**Cross-Linked Magnetic Chitosan/Activated Biochar for Removal of Emerging Micropollutants from Water: Optimization by the Artificial Neural Network**

#### **Amin Mojiri 1,\*, Reza Andasht Kazeroon <sup>2</sup> and Ali Gholami <sup>3</sup>**


Received: 18 February 2019; Accepted: 13 March 2019; Published: 17 March 2019

**Abstract:** One of the most important types of emerging micropollutants is the pharmaceutical micropollutant. Pharmaceutical micropollutants are usually identified in several environmental compartments, so the removal of pharmaceutical micropollutants is a global concern. This study aimed to remove diclofenac (DCF), ibuprofen (IBP), and naproxen (NPX) from the aqueous solution via cross-linked magnetic chitosan/activated biochar (CMCAB). Two independent factors—pH (4–8) and a concentration of emerging micropollutants (0.5–3 mg/L)—were monitored in this study. Adsorbent dosage (g/L) and adsorption time (h) were fixed at 1.6 and 1.5, respectively, based on the results of preliminary experiments. At a pH of 6.0 and an initial micropollutant (MP) concentration of 2.5 mg/L, 2.41 mg/L (96.4%) of DCF, 2.47 mg/L (98.8%) of IBP, and 2.38 mg/L (95.2%) of NPX were removed. Optimization was done by an artificial neural network (ANN), which proved to be reasonable at optimizing emerging micropollutant elimination by CMCAB as indicated by the high R2 values and reasonable mean square errors (MSE). Adsorption isotherm studies indicated that both Langmuir and Freundlich isotherms were able to explain micropollutant adsorption by CMCAB. Finally, desorption tests proved that cross-linked magnetic chitosan/activated biochar might be employed for at least eight adsorption-desorption cycles.

**Keywords:** chitosan; diclofenac; ibuprofen; magnetic biochar; naproxen

#### **1. Introduction**

Emerging micropollutants or organic micropollutants exist in the environment at trace concentrations, and their impact on the human health and the environment are presently unknown. These pollutants are contained in polycyclic aromatic hydrocarbons (PAH), personal care products, pharmaceuticals, pesticides, industrial chemicals, and metallic trace elements [1]. Pharmaceutical micropollutants are commonly found in various environmental compartments. The growing use of pharmaceuticals is raises questions regarding their potential risk to human health, the environment, and water quality [2]. Diclofenac, ibuprofen, and naproxen are non-steroidal anti-inflammatory drugs (NSAIDs), which are a commonly consumed class of pharmaceuticals [3]. All pharmaceuticals belonging to this group are acidic in nature with pKa values in the range of 3–5 [4].

Among the different non-steroidal anti-inflammatory drugs, diclofenac is widely applied. Diclofenac (Figure 1; 2-((2,6-dichlorophenyl)amino)phenylacetic-acid) has been stated to cause chronic results such as renal and gastrointestinal tissue damage in some vertebrates [5]. Ibuprofen (Figure 1;

2-(4-Isobutylphenyl)propionic acid) is applied for the treatment of pain and inflammation and dropping of a fever [6]. Naproxen (Figure 1) enters aquatic environments chiefly over the effluents of wastewater treatment plants. It is categorized as a high-priority pharmaceutical. Naproxen might affect living organisms and diminish the biodiversity of natural environmental communities because of its biological activities [7]. Mostly, conventional wastewater treatment methods fail to eliminate pharmaceuticals totally from the water [2]. One of the most promising ways to remove emerging micropollutants is by using adsorbents. Biochar and chitosan are low-cost adsorbents which have been previously used in the literature to remove micropollutants from water [8,9].

**Figure 1.** Structures of the studied organic micropollutants [10,11].

Chitosan is one of the biopolymers that is derived from chitin; chitin is a natural amino polysaccharide and is composed primarily of repeating β-(1,4)-2-amino-2-deoxy-d-glucose (or d-glucosamine) units. The benefits of chitosan include its low cost, ease of polymerization and functionalization, and good stability [12]. Amouzgar and Salamatinia [8] stated that chitosan has certain capabilities for removing emerging micropollutants from water. Another low-cost adsorbent is biochar.

Thermochemical decomposition procedures transform biomass materials to syngas, bio-oil, and biochar. Biochar is low cost, environmentally friendly, and can be applied for a variety of purposes [13]. Quesada et al. [14] stated that using biochar as a low-cost material is a promising way to eliminate pharmaceuticals from wastewater. Sizmur et al. [15] stated that the activation process improves the surface area and porosity of biochar, so its adsorption capacity might be increased. Hence, this study aimed to produce a new cross-linked magnetic chitosan/activated biochar to remove emerging micropollutants and to optimize the removal efficiency using an artificial neural network (ANN). This experiment design and its optimization process have not been previously reported in the literature.

#### **2. Materials and Methods**

#### *2.1. Materials*

In this study, biochar was extracted from agricultural residues. Chitosan (medium molecular weight; code: 07947-52), diclofenac sodium (DCF; C14H10Cl2NNaO2; 98%; molecular weight = 294.05 g/mol), ibuprofen (IBP; C13H18O2; 98%; molecular weight = 206.3 g/mol), and naproxen (NPX; 98%; C14H14O3; molecular weight = 230.2 g/mol) were obtained from Sigma–Aldrich in the analytical purity and applied in the experiments directly without any further purification. Chloroform, acetone, and methanol (99.5% mass purity) were from Merck.

#### *2.2. Producing Cross-Linked Magnetic Chitosan/Activated Biochar (CMCAB)*

Based on the method by Liu et al. [16] in the first step, magnetic fluid was prepared by a co-precipitation technique. Fe2+ and Fe3+ (molar ratio 2:3) solution was placed into a beaker using a stirrer at 55 ◦C, then NaOH solution was added dropwise with continuous stirring for almost 15 min until the pH got to 9.0. After altering the temperature of the reaction vessels to 65 ◦C, 0.8 mL Tween 80 was augmented into the mixture using a stirrer for 30–40 min, and the pH value was adjusted to 7.0. After that, the product was washed with distilled water three times and was dispersed in an ultrasonic device for 40 min. Finally, the solution was diluted to gain magnetic fluid (40 g L<sup>−</sup>1).

In the second step, the activated biochar was produced. Biochar extracted from agricultural residues was done by an activation process with 4 M NaOH for 2 h and then dried for 12 h at 105 ◦C. Then, the biochar was separated from the NaOH solution via a Buchner filter funnel, heated at 800 ◦C for 2 h under a 2 L/min nitrogen gas flow, and then let to cool at a rate of 10 ◦C /min. The activated biochar was washed consecutively with deionized (DI) water and 0.1 M HCl to attain pH 7 and dried again at 105 ◦C. As a final point, the activated biochar was crushed and sieved through a 200-mesh (74 μm) sieve [17].

Finally, to achieve the cross-linked magnetic chitosan/activated biochar, 5.0 g of chitosan was dissolved in 250 mL 2% acetic solution with stirring. Next, 25 mL of magnetic fluid was added dropwise into the solution with constant stirring for 30 min in a water bath at 50 ◦C. Then, 5.0 g activated biochar was augmented with continuous stirring for another 60 min. Afterward, 6 mL of glutaraldehyde was injected into the reaction system to produce a gel and the pH of the reaction system was adjusted to 8.0–10.0. As a final point, the mixture was retained in a water bath for 1 h. The precipitate was washed till the pH touched about 7 and was dried at 60 ◦C and sieved [16]. Table 1 shows the features of the cross-linked magnetic chitosan/activated biochar (CMCAB). The CMCAB features were monitored by the Autosorb (Quantachrome AS1wintm, version 2.02, Quantachrome Instruments, Boynton Beach, FL, USA). In terms of the BET technique, the specific surface area and pore size distribution of CMCAB with the specific surface area and pore size distribution analyzers were determined under the conditions of liquid nitrogen temperature. The zeta potential of the CMCAB was analyzed by the zeta potential meter (Zetasizer nano-ZS90, Malvern Panalytical Ltd, Malvern, UK) at 25 ◦C in different pH (3–9).


**Table 1.** Characteristics of the cross-linked magnetic chitosan/activated biochar (CMCAB).

#### *2.3. Producing the Synthetic Aqueous Solution and Experiment Design*

Stock solutions of organic micropollutants were prepared in acetone, chloroform, or methanol as described by Sühnholz et al. [18]. In this study, the initial concentration of organic micropollutants ranged from 0.5 mg/L [19] to 3 mg/L [20]. The pH was varied from 4 to 8 [21]. Based on preliminary experiments, the adsorption time (h) was fixed at 1.5, which is in line with selected ranges by Kim et al. [9]. Based on preliminary experiments, the adsorbent dosage was fixed at 1.6 g/L, which is in line with the findings of Wu et al. [22]. Based on preliminary experiments, each run was carried out at room temperature (25 ± 1 ◦C) using a shaker with 300 rpm shaking speed for all conditions [17,23]. A schematic of the current study is shown in Figure 2.

**Figure 2.** Schematic of experiments.

#### *2.4. Analytical Techniques*

All analytical methods were conducted on the basis of the standard methods [24]. The concentrations of emerging micropollutants were tested via ultraviolet spectra and measured by a high-pressure liquid chromatography (HPLC) (LC-20AT, Shimadzu International Trading (Shanghai) Co., Ltd., Tokyo, Japan). The analytical techniques for diclofenac (DCF), ibuprofen (IBP), and naproxen (NPX) were obtained from the literature [25]. The applied mobile phase contained a mixture of acetonitrile and 0.2% formic acid in water (60:40, v/v) at a flow rate of 0.8 mL min<sup>−</sup>1. The concentrations of DCF, IBP, and NPX were tested using a UV detector at the wavelengths of 200, 200, and 230 nm.

#### *2.5. Optimization Analysis using an Artificial Neural Network (ANN)*

The percentage of micropollutants (MP) eliminated from the solution was estimated using Equation (1)

$$\text{Removal}\,\%\,=\,\frac{\text{C}\_{\text{i}}-\text{C}\_{\text{f}}}{\text{C}\_{\text{i}}}\times 100\tag{1}$$

The initial concentration of MP and the final concentration of MP are denoted by Ci and Cf, respectively.

MATLAB R2015a software (R2015a, Mathsworks, Natick, MA, USA) was applied to model the adsorption procedure on the basis of an ANN. Figure 3 displays the topology for the ANN and the variation of parameters in this study. The two neurons in the input layer represent pH (4–8) and micropollutant concentration (0.5–3 mg/L). There were four neurons in the hidden layer and one neuron in the output layer (removal efficiency) for modeling each micropollutant elimination. A total of 50 experimental results applied to model the network were divided randomly into training (60%), validation (20%), and test (20%) sets [26]. The ANN performance was defined based on the values of the mean squared error (MSE) and coefficient of determination (R2). They were respectively evaluated using Equations (2) and (3). Levenberg–Marquardt (LM) was applied to train the model, and validation was stopped when the maximum validation failures were equal to zero.

$$\text{MSE} = \frac{1}{\mathbf{N}} \sum\_{i=1}^{N} (\left| \mathbf{y}\_{\text{prd},i} - \mathbf{y}\_{\text{exp},i} \right|)^2 \tag{2}$$

$$\mathbf{R}^2 = 1 - \frac{\sum\_{i=1}^{N} \left( \mathbf{y}\_{\text{prd, i, y}} \mathbf{y}\_{\text{exp, i}} \right)}{\sum\_{i=1}^{N} \mathbf{y}\_{\text{prd, i}} - \mathbf{y}\_{\text{m}}},\tag{3}$$

*Water* **2019**, *11*, 551

In Equations (4) and (5), yprd,i refers to the predicted value using the ANN model, yexp,i is the experimental value, N is the number of datapoints, and ym indicates the average of the experimental values.

**Figure 3.** Schematic of an artificial neural network (ANN) design.

#### *2.6. Adsorption Isotherm Study*

Batch adsorption studies were done via different dosages (1–7 g/L) of the CMCAB in a fixed MP concentration (2.5 mg/L), pH (6), and adsorption time (30 min). Beakers with working volumes of 100 mL were shaken at 300 rpm for 30 min.

The capacity of adsorption (mg/g) was estimated via the following Equation (4) [27]:

$$\mathbf{q}\_{\rm c} = \frac{(\mathbf{C}\_0 - \mathbf{C}\_{\rm cq})\mathbf{V}}{\mathbf{m}\_{\rm s}},\tag{4}$$

where the initial micropollutant (MP) concentration is denoted by qe, Ceq is the MP concentration (mg L−1) at equilibrium, the volume of solution (L) is represented by V, and ms is the mass of the adsorbent (g).

#### *2.7. Regeneration and Desorption Study*

Regeneration studies were carried out to monitor the economic usability of the CMCAB adsorbent. The adsorbent was regenerated by soaking in 100 mL methanol for 2–3 h in batch experiments and then washed using distilled water in order to consider the desorption and regeneration of the CMCAB. Eight adsorption/desorption cycles were carried out. After every cycle, the residual concentration of MPs was monitored [28].

#### **3. Results and Discussion**

The efficiency of the removal of emerging micropollutants via cross-linked magnetic chitosan/activated biochar (CMCAB) is shown in Table 2. Figure 4 shows the FTIR results of CMCAB.

In the FTIR results of chitosan (Figure 4a), peaks 3398 and 2913 can be attributed to O–H and C–H, respectively [29]; peak 1613 may be related to C = O [30]. N–H and CH–OH could explain peaks 1584 and 1401, respectively [29]; and peak 837 is attributed to CH groups [30]. In the FTIR results of activated biochar (Figure 4b), peaks 3207 and 2981 are attributed to O–H and C–H, respectively, while peaks 1608 and 1513 may be related to C = O and C = C, respectively [31]. C–O and O–H could be responsible for peak 1201 [31], and peak 842 is attributed to C–H groups [30]. In the FTIR results of the CMCAB (Figure 4c), peaks 3496 and 2915 are attributed to –OH (or –NH) and C–H, respectively [29]. Peaks C = N and C–O could explain peaks 1638 and 1043, respectively [32,33] and peaks 771 and 573 are attributed to Fe–O [32,33]. The zeta potential of CMCAB is shown in Figure 5. Based on Figure 5, the zeta potential of CMCAB was positive in pH (3) to (5) it is in line with finding of Liu et al. [16] and Zhang et al. [33]. Zeta potentials (mV) were 19, 16 and 1 in pH (3), pH (4) and pH (5), respectively. After that zeta potential became negative which could be supported by findings of Zhang et al. [33]. Zeta potentials (mV) were −5, −6, −8 and −11 in pH (6), pH (7), pH (8), and pH (9), respectively. It should be mentioned that the zero point during the zeta potential testing for CMCAB was reached at 5.2 of pH, which could be supported by findings of Liu et al. [16].

**Figure 4.** FTIR images of chitosan (**a**), activated Biochar (**b**), and cross-linked magnetic chitosan/activated biochar (CMCAB) (**c**).

**Figure 5.** The zeta potential of CMCAB in pH (3–9).


**Table 2.** Elimination of diclofenac (DCF), ibuprofen (IBP), and naproxen (NPX) by CMCAB.


**Table 2.** *Cont*.

#### *3.1. Emerging Micropollutants Removal*

Based on Table 2 and Figure 6a, the maximum removal of diclofenac (DCF) was 96.4% (2.41 mg/L) at pH 6 and an initial concentration of 2.5 mg/L, while the minimum removal of DCF was 38.7% (0.77 mg/L) at pH 8 and an initial concentration of 2 mg/L. Liang et al. [34] reported 70% DCF removal via magnetic amine-functionalized chitosan. Lonappan et al. [35] reported 42% to 98% DCF removal in the presence of a high dosage of biochar microparticles (2–20 g/L). Based on Table 2 and Figure 6b, the optimum elimination of ibuprofen (IBP) was 98.8% (2.47 mg/L) at pH 6 and an initial concentration of 2.5 mg/L, and the minimum removal of IBP was 40.2% (1.20 mg/L) at pH 8 and an initial concentration of 3 mg/L. Chakraborty et al. [36] removed 82% to 91% of IBP via bi-directional activated biochar over high contact time (12–18 h). Paradis-Tanguay et al. [37] removed 70% of IBP using chitosan/polyethylene oxide (PEO) electrospun nanofibers. Based on Table 2 and Figure 6c, the maximum removal of naproxen (NPX) was 95.2% (2.38 mg/L) at pH 6 and an initial concentration of 2.5 mg/L, while the minimum removal of NPX was 38.7% (0.77 mg/L) at pH 8 and an initial the concentration of 2 mg/L. Jung et al. [38] reported 97% NPX removal via a combined coagulation/biochar method. Based on Table 2, the removal effectiveness slightly increased with increasing initial concentration of emerging micropollutants from 0.5 mg/L to 2.5 mg/L.

As shown in Table 2 and Figure 6, the elimination efficiencies of the emerging micropollutants increased with increasing the pH from 4 to 6, and maximum micropollutant removal occurred at pH 6, whereas at pH 5–6, the net surface charge is positive and ion repulsion still exists [39]. Then, the removal effectiveness decreased from pH 6 to 8. Gu et al. [40] reported that the pH of a solution has a significant impact on the adsorption procedure because the surface charge of the adsorbent might be changed in the varied pH. Rafati et al. [41] reported that the maximum removal of emerging micropollutants using an adsorption method was reached at pH 6. The diminishing competition of H<sup>+</sup> ions at increasing pH improved the adsorption to reach the maximum removal at pH 6. Besha et al. [42] expressed that elimination of acidic pharmaceuticals such as ibuprofen, naproxen, and diclofenac might be enriched at slightly acidic pH; this is probably because of the hydrophobicity of these compounds.

**Figure 6.** Removal efficiencies for diclofenac (**a**), ibuprofen (**b**), and naproxen (**c**).

#### *3.2. Optimization using an ANN*

Artificial neural networks (ANN) are computer techniques on the basis of models of the human brain's biological activities, such as the capability to learn, think, solve issues and remember. Neural network models contain weights and neurons. The neural network contains a combined structure comprising an input layer, intermediate layer (hidden layer), and an output layer. Each layer contains of simple processing features called neurons. The mean square error (MSE) and R2 values (Table 3) for DCF, IBP, and NPX elimination are shown in Table 3. Figure 7 indicates the best setting of the ANN. Figure 8 displays the change in the MSE values by Levenberg–Marquardt (LM) through selecting various functions such as pure linear, transig, and log sigmoid. This figure also specifies that the training was completed after 68, 25, and 34 epochs for DCF (a), IBP (b), and NPX (c), respectively. These consequences also proved that the ANN model was well-trained at the end of the training phase [43,44].

The high values of R2 (Figure 9) indicated an excellent agreement between the ANN predicted data and the actual data [43].



**Figure 7.** Artificial neural network (ANN) settings for the best model for diclofenac (**a**), ibuprofen (**b**), and naproxen (**c**).

**Figure 8.** Mean square error (MSE) versus the number of epochs for diclofenac (**a**), ibuprofen (**b**), and naproxen (**c**).

**Figure 9.** *Cont*.

**Figure 9.** Model prediction versus experimental values for the optimum topology for diclofenac (**a**), ibuprofen (**b**), and naproxen (**c**).

#### *3.3. Adsorption Isotherm*

#### 3.3.1. Langmuir Isotherm

The mathematical expression of this isotherm is presented in the following Equation (5):

x <sup>m</sup> <sup>=</sup> abCe (1 + bCe) , (5)

where <sup>x</sup> <sup>m</sup> corresponds the mass of adsorbate adsorbed/unit mass of adsorbent (mg adsorbate per g adsorbent), a and b denote empirical constants, and Ce denotes the equilibrium concentration of the adsorbate in the solution following adsorption (mg/L) [27].

Table 4 and Figure 10 display the details of the Langmuir isotherm studies. The R<sup>2</sup> values were 0.932, 0.962, and 0.893 for DCF, IBP, and NPX removal, respectively. Based on Figure 10, with the decrease in values of (1/Ce), the values of (1/(x/m)) were increased. The high R2 values show that elimination of DCF, IBP, and NPX could be explained by the Langmuir isotherm. For the DCF elimination using the Langmuir isotherm model, the values of *b* and Q (mg/g) were 0.77 and 22.1, respectively. Jodeh et al. [45] reported Q = 22.2 during DCF removal using an adsorption method. For IBP removal using the Langmuir isotherm model, the values of *b* and Q (mg/g) were 0.64 and 21.2, respectively. For NPX elimination using the Langmuir isotherm model, the values of b and Q (mg/g) were 0.76 and 33.3, respectively. Values of Qm = 21.7, b = 0.75, and R2 = 0.8 were reported by Sun et al. [11] and are in line with the results of the current study. Sun et al. [11] reported Qm = 33.6, b = 0.75 and R2 = 0.97 during NPX removal via an adsorption method, which are also in line with the results of the current study.

