxides being noticed in the interval of polymer doses between 0.14 mg L −1 and 0.24 mg L −1 after 120 min settling time. It seems that the presence of the FeO particles helps the polymer to remove TiO 2 particles; in the absence of TMAP 0.7 -P, the residual absorbance values of the two oxides in the mixture were 56.28% (FeO) and 51.13% (TiO 2 ), after the same settling time.
Additional evidence for the role played by the polymer in the separation of the above mixtures was obtained with the help of the particle s surface morphology (Figure 9a-d).
both metal oxides) on suspensions containing mixtures of FeO/TiO2 (w/w, 1/1) has been checked (Figure 8b). It turned out that the polymer sample was a very good separation agent for the suspension containing the mixture of FeO and TiO2, a removal efficiency of about 90% of both metal oxides being noticed in the interval of polymer doses between 0.14 mg L −1 and 0.24 mg L −1 after 120 min settling time. It seems that the presence of the FeO particles helps the polymer to remove TiO2 particles; in the absence of TMAP0.7-P, the residual absorbance values of the two oxides in the mixture were 56.28% (FeO) and 51.13% (TiO2), after the same settling time.
Additional evidence for the role played by the polymer in the separation of the above mixtures was obtained with the help of the particleʹs surface morphology (Figure 9a-d). The scanning electron micrographs reveal that, in the absence of TMAP 0.7 -P, both mixtures contain separated particles of different sizes and shapes (Figure 9a,c), while bigger aggregates are formed after the interactions with polycation chains (dose op : 0.6 mg L −1 and dose op : 0.2 mg L −1 , respectively) (Figure 9b,d), proving once more the role of TMAP 0.7 -P in the particles aggregation and separation.