**Table 4.** Langmuir and Freundlich isotherms study for DCF, IBP, and NPX removal by CMCAB.

**Langmuir Isotherm Freundlich Isotherm**

**Figure 10.** Langmuir isotherm regressions for diclofenac (**A**), ibuprofen (**B**), and naproxen (**C**).

#### 3.3.2. Freundlich Isotherm

The Freundlich isotherm defines the adsorption equilibrium as follows (Equation (6)):

$$\mathfrak{q}\_{\mathrm{m}} = \mathbb{K}\_{\mathrm{f}} \mathbb{C}\_{\mathrm{e}}^{1/\mathrm{n}} \tag{6}$$

where Kf is a fixed variable representing the relative adsorption capability of the adsorbent (mg1−(1/n)/L1/n/g−1), and n is a fixed variable signifying adsorption intensity [27].

Table 4 and Figure 11 display the details of the Freundlich isotherm studies. The R2 values were 0.943, 0.934, and 0.988 for DCF, IBP, and NPX removal, respectively. The high R2 values show that removal of DCF, IBP, and NPX could fit the Freundlich isotherm. Based on Figure 11, with the increase in values of Log(Ce), the values of Log(x/m) were decreased.

The Freundlich capacity factor (K) and 1/n were 30.27 and −17.27, respectively, for DCF removal. Values of Kf in the range 7.6–63.6 and R2 in the range 0.92–0.96 were reported by Sathishkumar et al. [46] for DCF removal via an adsorption method. The Freundlich capacity factor (K) and 1/n were 54.57 and −19.41, respectively, for IBP removal. Coimbra et al. [47] reported a Kf value of 55.30 and R<sup>2</sup> of 0.98 for IBP removal by an adsorption method. The Freundlich capacity factor (K) and 1/n were 16.94 and −12.26, respectively, for NPX removal. Mojiri et al. [48] stated that higher 1/n values indicate that the adsorption bond is weak. Increasing the log (Ce) caused decreasing the log <sup>x</sup> m . Thus, 1/*n* (the slope of the line) is negative [48].

*Water* **2019**, *11*, 551

**Figure 11.** Freundlich isotherm regressions for diclofenac (**A**), ibuprofen (**B**), and naproxen (**C**).

#### *3.4. Regeneration and Desorption Study*

Regeneration of adsorbents is a vital procedure in wastewater treatment to decrease the processing cost. Various regeneration methods have been applied for desorption studies, including thermal regeneration and chemical regeneration. Nevertheless, it is vital to select the appropriate pH and desorbents (such as inorganic desorbents NaOH, H2SO4, and HCl or organic desorbents ethanol, methanol, and acetic acid) for the chemical desorption procedure [49]. Emerging micropollutants are highly soluble in alcohols due to the presence of hydroxyl groups. Moreover, the low molecular weight alcohols may enrich the effectiveness of emerging micropollutants desorption. Alizadeh Fard and Barkdoll [50] stated that NaOH and HCl could not efficiently desorb micropollutants. In addition, they also stated that methanol's restoration capacity is higher than ethanol's. In this study, after eight (Figure 12) cycles with an initial concentration of 2.5 mg/L, the removal effectiveness of the cross-linked magnetic chitosan/activated biochar remained almost unaffected.

**Figure 12.** Regeneration results of CMCAB during removal of DCF (**A**), IBP (**B**) and NPX (**C**).

#### **4. Conclusions**

Diclofenac (DCF), ibuprofen (IBP), and naproxen (NPX) are anti-inflammatory drugs, which are a frequently consumed class of pharmaceuticals. As these are emerging micropollutants, we evaluated their removal using cross-linked magnetic chitosan/activated biochar (CMCAB). An artificial neural network (ANN) with two independent factors—pH (4–8) and micropollutant concentration (0.5–3 mg/L)—was applied to optimize the elimination efficiency. The main conclusions of this new research are listed below:


**Author Contributions:** A.M. was responsible for setting up the experiments, completing most of the experiments, and writing the initial draft of the manuscript; R.A.K. and A.G. modified the manuscript and contributed to the literature search.

**Funding:** This research received no external funding.

**Acknowledgments:** The author would like to express their gratitude to the Institute for Infrastructure Engineering and Sustainable Management (IIESM), University Technology Mara (UiTM), Malaysia.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Inactivation and Loss of Infectivity of Enterovirus 70 by Solar Irradiation**

#### **Muhammad Raihan Jumat and Pei-Ying Hong \***

Biological and Environmental Science and Engineering Division (BESE), Water Desalination and Reuse Center (WDRC), King Abdullah University of Science and Technology (KAUST),

Thuwal 23955-6900, Saudi Arabia; raihan.jumat@kaust.edu.sa

**\*** Correspondence: peiying.hong@kaust.edu.sa; Tel.: +966-12-808-2218

Received: 28 November 2018; Accepted: 24 December 2018; Published: 2 January 2019

**Abstract:** Enterovirus 70 (EV70) is an emerging viral pathogen that remains viable in final treated effluent. Solar irradiation is, therefore, explored as a low-cost natural disinfection strategy to mitigate potential concerns. EV70 was exposed to simulated sunlight for 24 h at a fluence rate of 28.67 J/cm2/h in three different water matrices, namely, phosphate-buffered saline (PBS), treated wastewater effluent, and chlorinated effluent. In the presence of sunlight, EV70 decreased in infectivity by 1.7 log, 1.0 log, and 1.3 log in PBS, effluent, and chlorinated effluent, respectively. Irradiated EV70 was further introduced to host cell lines and was unable to infect the cell lines. In contrast, EV70 in dark microcosms replicated to titers 13.5, 3.3, and 4.2 times the initial inoculum. The reduction in EV70 infectivity was accompanied by a reduction in viral binding capacity to Vero cells. In addition, genome sequencing analysis revealed five nonsynonymous nucleotide substitutions in irradiated viruses after 10 days of infection in Vero cells, resulting in amino acid substitutions: Lys14Glu in the VP4 protein, Ala201Val in VP2, Gly71Ser in VP3, Glu50Gln in VP1, and Ile47Leu in 3Cpro. Overall, solar irradiation resulted in EV70 inactivation and an inhibition of viral activity in all parameters studied.

**Keywords:** enteric virus; remediation technology; water quality

#### **1. Introduction**

Climate change, urbanization, and increasing global population have placed considerable pressure on freshwater supplies [1–3]. Wastewater can be used as an alternative water resource for agriculture irrigation and aquifer recharge but would first require appropriate treatment in wastewater treatment plants (WWTPs). WWTPs act as engineered barriers to treat municipal wastewater to a quality that is sufficiently safe for reuse. In most WWTPs, the final treatment step typically includes the use of chlorine as a disinfectant to reduce the biological activity of remnant pathogens present in the treated effluent [4]. However, each pathogen reacts differently to different disinfectants, and a single disinfection strategy is rarely effective against all pathogens [5]. For example, a WWTP utilizing chlorination as a disinfection strategy was able to inactivate human adenoviruses but not enteroviruses fully from wastewater [6]. Additional disinfection strategies, particularly those that are low-cost and easily accessible, may have to be deployed to further inactivate remnant viral contaminants.

Solar irradiation is a freely accessible, low-cost biocidal strategy that is abundant in many tropical countries and can be used to circumvent this need. The biocidal effect of sunlight works through the effects of ultraviolet A (UV-A) and ultraviolet B (UV-B). UV-A, of wavelengths 320–400 nm, is absorbed by molecular chromophores which, in turn, generate reactive oxidative species (ROS). ROS induce damage to cellular membranes, proteins and nucleic acids, rendering viruses and other pathogens inactive. UV-B, of wavelengths 280–320 nm, functions directly through absorption by nucleic acids and proteins. UV-B can also affect pyrimidines directly, inducing mutagenic and genotoxic effects in the genomes of microbes [7,8].

Several studies have documented the effects of irradiation on viruses. However, the dosage required for viral inactivation varies widely with viral species, particle size, genome type, length, and polarity [9–12]. For instance, numerous studies have investigated the effects of solar irradiation on members of the *Picornaviridae* family, which contain a single positively stranded RNA genome [13]. Heaselgrave et al., reported a 4-log inactivation of polioviruses with solar irradiation ranging from 198 to 1224 J/cm2 [14,15]. In contrast, Coxsackie viruses required 117–198 J/cm<sup>2</sup> of solar irradiation for a 4-log inactivation, while ECHO viruses required 50–60 J/cm2 for a 2-log reduction [15,16].

The variation in solar intensity required to inactivate different RNA viruses within the same family shows that susceptibility of viruses to solar irradiation differs at the species level. A species within the *Picornaviridae* family that has not been studied in this aspect is enterovirus 70 (EV70). These viruses are mainly transmitted by the fecal-oral route and cause gastroenteritis. However, it can cause other symptoms, which include hemorrhagic conjunctivitis, diabetes (through infection of islet cells), and central nervous system complications [17–20]. These viruses are acid and heat stable, allowing for their survival in the gastrointestinal tract but inadvertently conferring persistence in WWTPs [21,22]. Infectious EV70 has been detected in the effluents of several WWTPs globally [6,23–25]. This indicates that the existing disinfection procedures employed are not adequate to provide safe water for reuse, and there exists a need to explore the efficacy of solar irradiation as a possible additional disinfection strategy against enterovirus 70 (EV70).

In this study, EV70 was exposed to simulated sunlight irradiation for 24 h at a fluence rate of 28.67 J/cm2/h. Aliquots of the virus were harvested at specific time points followed by determination of its infectious titer and RNA concentration. We employed a focus forming assay to overcome the inability of EV70 to replicate well in cell culture [6,26]. To determine if any damage was incurred on the capsid, viruses were assayed for their binding ability to Vero cells. The viral growth kinetics were also assayed by counting the foci generated over a nine-day infection period. Ten days after infection, the genomes of EV70 were sequenced. The assays revealed that irradiated viruses had inhibited replication and binding and harbored nonsynonymous nucleotide substitutions compared to dark-control viruses. Viruses suspended in a wastewater matrix also experienced a significant reduction in viral activity upon exposure to solar irradiation, albeit not as pronounced as that observed when suspended in a saline buffer. Interestingly, all of the irradiated viruses in this study failed to replicate in cell culture, providing a strong endorsement of sunlight as a low-cost natural disinfection strategy.

#### **2. Materials and Methods**

#### *2.1. Cells and Viruses*

Enterovirus 70 (EV70) was purchased from American Type Culture Collection (ATCC VR-836, Manassas, VA, USA) and propagated in human embryonic kidney (HEK) 293T cells (ATCC CRL-3216). HEK 293T cells were maintained in 75 cm<sup>2</sup> flasks (Corning Incorporated, Corning, NY, USA) in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum and 1× penicillin and streptomycin (growth medium) (Corning Incorporated, Corning, NY, USA). For the infection study, the HEK 293T cells were seeded in 175 cm2 flasks till confluency and inoculated with EV70 diluted in 10 mL of DMEM supplemented with 2% FBS and 1× penicillin and streptomycin. Flasks were incubated for 1 h at 37 ◦C and 5% CO2. After incubation, another 10 mL of DMEM supplemented with 2% FBS (virus infection media) was added, and the flasks were returned to the incubator. The cells were observed daily for cytopathic effect (CPE). Once CPE was observed, cells were harvested in the supernatant and pelleted by centrifuging at 2000× g for 10 min. The cells then underwent three freeze-thaw cycles to release any intracellular viruses. The lysate was collected in the supernatant harvested earlier. Next, 30% polyethylene glycol (PEG) 8000 in 0.4 M NaOH was added to the total volume of the lysate and supernatant to a final concentration of 15%. This mixture was stirred

at 150 rpm overnight at 4 ◦C. Viruses were pelleted by centrifuging at 10,000× g for 30 min. The pellet was resuspended in 50 mL of sterile 1× phosphate-buffered saline (PBS) or 0.45 μm-filtered effluent or chlorinated effluent wastewater collected from the Wastewater Treatment Plant of King Abdullah University of Science and Technology (KAUST) [6]. Physical parameters of these collected wastewaters are shown in Table S1. Resuspended viruses were immediately used for solar inactivation.

#### *2.2. Simulated Solar Inactivation Trials*

Six milliliters of PEG-purified viruses was dispensed into each of the six microcosms. Each microcosm was made from 5 mL glass beakers (solution depth: 2.7 cm) wrapped in black duct tape to prevent unwanted light penetration. Each microcosm contained a magnetic stirrer and was covered at the top with aluminum foil, in the case of dark controls (n = 3), or left open (n = 3). Each of these microcosms was then placed in 50 mL beakers containing 20 mL of water that served to regulate the temperature at 20 ◦C. The 50 mL beakers of the dark controls were also covered with aluminum foil, while the irradiated microcosms were covered with a glass filter (Newport Corporation, Irvine, CA, USA) that allowed light of wavelengths ≥280 nm to pass through. Each beaker was then placed in an Atlas Suntest® XLS+ photostimulator (Atlas, Chicago, IL, USA) equipped with a xenon arc lamp. The solar irradiation was measured by a spectroradiometer (ILT950, International Light Technologies, Peabody, MA, USA) to determine the UV irradiance as described previously [27]. The irradiance rate at 280–700 nm was ca. 28 J/cm2/h. The UV irradiance provided by the solar simulator approximated the irradiance measurements of direct noon sunlight measured at two locations on the KAUST campus [27]. Viruses (500 μL) were harvested from the irradiated samples and dark controls after 2 h, 4 h, 6 h, 8 h, 10 h, 12 h, 14 h, 16 h, 20 h, and 24 h (n = 3) of solar inactivation. Virus samples were stored at −80 ◦C until use.

#### *2.3. Virus Inactivation Kinetics Evaluated by Means of Infectious Assay*

Vero cells CCL-81 (American Type Culture Collection, Manassas, VA, USA) were seeded onto 96-well plates at 2 × <sup>10</sup><sup>4</sup> cells/well in growth media overnight at 37 ◦C with 5% CO2. Cells were infected with viruses harvested at all time points from the solar inactivation experiment described above. Briefly, viruses were serially diluted (100–10−4), and 50 μL of the diluted inoculum was added to a well with confluent Vero cells for 1 h at 37 ◦C with 5% CO2. The inoculum was aspirated and replaced with DMEM supplemented with 100 μL of virus infection media. Cells were then returned to the incubator for 16 h. After incubation, cells were washed with 1× sterile PBS and subsequently fixed and permeabilized with 100 μL of ice-cold methanol-acetone (50%-50%) for 10 min. Cells were then washed with 1× PBS three times and incubated with 20 μL of anti-EV antibody (Merck, Cat No. 3321) for 1 h at 37 ◦C. Cells were then washed with 1× PBS three times and incubated with 20 μL of anti-Mouse (Ms) IgG-FITC (Merck) for 1 h at 37 ◦C. Cells were washed again with 1× PBS before adding 100 μL of 1× PBS to each of the wells after the last wash. Wells were observed under an epifluorescence microscope for foci (single infected cells). Foci were counted, and the viral titer was estimated using Equation (1):

$$\text{FFU/L} = f \times \frac{10^6 \,\mu\text{L}}{50 \,\mu\text{L}} \times \frac{1}{\text{Dilution Factor}}\tag{1}$$

where FFU/L is focus-forming units per liter and *f* is the number of fluorescently labeled cells. The infectious titer of irradiated samples was compared against the dark control. Each viral dilution was inoculated in two wells of a 96-well plate, and titration was carried out in triplicate (n = 3). The results from this focus-forming assay were converted to log and natural log (ln) curves by calculating log(N*t*/N0) and ln(N*t*/N0), respectively, where N*<sup>t</sup>* = virus titer at time *t* and N0 = virus titer at 0 h (start of experiment). The slopes of the dark-control and irradiated samples (*k*) were calculated from the ln curves. Prior to the solar inactivation experiments, the absorbance value at 280–700 nm of the PEG-purified virus was determined using a UV-3600 UV–VIS spectrometer

(Shimadzu, Kyoto, Japan). The readings were used to generate correction factors that were applied to the slopes of the decay curves prior to half-life calculations and statistical comparison. The half-lives for each experiment, or the durations needed to reduce the viral titer by half, were calculated using the first-order kinetics Equation (2):

$$\ln(\text{N}\_t/\text{N}\_0) = -k^\*t\tag{2}$$

where *k*\* is the corrected slope of the inactivation curve and *t* is time. Statistics were carried out by simple linear regression analysis of the log(N*t*/N0) values of dark-control and irradiated samples.

#### *2.4. Virus Inactivation Kinetics Evaluated by Means of RNA Concentration Decay*

RNA was extracted from viruses harvested at each time point throughout the solar inactivation trials by an RNeasy Mini kit (Qiagen, Hilden, Germany). Viral RNA was eluted in 35 μL of water, and the concentration of extracted RNA was determined using the Qubit® single-stranded RNA assay kit with the Qubit® fluorometer (Thermo Fisher Scientific, Carlsbad, CA, USA). Decay kinetics and statistical comparisons were carried out similarly as described in the previous section. Specifically, the RNA concentration obtained at the start of the experiment was defined as N0 and was expressed in log(N*t*/N0) and ln(N*t*/N0) equations, where N*<sup>t</sup>* is the RNA concentration of the virus sample harvested at time *t* and N0 is the RNA concentration harvested from 0 h (start of experiment).

#### *2.5. Growth Curve Analysis*

Vero cells were seeded onto 96-well plates at a cell density of 2 × 104 cells per well in growth media overnight at 37 ◦C with 5% CO2. Cells were then inoculated with 50 μL of serial dilutions (100–10−4) of T0, D24 or L24 at the multiplicity of infection of 0.1, where T0 is presolar-irradiated EV70 (wild-type), D24 is the dark-control EV70 post solar irradiation, and L24 is EV70 that had undergone 24 h of simulated solar irradiation. Cells were incubated at 37 ◦C with 5% CO2 for 1 h for viral absorption. Each dilution was added to two wells of a 96-well plate, and the titration for each virus sample was done in duplicate (n = 2). Virus inoculum was removed, replaced with viral infection medium, and incubated for infection to take place. Cells were fixed with ice cold methanol-acetone (1:1 *v/v*) for 10 min at 1, 3, 5, 7, and 9 days post-infection (dpi). Cells were subsequently labeled with anti-EV and anti-Ms IgG-FITC antibodies, and the corresponding viral titers were calculated as described above.

#### *2.6. Virus Absorption Assay*

Vero cells were seeded in 6-well plates at a cell density of 1.2 × <sup>10</sup><sup>6</sup> cells per well in growth media overnight at 37 ◦C with 5% CO2. Cells were then inoculated in 100 μL of T0, D24 or L24 diluted in 500 μL of 1× sterile PBS for 1 h at 4 ◦C. After binding, cells were washed with 1× sterile PBS and subsequently scraped and collected in 400 μL of 1× sterile PBS. This cell suspension went through three rounds of freeze-thaw to release bound viruses and were then serially diluted (100–10−4) and titered in a similar manner as described in Section 2.3.

#### *2.7. EV70 Genome Sequencing*

Vero cells were seeded in the same conditions as for the viral absorption assay described in Section 2.6. Cells were then inoculated with similar titers of T0, D24 or L24 in 500 μL of 1× sterile PBS for 1 h at 37 ◦C with 5% CO2. The viral inoculum was replaced with DMEM supplemented with 2% FBS and 1× penicillin and streptomycin. Cells were placed in the incubator for 10 days. At 10 days post-infection (dpi), cells were washed, scraped and collected in 400 μL of 1× sterile PBS. RNA was extracted from the cells by the RNeasy Mini kit (Qiagen, Hilden, Germany). RNA was used as the template for fragment Polymerase Chain Reaction (PCR) where the genome was amplified into 18 overlapping fragments of 750 base pairs. Reaction mixes were prepared by adding 25 μL of 2× RT Buffer, 1.5 μL of forward and reverse primers at 10 μM concentration, 1 μL of Life Technologies

SuperScriptTM II RT enzyme (Thermo Fisher Scientific, Carlsbad, CA, USA), 18 μL of water and 0.83 μL of template RNA. Primers for each of the fragments are listed in Table S2.

Fragments were amplified by touchdown PCR, which included cDNA synthesis at 55 ◦C, for 30 min, an initial denaturation step of 94 ◦C, for 2 min, followed by 15 cycles of 94 ◦C for 15 s, annealing at 52 ◦C for 30 s, and extension at 68 ◦C for 100 s, with the annealing temperature decreasing by 1 ◦C with each cycle. This was followed by another 30 cycles of 94 ◦C for 30 s, annealing at 48 ◦C for 30 s, and extension at 68 ◦C for 100 s. A final extension at 72 ◦C for 5 min was performed. Amplicons were run on a 1.2% agarose gel and visualized by SYBR Green (Thermo Fisher Scientific, Carlsbad, CA, USA). Bands corresponding to ~750 bp were extracted using the Wizard® SV Gel and PCR Clean-Up system (Promega, Fitchburg, WI, USA). Purified PCR products were sent to the KAUST Genomics Core lab for Sanger sequencing. PCR sequences were aligned using the SeqMan program of the DNASTAR's Lasergene software package (DNASTAR, Madison, WI, USA). Aligned contigs were saved as consensus sequences for each of the viral samples. Consensus sequences were aligned in BioEdit Sequence Alignment Editor [28] and translated in silico. The amino acid sequences of each generated genome were submitted to the Phyre2 web Portal for 3D structure prediction [29]. The .pdb files generated from Phyre2 were visualized in PyMOL [30].

#### **3. Results**

#### *3.1. Viral Inactivation Upon Solar Irradiation*

The infectious capacities of irradiated EV70 in phosphate-buffered saline (PBS), as well as in effluent and in chlorinated effluent wastewater matrices were assayed against dark-control EV70 (Figure 1A). After 24 h of irradiation, at a fluence of 688 J/cm2, the infectivity of dark-control EV70 was reduced by 0.18 ± 0.07 log, 0.11 ± 0.003 log and 0.12 ± 0.03 log in PBS, effluent and chlorinated effluent matrices, respectively (*kobs* = 0.014, 0.010 and 0.009) (Figure 1A and Table 1A). One-way ANOVA revealed no significant difference between the decay constants of the dark controls in the three different matrices (*p* > 0.1). Linear regression analysis of the slopes of the dark control showed no positive correlation with respect to time, suggesting that dark-control viruses were relatively stable in each of the matrices over a 24 h period (*p* > 0.01).

**Figure 1.** Decay curves of EV70 under simulated solar irradiation (Y-Y-Y) compared to dark-control EV70 (Y- Y- Y) over 24 h for viruses suspended in phosphate-buffered saline (PBS) (n = 3), effluent matrix (n = 2), and chlorinated effluent matrix (n = 2). The irradiance rate at 280–700 nm was 27.86 J/cm2/h. (**A**) Viruses were harvested at each time point and subsequently tittered. (**B**) RNA was extracted from the viruses harvested at each time point and subsequently quantified.

**Table 1.** Decay kinetic constants and half-life of (**A**) EV70 infectivity, and (**B**) RNA concentration, under simulated solar irradiation for 24 h in phosphate-buffered saline (PBS) (n = 3), effluent wastewater matrix (n = 2), and chlorinated effluent wastewater matrix (n = 2). *kobs* = decay constant. *t*1/2 = half-life of decay.


In contrast, irradiated EV70 reduced in infectivity by 1.7 ± 0.2, 1.0 ± 0.1, and 1.3 ± 0.3-logs in PBS, effluent and chlorinated effluent with decay constants of 1.4, 0.9 and 1.0 (Figure 1A and Table 1A). *t*-test analysis between the decay constants of the dark-control and irradiated samples showed that the decay within each matrix was significant (*p* < 0.05). One-way ANOVA of the decay constants of the irradiated samples revealed that the decay in each of the matrices was significantly different from the others, with the decay in PBS being the fastest (*t*1/2 = 30 ± 3 min), followed by the decay in chlorinated effluent (*t*1/2 = 41 ± 3 min), then effluent (*t*1/2 = 47 ± 5 min) (*p* < 0.05) (Figure 1A and Table 1A).

#### *3.2. RNA Decay*

After the same dose of simulated solar irradiation, the RNA concentration from dark-control EV70 decreased by 0.3 ± 0.1 log, 0.1 ± 0.02 log, and 0.2 ± 0.04 log in PBS, effluent and chlorinated effluent matrices, respectively (*kobs* = 0.002, 0.007, and 0.013). One-way ANOVA revealed that the decay constants of the dark control did not differ between the three matrices (*p* > 0.05) (Figure 1B and Table 1B). Linear regression analysis of the dark control showed a positive correlation with respect to time, suggesting that RNA of the dark-control samples was not stable in any matrix (*p* < 0.05).

The RNA concentrations of the irradiated samples decayed by 1.1 ± 0.1 log, 0.5 ± 0.1 log, and 0.4 ± 0.1 logs in PBS, effluent and chlorinated effluent, respectively (*kobs* = 0.77, 0.48, and 0.29) (Figure 1B and Table 1B). Within each matrix, the decay constant of the irradiated samples differed significantly from the dark control, suggesting that simulated solar irradiation sped up RNA decay in EV70 (*p* < 0.05). The decay constants of the irradiated samples did not differ significantly from each other (*p* > 0.05), but RNA from EV70 in PBS decayed the fastest (*t*1/2 = 57 ± 14 min), followed by EV70 in effluent wastewater matrix (*t*1/2 = 88 ± 9 min) and then EV70 in chlorinated effluent wastewater matrix (*t*1/2 = 145 ± 7 min) (Table 1B).

#### *3.3. Irradiated EV70 Displays Inhibited Viral Replication*

To study the replication kinetics of irradiated viruses, Vero cells were infected with the same titer of T0, D24 or L24 for nine days. Focus-forming units were counted throughout this period to produce the growth curves presented in Figure 2. Sixteen hours post-infection (hpi), T0 in PBS replicated to 3.9 × <sup>10</sup><sup>3</sup> ± 8.1 × 102 FFU/mL, and D24 in PBS replicated to 4.7 × <sup>10</sup><sup>3</sup> ± 1.1 × 103 FFU/mL. Both T0 and D24 peaked on the fifth day post-infection (dpi) at titers of 7.7 × 104 ± 4.1 × 103 FFU/mL and 6.3 × 104 ± 1.0 × 104 FFU/mL, respectively. A titer of 2.9 × 104 ± 1.0 × 104 FFU/mL and 3.2 × <sup>10</sup><sup>4</sup> ± 1.4 × <sup>10</sup><sup>4</sup> FFU/mL was observed for both these viruses at 9 dpi, respectively. This apparent reduction in titer was probably due to the detachment of infected cells from the monolayer nine days after infection. In contrast, L24 in PBS replicated to 6.4 × 103 ± 1.3 × <sup>10</sup><sup>3</sup> FFU/mL 16 hpi and remained relatively similar over the 9-day period. Nor did L24 display a peak at 5 dpi, as seen in T0 and D24 in PBS (Figure 2A).

**Figure 2.** Growth kinetics of nontreated EV70 (T0 Y), dark-control EV70 (D24 *3*), and simulated-solar-irradiated EV70 (L24 Y). Confluent monolayers were infected with similar concentrations of each of the viral samples, and the foci formed over a nine-day period were enumerated. (**A**) EV70 resuspended in phosphate-buffered saline (PBS) (n = 3). (**B**) EV70 resuspended in effluent wastewater matrix (n = 2). (**C**) EV70 resuspended in chlorinated effluent wastewater matrix (n = 2).

T0 in effluent matrix replicated to 1.1 × 104 ± 2.6 × 102 FFU/mL at 16 hpi and peaked to 4.5 × <sup>10</sup><sup>4</sup> ± 4.7 × 103 FFU/mL at 5 dpi, and this titer reduced to 3.1 × 104 ± 4.5 × <sup>10</sup><sup>3</sup> FFU/mL at 9 dpi. D24 in effluent matrix exhibited a similar growth pattern, with a titer of 1.2 × 104 ± 1.4 × 103 FFU/mL at 16 hpi, 5.1 × <sup>10</sup><sup>4</sup> ± 5.2 × <sup>10</sup><sup>3</sup> FFU/mL at 5 dpi and 3.6 × <sup>10</sup><sup>4</sup> ± 9.1 × 103 FFU/mL at 9 dpi. L24 in effluent matrix replicated to 1.1 × 104 ± 1.8 × <sup>10</sup><sup>3</sup> FFU/mL at 16 hpi. On 5 dpi, when T0 and D24 replicated to peak titers, L24 in effluent matrix only displayed a titer of 1.1 × 104 ± 5.1 × 103 FFU/mL. The titer of L24 in effluent matrix did not exceed the titer displayed at 16 hpi throughout the course of the experiment (Figure 2B).

In the chlorinated effluent wastewater matrix, T0 replicated to 1.5 × 104 ± 1.4 × 102 FFU/mL at 16 hpi and peaked at approximately 4.0 × 104 FFU/mL approximately 130 hpi. It reached 3.0 × <sup>10</sup><sup>4</sup> ± 2.5 × 103 FFU/mL at 9 dpi. D24 in chlorinated effluent replicated to 1.3 × 104 ± 1.3 × 103 FFU/mL at 16 hpi and peaked at 4.2 × 104 ± 8.1 × <sup>10</sup><sup>3</sup> FFU/mL at 7 dpi, before finally reaching a titer of 2.0 × 104 at 9 dpi. In contrast, L24 in chlorinated effluent replicated to 1.3 × 104 ± 1.6 × 103 FFU/mL at 16 hpi and decreased to 2.6 × 103 ± <sup>9</sup> × 101 FFU/mL by 9 dpi (Figure 2C). T0 and D24 in chlorinated effluent peaked later than in PBS or effluent, possibly due to the presence of residual disinfection byproducts that may have a toxic effect on mammalian cell lines [31].

#### *3.4. Irradiated EV70 Displayed Reduced Binding Capability*

Untreated EV70 (T0) and dark-control EV70 harvested at 24 h post-irradiation (D24) displayed similar binding affinities to Vero cells in cell culture in all three matrices (Figure 3) (*p* > 0.05). EV70 in PBS displayed the greatest reduction in binding among the three matrices, 2.6 ± 0.3 log, followed by EV70 in effluent matrix at 1.8 ± 0.3 log and EV70 in chlorinated effluent matrix at 1.3 ± 0.2 log. One-way analysis of variance (ANOVA) revealed that the log reduction values (LRV) of the irradiated samples were significantly different from each other (*p* < 0.01), suggesting that the matrix affected the binding affinity of EV70 to Vero cells in cell culture (Figure 3).

**Figure 3.** Log reduction values of the binding affinity of dark-control viruses (D24) and solar-irradiated EV70 (L24) with respect to untreated EV70. *Y*-axis represents log reduction value (LRV) with respect to nontreated EV70. D24: dark-control EV70 viruses that were placed in the solar simulator for 24 h but kept in the dark. L24: EV70 viruses that were exposed to simulated solar irradiation for 24 h. The irradiance rate at 280–700 nm was 28 J/cm2/h. PBS, n = 3; effluent, n = 2; chlorinated effluent, n = 2.

#### *3.5. Irradiated Viruses Select for Five Nonsynonymous Mutations*

D24 and L24 from all three water matrices were infected in Vero cells for 10 dpi. Viral RNA was extracted and amplified into 18 overlapping fragments for Sanger-based sequencing. Figure 4 shows the alignment of the nucleotide sequence and the in silico-translated amino acid sequence. Nucleotide numbers here start at the first ATG of the coding sequence of the reference strain EV70 J670/71 (GenBank D00820.1) [32]. D24 in all matrices and in both replicates in PBS yielded the same nucleotide sequence, showing that any sequence difference seen in L24 was an effect of solar irradiation. L24 in PBS displayed five nonsynonymous nucleotide substitutions (Figure 4). A40G and C809T were observed in the VP4 and VP2 genes, respectively. These mutations resulted in conserved-amino-acid substitutions: Lys14Glu and Ala201Val (Figure 4A,B). Both these changes occurred in unstructured motifs of their respective proteins (Figure S1). G1171A was observed in the VP3 gene, which caused a nonconserved-amino-acid substitution of Gly71Ser (Figure 4C). However, this mutation maintained the β-sheet structure of this protein (Figure S1). G1810C was observed in the VP1 gene, resulting in Glu50Gln substitution, which occurred in an unstructured region of the protein (Figure 4D and Figure S1). A4801C in the 3Cpro gene resulted in the conserved-site substitution of Ile47Leu (Figure 4E). No structural changes were observed due to this mutation (Figure S1). Out of these five nonsynonymous mutations, A40G was also seen in L24 in chlorinated effluent matrix (Figure 4A). L24 in both wastewater matrices harbored a synonymous mutation of G4698A in the 3Cpro gene, which was not observed in L24 in PBS (Figure 4F).


**Figure 4.** Genome sequence analysis of D24 and L24 (in PBS, effluent and chlorinated effluent) 10 days post-infection (dpi) in Vero cells. The nucleotide sequence of D24 in PBS is taken as the reference strain in this alignment (top row). The second (and other even-numbered) rows represent the predicted amino acid sequence. Identical nucleotide sequences are represented by a dot (**.**). Six nonsynonymous mutations were observed in the genes coding for (**A**) VP4, (**B**) VP2, (**C**) VP3, (**D**) VP1, and (**E**,**F**) 3Cpro. The number of experiments performed was n = 2 for EV70 in PBS, n = 1 in effluent (Eff) andn=1 in chlorinated effluent (Chl Eff). Sequence data could only be obtained from one experimental run for solar inactivation in wastewater matrices. Scale (top row) represents the nucleotide sequence of EV70 strain J670/71 (from NCBI reference D00820), with position 1 corresponding to the first nucleotide of the open reading frame.

#### **4. Discussion**

Earlier observations of viable and infective viruses in post-treated effluent provided the main impetus for this study [6], as their presence can complicate the reuse of reclaimed waters. To circumvent viral risks, chlorine disinfection is typically performed at the last step of a wastewater treatment process. However, chlorine works with varying effectiveness against different types of viruses [5]. This led to the suggestion of including combinations of various disinfection processes in a single WWTP. However, retrofitting different modular units of disinfection processes may incur additional operating costs. Solar disinfection of treated wastewater was therefore studied to provide a natural, low-cost and abundant disinfection strategy to further inactivate remnant viruses present in the reclaimed waters.

Specifically, EV70 was chosen as a model organism in this study, as infectious enteroviruses were previously found after wastewater treatment in concentrations approximating the infectious dose [6,33]. EV70 has not been studied extensively for its susceptibility to disinfectants due to its lack of plaque-producing capability in cell culture. To overcome this hurdle, a focus-forming assay was employed, which measured viral titer by fluorescently labeling virus-infected cells with virus-specific antibodies. This technique also required shorter duration compared to a traditional plaque assay [26].

We observed that EV70 in PBS experienced a 1.7-log reduction in infectivity after a dose of 688 J/cm2 (Figure 1). This is consistent with the finding that poliovirus type 2 experienced a 4-log reduction with a simulated solar irradiation of 1224 J/cm2, which is equivalent to a 2-log reduction at approximately 612 J/cm2 [14]. Both EV70 and polioviruses are from the *Picornaviridae* family and have similar sizes (approximately 30 nm in diameter), capsid structures and genome lengths (EV70: 7200 nt, poliovirus: 7500 nt) [34,35]. In contrast, other members of *Picornaviridae* require differing doses of solar irradiation to achieve a similar reduction in infectivity. For example, Coxsackie viruses require approximately 58.5–99 J/cm2, and ECHO viruses require 50–60 J/cm2 of solar irradiation to achieve a 2-log reduction [15,16]. Both Coxsackie and ECHO viruses have similar sizes (28 nm and 24–30 nm, respectively) and genome lengths (approximately 7400 nt and 7500 nt, respectively) to EV70 [36–38]. The data presented in this study agree with earlier studies that infer the need for varying solar fluence to inactivate different viral species. While the structures of viruses are generally similar within a family, species might differ in protein folding and genome secondary structure, which give rise to differences in susceptibility to solar irradiation [39].

*Picornaviridae* have a positively stranded RNA genome that is directly translated by host-cell ribosomes [13]. Here, damage to the genome was indicated by the decay in the RNA concentrations in the presence of solar irradiation (Figure 1B). In addition, the reduction in binding capacity of L24 indicated conformational damage to the capsid, stopping it from recognizing the viral receptor on the Vero cells (Figure 3). This reduction in binding was of a larger magnitude than the reduction in infectivity as seen in Figure 1A at 24 h across all three matrices. Since receptor binding is the first step in a virus replication cycle, any irradiation-induced damage to the capsid could result in the inability of the capsid to recognize the receptor on host cells. Hence, this decrease in binding of L24 was most likely the primary cause for the decrease in infectious capacity.

Although EV70 with a damaged capsid may have a reduced binding capability, replication would theoretically still be possible if the interior structure of the viral particle remained undamaged. To test this, we performed a growth curve analysis, which showed that L24 was unable to replicate to similar titers as T0 or D24 even after 9 dpi in all three water matrices (Figure 2). This was observed despite the similar multiplicity of infection between T0, D24, and L24. This information indicates an inability of solar-irradiated EV70 to replicate as effectively as wild-type or dark-control viruses. The capsids of *picornaviruses* undergo a dramatic antigenic alteration before the virus uncoats [40]. Translation is then initiated by the internal ribosomal entry sites in the 5 untranslated region of the genome, which is composed of five stem-loops (II-VI) [41,42]. Viral translation is also promoted by the binding of host-cell IRES trans-acting factors, such as FBP1-3, hnRNP K and hnRNP A, which recognize the 5 untranslated region of the viral genome [42–45]. Solar irradiation could induce structural damage to the capsid and genome of EV70, leading to reduced binding and replication capacity. Not only would a structurally damaged capsid fail to bind to the host-cell receptor, but it might fail to undergo the antigenic alteration necessary for uncoating to occur [40]. UV irradiation promotes RNA-protein cross-linking [46]. The formation of covalent bonds between the EV70 genome and the capsid might affect the release of the RNA out of the capsid during the uncoating process. The integrity of the cloverleaf and stem loop structure present in the 5' untranslated region of the EV70 genome might be negatively affected by solar irradiation. A disintegration of structure in this region of the genome may prevent successful docking of the ribosome and other host-cell translation initiation factors. Lastly, owing to the structural damage to the genome, translation might not proceed as efficiently as in wild-type viruses, producing proteins which might not support viral replication.

To elucidate if mutations did indeed occur in key proteins of EV70, we sequenced the coding region of the genome. Initially, fragment PCR of viruses directly sampled after 24 h of solar irradiation was performed (data not shown). However, this did not yield sufficient concentrations of PCR amplicons for sequencing. To overcome this technical constraint, Vero cells were infected with L24 or D24 for 10 days, and the viral RNA, which had amplified in the course of the infection, was extracted and sequenced. The sequence of the L24 viral genome derived from this experiment is, hence, not a direct product of solar irradiation but was selected for 10 dpi. This genome could be viewed as an 'escape mutant', being the only sequence that had replicated enough to be amplified by PCR. However, this sequence was still unable to replicate as effectively as T0 or D24 (Figure 2).

The irreproducibility of nucleotide substitutions between trials 1 and 2 for L24 in PBS indicate that solar irradiation induces mutations in a random manner. However, four out of the six mutations listed occurred in the capsid genes, which are at the 5 end of the genome. Positions 40 and 4801 also showed mutations in two of the four irradiated samples (Figure 4A,F). These findings might suggest that the capsid genes, as well as position 4801 in the 3Cpro gene, are more prone to mutation by solar irradiation compared to the rest of the genome.

The structure of the capsid of bovine enterovirus (BEV), a *picornavirus*, has been determined [47]. The structural proteins of BEV share 48% identity with EV70 [48], and its tertiary structure is collinear with other enteroviruses [49]. Comparisons with the amino acid sequence of BEV's capsid reveal that the amino acid substitutions of EV70 listed in this study did not occur in any of the known functional motifs. However, an earlier study showed that an introduction of a single amino acid substitution at five different positions in the capsid genes resulted in a change in viral tropism [50]. These proteins constitute the capsid and form the depression known as the 'canyon', which recognizes the cellular receptor DAF/CD55 for attachment to the host [51,52]. VP1, which is the most exposed protein of the capsid of *Picornavirus* [53], forms a hydrophobic pocket that allows for myristic acid binding [54] and is believed to be involved in the binding of metal ions [55,56]. VP1 is also believed to have a role in the uncoating of the virus particle [50]. Even though Glu50Gln in VP1 occurred in an unstructured motif (Figure S1), the substitution might alter the charge of VP1.

EV70 with a glutamic acid instead of a lysine at position 14 of VP4 protein replicates poorly in HeLa cells [50]. In this current study, L24 in PBS and chlorinated effluent displayed this substitution (Figure 4), which may have accounted for the poor replication. It is likely that this mutation resulted in a change in the charge of the overall protein, as lysine is typically positive at neutral pH while glutamic acid is negatively charged. This would have resulted in poor binding of EV70 to the host cells. Similarly, even though the amino acid substitution Gly71Ser did not affect the folding of the β-sheet in VP3 (Figure S1), the overall polarity of the protein might have been affected owing to the polar nature of serine as opposed to glycine. Both these substitutions might have synergistically affected viral function.