Effect of the Pullulan Derivative/FeO Flocs on the Reduction of the Fungicide Bordeaux Mixture Particles Content from Synthetic Wastewater
It is well known that the flocculation process generates different types of sludge that, apart from a large amount of water, could contain the removed contaminants, the coagulants (aluminum-or iron-based salts)/flocculants (polymers), etc., depending on the composition of the wastewater treated and type of the treatment method applied. These, in their turn, could contaminate the groundwater, the land, etc., challenging, thus, the scientific community to find solutions for proper wastewater sludge management (disposal, regeneration, reuse). As some of them are expensive or difficult to apply (landfills, incineration, compression into building blocks, digestion) [32][33][34][35], some other cost-effective and environmentally friendly options have been investigated. Thus, the use of sewage sludge and that generated by various wastewater treatment plants, as adsorbents, has become an alternative to chemicals to remove heavy metals, phosphorus/phosphates, dyes, phenolic compounds, pesticides, etc. [36][37][38][39][40]. Regarding pesticides, one of the most dangerous types of contaminants for the whole environment (groundwater and surface water, atmosphere, soil), fertilizer and steel industry wastes for removal of 2,4-D and carbofuran pesticides [41], alum sludge for removal of glyphosate [42], activated sludge followed by pine bark adsorption for lindane and heptachlor [43], etc., have been used. Quite recently, in a preliminary test, some flocs, obtained in the flocculation process of zirconium silicate (kreutzonit) particles by grafted cationic derivatives based on pullulan, exhibited remarkable efficacy in the removal of the Bordeaux mixture particles, a fungicide widely used to combat diseases of fruits, grapes, ornamental plants, etc., from simulated wastewater; a removal efficiency higher than 95% (residual BM content of 25 mg L −1 ) at an optimum amount of flocs (10 g flocs/1000 mL BM suspension) was found [14]. Both this result and the lack of information regarding the use of polymer/contaminant flocs, as adsorbents for pesticides, motivated us to also test the ability of the FeO/TMAP x -P flocs to reduce the BM particles content from a model suspension.
For the separation tests, performed according to the method already reported by Ghimici and Constantin [14], flocs of TMAP 0.7 -P/FeO, dried for several days at room temperature and 1 day under vacuum, and a BM sample (MIF type -IQV, Spain) (a combination of 20% copper sulfate, slaked lime (Ca(OH) 2 ), and water) have been used. Different amounts of FeO/TMAP 0.7 -P flocs were added under stirring (200 rpm) to 50 mL BM particle suspension placed into 100 mL beakers (c, w/w) = 0.5 g L −1 ), previously sonicated for 15 min. Stirring of the mixture was continued for a different time when the flocs amount was kept constant and 1 h for the experiments when this varied. In order to evaluate the BM removal percent, supernatant absorbance measurements were performed after the established settling times of the suspension. Figure 10a depicts the impact of both the stirring and settling time on the BM particle separation when 0.1 g flocs were used as adsorbent. The stirring time had no effect on the separation process, the residual BM absorbance values being constant over 20 h of stirring, at the same settling time. On the other hand, the BM particle removal increased with the settling time, irrespective of the period of stirring. The highest removal efficiency increase was observed after the settling time of 20 h (around 50%), with further increase in this parameter leading to an increase in removal efficiency to a lesser extent, up to 65% (48 h) and 70% (72 h). Given the finding above, Figure 10b shows only the data collected after 20 h and 48 h, for the experiments with different amounts of flocs. The flocs amount was in the range of 0.1 g/50 mL to 0.7 g/50 mL suspension of BM. The plotted data show a pronounced downtrend of the fungicide content in the model suspension with the flocs amount augmentation up to 0.5 g of flocs/50 mL suspension when a residual BM absorbance of around 10% (corresponding to a residual BM concentration of about 0.049 g L −1 ) was observed; the calibration curve for the BM suspension was not shown here. For the higher amounts of flocs, the residual BM absorbance values increase, which means the decline in removal efficacy of FeO/TMAP0.7-P flocs. In order to explain these results, we performed additional experiments where a solution of TMAP0.7-P or FeO particles was used for the BM suspension separation. As Figure 11 shows, the polymer sample has shown a very good separation efficiency for the fungicide particles, namely 94% (polymer dose 6 mg L −1 ), after 2 h settling. On the other hand, the BM particle removal increased with the settling time, irrespective of the period of stirring. The highest removal efficiency increase was observed after the settling time of 20 h (around 50%), with further increase in this parameter leading to an increase in removal efficiency to a lesser extent, up to 65% (48 h) and 70% (72 h). Given the finding above, Figure 10b shows only the data collected after 20 h and 48 h, for the experiments with different amounts of flocs. The flocs amount was in the range of 0.1 g/50 mL to 0.7 g/50 mL suspension of BM. The plotted data show a pronounced downtrend of the fungicide content in the model suspension with the flocs amount augmentation up to 0.5 g of flocs/50 mL suspension when a residual BM absorbance of around 10% (corresponding to a residual BM concentration of about 0.049 g L −1 ) was observed; the calibration curve for the BM suspension was not shown here. For the higher amounts of flocs, the residual BM absorbance values increase, which means the decline in removal efficacy of FeO/TMAP 0.7 -P flocs. In order to explain these results, we performed additional experiments where a solution of TMAP 0.7 -P or FeO particles was used for the BM suspension separation. As Figure 11 shows, the polymer sample has shown a very good separation efficiency for the fungicide particles, namely 94% (polymer dose 6 mg L −1 ), after 2 h settling. observed; the calibration curve for the BM suspension was not shown here. For the higher amounts of flocs, the residual BM absorbance values increase, which means the decline in removal efficacy of FeO/TMAP0.7-P flocs. In order to explain these results, we performed additional experiments where a solution of TMAP0.7-P or FeO particles was used for the BM suspension separation. As Figure 11 shows, the polymer sample has shown a very good separation efficiency for the fungicide particles, namely 94% (polymer dose 6 mg L −1 ), after 2 h settling. Figure 11. The residual BM absorbance (%) dependence on the polymer dose (TMAP0.7-P); cip = 1 g L −1 ; cBM = 0.5 g L −1 . Figure 11. The residual BM absorbance (%) dependence on the polymer dose (TMAP 0.7 -P); c ip = 1 g L −1 ; c BM = 0.5 g L −1 .
Regarding the implication of the FeO particles in the separation process, the tests have shown residual BM absorbance values around 48% (20 h settling time) and 33% (48 h settling time) for particles FeO amount/50 mL BM suspension between 0.01 g and 0.03 g. Given that in the absence of the metal oxide particles, the residual absorbance value for fungicide were almost the same, 49.19% (20 h settling time) and 36.33% (48 h settling time), one may assume that there are no interactions between the fungicide and the FeO particles. From these results, one may surmise that the polymer chains on the flocs surface have a decisive role in the fungicide separation process. Here, we have to remind the reader that the main flocculation mechanism implied in the FeO particle removal by the pullulan derivatives TMAP x -P is the charge patch one (see Section 2.1). It is possible that the polymer chains adsorbed on the FeO particles have free positive charges which could attract the negative species of BM (for example, SO 4 2− ) (ζ = −20 mV), entailing the fungicide particle separation. Moreover, the interactions between the polymer chains and the above-mentioned contaminant could be also favored by the Cu 2+ ions affinity to the amine/carbamate groups within its chemical structure (Scheme 1, Section 3.1. Materials).
Supplementary evidence for the BM particle retention by the FeO/TMAP 0.7 -P flocs was obtained by both particle size distribution ( Figure 12a) and EDX (Figure 12b,c)  Regarding the implication of the FeO particles in the separation process, the tests have shown residual BM absorbance values around 48% (20 h settling time) and 33% (48 h settling time) for particles FeO amount/50 mL BM suspension between 0.01 g and 0.03 g. Given that in the absence of the metal oxide particles, the residual absorbance value for fungicide were almost the same, 49.19% (20 h settling time) and 36.33% (48 h settling time), one may assume that there are no interactions between the fungicide and the FeO particles. From these results, one may surmise that the polymer chains on the flocs surface have a decisive role in the fungicide separation process. Here, we have to remind the reader that the main flocculation mechanism implied in the FeO particle removal by the pullulan derivatives TMAPx-P is the charge patch one (see Section 2.1). It is possible that the polymer chains adsorbed on the FeO particles have free positive charges which could attract the negative species of BM (for example, SO4 2− ) (ζ = −20 mV), entailing the fungicide particle separation. Moreover, the interactions between the polymer chains and the abovementioned contaminant could be also favored by the Cu 2+ ionsʹ affinity to the amine/carbamate groups within its chemical structure (Scheme 1, Section 3.1. Materials).
Supplementary evidence for the BM particle retention by the FeO/TMAP0.7-P flocs was obtained by both particle size distribution ( Figure 12a) and EDX (Figure 12b,c) measurements. Thus, the higher size of the flocs (D50 = 40 ± 2.36 μm) formed after the retention of BM particles (Figure 12a) compared to that of TMAP0.7-P/FeO (8.32 ± 0.39 μm, Figure 2) and the presence of Cu ions on the FeO/TMAP0.7-P/BM floc surface demonstrated by EDX (Figure 12c) enforce the ability of the FeO/TMAP0.7-P to separate BM particles from suspension. Thus, the higher size of the flocs (D50 = 40 ± 2.36 µm) formed after the retention of BM particles (Figure 12a) compared to that of TMAP 0.7 -P/FeO (8.32 ± 0.39 µm, Figure 2) and the presence of Cu ions on the FeO/TMAP 0.7 -P/BM floc surface demonstrated by EDX (Figure 12c) enforce the ability of the FeO/TMAP 0.7 -P to separate BM particles from suspension.
Although TMAP 0.7 -P/FeO flocs exhibited a slightly lower removal efficiency compared to that shown by the grafted pullulan derivatives/kreutzonit flocs, this result could give useful information regarding the beneficial use of the pullulan-based flocculants/FeO flocs for removal of the BM particles from simulated suspensions. We have to mention here that this study is the first step of a series of investigations where the flocs resulting from the flocculation of different inorganic particles (other metal oxides, clays) will be tested as adsorbents for other toxic organic contaminants (pesticides, dyes, etc.) from wastewater.