In addition to assessing changes in the capsid proteins, the 3Cpro protein of *Picornaviridae* was also assessed since this protein displays a multitude of functions in the infected cell. Initially shown to be a protease that cleaves the functional proteins from the polyprotein precursor, 3Cpro also cleaves host-cell proteins to shut down host-cell transcription, translation, and nucleo-cytoplasmic trafficking and promote apoptosis (reviewed in [57]). There exist four main functional domains in the 3Cpro protein: the *N*-terminal domain (aa 12–13), the central domain (aa 82–86), the β-ribbon (aa 123–133) and the C-terminal domain (aa 154–156) [58–60]. The amino acid substitution Ile47Leu occurred in between the *N*-terminal domain and the central domain, an area that lacks any known function (Figure 4). It is important to note that this substitution did not affect the integrity of the β-sheet motif of this protein (Figure S1). However, further investigations into this amino acid substitution should be carried out to determine if the function of the 3Cpro protein is altered. If this substitution results in a change in the function of the protein, this could explain the reduced ability of L24 to replicate to high titers, as seen in cells infected with T0 and D24 (Figure 2). This could be a result of inadequate cleavage of the viral polyprotein or inadequate suppression of host-cell factors, allowing for the host cell to overcome the viral replication machinery.

In addition to observing a significant impact on the viral infectivity and persistence due to solar irradiation, we also observed that viral inactivation occurred at a slower rate when the viruses were present in wastewater matrix. This concurs with earlier observations [61–63]. Furthermore, out of the six nucleotide mutations found, only 2 were seen in L24 in wastewater matrices, A40G and G4698A, while L24 in PBS had 5 mutations (Figure 4). This indicates that viruses in the wastewater are less susceptible to UV-B [12,16]. Effluent and chlorinated effluent wastewaters used in this study had a total organic carbon (TOC) concentration of 4.2 mg/L and 5.2 mg/L, respectively, while PBS had undetected levels of TOC, as expected (Table S1). These organic compounds can act as radical scavengers [64,65], reduce light intensity [66], or encapsulate viruses with a protective organic coating that makes them more resistant to external environmental stressors when present in wastewaters. The latter has been alluded to by the findings that non-enveloped viruses are stable in wastewaters [67,68]. Alternatively, the high alkalinity in wastewater might favor the reaction between bicarbonates and hydroxyl radicals formed upon solar irradiation. This reaction results in the generation of CO3 - which reacts slower with organic molecules compared to O2 radicals [27,69,70].

While these reasons could explain the slower inactivation rates of EV70 in wastewater matrices, it is important to note that irradiated viruses, irrespective of matrix, all failed to propagate in cell culture (Figure 4). This indicates that solar irradiation successfully inhibits viral replication in cell culture, preventing the generation of infectious viral progeny in all three water matrices evaluated in this study. This strongly suggests that solar irradiation modifies the replication capacity of EV70 to the point that it might not pose a significant public health threat. Although the data from this study suggest that solar irradiation may serve as a good disinfection technique, its efficacy may be lower in turbid waters due to lower solar penetration and higher light-scattering effect. Operators would also need to create a holding tank that is shallow enough to allow for proper solar penetration and irradiation. This would not be feasible in densely populated places with limited land space. Hence, the use of solar irradiation as an effective, natural, and low-cost disinfection strategy against EV70 would only be feasible for use in low-turbidity waters, presumably in permeates after membrane filtration processes, and in places unconstrained by land availability.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4441/11/1/64/s1, Figure S1: Predicted structure of EV70, Table S1: Physical parameters of the matrices used in this study, Table S2: Sequences and names of primers used in this study.

**Author Contributions:** M.R.J. and P.-Y.H. conceived and planned the experiments. M.R.J. carried out the experiments and performed the analysis. M.R.J. and P.-Y.H. wrote the manuscript together. P.-Y.H. provided reagents and materials. All authors have read and approved the manuscript.

**Funding:** This research was supported by the KAUST baseline funding BAS/1/1033/-01-01 awarded to P.-Y.H.

**Acknowledgments:** The authors would like to thank George Princeton Dunsford for access to the KAUST wastewater treatment plant, Moustapha Harb for providing sampling assistance, Nada Al Jassim for training required to operate the solar simulator and Noor Zaouri for assistance in the LC-OCD analysis.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Potential Use of** *Dimocarpus longan* **Seeds as a Flocculant in Landfill Leachate Treatment**

#### **Hamidi Abdul Aziz 1,2,\*, Nor Aini Rahim 1, Siti Fatihah Ramli 1, Motasem Y. D. Alazaiza 1, Fatehah Mohd Omar <sup>1</sup> and Yung-Tse Hung <sup>3</sup>**


Received: 19 October 2018; Accepted: 14 November 2018; Published: 16 November 2018

**Abstract:** Landfill leachate is a highly polluted and generated from water infiltration through solid waste produced domestically and industrially. In this study, a coagulation–flocculation process using a combination of Polyaluminium chloride (PACl) as a coagulant and *Dimocarpus longan* seed powder (LSP) as coagulant aid was used in treating landfill leachate. LSP has been tested as the main coagulant and as coagulant aid with PACl. As the main coagulant, the optimum dosage and pH for PACl were 5 g/L and 6, respectively, with removal efficiencies of 67.44%, 99.47%, and 98% for COD, SS, and color, respectively. For LSP as the main coagulant, results show that LSP is not effective where the removal efficiencies obtained for COD, SS, and color were 39.40%, 22.20%, and 28.30%, respectively, with the optimum dosage of 2 g/L and pH 4. The maximum removal efficiencies of COD, SS, and color were 69.19%, 99.50%, and 98.80%, respectively, when LSP was used as coagulant aid with PACl. Results show that using LSP as coagulant aid was found to be more effective in the removal of COD, SS, and color with less PACl dosage. The PACl dosage was decreased from 5 to 2.75 g/L when LSP was used as a coagulant aid. Cost estimation for using PACl alone and using LSP as the coagulant aid showed a reduction in the cost of approximately 40% of the cost of using PACl alone. Overall, this study confirmed the efficiency of LSP to be used as a natural coagulant aid in leachate treatment.

**Keywords:** *Dimocarpus longan* seeds; leachate treatment; coagulant–flocculation; polyaluminium chloride

#### **1. Introduction**

Landfill is the most widely accepted and prevalent methods for municipal solid waste (MSW) disposal in developing in many countries around the world due to its inherent forte in terms cost saving and simpler operational mechanism [1]. Environmental pollution caused by the landfill leachate has been one of the typical dilemmas of landfilling method [2]. Leachate is the liquid produced when water percolates through solid waste and contains dissolved or suspended materials from various disposed materials and decomposition process. It is often high-strength wastewater with extreme pH, chemical oxygen demand (COD), biochemical oxygen demand (BOD), inorganic salts and toxicity [3,4]. Its composition differs over the time and space within a particular landfill, influenced by a broad spectrum of factors such as waste composition, landfilling practice (solid waste contouring and compacting), local climatic conditions [5], landfill's physicochemical conditions, biogeochemistry and landfill age [6]. The composition and characteristics of the landfill leachate are the main factors determine the choices of the treatment method [7].

To protect groundwater aquifer and adjacent surface water from leachate contamination, the handling and treatment of leachate must be meticulously designed to minimize its potential adverse impacts [8]. Biological treatment is an environmentally friendly method that can be applied for the treatment of young or freshly produced leachate [9]. However, it is ineffective for leachate from older landfills that usually contain high COD and ammonium content, low biodegradability (high COD/BOD ratio) and multiple heavy metal ions [10]. In contrast, physical and chemical methods are more effective for older leachate as compared to young leachate [11,12]. Sometimes, the quantity of short-term leachate could be difficult to predict since it mainly depends on precipitation; however, the quantity of long-term can be predicted more accurately [13].

In general, generated leachate at early stages of waste decomposition is highly rich in BOD5 and contains a high amount of biodegradable and nonbiodegradable materials such as volatile fatty acids [12]. However, the stabilized leachate from old landfills is often highly polluted with non-biodegradable organic substances, such as fulvic substances and humic-like, which are measured as COD [14]. Furthermore, the stabilized leachate contains a large quantity of inorganic substances, especially ammonium-nitrogen (NH3-N) [15], which results from the hydrolysis and fermentation of nitrogen-containing fractions of biodegradable refuse substrates. When the bioreactor landfills contain contaminant leachate, collection and in situ recirculation for acceleration of decomposition of readily available organic fractions of wastes, leachate NH3-N concentrations may accumulate to produce higher levels as compared to the traditional landfills [16].

Coagulation–flocculation is one of the major chemical methods being used for leachate pretreatment [17]. These methods are applied to remove suspended solids and recalcitrant substances such as humic acid, fulvic acid, or undesirable compounds like heavy metals, absorbable organic halides (AOX) and polychlorinated biphenyls (PCBs) from the leachate [18]. This simple method outshines other advanced technologies like membrane and chemical oxidation technologies in terms of leachate pre-treatment application. The treatment mechanism of this method mainly consists of charge neutralization between negatively charged colloids and cationic hydrolysis products followed by an amalgamation of impurities through flocculation [19]. Total suspended solids (TSS), as well as colloidal particles, are the main parameters removed from this process [20]. The major component of colloid particles is organic compounds. Coagulation is generally defined as destabilization of a colloidal suspension or solution by neutralizing the forces that keep them apart. Cationic coagulants give a positive electric charge that reduces the negative charge of the colloids in solution. As a result, particles of colloid to compose large particles (flocs). Usually, fast mixing is required to disperse the coagulant throughout the solution [21].

The chemistry of coagulation–flocculation is mainly based on the electrical characteristics. The majority of the particles present in the leachate are negatively charged (−30 to −40 mV) [22]. Therefore, they tend to replace each other. Most of the coagulant chemicals are usually used to neutralize the negative charge on colloidal particles to prevent the repelling of these particles with each other [4]. The quantity of coagulant that will be added to the leachate is related to the zeta potential which is known as the electrical potential reflecting the voltage difference between the diffuse layer boundary and the dispersant [23]. Therefore, in the case of large zeta potential, more coagulant is needed. The coagulants have positive charges which attracted to the negative particles in solution; hence the combination of negative and positive charges results in a neutral charge and turn the particles no longer repel each other [4].

The commonly used commercial coagulants are aluminum sulfate (alum), polyaluminum chloride (PACl), ferrous sulfate, ferric chlorosulphate, and ferric chloride [24]. Inorganic coagulants are generally effective however there are some drawbacks related to the high amount of metal ions in sludge [25] while on the other hand, natural coagulants were found to produce relatively low sludge volume and are safe to humans when compared to the inorganic coagulant [26]. Natural coagulants have

been widely applied in wastewater treatment [27] but it is still not used widely in landfill leachate treatment despite that they are in abundant quantities, relatively less expensive, and environmentally friendly [28].

In Malaysia, there has been a recent upsurge in the food industry where a large number of solid wastes was generated annually and especially *Dimocarpus longan* seeds, which are disposed from food manufacturing factories. *Dimocarpus longan* belongs to Sapindaceae family and goes by many scientific names: Nephelium *Dimocarpus longan* Camp, *Dimocarpus longan* Lam, and Euphoria *Dimocarpus longan* Strand. Currently, *Dimocarpus longan* is consumed as fresh and processed fruits while the seeds, which account for about 17% of the fresh weight of whole fruits, are discarded as waste or burned as fuel [29]. The seeds have been found as a rich source of antioxidant phenolic compounds that promising as functional food ingredients or natural preservatives. Soong and Barlow [30] reported that *Dimocarpus longan* seeds contained high levels of corilagin, gallic acid, and ellagic acid, which have been proven to acquire strong free radical-scavenging activity [31]. The seeds have been shown earlier to contain the hydrolysable tannins (ellagitannins) corilagin and acetonyl-geraniin [32]. Corilagin has been extensively studied for its pharmacological activities in the extract of plants such as Acer nikoense and Phyllanthus amarus and also as a pure isolated compound.

No actual data are available on the production and area of *Dimocarpus longan* in Malaysia. It is mainly cultivated in Penang and Kedah. Obtaining precise data on the production and acreage of this species is relatively difficult due to its small production. Based on the author's knowledge, there are no published studies in the literature regarding the usage of *Dimocarpus longan* seed powder (LSP) in wastewater and landfill leachate treatment. The main goal of this study is to investigate the applicability of composite coagulant made from LSP as a natural coagulant in removing color, COD and Suspended Solids (SS) from stabilized leachate. The main objectives of the study are; (i) to determine the optimum pH and dosage of polyaluminum chloride (PACl) and LSP as the main coagulant in removing COD, SS and color; and (ii) to determine the efficiency of LSP as coagulant aid and PACl as the main coagulant in removing COD, SS, and color.

#### **2. Materials and Methods**

#### *2.1. Leachate Sampling and Characterization*

Landfill leachate samples were collected from Alor Pongsu Landfill Site (APLS) in Bagan Serai, Perak, Malaysia from January through April 2018. APLS is classified as an anaerobic stabilized landfill. APLS started its operation in the year of 2000. Since its operation started, the landfill received approximately an average of 660,000 metric tons of solid waste per year, which is roughly 200 metric tons per day [33]. The site covers an area of 10 acres of palm oil plantation. Sampling was carried out using the grab sampling method while preservation was done according to Standard Methods for the Examination of Water and Wastewater [34]. The initial characteristics of the six leachate samples obtained were as contained in Table 1. All samples were kept in HDPE (high-density polyethylene) containers with sealed caps. Samples were transported to the laboratory within 1 h and stored in a cold room at 4 ◦C to minimize biological and chemical reactions prior to any treatability study. Before experiments, leachate samples were conditioned by putting them at room temperature for 2–3 h and homogenized by manual agitation. During the study, the leachate samples were characterized before and after each treatment. The samples were characterized in terms of turbidity, pH, suspended solids (SS), color, COD, manganese (Mn2+), copper (Cu2+), iron (Fe3+), zinc (Zn2+), phosphate (PO4 <sup>3</sup>−) and ammonia-nitrogen (NH3-N). All the analytical procedures were performed according to the Standard Method of Water and Wastewater [34]. pH was measured using a portable pH meter (CyberScan pH 510, Eutech, Singapore). Turbidity was measured using a turbidimeter (HACH 2100 N, HACH, Singapore). COD was measured using colorimetric method (5220-D). Heavy metals, NH3-N, and color were measured using a spectrophotometer (DR/2800, HACH, Singapore).


**Table 1.** Characteristics of raw leachate.

#### *2.2. Preparation of Dimocarpus Longan Seed Powder (LSP)*

The extraction method was adapted from Katayon et al. [35] with some modification by heating distilled water for 30 min with 100 ◦C using a hot plate and stirrer. Firstly, 500 g of fresh *Dimocarpus longan* seeds were peeled and separated from its aril. Then, they were washed using tap water to remove dirt before being air-dried for 48 h. After that, the layer, which is called the seeds coat covering the seeds, was removed. The seeds were again air-dried for another 48 h to ensure that it was completely dry before turning into powder form. The dried seeds were ground using a ring mill for 15 s until it became a fine powder. Finally, the seed powder was kept in a dry place to be used for experiments to be carried out later.

#### *2.3. Coagulation–Flocculation*

The current study investigated the coagulation–flocculation process using a combination of PACl as coagulant and LSP as a coagulant aid. A hydrolyed solution of PACl with the formula of [Al (OH)x Cly] (where x is in the range 1.35–1.65, and y = 3 − x) and pH 2.3–2.9 due to the presence of hydrochloric acid was supplied by Hasrat Bestari Sdn Bhd, Penang, Malaysia. An 18% solution of PACl was used as a stock solution throughout the experiments. Coagulation–flocculation experiments were carried out using jar test apparatus (SW6 Stuart Bibby Scientific Limited, Staffordshire, UK). Leachate samples were allowed to reach room temperature (approximately 3 h) before testing, and they were also thoroughly agitated to resuspend any settled solids. The leachate sample volume per beaker was 500 mL. The time and speed for rapid and slow mixing were set with an automatic controller. The jar test consisted of three subsequent stages: (1) rapid mixing stage with speed of 120 rpm for 3 min, (2) slow mixing stage with speed of 20 rpm for 15 min, and (3) final settling time for 45 min. During rapid mixing, the coagulant was added into the beakers while the impellers were maintained at fast speed. After a certain rapid mixing period, the stirrers were set to a slower speed for another period of time. After that, the stirrers were stopped, and the samples were left for final settling. Then, the samples were withdrawn using plastic syringe from 10 cm below the surface for the analytical determinations. Analyses were undertaken in triplicates. A 500 mL of leachate samples were filled into six beakers and agitated simultaneously while varying the rotational speed and allowing simulation of different mixing intensities and resulting flocculation process [36].

A preliminary coagulant performance study using jar test was conducted to determine optimum pH and dosage for PACl and LSP. It was noteworthy that the preliminary optimum pH studies were performed first with controlled coagulant dosages and the results were carried over to the preliminary optimum dosage studies as controlled pH since the pH would have a major impact on coagulant dosage. Different dosages and pH were investigated in this study for PACl and LSP as coagulant aid

for removing COD, color, and TSS. The examination of pH effect was performed by adjusting the pH value of leachate samples between 5 and 9 using solutions of 0.1 N sulphuric acid (H2SO4) and 0.1 N sodium hydroxide (NaOH). The removal efficiency was investigated by using LSP as coagulant aid and PACl. Zeta potential test was conducted to enhance the results of the jar test and justify the removal mechanisms of the coagulation process. Zeta potential can present a measure of the net surface charge on the particle and potential distribution at the interface. Consequently, zeta potential serves as an important parameter in the description of the electrostatic interaction between particles in dispersed systems and the properties of the dispersion as affected by this electrical phenomenon [37]. In this study, the surface charge was evaluated by using Malvern Zetasizer Nano ZS. Measurements were taken at 25 ◦C with distilled water as the dispersal medium.

#### **3. Results and Discussion**

#### *3.1. Characteristics of Leachate*

The physicochemical parameters of landfill leachate are listed in Table 1. The leachate is categorized as stabilized leachate since its BOD5/COD ratio < 0.1. The BOD5/COD ratio indicates the degree of biodegradation and landfill age. For example, young leachate has BOD5/COD ratio up to 0.83 during the acidogenic phase and decrease to 0.05 for old landfills during methanogenic phase [38]. The low BOD5 and BOD5/COD values for stabilized leachate agreed with the literature [11,15]. The high concentration of SS (745 mg/L) indicated the presence of organic and inorganic solids. A considerable concentration of ammonia nitrogen was found which is attributed to the decomposition of nitrogenous substances in refuse and the release of soluble nitrogen from solid wastes [15]. The dissolved organics mainly contributed a greater concentration of color (5517 Pt-Co). These organic compounds may be present in the form of recalcitrant material mainly composed of humic-like substances. A low value of BOD5 means low biodegradability while the presence of high concentration of NH3-N indicates high leachate toxicity [39].

#### *3.2. Characteristics of LSP*

Figure 1 illustrates the particle size distribution of LSP using Mastersizer analysis (Malvern Panalytical Ltd., Westborough, MA, USA). Results show that d10, d50, and d90 were recorded at 5.317 μm, 13.087 μm, and 32.460 μm, respectively. Fourier transformed infrared spectroscopy (FTIR) was used to investigate the structure of LSP and the analysis of their functional groups as shown in Figure 2. The FTIR spectrums of LSP show a weak intensity at 3435 cm<sup>−</sup>1, due to the O-H stretching and also overlap with a primary amine and aliphatic primary amine due to N-H stretching. At 2989 cm−<sup>1</sup> the functional group is under carboxylic acid which is bonded by strong O-H stretching. This band also overlaps with medium C-H stretching under a functional group of an alkane. Under wavenumber 2591 cm−1, Aldehyde with a medium bond of C-H, ariel together with Thiol is weak in the intensity of S-H stretching. A weak aromatic compound with C-H bonding was found in 1867 cm−1. Alkene shows at 1639 cm−1, with strong and medium bond due to C=C stretching. At 1526 cm−<sup>1</sup> the functional group is under nitro compound which is bonded with strong N-O stretching while at 1276 cm−<sup>1</sup> there is a strong intensity due to the stretching of C-F bond under fluor compound function. At 1136 cm−<sup>1</sup> wavenumber, there is a strong intensity in stretching of sulfone with a strong bond of S=O. At wavenumber of 1020 cm−1, strong intensity of C=O under a functional group of alkyl aryl-ether, medium intensity of C-N stretching with a functional group of amine and also a strong intensity of C-O with vibration group on stretching and vinyl ether functional group. Surface morphology for LSP was investigated using a scanning electron microscope (SEM) with different magnifications as shown in Figure 3. They grouped the oval granular of LSP together and formed into a clod of an elliptical. It also had a cloudy or velvety like coating surface. SEM shows that the surface texture of the LSP was rough and there was an accumulation of fine particles with irregular geometric shapes spotted on the surface.

**Figure 1.** Particle size distribution of LSP (Longan seeds powder).

**Figure 2.** Fourier-transform infrared (FTIR) analysis of LSP.

(**a**)

(**b**)

#### *3.3. Determination of Zeta Potential and Particle Size od LSP as a Function of pH*

Figure 4 illustrates the zeta potential of LSP in conjunction with pH. It had a negative charge over the same pH range and reached the point of zero charges (PZC) at pH 7.

LSP has a negative level at surface starting from pH 2 to pH 6. However, at pH 7, the surface charges for LSP had become absolutely neutral and gradually turn negative when it was at pH 12. Under this condition, it could be said that the LSP was anionic coagulants, and the main mechanism governing the aggregation of the constituent was bridging [40]. Figure 5 illustrates the zeta potential and z-average particle size variation in conjunction with pH. It had a negative charge over the same pH range and reached PZC at pH 7.

**Figure 5.** Zeta potential and z-average particle size variation with different pH values (PS is the particle size, ZP is the zeta potential).

#### *3.4. PACl as a Main Coagulant (Optimum pH and Dosage)*

The preliminary coagulant performance study for PACl showed that the PACl is very effective as a main coagulant at pH 6 (data not shown). Therefore, the process of the determination of the optimum dosage for PACl was conducted by adjusting the pH value to pH 6. Figure 6 shows the major range of PACl dose to remove pollutants at a constant value of pH 6.

**Figure 6.** The effect of PACl in major range of dosage on the removal of COD, color and SS at pH 6.

The observations in Figure 6 show that the highest pollutant removal by PACl was in the range of 4.25–7.5 g/L PACl dosage. By taking this range as a reference, further coagulation tests on removal was conducted within this major range to get the optimum coagulant dose for PACl. The dosages varied from 4.25 to 7.5 g/L at a constant pH 6. Figure 7 shows the effect of PACl major range of dosage on the removal of COD, color, and SS at a constant pH 6. Based on the results, the optimum dosage of PACl was 5.0 g/L which achieved highest removal efficiencies of 67.44%, 98.73%, and 99.47% for COD, color, and SS, respectively.

The usage of PACl provided a better removal of color and SS. At the beginning of the experiments, leachate samples have an initial black color due to the presence of a humic substance [41]. The most effective removal of color occurs at the dosage of 5.0 g/L with 98.73%. The removal efficiency decreased gradually even though PACl dosage still added. This observation is most likely because when a large amount of coagulant dosage has been added over the optimum dosage, the surface of the particle's charge reversed due to of continuous absorption of mono and polynuclear hydrolysis species of PACl. The colloidal particles cannot be removed by perikinetic flocculation as they became positively charged particles [42]. The charge neutralization theory can explain this behavior. When a coagulant is added to the landfill leachate at optimum pH, colloid destabilization occurred when positively charged metal ions encounter with negatively charged colloids neutralizes the charge. The removal of particles will only take place more effectively when the more metal-based coagulant is added, as explained by the Schulze–Hardy rule [43]. As a result, when an extra dosage is added, colloids start to absorb the excessive positive charges which remain in the solution and become positively charged. Therefore, the electrical repulsions between positively charged colloids and metal ions occur. The colloids become stable again as the result, weaken the ability of coagulant to remove contaminants. According to Baghvand et al. [44], overdosing of coagulant will disturb the development process. Thus, the right amount of dose should be added to any wastewater treatment.

COD removal for PACl recorded a high removal of 67.44% at 5.0 g/L of PACl. This is because at a higher concentration of coagulant dosage, the flocs produced have a good consistency and in a better structure than at a lower dosage [45]. Below the optimum dosage, the removal of COD by PACl is not effective due to the fact that at lower concentrations of dose, a smaller floc is produced, and it influenced the velocity of the sludge [46]. SS also has the same pattern of removal, where the highest removal of SS was 99.47% at 5.0 g/L of PACl.

**Figure 7.** The effect of PACl major range of dosage on the removal of COD, color and SS at a constant pH 6.

#### *3.5. Removal Efficiency of SS, Color, and COD Using LSP*

A preliminary coagulant test has been conducted using LSP to determine the optimum pH. From the observation, LSP showed its effectiveness at pH 4 (data not shown). Therefore, the subsequent tests were carried out at pH 4. Different dosages of LSP were used to investigate its effect on the removal of color, SS, and COD at constant pH 4 as shown in Figure 8. Results show that the maximum removal efficiencies of color, SS, and COD were 28.3%, 11.2% and 15.1%, respectively at 2 g/L of LSP dosage.

**Figure 8.** The effect of LSP in major dosages on the removal of color, SS, and COD at a constant pH 4.

Therefore, the optimum dosage of LSP was 2 g/L. It was found that at a higher concentration of LSP, the removal efficiencies of color, SS and COD was decreasing gradually until 5 g/L dosage. Thereafter, the removal efficiencies fluctuated for COD and color and remained at zero for SS.

#### *3.6. PACl as the Main Coagulant with LSP as a Coagulant Aid*

Optimum dosage of PACl (5 g/L) was added to different dosages of LSP by fixing the same conditions of jar test for the slow and rapid mixing followed by settling for 2 h at the optimum pH 6 for PACl. LSP dosage as coagulant aid, (0 g/L as a control sample) with 0.5, 1.0, 2.0, 4.0, 6.0, 8.0 and 10.0 g/L were used for pollutant removal and for reducing the dosage of primary coagulant.

Figure 9 shows the effect of LSP as a coagulant aid to remove COD, color, and SS at a different concentration of PACl. In the PACl predetermined test, 5.0 g/L was used as optimum dosage and resulted in removal efficiencies of 67.44%, 98.73%, and 99.47% for COD, color, and SS. However, when LSP was used as a coagulant aid, the results were better than the use of only PACl in terms of reducing the dosage of PACl and increase the removal efficiencies. The performance of COD removal has been improved from 67.44% to 69.19% at 2.75 g/L of PACl with the combination of 2 g/L of LSP as a coagulant aid when compared with 5 g/L PACl alone. In the coagulation test for the optimum dosage, it was found that PACl alone was able to remove 98% color at 5.0 g/L concentration. From the graph, it shows that as LSP dose increased from 1.0 to 10.0 g/L; the removal rates of color also almost similar for all LSP dosages. The maximum removal efficiency of color (98.80%) was obtained at LSP dosage of 2 g/L and 2.75 g/L PACl. Furthermore, the highest removal of SS was detected at the same combination of 2 g/L LSP and 2.75 g/L PACl with 99.50% removal efficiency.

**Figure 9.** The effect of LSP as coagulant aid on the removal of color, SS, and COD at a constant pH 6 and using 2.75 g/L of PACl dosage.

Figure 10 shows a comparison between the efficiency of using PACl in landfill leachate treatment alone with two different dosages (5 g/L and 2.75 g/L) and using a combination of PACl as a coagulant in conjunction with LSP as a coagulant aid. From the graph, it is clear that the removal efficiency of COD, color, and SS was better when the PACl was used in conjunction with LSP as a coagulant aid. The dose of metal coagulant can be reduced without affecting the removal performance when polyelectrolyte is used as a coagulant aid because polyelectrolyte has higher charge density and molecular weight which act as an important role in coagulation. This is due to the addition of coagulant aid could help to form bigger flocs and produced more particles sediment, thus increasing the sedimentation rate [47]. Formation of flocs became quicker when LSP is used as a coagulant aid.

**Figure 10.** Comparison of using PACl alone and PACl in conjunction with LSP as a coagulant aid.

#### *3.7. Coagulant Cost Estimation*

Generally, the leachate treatment cost depends on factors, such as landfill design, the quantity of leachate, level or degree of treatment needed, and final removal method for residues and effluent. Obtaining data on the cost of leachate treatment is difficult because it requires the cooperation of the company in charge. Therefore, on the basis of the chemicals used, the costs of both coagulants are estimated. Table 2 showed the cost comparisons when the LSP was used as the coagulant aid. From the comparisons, a cost of RM 7800 was found when 5 g/L of PACl was used alone, whereas the use of 2 g/L of LSP as coagulant aid only cost RM 4646, a reduction of approximately 40% of the cost.



### **4. Conclusions**

The application of the coagulation–flocculation process to landfill leachate was examined in this study using longan seed powder (LSP) as a natural coagulant aid. LSP was not effective when used as a main coagulant. Compared to when PACl was used alone, a slight improvement of COD, color, and SS removal efficiencies were obtained when PACl was used as a main coagulant and LSP as a coagulant aid. The maximum removal efficiencies of COD, color, and SS were 69.19%, 98.80%, and 99.50%, respectively. In addition, using of LSP as coagulant aid was able to reduce the PACl dosage from 5 g/L when used alone to 2.75 g/L when used with conjunction of LSP. A cost estimation for using LSP as a coagulant aid showed a reduction in the cost of using PACl alone of approximately 40% from the cost of PACl. Overall, this study confirmed that LSP is an effective material to be used as a natural coagulant aid for landfill leachate treatment.

**Author Contributions:** H.A.A. conceived and designed the experiments, N.A.R. performed the experiments, M.Y.D.A. wrote and revised the paper, S.F.R. and F.M.O. conducted the data analysis, and Y.-T.H revised and proofread the paper.

**Funding:** This work was funded by Universiti Sains Malaysia under iconic grant scheme [Grant No. 1001/CKT/870023] for research associated with the Solid Waste Management Cluster, Engineering Campus, Universiti Sains Malaysia.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **The Application of Modified Natural Polymers in Toxicant Dye Compounds Wastewater: A Review**

#### **Siti Aisyah Ishak 1, Mohamad Fared Murshed 1,\*, Hazizan Md Akil 2, Norli Ismail 3, Siti Zalifah Md Rasib <sup>2</sup> and Adel Ali Saeed Al-Gheethi <sup>4</sup>**


Received: 12 June 2020; Accepted: 11 July 2020; Published: 17 July 2020

**Abstract:** The utilization of various types of natural and modified polymers for removing toxicant dyes in wastewater generated by the dye industry is reviewed in this article. Dye wastewater contains large amounts of metals, surfactants, and organic matter, which have adverse effects on human health, potentially causing skin diseases and respiratory problems. The removal of dyes from wastewaters through chemical and physical processes has been addressed by many researchers. Currently, the use of natural and modified polymers for the removal of dyes from wastewater is becoming more common. Although modified polymers are preferred for the removal of dyes, due to their biodegradability and non-toxic nature, large amounts of polymers are required, resulting in higher costs. Surface-modified polymers are more effective for the removal of dyes from the wastewater. A survey of 80 recently published papers demonstrates that modified polymers have outstanding dye removal capabilities, and thus have a high applicability in industrial wastewater treatment.

**Keywords:** natural and modified polymer; biodegradability; toxicant dyes; industrial wastewater treatment

#### **1. Introduction**

The wastewater generated from different manufacturing processes poses serious problems for organisms and aquacultures, due to the high toxicity of these wastes, which contain different types of pollutants, such as plastic, leather, ink, fabric, palm oil, soap, pulp, and paper. These wastes are disposed of directly (with partial treatment) into the environment and natural water systems. Short-term exposure to these pollutants causes tremors and nervous system disorders, while long-term exposure causes thyroid dysfunction, weight loss, and generalized hypoxia [1]. Therefore, the treatment of the dye-containing wastewater before the final disposal is an urgent matter, not only to meet international standards, but also to protect the biodiversity in nature and to ensure the availability of pure water for future generations. The dye-containing wastewaters have been of great interest to researchers during the past several years, primarily due to the high tectorial values of the dyes, where the discharge of less than 1 ppm of a dye into the water might cause significant changes in the water's physical and chemical characteristics. The traditional treatment methods used for the treatment of wastewaters depend mainly on chemical, physical, and biological processes (Table 1), which contribute effectively to improving the quality of the effluent parameters, such as the chemical oxygen demand (COD), biochemical oxygen demand (BOD), total suspension solids (TSS), and turbidity. Unfortunately, these methods are insufficient to remove the dyes from the wastewater. Coagulation/flocculation is a potential alternative, and a highly efficient method of removing dyes from dye-containing wastewaters.



The conventional coagulation/flocculation process, using inorganic polymers (synthetic or semi-synthetic), such as alum and ferrous sulphate (FeSO4), could increase the environmental pollution levels by introducing non-biodegradable compounds [16]. Therefore, many researchers have shifted to using natural coagulants for wastewater treatment due to the advantages of these coagulants over chemical agents, particularly their low toxicity, low residual sludge production, and biodegradability [17]. Natural coagulants are of great interest to scientists, since they are natural, low-cost products, characterized by their environmentally friendly behaviour, and are presumed to be safe for human health [18]. However, in many of these studies, the utilization of natural coagulants was associated with the addition of natural polymers, in order to enhance the floc size by attracting smaller particles to generate much larger flocs, and, in some of the studies, natural polymers were used, without adding any coagulant, due to the high efficacies of natural polymers in the flocculation process (direct flocculation) [19,20].

The use of natural polymers (plant or animal sources) is a promising method for treating wastewater and removing dyes, due to the chemical structure and the composition of the polymers, such as the presence of many functional groups, which contribute effectively towards the removal of dyes from the wastewater. In addition, natural polymers are non-toxic, low-cost, renewable, biodegradable, and biocompatible [21]. Natural polymers are synthesized from plant products, such as starch, guar gum, gum acacia, locust bean gum, pectin, nirmali seeds (Strychnos potatorum), and drumstick trees (Moringa oleifera), as well as from non-plant sources, such as alginates, carrageenans, chitin, chitosan, bacteria, algae, and fungi [22–28]. Nonetheless, in many cases, natural polymers are not sufficient to remove the dyes from highly complex dye-containing wastewaters, containing different types of pollutants, such as heavy metals, which have a negative effect on the attraction of dyes to natural polymers. Therefore, these polymers should be subjected to a modification process, involving chemical or physical treatment, in order to increase their efficiency in removing dyes from complex wastewaters. It is vital to modify the polymers according to the target application with tailor-made specifications, designed using blending, grafting, curing or derivatization methods. The natural polymer can be chemically modified by mineral acids, bases, salts of weak acids, enzymes, acetylation, saponification, concentrated ammonium systems, and primary aliphatic amines [29]. The physical modification process includes the blending of two or more types of the polymer at an ambient temperature or elevated temperature. Polymer grafting involves the monomer being covalently bonded onto the polymeric chain, which requires a longer time compared to curing. Curing forms, a coat of oligomers mixture onto the substrate using physical forces. In derivatization, the substitution of a simple molecule with a reactive group on the polymeric chain occurs to provide additional functional groups.

The current review article discusses the application of modified natural polymers in the removal of dyes from dye-containing wastewaters. The characteristics of natural polymers from plant and non-plant sources are reviewed. The feasibility of using modified natural polymers as an alternative technology for the removal of dyes from wastewater is investigated. The main aim of this article is to summarize the characteristics of dye-containing wastewaters, as well as the recent research concerning the application of modified natural polymers for the removal of dyes from different wastewaters. A comparison of several publications on the application of natural polymers has been compiled for this purpose. The authors recommend that the reported removal capacities of natural polymers be taken as a response to specific conditions, instead of maximum removal capacities.

#### **2. Characteristics of Dyes**

Over 100,000 types of commercial dye, for a total of more than 7 <sup>×</sup> 105 tons of dyestuff, are produced by the textile industries around the world annually [30]. Dyes found in the wastewater are primarily used in industrial activities, such as the textile industry and food processing. These wastes contain different types of chemicals, such as dyestuff, bleaching agents, finishing chemicals, starch, thickening agents, surface active chemicals, wetting and dispensing agents, as well as metal salts, which are used during each stage of textile production [31]. Several types of heavy metals have negative impacts on human health, as listed in Table 2 [32].


**Table 2.** Classes of di fferent dyes [32].

239

Dyes are classifieds into anionic dyes, cationic dyes, and non-ionic dyes. Cationic dyes are a category of basic dyes, while anionic dyes are known as disperse dyes, and are comprized of acid dyes, as well as direct and reactive dyes. Cationic dyes are water-soluble, with a positive charge and high color visibility. Anionic dyes carry a negative charge and differ from cationic dyes in terms of water-solubility, structure, and ionic substituents [30]. Several different types of dye have been universally utilized in textile industries, such as azo, triphenylmethane, perylene, anthraquinone, and indigoid dyes [33]. The different types of dyes are listed in Table 3 [34]. It can be noted that many of the dyes are used in different industrial applications.



In addition, the dyes are subjected to chemical and physical processes during their application, which might result in the production of different, unknown secondary chemical substrates in the generated wastewater. This point is of critical concern among scientists, since new unknown chemicals are being released into the environment. Thousands of types of synthetic dyes are commercialized to obtain multicolor fabrics [32]. It is estimated that the concentration of dye effluent can be in the range of 10 to 250 mg/L [33]. However, it depends on the specific dye industry. The highest concentration of dye effluent recorded from the reactive dye industry reached 7000 mg/L [35]. Another concern associated with the traditional treatment of the dye-containing wastewater lies in the dye's unknown degradation pathway, which has resulted in the release of secondary toxic by-products into the environment and the natural wastewater, rather than being removed from the partially treated dye wastewater. The formation of chlorinated compounds and phthalic acid esters (PAEs) in the treated dye wastewater has been reported in the literature, for example, [36]. Dyes released into the environment can have acute effects on organisms, based on their level of toxicity [37]. The presence of dyed water can even be at concentrations as low as 1 mg/L [38]. This may affect the amount of light penetrating the water, reducing photosynthesis rates. Thus, it is compulsory for raw wastewater to be treated before being released into the environment. The standards and the guidelines for the disposal of textile wastewater are illustrated in Table 4. Most of the countries listed showed similar numbers for permissible pollutant concentrations to those listed by the United Sates Environmental Protection Agency (US EPA), with the exception of Nigeria. These pollutants were SS, Hg, As, Cn, Cu, Mn, Sn, Zn, B, Fe, Ag, Al, Se, Ba, F, formaldehyde, phenol, sulphide, oil and grease, ammoniacal nitrogen, and color. This shows that the environmental regulation in Nigeria is not strict, at least in terms of water and wastewater sustainability. Amendments to the regulations should be considered, in order to provide clean water for future generations. Jordan and Bangladesh did not list the permissible concentrations of certain pollutants, such as Cr3, Sn, Ag, Al, Ba, formaldehyde, and color, according to their regulations. One possible reason is that these countries follow US EPA guidelines only for several ion's discharge limits. The list shows the different concentrations of ion discharge limits except for As, Pb, Ag, Al, Se, and Ba.

**Table 4.** Permissible standard limits for industrial wastewater in di fferent countries.



**Table4.***Cont.*

#### **3. Application of Polymers in Wastewater Treatment**

A polymer, whether grafted as a polysaccharide base, is highly efficient at binding and linking particles to itself and vice versa during collisions, resulting in the formation of larger, more settled flocs [45]. Polymers used in flocculation and coagulation might be inorganic or organic, and might be generated from natural resources, such as tannin, pectin, sodium alginate, chitosan, cellulose, gums and mucilages, which are derived from polysaccharides and proteins [46], or synthesized, such as acrylamide based poly-(2-methacryloyloxyethyl)-trimethylammonium chloride [47–49]. However, most of the previous studies focus extensively on the utilization of natural polymers, due to their high biodegradability. Many of the natural polymers discussed in the literature have been extracted from Moringa oleifera, Strychnos potatorum, Pseudomonas plecoglossicida, Spirogyra sp., and Aspergillus niger [50–54]. The extraction of polymers from agro-waste, such as guar gum, pectin, tannin, and locust bean gum, is explored due to these being environmentally safe, natural compounds from renewable resources, and not producing unintended hazardous wastes [55]. The high efficiency of the natural polymers in the removal of dyes from wastewater lies in the presence of different functional groups, such as the carboxyl, hydroxyl, phosphate, amine functional groups, which can bind to the cationic charges on the dye molecules by using electrostatic force [56]. Moreover, the renewable resources are abundant, and the aspect of biodegradability attracts many researchers [57].

The studies concerning the application of natural polymers in water and wastewater treatment are listed in Table 5. Most of them were applied as adsorbents or flocculants during the coagulation and flocculation process. Natural polymers usually work best in an acidic environment. However, some studies used a pH of 8 during the adsorption, to remove the methylene blue dye with the aid of acrylic acid [58]. About 20 mg of effective material was able to remove the dye color up to 45%. The highest removal rate, at 99.2%, was accomplished at pH 2 using pectin in 34.32 mg/L of Crystal Ponceau 6R dye [59]. A natural polymer extracted from animal waste, known as chitosan, was also able to remove 99% of the Duasyn Direct dye at pH 3.4, in combination with other materials, such as polyacrylamide and bentonite, as coagulants [60]. The cellulose/polyaniline (Ce/Pn) nanocomposite removed more than 90% of Remazol Brilliant Blue R (RBBR), Reactive Orange 16 (RO), Remazol Brilliant Violet 5R (RBVR), and Reactive Black 5 (RB) from the synthetic Remazol dye effluent. In contrast, 70.23% and 80.78% of the RBBR dyestuff was removed by using a chitosan-poly (acrylic acid) conjugate in an acidic environment, at pH 4 and pH 5, respectively [61]. RBBR dyes were successfully removed (100%) by using a combination of chitosan and cross-linked chitosan.