Materials
Cationic pullulan derivatives with three different contents of pendant quaternary ammonium groups (TMAPx-P) were prepared by a two steps reaction (Scheme 1): Firstly, pullulan (P) (Mw = 200 Kg mol −1 purchased from Hayashibara Lab. Ltd. (Okoyama, Japan)) was modified by the amidation reaction with 3-dimethylamino-1-propylamine (DMAPA) in dimethyl sulfoxide (DMSO) and in the presence of N,N'carbonyldiimidazole (CDI) as activator and 4-dimethylaminopyridine (DMAP) as a catalyst, according to Constantin et al. [7,8]. Then, the pullulan derivatives containing three different amounts of tertiary amine groups (DMAPAx-P) were used as precursors in the synthesis of quaternary ammonium derivatives possessing methyl groups (TMAPx-P) [9]. Briefly, a methyl iodide solution (MeI) (18%, w/v; MeI/DMAPA molar ratio = 3/1) in DMF was dropwise added to the DMAPAx-P solution in DMF (5%, w/v) at 20 °C for 15 min. and the quaternization reaction was continued at 50 °C for 24 h. The obtained product (TMAPx-P) was precipitated in acetone. The precipitate was dissolved in 15% (w/v) NaCl solution in order to replace the iodide ions with chloride ions followed by dialysis with deionized water for 3 days to remove inorganic materials and then freezedried.
Following these procedures, pullulan derivatives with various degrees of substitution (DS) with quaternary ammonium groups (number of amino groups per anhydroglucose unit in pullulan), namely 0.2, 0.4, and 0.7, were successfully obtained (chemical structure in Scheme 1). Scheme 1. Synthesis route and chemical structure of TMAPx-P derivatives.
The polymers code is TMAPx-P, where P means pullulan, TMAP is trimethylammonium propyl carbamate group, and x = DS. Table 3 summarizes the synthesis parameters and some characteristics for the pullulan derivatives. Scheme 1. Synthesis route and chemical structure of TMAP x -P derivatives.
The polymers code is TMAP x -P, where P means pullulan, TMAP is trimethylammonium propyl carbamate group, and x = DS. Table 3 summarizes the synthesis parameters and some characteristics for the pullulan derivatives. FeO particles are commercially available in packs with 150 g (Kynita SRL, Valcea, Romania). TiO 2 particles, crystalline form anatase, were supplied by Sigma-Aldrich (Steinheim, Germany).

Methods
The simulated suspensions of metal oxide particles were prepared in distilled water, as follows: (i) FeO model dispersions: concentration (c FeO ): 0.2 g L −1 , 0.4 g L −1 , 0.6 g L −1 , pH = 8.34, zeta potential (ζ) = −26.7 mV; (ii) TiO 2 model dispersions: c TiO2 : 0.05 g L −1 , pH = 5, ζ = −38.1 mV. The polymer solutions with different concentrations (c ip , g L −1 ) = 0.1, 0.3, and 1) were also prepared in distilled water and stabilized at room temperature for 1 day before use. The flocculation experiments were carried out according to the method already reported by Ghimici et al. [27]. Before starting the tests in a Cole Parmer stirrer/hotplate with 9 places, at room temperature, the dispersions of inorganic particles were sonicated for 15 min (ultrasonicator SONICS VCX 750) and then placed into 100 mL beakers (50 mL each). Different volumes of pullulan derivatives solution were added to metal oxide suspensions under stirring (500 rpm) which was kept for another 3 min. Then, the stirring speed was decreased to about 200 rpm for 15 min after which the suspension was left to settle for 120 min (FeO particles) and 1200 min (TiO 2 ones). The above periods of settling time, called in the paper the optimum settling time (OST) (period of time after which the residual absorbance of the supernatant (%) remained almost constant) was established, in previous experiments. Subsequently, absorbance measurements (spectrophotometer SPECOL 1300 Analytik Jena, Jena, Germany) at λ = 635 nm for FeO and 400 nm for TiO 2 particles) and zeta potential ones (Zetasizer Nano-ZS, ZEN-3500 model (Malvern Instruments, Malvern, England)) were performed on the supernatant sample (10 mL). The separation of metal oxide particles from suspension in the absence of the pullulan derivatives under the selected experimental conditions (pH, concentration of polymer and metal oxide particles, etc.) has been also evaluated.
The metal oxides particles removal was given as a percent of the initial absorbance recorded for the inorganic particle suspensions without polymer, at time zero (residual metal oxides particles absorbance (%)): residual metal oxides particles absorbance (%) = A f A i × 100 where, A i = the absorbance of the metal oxides suspension in the absence of polymer; and A f = the absorbance of supernatant after the addition of the pullulan derivative. The size distribution of the FeO particles/flocs before and after the treatment with polycations has been determined with Laser Particle Size Analyzer-Partica LA-960V2-HORIBA (Japan) (D50, µm). In addition, the surface morphology analysis was performed on FeO particles and some flocs using a Verios G4 UC Scanning Electron Microscope (Thermo Scientific, Brno, Czech Republic) equipped with energy dispersive X-ray (EDX) analyzer (Octane Elect Super SDD detector, Mahwah, NJ, USA); SEM analysis was carried out in high vacuum mode using a secondary electron detector (Everhart-Thornley Detector, Thermo Scientific, Brno, Czech Republic) at an accelerating voltage of 5 kV.