Natural adsorbents can also minimize the reaction time. Studies of alginates showed a 50% removal of methylene blue and methyl orange, which only required 10 and 17 min to achieve, respectively [62]. Previously, researchers used an alum and Acanthocereus tetragonus (a cactus species) to treat synthetic water, containing 100 to 500 ppm of Congo Red and Direct Blue dye [62]. The results indicated that using A. tetragonus as a coagulant resulted in 90% (up to 96%) color removal, while only 80% of color was removed by the alum. In addition, two types of plant organisms, namely Moringa oleifera seeds and Grewia venusta peel, were investigated for the treatment of synthetic dyes, namely, indigo carmine (reactive dye) and methyl orange dye [63]. The M. oleifera seeds were more effective than the G. venusta peel, with 99% and 85% dye removal, respectively. Both plants also showed optimum performances in acidic environments. The G. venusta peel, however, required harsh acidic conditions, at pH 2, to achieve the highest removal rate. In actual textile wastewater, plant extraction was not able to decolorize the pollution when using a single coagulant. Additional assistance was required to bind the dye particles together. Inorganic iron was added to the okra mucilage at the optimum pH of 6, in order to remove 93.57% of the colorant [64]. Textile wastewater is known to be a complicated waste to process, due to the presence of both cationic and anionic charges.


Treatment of several types of wastewater using natural polymers.

**Table**

**5.**

#### *Water* **2020**, *12*, 2032


**Table 5.** *Cont.*

Various monomers are tabulated and listed, along with some details on the dye removal rates from previous studies, in Table 6. Cationic monomers, such as poly-(2-methacryloyloxyethyl)-trimethylammonium chloride (PDMC), diallyldimethyl ammonium chloride, diethanolamine, and polyethylenimine, were very promising in thermal conditions and certain solvents. Grafting by using the opposite charges on the polymer's main backbone has a greater potential to increase the dye removal rate. A previous study showed an amphoteric grafting branch on chitosan using two monomers, known as carboxymethyl and poly-(2-methacryloyloxyethyl)-trimethylammonium chloride, for the removal of different charges of a dye molecule (cationic and anionic dyes) [68]. The existence of a quaternary ammonium group increased the cationic charges in the acidic environment, but a weak anionic character was displayed when the pH was above its isoelectric point. As the result, the bridging between these polymers and dyes became stronger. Some monomers made the polymer sensitive to pH in both acidic and basic environments. Diallyldimethyl ammonium chloride was grafted onto carboxymethyl cellulose (CMC), which was already cross-linked with mono-chloroacetic acid (MCA) and epichlorohydrin (ECH), and demonstrated a good performance in a methylene blue dye reduction, at over 98.54% removal in acidic conditions, and 83.07% removal in basic conditions, within a mere 20 min [69]. Using a PDMC monomer grafted onto carboxymethyl chitosan resulted in a 90% removal of Acid Green 25 at pH 4 and 98% of Basic Bright Yellow at pH 11. The quaternary ammonium salt was attracted to the anionic Acid Green 25 dye, producing a strong electrostatic attraction in an acidic environment, which is known as the neutralization effect. The presence of anionic charges on carboxymethyl chitosan assisted with the binding between the molecule of dye, known as Basic Bright Yellow dye, and anionic charges in the alkaline phase. A study on the alteration of bentonite with polyethylenimine to form an electrostatic charge and a hydrogen bond with the Amino Black dye, was successfully executed in an acidic environment (pH 3), at an adsorption rate of 264.5 mg/g [70]. The results showed that the mixture, containing 30% of polyethylenimine and 70% bentonite, was successfully grafted onto a molecule of (3-Glycidyloxypropyl) trimethoxysilane, with the aid of an epoxy bond, over a period of 24 h at 60 ◦C, in the presence of nitrogen gas.



The next type of monomer, known as non-ionic monomers (such as acrylamide), have been widely grafted onto many types of polymers. One of the important steps for improving the efficiency during water treatment was to select the polymer based on the best grafting ratio. An ultrasound-assisted method was applied in order to increase the grafting efficiency since it could reduce the polymerization time. A previous study showed that sodium alginate (SAG) grafted with polyacrylamide (PAM) was able to achieve a high color removal with a grafting efficiency of 75% [72]. The highest efficiency was attained at pH 10, with a 99% adsorption of methylene blue. Recently, an in-situ ultrasonic wave-assisted polymerization was explored as a substitute for proper physical emulsion mixing. In one of the studies, the ultrasonic system was applied to anionic monomers, such as acrylic, triochloroacetic acid, and poly (glycidyl methacrylate), and showed an affinity with cationic dyes. In addition, a monomer can also help increase the pure water flux by using an acrylic monomer to reduce the graft density, which can enhance the hydrophilicity. The fabrication of polypropylene composite hollow fibre membranes with acrylic monomers demonstrated good dye retentions, with a 99.5% and 98.7% removal of Congo Red and methylthionine chloride, respectively [75]. A recent study explored the modification of a magnetic adsorbent, using poly (glycidyl methacrylate (PGMA)) microspheres, cross-linked with ethylene glycol dimethacrylate (EGDMA), in the decolorization of dyes [73]. The polymerization techniques were modified, to some extent. A modified multi-step swelling polymerization method was employed with iminodiacetic acid (IDA) used to produce carboxyl groups, and the magnetic traits were successfully embedded inside the microsphere's pore using in-situ chemical co-precipitation. The coating microspheres were then exposed to ultraviolet (UV) radiation. The results indicated a good adsorption rate, and the decolorization rate reached 98.5%. The decolorization efficiency was more than 80%, despite the adsorption–desorption cycle being run ten times. Other monomers, such as trichloroacetic acid, remove 99% of cationic dyes, such as malachite green and rhodamine B [74]. High adsorption capacities were recorded for rhodamine B, at 222.6 mg/g, and 190.6 mg/g for malachite green. A low adsorption rate was detected for anionic orange dye, at 40 mg/g, due to a smaller number of cationic charges on the adsorbent.

Current research focuses on the extraction of microbial polymers, since it is easy to carry out, and cost effective for industries. In order to obtain highly effective polymers for use in the removal of dyes, researchers have focused on the polymers generated from indigenous microbes, such as Pseudomonas pseudoalcaligenes, Pseudomonas plecoglossicida, and Staphylococcus aureus, in fawn dyes, mediblue, whale dyes, and mixed dyes [76], because these dyes have a high resistance to decolorization. This is possibly due to the acidic nature of these dyes, which makes it difficult for them to be absorbed by microbial polymers [77]. The indigenous microbes might have adopted and developed a resistance mechanism in order to survive in these dyes, therefore, these organisms exhibit a high dye removal efficiency. In a recent study, one bacteria species (Brevibacillus laterosporus) and one yeast species (Galactomyces geotrichum) were immobilized in a stainless-steel sponge and in polyurethane foam. The microbial consortia successfully decolorized 50 mg/L of Remazol Red dye in the stainless-steel sponge and in the polyurethane foam in 11 h and 15 h, respectively [78]. Immobilization by using calcium alginate and polyvinyl alcohol produced more consistent results but required more time to complete the decolorization process at 20 h in the stainless-steel sponge and 24 h in the polyurethane foam. In another study, nine different bacterial strains from textile wastewater and sludge were isolated, which resulted in one Planococcus sp. with decolorization abilities being found in textile wastewater [79]. The Planococcus sp. decolorization ability was increased to 78% by combining it with a 55% peptone and a 60% dextrose solution (in a nitrogen and a carbon source, respectively).

#### **4. Graft Polymer (Coagulant**/**Flocculant)**

New developments in the polymer research have drawn attention towards graft polymers, also known as grafted copolymers. The advantages of graft polymers include their non-toxicity, high biodegradability in nature, and low cost (Figure 1). Moreover, their high molecular weights, as well as the existence of new branching on the molecular chains, makes them more suitable for removing dyes from wastewater. Natural graft polymers are defined as additional polymers inserted into the backbone of a natural polymer, in order to alter the molecular chain. The alteration extends the natural polymer's length, thus improving the adsorption of molecules with opposite charges in the solution [80]. Moreover, the existing branches of natural polymers have been modified in many studies, by inserting an acetyl group into the chitosan, resulting in more functional groups being added to the polymeric chain, thus improving the absorbance capacity of the polymers for azo dyes from wastewater [81]. The formation of the carboxylic group on the chitosan's polymeric chain was able to remove cationic and anionic dyes, which in this case involved the synthetic methylene blue and methyl orange dyes. As reported by the previous study, the grafted surface of graphene oxide showed a good potential for dyes due to the presence of carboxylic and hydroxyl groups that produce a colloidal dispersion in the aqueous medium due to its hydrophilic nature [82]. To conclude, the existence of the carboxylic group was indeed helpful in removing the dye particles.

**Figure 1.** Benefits of grafting a natural polymer.

In order to better understand graft polymers, further studies on their mechanisms are required. The two mechanisms involved in graft polymers are charged neutralization and bridge aggregation. Several insoluble complexes form at higher speeds during rapid mixing, indicating that neutralization has occurred. Subsequently, bridging occurred, resulting in the aggregation of the insoluble complexes and contributing to the formation of larger flocs, due to the increase in the molecular weight [83]. The links in the large flocs settled down rapidly, followed by those in the smaller flocs. A study showed that the bridging effect was more beneficial to the flocculation of grafted natural polymers than for linear polymers [84]. Figure 2 demonstrates the differences in natural polymer branching before and after modification [83]. Different particles are adsorbed onto the grafted natural polymer's chain to form bridges, which link to the opposite charges and occupy freer binding sites. Natural graft polymers carry more binding sites as the length of the modified natural polymer is longer than the original length [85].

**Figure 2.** (**a**) Linear polymer represents the coil form bound to the oppositely charged particles, while (**b**) is the graft polymer presenting the comb form, able to bind to a large number of oppositely charged particles in the solution [83].

#### *4.1. Factors A*ff*ecting the E*ffi*ciency of Polymer Coagulants*

#### 4.1.1. Type of Coagulant

The current trends in wastewater coagulation research focus mainly on the utilization of biodegradable polymers, such as chitosan, Moringa oleifera, nirmali seeds, and cactuses [86]. Natural polymers are used as coagulants because they are derived from renewable resources, biodegradable, cheap, non-toxic, and able to minimize the sludge production at the end of the treatment, as well as prevent the health risks associated with the utilization of alum, such as Alzheimer's disease [87]. Nonetheless, the study authors stated that synthetic organic polymers, such as diallyldimethyl ammonium chloride, as well as copolymers of quaternized dimethylaminoethyl acrylate or methacrylate, exhibited a better performance in the coagulation processes, due to charge densities and molecular masses [88]. The advantages of the synthetic organic polymers include the slow degradation activity, compared to that of the natural polymers with a longer life span, as well as the high charge density and molecular mass, which make them effective coagulants [89]. The disadvantages of synthetic polymers were not only their hydrophilic nature, but also their high cost [90]. This resulted in extra care being required during storage and transportation. Other drawbacks included limited molecular weight and dosage scale, which narrowed down the application range, and the presence of poisonous monomers, which are non-biodegradable and can be hazardous to the environment [91]. Although the wastewater treatment becomes more effective with an increase in dosage, problems arise when synthetic polymers are used due to unreacted chemicals making up the monomer unit (e.g., formaldehyde), as well as unreacted monomers (e.g., diallyldimethylammonium chloride and acrylamide) and reaction by-products.

#### 4.1.2. Coagulant Concentrations and Mixing Conditions

A sufficient amount of coagulant should be dispersed completely in the wastewater, with the optimum mixing speed and time, in order to get the maximum contact between the coagulant and the suspended particles. One of the main factors affecting a coagulant's efficiency is its concentration. It has been demonstrated that coagulants with a cationic charge neutralize the suspensions and destabilize the colloids. Increasing a polymer's concentration improves its performance, however, a high dosage has a negative effect on the coagulation processes and restabilizes the colloids, reversing the charge and reducing the removal rate [92]. A previously published study proved that increasing the organic coagulant concentration increased the COD removal rate, which, in this case, involved a combination of Moringa oleifera and potassium chloride. Using aluminium sulphate had the opposite effect [93]. The removal of organic matter declined as the concentration of hydrolysed metal salt increased. The hydrolysed metal salt can contribute to coagulation by adsorption [94]. However, the reversal of charge on the colloidal particle is the reason why the removal rate decreased. The aim of adding the salt to the natural polymer was to aggregate the particles, by means of a double layer compression [10]. The excessive addition of salt may also reduce the removal rate, due to the reduction in protein solubility (the salting-out effect) and the effect of hydration [95].

On the other hand, the mixing processes can also increase the removal rate efficiencies. The two phases that involve the mixing parameters in the coagulation are rapid mixing and slow mixing. The speed limit and the duration of time also play a role. The purpose of the rapid mix phase is to disperse the coagulant well, in order to stimulate particle collisions by using a power paddle. Rapid mixing was applied in a study, at speeds ranging from 80 rpm to 400 rpm and for time periods ranging from 0.1 min to 8 min. By varying rapid mixing, the formation, the breakage, and the regeneration of floc can be determined. A study showed that each coagulant has its own rapid mix condition [96]. An aluminium-based coagulant, for example, required the minimum time period to form a larger floc during the rapid mix phase, but the maximum time was needed when using a cationic polyelectrolyte [97]. Somehow, the regrowth of flocs after longer time periods was possible when using the cationic polyelectrolyte. The alum-based coagulant, on the other hand, had an irreversible effect on the floc recovery after the breakage. In agreement with other studies, mixing at 120 rpm for half a minute is required, in order to form larger flocs of highly turbid water [98]. The study also found that rapid mixing was connected to slow mixing, due to the flocs' resistance throughout the slow mixing phase. The time requirement during mixing did not appear to be the primary factor in removal effectiveness. The study found that the efficiency of coagulation, in terms of color and turbidity, had an indirect effect on the time [99]. However, the application of these factors during the process can provide some additional data. The following phase of slow mixing can take place at speeds ranging from 10 rpm to 60 rpm, for time periods ranging from approximately 5 min to 30 min. The investigation indicated that a slow mixing intensity had a beneficial effect on the charge neutralization coagulation when compared to sweep flocculation. Based on the optimum conditions, longer time periods are required during slow mixing, in order to produce larger flocs. During charge neutralization, an extended slow mixing phase can help to boost the coagulation performance, in case of inadequate rapid mixing [100]. This is in contrast with the sweep flocculation mechanism, where a shorter period of time is required for the slow mixing phase if the rapid mixing is excessively long. The mixing process should be followed by a sufficiently long settlement process [101]. Having larger flocs can shorten the time required for the settlement. Thus, the formation of larger flocs is important in the coagulation process.

#### 4.1.3. Functional Groups

Polymeric coagulants contain several types of functional groups with negative charges, such as hydroxyl (OH−), amine (NH3−), phosphate (PO4−), and carboxyl (COO−). These groups have bridging effects on the particles with opposite charges in the water and the wastewater [102]. In nirmali seeds, the presence of OH− groups along the galactomannan and the galactan molecular chains provide abundant attachment sites for interparticle bridging. The presence of the hydroxyl group, indicated by a broad wave between 3100 and 3500 cm<sup>−</sup>1, can be observed for natural polymers. Previous studies on different polymer analyses of functional groups are summarized in Table 7.


**Table 7.** Summary of the functional groups on selected polymers.

Biopolymeric flocculants, such as pectin, have similarly negative charged particles (OH and COO) [109] with longer chains, resulting in electrostatic repulsion, due to the chains stretching out [110,111]. Studies on Polydiallyldimethylammonium chloride (PDADMAC) grafted onto a locust bean gum showed the adsorption of amino groups reacting with anionic dye particles, indicated at 1474 and 3022 cm−<sup>1</sup> on the spectrum, increased with time, until the adsorption equilibrium was reached at approximately 600 min [112]. Other studies on the existence of phenolic groups in tannin structures showed these groups undergoing deprotonation to produce phenoxide effortlessly, thus expanding the oxygen atom's electron density [103]. The efficiency of the coagulation process is enhanced by the presence of additional phenolic compounds in the polymer structure [81]. The studies have shown that phenolic and amine groups are accessible in commercial tannin, which is a cationic polymer, with single tertiary amine groups present in each monomer [113]. Natural polymers can be grafted onto the surface of graphene oxide (GO) for the removal of dyes as GO has show a good potential for the removal of dye from aqueous solutions, as investigated by experimental and computational methods [82]. The computational evidence was illustrated using visual molecular dynamic programmes as can be seen in Figure 3. The surface area of graphene oxide posed a more negative charge in low pH with the absence of salt, which resulted in a protonated carboxylic group since no sodium ion (Na+) can attach to the surface. Otherwise, an anionic group (CH3COO<sup>−</sup>, SCN<sup>−</sup>, SO4 <sup>2</sup><sup>−</sup>, NO3 −) was able to attach to the hydrophobic graphene oxide. A previous study claimed that the higher surface area assisted the graphene oxide's ability to interact with direct blue Indosol dyes (a subcategory of anionic dyes) at pH < 4, due to several factors, such as the presence of hydroxyl, carboxylic, and oxygen groups, as well as the active sites on the dye particles, which may have resulted in stronger chemical bonds between the graphene oxide and the dyes [114]. Instead, Direct Red 81 was added to solutions with higher pH values (>7.5), resulting in nearly 100% removal. The new hydroxyl group will interact with amine groups from dye particles in basic conditions, increasing the adsorption observed in studies. In addition, multiple layers of neutralized sulphonate groups in the dye are formed, due to the electrostatic interaction between the oxygen-containing functional groups on the graphene oxide. In addition, the pi interaction and the bonding with the hydrogen molecules also help to increase the adsorption rate.

**Figure 3.** Visual molecular dynamics simulation of the attachment between graphene oxide (GO): (**a**) CH3COO<sup>−</sup>; (**b**) SCN−; (**c**) SO4 <sup>2</sup><sup>−</sup> and (**d**) NO3 −. Color coding: C-cyan, O-red, N-dark blue, H-white, S-yellow, Na-light blue [82].

#### 4.1.4. Molecular Weight (MW) of the Flocculant/Coagulant Aid

Different polymers have different molecular weights, based on their molecular structures. The molecular weight of the flocculant added after the coagulant has a significant impact on the process. Polymers with higher molecular weights contribute to efficient toxin removal, due to the mechanisms involved, such as charge neutralization, bridging, and electrostatic patch. Negatively charged particles are destabilized by the Van der Waals forces, due to the presence of elements with high molecular weights and numerous positive charges [110]. As soon as the Van der Waals forces are balanced out by repulsive electrostatic forces, flocs begin to develop. In addition, larger loops and ends form as the molecular weight increases; consequently, more sites are available to attract suspended particles [115]. Polymers with a minimum molecular weight of 800,000 daltons are more suitable for bridging [116]. A study of anionic polyacrylamides with different molecular weights indicates that a larger equivalent

size (resulting from a higher molecular weight being added at an extremely rapid mixing rate), results in a faster settling time [117]. However, a higher dosage is required to increase the density and to shorten the settling time, by using a lower molecular weight. Polymers with higher molecular weights and more branched chains demonstrate a better color removal performance and faster settling rates than linear chains with small numbers of active sites for the pollutant to bind [118]. A high resistance and large flocs simulated the separation of the pollutant particles out of the solution and their subsequent sinking to the bottom. The adsorption of the target pollutant requires a strong adhesion between solid–liquid interfaces, which is influenced by the polymer's molecular weight [119]. In another study, lignin- [2-(methacryloyloxy) ethyl] trimethyl ammonium chloride (METAC), with a high molecular weight, was more successful at removing Reactive Orange 16 than Reactive Black 5, though both dyes are anionic azo dyes [120]. The polymerization of polyacrylamides, with gelatine used as a tested stabilizer, shows a lower dye removal potential when using a low molecular weight grafted polymer [121]. The Congo Red dye removal was compared when using a copolymer (PAB) before and after grafting with dextran (DAB), and the experiment showed a 68.1% removal rate when using DAB, compared to 40.9% without the DAB, due to the higher molecular weight of the DAB grafted polymer [122]. Using a commercial polymer, such as polyaluminium chloride (PAC), achieves a 48% removal rate. This indicates that the molecular weight does influence the effectiveness of the coagulation, as well as the flocculation performance.