Conclusions
The separation of FeO and TiO 2 particles from synthetic wastewater containing the metal oxides alone or a mixture of them, as well as different salts and kaolin by cationic pullulan derivatives with different content of N,N,N-trimethylammonium propyl carbamate chloride (TMAPx-P), has been investigated. The main results of the present investigations may be outlined as follows:

•
The residual metal oxides absorbance values changed with the addition of different doses of pullulan derivatives, irrespective of the particle s dispersion medium composition.

•
The dose op values decreased with increasing ionic group content from 1 mg L −1 .
(TMAP 0.2 -P) to 0.14 mg L −1 (TMAP 0.7 -P) in the case of the FeO particles and from 3 mg L −1 (TMAP 0.2 -P) to 1.4 mg L −1 (TMAP 0.7 -P) in the case of the TiO 2 ones. • For the FeO particles removal, higher dose op values were recorded for the initial polymer concentration, c ip located either in the dilute regime (c ip = 0.1 g L −1 − dose op = 0.5 mg L −1 ) or in the semidilute one (c ip = 1 g L −1 − dose op = 0.6 mg L −1 ) than that recorded for c ip close to c* (c ip = 0.5 g L −1 − dose op = 0.3 mg L −1 ). • A noticeable increase in the dose op (0.6 mg L −1 ) and optimum dose interval (0.14 mg L −1 and 0.8 mg L −1 ) was found for the higher pH value than that of the natural one and the high particle suspension concentration, c Feo = 0.6 g L −1 , respectively.

•
The zeta potential and the particle aggregate size measurement results pointed to the patch flocculation mechanism.

•
The pullulan derivatives proved to be good flocculants for the mixture of FeO and TiO 2 particles (w/w, 1/1) as well as for the FeO particles from suspensions containing mixtures of salts and kaolin particles.

•
The FeO/TMAP 0.7 -P flocs exhibit good performance in the removal of the fungicide Bordeaux mixture particles from simulated wastewater (residual BM content of 0.049 g L −1 ) at the optimum amount of particles (50 g flocs/1000 mL BM suspension).  Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.