#### 4.1.5. Type of Charge Density

A polymer's charge density is categorized into low, medium, and high, based on the percentage mole of the ionic group (10%, 25%, and 50–100%, respectively) [123]. A low charge density enhances the polymer's bridging effect. The effectiveness of the charge density ranges from 5% to 15% and can improve shear resistance with higher molecular weights. Introducing graft copolymers results in an improved stability and extends the biodegradability, to some extent [124]. In addition, re-flocculation is incomplete when the charge density is below 12. The optimum flocculant concentration has been found to be dependent on the ionic strength [125]. Adding cationic charge polymers significantly enhances the coagulating capabilities. This effect increases in the order of monovalent < bivalent < trivalent. For example, adding the trivalent ion results in a stronger floc structure, and increases the floc size, the density and the shear resistance to a greater extent compared to monovalent ions [126]. One study, comparing four different lignin-based polymers extracted from pulping sludges, showed an excellent removal rate of disperse dye wastewater when higher-charge-density polymers were used. Particles form larger agglomerates when high-charge-density polymers are utilized, while looser molecules, despite the abundance of active sites for the dye molecule to attach, undergo slower reactions, as a consequence of the steric bulk [118].

#### **5. Removal of Dyes by Using Modified Natural Polymers**

Modified natural polymers have gained more attention in the recent years. Different types of compounds have been used to modify natural polymers for the purpose of removing dyes from wastewaters—for example, ammonia, formaldehyde, zinc oxide nanoparticles, lignosulfonate, carboxymethylstarch, polyacrylamide and various monomers [127]. The studies concerning modified natural polymers are presented in Table 8.



*Water* **2020** , *12*, 2032




**Table8.***Cont.*

Previously, researchers would come up with effective grafting procedures involving natural polymers. For example, an amine-modified tannin gel effectively removed brilliant green (at 94.05%) in neutral pH conditions, using the external surface adsorption mechanism [47]. In another study, a tannin-based polymer was grafted onto tannin extracted from Schinopsis balansae and Acacia mearnsii de Wild, by running a Mannich base reaction [139]. A tannin extract, modified with Clarotan and diethanolamine, was first acknowledged for its potential use as a dye or a surfactant in wastewater and river water [71]. Based on the wastewater simulation test, about 92% of the dye could be removed from a 100 mg/L stock solution, and the surfactant concentration could be reduced from 50 to 7.5 mg/L using a 150 mg/L stock solution. Modified tannin can perform well in water and wastewater treatment processes, as shown by previous studies [140,141]. Another study, focusing on lead adsorption, showed the ability of tannin-based hydrogels to absorb metal elements on their surfaces well [131]. When grafted onto a well-known chitosan, (3-chloro 2-hydroxypropyl) trimethylammonium chloride was used to reduce 1000 mg/L of the melanoidin dye at pH 3 using a 3 g/L dose chitosan-g-CHPTAC, which was able to remove up to 76.2% of the color and 90.14% of the turbidity [130]. A study by Sanghi et al. [67] focuses on grafting polyacrylamides onto different types of polysaccharides, such as amylose, amylopectin, starch, tamarind kernel, guar gum, glycogen, and chitosan. Grafted glycogen, having the highest molecular weight (6.81 × 106 g/mol) and radius of gyration, was the most effective, adsorbing 96.2% of the dyes, due to more branching taking place on the glycogen's backbone. This showed that the molecular weight is related to the effectiveness of the treatment. In a previous study, carboxyl methyl chitosan-graft-polyacrylamide (CMC-g-PAM) exhibited a high efficiency (above 90%) in removing anionic and cationic dyes. An investigation of ternary graft polymers revealed a high decolorization performance when using grafted chitosan for removing anionic and neutral dyes but demonstrated a low efficiency for the removal of cationic dyes [132].

Recent studies have demonstrated that acrylamide grafted onto sodium alginate successfully removes the methylene blue dye, with a removal rate of 99%. The sodium alginate alone did not remove any of the dye at all. However, there are some limitations of using modified natural polymers to treat real textile wastewater, due to the presence of different organic and inorganic complex chemicals. For example, grafted carboxymethyl starch was only able to remove 88.18% of the detected color at 520 nm [28]. The percentage removal of a synthetic dye can reach nearly 100%, but not in real textile wastewater. Grafted cellulose was used for the treatment of silk printing and dyeing wastewater, achieving a 95.7% removal of the COD [136]. The attachment of cellulose onto hyperbranched polyethylenimine resulted in a highly effective removal of ammonia nitrogen, total iron, and total phosphorus, at the original pH. The effluent pH did not require any adjustment, being approximately neutral. This characteristic is an additional benefit of polymer grafting, as the pH values during coagulation and flocculation treatments are independent of one another. Another study reported a wide pH range (from pH 5 to pH 9) available for the treatment of Acid Blue 113 and Reactive Black 5, which happened to remove more than 90% of the dye colors [135]. In addition, increasing the grafting ratio also increases the floc size and compactness, resulting in a lower dosage being required, in order to achieve a higher color removal efficiency. A cellulose-based flocculant, combined with poly-(2-methacryloyloxyethyl)-trimethylammonium chloride, showed the best removal rate for an anionic dye (97.3%). Increasing the grafting ratio also improved the color removal rate. Moreover, other advantages of grafted polymers depend on the dosage applied during the process, which can be reduced up to 50% in synthetic wastewater using PAFC-Starch-g-p (AM-DMDAAC) [138]. With increasing environmental awareness, a recent study demonstrated the outstanding biodegradability of cellulose, extracted from bamboo pulp, grafted with polyacrylamide, at 66.5% and 67.6% after 45 d and 90 d, respectively, in a soil-extracting solution [137]. It was also successful at removing organic dyes, such as cationic and disperse dye solutions, with an average removal rate of 97%. Overall, grafted natural polymers showed a superior dye removal performance, compared to natural polymers on their own.

#### **6. Conclusions**

Over the past several decades, high levels of toxicants have been produced as the result of dye wastewater treatment involving harmful chemicals. Even though factors like turbidity, color, COD, BOD, and the levels of heavy metals have been reduced to meet the permissible standards, the sludge produced as the result of the treatment still comes into contact with toxic materials. Thus, an effective solution is required in order to improve water quality. Grafted natural polymers could replace commercial polymers, with an additional incentive of reduced costs. The characteristics of effective coagulants can be enhanced by using specific types of polymers, the concentration and mixing conditions, functional groups, higher molecular weight, and charge density according to the target dyes pollutant. The chemical modification of polymers provided the opportunity to explore beyond conventional applications. A thorough understanding of the polymer and its chemical modification has a vast potential to be the future trend for the use of cosmetics, pharmaceutical, food, leather, paper, and textile industries.

**Author Contributions:** Conceptualization, S.A.I., H.M.A. and A.A.S.A.-G.; methodology, S.A.I. and S.Z.M.R.; resources, S.A.I.; writing—original draft preparation, S.A.I.; writing—review and editing, S.A.I.; supervision, M.F.M., A.A.S.A.-G., and N.I.; funding acquisition, M.F.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by RUI grant no. 1001/PAWAM/814259.

**Acknowledgments:** The authors would like to thank the Malaysian government for providing Mybrain15 scholarship and thankful for the kind support of staff of the Universiti Sains Malaysia.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


#### *Water* **2020**, *12*, 2032


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Metallic Iron for Environmental Remediation: Starting an Overdue Progress in Knowledge**

**Rui Hu 1,\*, Huichen Yang 2, Ran Tao 2, Xuesong Cui 1, Minhui Xiao 1, Bernard Konadu Amoah 1, Viet Cao 3, Mesia Lufingo 4, Naomi Paloma Soppa-Sangue 1,5, Arnaud Igor Ndé-Tchoupé 6, Nadège Gatcha-Bandjun 7, Viviane Raïssa Sipowo-Tala 5, Willis Gwenzi 8,\* and Chicgoua Noubactep 2,\***


Received: 22 January 2020; Accepted: 21 February 2020; Published: 27 February 2020

**Abstract:** A critical survey of the abundant literature on environmental remediation and water treatment using metallic iron (Fe0) as reactive agent raises two major concerns: (i) the peculiar properties of the used materials are not properly considered and characterized, and, (ii) the literature review in individual publications is very selective, thereby excluding some fundamental principles. Fe0 specimens for water treatment are typically small in size. Before the advent of this technology and its application for environmental remediation, such small Fe0 particles have never been allowed to freely corrode for the long-term spanning several years. As concerning the selective literature review, the root cause is that Fe0 was considered as a (strong) reducing agent under environmental conditions. Subsequent interpretation of research results was mainly directed at supporting this mistaken view. The net result is that, within three decades, the Fe<sup>0</sup> research community has developed itself to a sort of modern knowledge system. This communication is a further attempt to bring Fe<sup>0</sup> research back to the highway of mainstream corrosion science, where the fundamentals of Fe0 technology are rooted. The inherent errors of selected approaches, currently considered as countermeasures to address the inherent limitations of the Fe<sup>0</sup> technology are demonstrated. The misuse of the terms "reactivity", and "efficiency", and adsorption kinetics and isotherm models for Fe<sup>0</sup> systems is also elucidated. The immense importance of Fe0/H2O systems in solving the long-lasting issue of universal safe drinking water provision and wastewater treatment calls for a science-based system design.

**Keywords:** adsorption capacity; decentralized water supply; electrochemical reaction; inconsistent view; sand filtration; wastewater treatment; zero-valent iron

#### **1. Introduction**

Metal corrosion is one of the most important problems in industry, transport and agriculture [1–4]. Understanding metal corrosion comprises its detection (e.g., analytical, visual), its monitoring (e.g., mass loss, H2 evolution) and its long-term characterization under various field conditions [4,5]. The corrosion research aims at determining the durability of metallic structures under operational conditions (e.g., oil and gas pipelines, tanks) and revealing the mechanisms of corrosion process [5,6]. This mechanism can be chemical, electrochemical or mixed [1]. Various tools have been used to characterize the corrosion resistance of different metals under various application conditions [4]. The overall result is the availability of integrated approaches to assess and predict the corrosion processes and thus the longevity of metallic structures (e.g., buried pipes) [3,4]. However, the frequency and sudden nature of metallic pipe failures worldwide indicate the inadequacy of current knowledge related to the longevity of buried metallic pipes.

Metallic iron (Fe0) used as a reactive material in subsurface permeable reactive barriers is comparable to iron pipes with three major differences: (i) corrosion is welcome because it is a rather useful process [7–9], (ii) a reactive wall is ideally permanently water saturated, and (iii) the length of used particles (<5 cm) is tiny compared to pipes which are up to 12 m in length. On the one hand, Fe<sup>0</sup> specimens used in water treatment comprise steel wool with thickness varying between 25 and 90 μm [10,11]. On the other hand, the length of these particles is comparable to the wall thickness of iron pipes (2–4 mm). There has been no real system analysis for remediation Fe0 materials with the aim to outline the differences making their peculiar characteristics. In addition, traceably deriving the longevity of remediation Fe<sup>0</sup> specimens from Fe0 pipes is impossible because of the differences highlighted.

The remediation Fe0 was termed as zero-valent iron (ZVI). This acronym is perhaps the first problem of this still innovative technology. A literature research with "zero-valent iron" as keyword would never reveal the ancient literature on Fe0 for water treatment [12–16]. Indeed, no research group until 2017 has referenced a single article from the ancient use of Fe<sup>0</sup> in water treatment [17,18]. In chemistry, metallic elements are characterized by their oxidation state, the one of Fe0 is zero (0). Upon oxidation, Fe<sup>0</sup> is transformed to FeII, FeIII or FeIV species. Under environmental conditions, only FeII and FeIII species are stable. The oxidation of Fe<sup>0</sup> to FeII is a redox process characterized by an electrode potential whose value is ™0.44 V [1]. According to the first principle of chemical thermodynamics, Fe0 can be oxidized by oxidizing agents from each redox couple having a higher electrode potential (E0 > ™0.44 V). Water (H2O or H<sup>+</sup>) is a relevant oxidizing agent for Fe0 under environmental conditions. The electrode potential for the redox couple H+/H2 is 0.00 V. Fe<sup>0</sup> immersed in (contaminated or polluted) water is corroded to form H/H2 and FeII (and mixed FeII/FeIII) species which are stand-alone reducing agents [19–26]. Clearly, it is not surprising that selected species undergo reductive transformations in an Fe0/H2O system [19,26–29]. The application of metallic iron in water treatment dates back to the 1890s (Table 1). Suspended matter then left to settle




**Table 1.** *Cont.*

Table 1 presents a summary of the historical applications of metallic iron in water treatment. The history of Fe<sup>0</sup> application in environmental remediation has been the subject of several papers by our research group. Comprehensive research on Fe0 for water treatment revealed that the generation of iron hydroxides and oxides (iron corrosion products or FeCPs) is the root cause of contaminant removal in Fe0/H2O systems [22,23,25,26,53,54]. For example, arsenic [55]*,* carbon tetrachloride [29]*,* chromium [56], fluoride [57,58], hexachloroethane [59] , methylene blue [60], methyl orange [61], Orange II [62], phosphate [63], selenium [64] and zinc [65] are all removed in Fe0/H2O systems despite their differences in charge, redox-reactivity and size.

A critical research review article from 2008 [54] established that adsorption, co-precipitation and size-exclusion were the fundamental mechanisms of contaminant removal in Fe0/H2O systems. Yet to date, 13 years later, researchers are still trying to establish the mechanisms by which aqueous contaminants are removed in the presence of Fe<sup>0</sup> [59,63,64,66,67]. Unfortunately, none of them has proven the alternative concept [53,54,68] wrong, and most of them are considering Fe<sup>0</sup> as the electron donor for observed transformations (electrochemical reaction) [69–71]. Moreover, "contaminant reduction" and "contaminant removal" are mostly randomly interchanged. Therefore, a better

argumentation is still needed to convincingly explain the in-depth knowledge on the mechanisms causing contaminant removal in Fe0/H2O systems. Another important application of Fe0 materials entails using Fe<sup>0</sup> to induce a pH shift and/or enhance microbial processes involved in methane production in anaerobic digesters [72–75]. This aspect is not considered herein as it is not focused on contaminant removal as discussed in Section 2.

This paper presents a profound analysis of the Fe0/H2O system and derives the leading causes and factors influencing its efficiency for water treatment. Relevant factors include: (i) the Fe<sup>0</sup> specimen including its form and size (intrinsic reactivity), (ii) the water chemistry including the nature of contaminants, the presence of dissolved O2 and the pH value, (iii) the contact time (flow velocity or mixing intensity), and (iv) the Fe<sup>0</sup> amount and its proportion in the reactive mixture (thickness of the reactive layer or number of columns). Contaminants are explicitly considered within water chemistry, one of the four main groups of factors influencing iron corrosion. The approaches conventionally adopted to investigate these factors, the major findings, the limitations and the knowledge gaps are presented. It is argued that no progress in knowledge is possible before the research community agrees on the key issue that Fe0 is not a reducing agent (under field conditions).

#### **2. The Fe0**/**H2O System**

#### *2.1. Overview of Fundamental Aspects*

There is a transfer of electrons from the Fe<sup>0</sup> body (solid state) to the Fe0/H2O interface whenever a piece of a reactive Fe<sup>0</sup> specimen is immersed in an aqueous solution (Fe0/H2O system) [1,5,6,76]. This occurs because Fe<sup>0</sup> is not stable under environmental conditions or because the redox couple H+/H2 (E<sup>0</sup> = 0.00) is higher than that of FeII/Fe<sup>0</sup> (E<sup>0</sup> = ™0.44) in the electrochemical series [1,22–26]. Equation (1) reveals that the oxidative dissolution of Fe0 by protons (H+) and Equation (1a) considers that protons are from water (H2O <sup>⇔</sup> H<sup>+</sup> + HO–). Fe(OH)2 from Equation (1b) tends to polymerize and precipitate but can also be oxidized to even lower soluble Fe(OH)3 by dissolved O2 for example (Equation (2a)). Fe(OH)2 and Fe(OH)3 are polymerized and further transformed to various hydroxides and oxides (FeCPs) (Equation (3)) [1,21,77]. The different iron corrosion products (FeCPs) depict different adsorptive affinities for dissolved species [77–79]. Equation (4) summarizes the process of aqueous iron corrosion.

$$\mathrm{Fe}^{0} + 2\,\mathrm{H}^{+} \Rightarrow \mathrm{Fe}^{2+} + \mathrm{H}^{2} \tag{1a}$$

$$\text{Fe}^{0} + 2\text{ H}\_{2}\text{O} \Rightarrow \text{Fe(OH)}\_{2} + \text{H}\_{2} \tag{1b}$$

$$4\text{ Fe}^{2+} + \text{O}\_2 + 4\text{ H}^+ \Rightarrow 4\text{ Fe}^{3+} + 2\text{ H}\_2\text{O} \tag{2a}$$

$$4\text{ Fe(OH)}\_{2} + \text{O}\_{2} + 2\text{ H}\_{2}\text{O} \Rightarrow 4\text{ Fe(OH)}\_{3}\tag{2b}$$

$$\text{Fe(OH)}\_{2}, \text{Fe(OH)}\_{3} \Rightarrow \text{FeO,Fe}\_{3}\text{O}\_{4}, \text{Fe}\_{2}\text{O}\_{3'}, \text{Fe(OH)}\tag{3}$$

$$\text{Fe}^{0} + \text{H}\_{2}\text{O} + \text{(O}\_{2}) \Rightarrow \text{H}\_{2} + \text{iron hydroxide and oxides} \tag{4}$$

In summary, Equation (4) recalls that immersing a reactive Fe0 in water can be universally used to generate H2, Fe<sup>2</sup><sup>+</sup> and various FeII, FeIII and FeII/FeIII hydroxides and oxides. Equation (1) demonstrates that Fe<sup>0</sup> is a scavenger of humidity (H2O), while Equation (2) demonstrates the O2 scavenging nature of Fe0. These two scavenging characteristics have been exploited in several industrial applications [80–82]. For example, Fe<sup>0</sup> is used as desiccant in food packaging [81]. In the Fe<sup>0</sup> remediation literature, the best illustration for the nature of the Fe0/H2O system as generator of FeCPs is perhaps the excellent research article by Furukawa et al. [83]. These authors used several analytical tools to demonstrate the presence of ferrihydrite, green rust, magnetite and lepidocrocite in an Fe0/H2O system. More importantly, they conclude that an Fe0/H2O system is a temporally and spatially heterogeneous geochemical environment. Concerning the spatial heterogeneity, Furukawa et al. [83] specified that magnetite (Fe3O4) is generated in the vicinity of Fe0, whereas ferrihydrite (Fe(OH)3) precipitates away from the Fe0 surface. This

conclusion corroborates ancient findings [84] and recalls that even under oxic external conditions, there is a progressive O2 depletion culminating into anoxic conditions in the vicinity of Fe0. In this context, Stratmann and Müller [85] clearly demonstrated that oxygen is reduced by FeII species within the oxide scale (chemical reaction), while Fe<sup>0</sup> is oxidized by water (electrochemical reaction).

This section has recalled that, before Fe0 complete depletion, an Fe0/H2O system is a dynamic and heterogeneous system containing Fe<sup>0</sup> and all its corrosion products (Equation (4)). Accordingly, even the most accurate measurements and the most precise observations are just a static snap-shots of dynamic processes within the system. The situation is exacerbated by the evidence that, the processes occur over an enormous range of timescales ranging from some few minutes or hours in laboratory investigations to months and years in field applications [86–88]. This communication insists on the fact that the frequency of discrepant reports in the scientific literature is rooted on an insufficient system analysis. In this regard, an accurate system analysis constitutes the theory of the system. The theory of the Fe0/H2O system, in turn, is like a guide to constrain the choice of the model. The theory that some contaminants are reduced by electrons from Fe<sup>0</sup> (electrochemical mechanism) is flawed as water, even present as humidity or moisture do corrode iron [1]. In fact, even deionized water corrodes iron [89]. Figure 1 depicts the interactions of contaminants with solid phases in a pure adsorbent system and the Fe0-based system.

**Figure 1.** Schematic diagram comparing the interactions of contaminants (black points) with solid phases in a pure adsorbent system (left) and the Fe0-based system (right). The red points represent iron corrosion products (FeCPs) which are either coated on solids or suspended in the pore solution.

#### *2.2. Oxide Scale on Fe<sup>0</sup> and the Decontamination Process*

The universal oxide scale on iron metal (at pH > 4.5) is still regarded by the majority of active researchers on remediation Fe<sup>0</sup> as a disturbing factor compromising the electron transfer from the metal body (reactivity loss) [90–92]. This view contradicts the evidence that a lag time between the start of experiments and reductive transformation is reported in the literature [93,94]. This time corresponds to the quantitative generation of FeCPs, which act as contaminant scavengers. This means that relevant reducing agents are generated in situ. Another problem of the Fe<sup>0</sup> literature is that "contaminant removal" and "contaminant reduction" are randomly interchanged, while no real mass balance of the contaminants has been presented [53,54,95]. On the contrary, contaminants that are not recovered are assumed to be chemically reduced [68–70].

A look at the mechanism of oxide scale formation reveals that it cannot be electronically conductive. In fact, the initial scale is very porous and cannot transfer electrons because air and water are not electronically conductive (an aqueous solution can be ionic conductive—electrolyte). In subsequent stages, available pores are filled with nascent FeCPs, but they are never uniform and the oxide scale is a mixture of iron hydroxides and oxides [5]. An oxide scale made up of Fe3O4 alone would have been electronically conductive. However, such an Fe3O4 scale cannot exist under natural conditions (immersed Fe0). All other FeCPs are at best semi-conductors and cannot relay electrons from Fe0 under natural conditions. Clearly, reports justifying the reductive efficiency of Fe0/H2O systems using the semi-conductive nature of FeCPs are mistaken [96]. This assertion encompasses the Fe0/pyrite/H2O system, whose efficiency is mainly justified by the semi-conductive nature of FeS species [97–99].

The oxide scale on Fe0 is definitively a diffusion barrier for all dissolved species, including the pollutants. It is also the contaminant scavenger such that electrochemical corrosion of immersed Fe0 induces the generation of contaminant scavengers and other reducing agents. Thus the generation of solid FeCPs is a necessary process which has the (perceived negative) side effect of being expansive. Thus, designing an efficient and sustainable system requires answering the question: how long can FeCPs be generated to satisfactorily treat water while keeping a reasonable hydraulic conductivity (permeability)?

#### *2.3. Chemical Aspects*

As discussed earlier (Section 2.1), Fe2<sup>+</sup> from Equation (1) is transformed to ferrous hydroxide (Fe(OH)2) and ferric hydroxide (Fe(OH)3) which have a strong tendency to form colloids of particles that normally carry a positive charge [62,100,101]. These minerals are further transformed to other FeII/FeIII minerals (e.g., Fe2O3, FeOOH, green rust) exhibiting different affinities to dissolved species. The nature of oxides in each individual system depends on the intrinsic reactivity of the used Fe<sup>0</sup> material and the environmental conditions [5,79]. For example, two different Fe<sup>0</sup> specimens corroding under the same environment will not necessarily produce the same iron oxides, because the composition of the oxide scale depends on the relative kinetics of Fe<sup>0</sup> dissolution and Fe hydroxide precipitation, which in turn depends on the solution chemistry, including the pH value, dissolved ions and the salinity [5,6,102]. On the other hand, in-situ generated free Fe2<sup>+</sup> are adsorbed to the surface of available minerals to form the so-called structural FeII with a reducing power far larger than that of the free Fe2<sup>+</sup> (E<sup>0</sup> < 0.77 V) and sometimes stronger than Fe<sup>0</sup> (E0 < –0.44 V) [103]. The availability of several reducing agents in the Fe0/H2O system, and especially from structural FeII, partly stronger than Fe<sup>0</sup> implies that the electrochemical series of metal alone cannot predict the chemistry of the system.

#### *2.4. Physical Aspects*

The volumetric expansive nature of iron corrosion is the most important physical phenomenon occurring in Fe0/H2O systems [104]. There is expansion because the parent metal (Fe0) produces in-situ both: (i) H2, occupying a volume about 3100 times larger [26,105], and (ii) each solid oxide and hydroxide is at least twice larger in volume than Fe<sup>0</sup> (Voxide > Viron) [106,107]. For example, the specific density of magnetite (Fe3O4) is about one half that of iron (Fe0). It implies that after corrosion a space twice larger than the initial space is occupied [108–110]. While it can be assumed that H2 escapes from each open system, no free expansion of oxides in porous systems (e.g., water filters, reactive walls) can be assumed [26]. External or internal free expansion occurs in metallic pipes [111,112] and on the walls of steel canister for radioactive waste repositories [110,113]. On the contrary, free expansion cannot be expected in steel-reinforced concrete structures [108,109]. Table 2 presents a comparison of iron corrosion parameters in Fe0 remediation to that of water pipes and reinforced concrete. Accordingly, considering expansive iron corrosion, which culminates into permeability loss is an essential design parameter for porous Fe0/H2O systems (Fe<sup>0</sup> filters). For each Fe0 filter, the temporal production of both H2 and oxide is decisive for the long-term efficiency and the permeability of the system [105,107,114,115].


**Table 2.** Comparison of iron corrosion parameters in Fe<sup>0</sup> remediation to that of water pipes and reinforced concrete.

Another key feature in investigating the remediation Fe0/H2O system is that, at pH > 4.5, the Fe<sup>0</sup> surface is permanently covered by an oxide scale. The oxide scale acts both as: (i) conduction barrier for electrons from the metal body, and (ii) physical barrier for dissolved species, including O2 and pollutants of concern [22–26,53,54]. The net result is that Fe<sup>0</sup> is oxidized by water (electrochemical reaction—Equation (1)), and O2 and dissolved contaminants are reduced by reducing species present in the oxide scale (FeII and FeII/FeIII species, H2) (chemical reaction). Again, any reasoning based on the electrochemical series of elements is a misuse of chemical thermodynamic, as the physics of the system, specifically the expansive nature of iron corrosion and it electronically non-conductive nature are simply ignored.

#### *2.5. Kinetic Aspects*

Aqueous corrosion of Fe<sup>0</sup> materials under environmental conditions (pH >4.5) is an electrochemical process involving iron dissolution at the anode and H2 evolution at the cathode (Equation (1)). This electrochemical reaction is accompanied by the formation of an oxide scale on Fe<sup>0</sup> which is not protective as a rule [6,116–118]. In general, oxide scale growth and its protectiveness depend primarily on the precipitation rate of iron hydroxides [5,102]. As the Fe<sup>0</sup> surface corrodes under the initial scale, corrosion continuously undermines the scale. Voids are created and are progressively filled up by the ongoing hydroxide precipitation. The relative rate of (i) Fe<sup>0</sup> oxidative dissolution and (ii) hydroxide precipitation in the Fe0 vicinity determine the protectiveness of the oxide scale. According to Nesic [5], when the rate of hydroxide precipitation exceeds the rate of Fe0 dissolution, a dense protective oxide scale is formed. Conversely, when Fe0 dissolution undermines the new oxide scale faster than hydroxide precipitation can fill in the voids, a porous and non-protective scale forms.

There are several factors influencing the iron corrosion rate, including the intrinsic reactivity of a material and the solution chemistry. The solution chemistry includes the presence of dissolved O2, contaminants and other (mostly) ubiquitous species. In particular, two different Fe0 specimens may exhibit different degree of protectiveness under the same operational conditions. The most important feature to consider is that the corrosion rate is never linear. Thus, a linear extrapolation of the initial corrosion rates of Fe<sup>0</sup> specimens in engineered systems can give an inaccurate estimation of its service life. However, for remediation systems, the non-linear nature of the corrosion kinetics implies a decrease in efficiency. Again, results from pipe corrosion or wall corrosion cannot be transferred to remediation Fe0/H2O systems. However, it is certain that the semi-permeable nature of the oxide scale allows significant and continuous corrosion over long time periods ("rust never rest"). In this context, Roh et al. [119] reported of buried iron pieces from World War I still corroding in the subsurface. Thus, the objective of the Fe<sup>0</sup> remediation is to couple this long-term corrosion with efficient contaminant removal [7,88,118,120,121].

#### *2.6. Investigating the Fe0*/*H2O System*

The extent and kinetics of Fe0 oxidative dissolution are the result of the following: (i) the nature of metal (intrinsic reactivity), and (ii) the interactions between Fe<sup>0</sup> and the environment in which it is placed [76,122–125]. Therefore, (i) a change of material, and/or (ii) a change in environment results in changes in the rate and the extent of corrosion. In this study, the influence of the environment is considered at micro-scale, specifically what is happening on the Fe<sup>0</sup> surface, in its vicinity or over short distances (within the oxide scale). The oxide/water interface is also considered, but the volume of the solution is not. The aqueous phase is regarded as a reservoir of pollutants and co-solutes, while being the principal Fe<sup>0</sup> oxidizing agent [126,127].

Laboratory and pilot-scale investigations are usually conducted in order to obtain reliable information on the interactions of metallic devices with particular operational environments [128,129]. Laboratory tests are designed to simulate some relevant field situations. Laboratory studies are mostly aimed at obtaining data in a more convenient way and in a shorter time [123,124,130]. They also provide mechanistic information for field applications. However, short-term laboratory experiments are always a simplification and this should be borne in mind when interpreting achieved results [76,129,131,132]. Given the diversity of operational parameters that have been proven to influence iron corrosion from individual studies, one can be overwhelmed by their number and the fact that each material is unique in its corrosion behaviour [11,124,125,133]. Therefore, a first attempt toward a systematic investigation of relevant influencing factors goes through the consideration of the electrochemical nature of aqueous iron corrosion.

Fe<sup>0</sup> is a good conductor of electricity, and its electrochemical aqueous corrosion depends on the conductive nature of the solution in which it is immersed [1,76]. Understanding the corrosion processes helps in selecting Fe<sup>0</sup> materials necessary for designing sustainable systems. Therefore, the best way to design a sustainable system is to consider the fundamentals of iron corrosion at the design stage. This section highlights the knowledge of corrosion principles and the importance of the environment and materials for field design. The four compartments necessary for continuous Fe<sup>0</sup> corrosion are: (i) an anodic region on Fe0 where the metal is oxidized and releases Fe2<sup>+</sup> (leaving two electrons behind), (ii) an electrolyte to transport released Fe2<sup>+</sup> away from the anode, (iii) a cathode where the simultaneous reductive transformation coupled to iron oxidation occurs, and (iv) the Fe0 body transporting electron from the anode to the cathode. In an Fe0/H2O system, the anode and the cathode are different sites on the same Fe<sup>0</sup> specimen. In the conventional remediation technology, the size of Fe0 particles is generally small (<5 mm). The four compartments must be electronically connected for the electrochemical process to proceed. This means that if Fe<sup>0</sup> is covered by an oxide scale, the scale must be conductive to warrant the transfer of electrons from the metal body to the oxidizing agent within the oxide scale or at the oxide/H2O interface.

A Fe0/H2O system is made up of two interfaces: (i) Fe0/oxide and (ii) oxide/H2O. Because the oxide scale is never electronically conductive and is a diffusion barrier to many species, only water can quantitatively reach the Fe<sup>0</sup> surface. The net result is that Fe<sup>0</sup> oxidative dissolution is an electrochemical reaction (water is reduced) but all other observed/reported chemical reduction occur within the oxide film or at the oxide/H2O interface [53,54,127]. For the Fe<sup>0</sup>/H2O remediation system, it means that contaminants are not reduced by electrons from Fe0. This has important implications on the operating principles of Fe0/H2O systems: first, using the electrochemical series to predict the reductive transformation of any species has been a mistake, and second, using the stoichiometry of any electrochemical reaction involving Fe<sup>0</sup> is also a mistake. Accordingly, contaminants are 'just' dissolved species, capable of modifying: (i) the conductive properties of the electrolyte (H2O), (ii) the ion conductivity of the oxide scale, and (iii) formation and the transformation of the oxide scale.

#### **3. An Overview of the Mistakes of Past E**ff**orts**

#### *3.1. Contaminant Removal Mechanisms*

The major mistake of the Fe<sup>0</sup> remediation literature has been to consider that relevant contaminants are reduced by an electrochemical reaction (electrons from the metal) [134–136]. This section summarizes the extent of the confusion using a paper by Kamolpornwijit and Liang [137]. This paper is selected because it considered past efforts in understanding the service life of Fe<sup>0</sup> filters and performed long-term experiments (>400 days). Kamolpornwijit and Liang [137] considered porosity loss during nitrate (NO3 <sup>−</sup>) removal in an Fe0 barrier. The following reactions were considered:

$$\text{Fe}^{0} + 2\text{ H}\_{2}\text{O} \Rightarrow \text{Fe}^{2+} + \text{H}\_{2} + \text{OH}^{-} \tag{5}$$

$$4\text{ Fe}^{0} + \text{NO}\_{3}^{-} + 7\text{ H}\_{2}\text{O} \Rightarrow 4\text{ Fe}^{2+} + \text{NH}\_{4}^{+} + 10\text{ OH}^{-} \tag{6}$$

$$5\,\mathrm{Fe^{0}} + 2\,\mathrm{NO\_{3}}^{-} + 6\,\mathrm{H\_{2}O} \Rightarrow 5\,\mathrm{Fe^{2+}} + \mathrm{N\_{2}} + 12\,\mathrm{OH^{-}}\tag{7}$$

$$10\text{ Fe}^{2+} + 2\text{ NO}\_3^- + 6\text{ H}\_2\text{O} \Rightarrow 10\text{ Fe}^{3+} + \text{N}\_2 + 12\text{ OH}^-\tag{8}$$

Equation (5) corresponds to Equation (1) and represents aqueous iron corrosion, an electrochemical reaction. Direct electron transfer from Fe0 to NO3 <sup>−</sup> yielding NH4 <sup>+</sup> (Equation (6)) or N2 (Equation (7)) are only possible if the oxide scale on Fe0 is electronically conductive. Because an electronically conductive oxide scale does not exist under immersed conditions, Equations (6) and (7) are wrong. In discussing their results, Kamolpornwijit and Liang [137] considered that the H2 volume they measured corresponds to the stoichiometry of Equation (5). Comparing this H2 volume to the N2 volume after Equation (7), they concluded that the contribution of water to the corrosion of iron is minimal. In chemical terms, this is erroneous as Equation (5) shows that 5 moles of Fe<sup>0</sup> release only 1 mole of N2. Whether N2 escapes from the system or not, the 5 moles of Fe2<sup>+</sup> are further oxidized and or precipitated within the Fe0 filter and filling the initial porosity. In reality, it is 10 moles of Fe<sup>0</sup> that are needed to release one mole of N2 (Equation (8)).

One major problem of the Fe<sup>0</sup> literature has been the discussion of the reaction mechanism without mass balance considerations [31,95,138]. Complete mass balance analysis comprises the one of iron, which is admittedly a difficult task as the system is highly dynamic. However, the analysis made herein for nitrate reduction clearly demonstrates the wrongness of the view that Fe<sup>0</sup> is corroded by NO3 −. On the other hand, the importance of FeCPs in inducing porosity loss is demonstrated. If the discussion of the porosity loss starts with the assumption that one mole of N2 corresponds to 10 moles of corroded and expanded Fe0, a better evaluation of changes in the porosity will be achieved. Kamolpornwijit and Liang [137] is also an excellent illustration on how in-situ generated FeCPs are not properly considered while exotic species like CaCO3 or the flow regime (convection versus laminar) are given key role in pore filling and permeability loss. In essence, if a contaminated water is rich in carbonates a pre-treatment unit for their removal can be used such that decontamination units are

HCO3 − free. Thus, given the expansive nature of iron corrosion, loss of permeability and porosity even occurs in aqueous systems without CaCO3.

Another important feature from Kamolpornwijit and Liang [137] is that, in case nitrate is reduced to NH4 <sup>+</sup>, regardless of the real mechanisms, NH4 <sup>+</sup> must be removed from the aqueous phase. This removal occurs by adsorption, co-precipitation and size-exclusion [53,54]. These three mechanisms represent the fundamental mechanisms of contaminant removal in Fe0/H2O systems [26,67]. It is essential to recall that chemical reduction and even chemical precipitation are not relevant contaminant removal mechanisms in the concentration ranges relevant to natural waters [111,139]. In particular, for safe drinking water provision, physical methods (e.g., adsorption, filtration, ion exchange) are always mandatory to cope with the stringent regulations. As an example, water defluoridation by chemical precipitation yields an equilibrium fluoride concentration of about 8 mg L−<sup>1</sup> according to the solubility of CaF2 [57,58]. This value is far larger than the maximum permissible contamination level of 1.5 mg L−<sup>1</sup> for drinking water. Coming back to the Fe0/H2O system, available results show limited removal extent for fluoride while various removal extents for several contaminants depicting no redox reactivity in the systems has been documented. For this reason, a more rational approach is to consider that all species can be removed regardless of their redox potential, and identify exceptions on a case by case basis. In this regard, recent results by Hildebrant et al. [140] have confirmed that fluoride removal is low, but less reactive materials (EDTA test) comparatively remove more fluoride. This last observation corroborates the view that there is no single Fe<sup>0</sup> material for all situations and calls for more systematic investigations to identify appropriate Fe<sup>0</sup> materials for specific remediation applications.

The lack of systematic investigations has also affected the selection of Fe<sup>0</sup> materials used for pathogen removal [141–146]. Using various Fe<sup>0</sup> materials and very different experimental conditions, discrepant results have been reported [141,146]. Lu et al. [142] can be regarded as a perfect reflection of the state-of-the-art knowledge. The same authors investigated the mechanism of nitrate reduction and iron cycling by an iron-reducing bacteria strain (strain CC76) and metallic iron. They reported that the strain CC76 was able to utilize Fe2<sup>+</sup> (from iron corrosion) as electron donor for the nitrate removal. More importantly, they observed that Fe<sup>0</sup> inhibited the growth of strain CC76 in the early stage of the operation. This observation corresponds to CC76 removal by Fe0. However, the authors also reported that after the initial stage, strain CC76 was able "to tolerate" the presence of Fe0, meaning that no or less removal was achieved. This phase corresponds to a decrease in the kinetics of iron corrosion as discussed above and has been described in microbially-influenced corrosion [2]. If we recall the work of Kamolpornwijit and Liang [137] primarily investigating nitrate removal in an abiotic Fe0/H2O system, it then becomes apparently clear that the major source of discrepancy in the literature is the insufficient system analysis. Like with chemical contaminants, all pathogens will be removed in a well-designed system. For a filtration system, the key questions to address are: (i) which Fe0 material, in what amount, and for which contaminant and water? and (ii) with which filtration depth, for which flow velocity and for which operational duration? Another key aspect to consider is the nature of the aggregates (e.g., gravel, MnO2, pumice, sand) to be mixed with Fe<sup>0</sup> and the Fe<sup>0</sup> proportion in the mixture [107,114]. These aspects are yet to be addressed in the Fe0 remediation literature, but are critical for the design and field application of the technology. Table 3 presents a synthesis of current knowledge including the mistakes, and proposed refinements for future studies on Fe<sup>0</sup> remediation.


**Table 3.** Summary of synthesis of current knowledge and proposed refinements for future studies on Fe0 remediation.

#### *3.2. Reactivity and E*ffi*ciency*

In the Fe<sup>0</sup> remediation literature the terms "reactivity" and "efficiency" are mostly randomly interchanged. This has introduced confusion in the evaluation of independent results. In an effort to resolve this confusion the notion of "electron efficiency" has been introduced [147,148]. The electron efficiency (in %) is defined as the proportion of electrons from Fe0 oxidation that is used for the target reduction reaction, for example for nitrate reduction by Kamolpornwijit and Liang [137] (Section 3.1). In this approach, electrons used to reduce water (Eq. 1) or dissolved O2 are considered as "excess electrons" or avoidable electron wastage [148]. In other words, the main reaction is devalued to a side effect. Contrary to this still prevailing approach [92], the present work and related ones [149–153] recall that aqueous iron corrosion (Equation (1)) is not an unwanted reaction leading to an extra consumption of Fe0, but Equation (1) produces FeII species and H2, which are responsible for the documented chemical reduction and FeCPs which are responsible for contaminant removal and permeability loss [105–107].

Each Fe0 material is characterized by its intrinsic reactivity while each system is characterized by its efficiency for water treatment [154–156]. According to Miyajima and Noubactep [156], the efficiency is the expression of reactivity in a given system. This means that changing the Fe0 material modifies the efficiency. Conversely, an Fe0 material proven efficient in a system could be useless in another

system. In the quest of more efficient systems for water treatment, several Fe0 materials and groups of materials (e.g., bimetallics, iron nails, nano-Fe0, scrap iron, sponge iron, steel wool) have been tested and used for water treatment (Table 1). These efforts have been rendered difficult by a mistaken system analysis [157,158] and the evidence that most comparative works are based on testing materials for the removal of individual contaminants [159–162]. Lufingo et al. [11] recently discussed approaches that can be considered as universal as they are based on the stoichiometry of Equation (1): (i) H2 evolution [163,164] and (ii) Fe2<sup>+</sup> production [11,159]. They concluded that iron dissolution in a dilute solution of 1,10-Phenanthroline (2 mM) (Phen test) is the best available tool to characterize the intrinsic reactivity of Fe0 materials (Section 4).

#### *3.3. Misuse of Adsorption Isotherms and Kinetic Models*

Several studies investigating contaminant removal by Fe<sup>0</sup> materials have applied isotherm and kinetic models initially developed for materials where contaminant removal occurs via adsorption [131, 132,165,166]. A detailed discussion of the various adsorption isotherm and kinetic models, including assumptions and equations are presented in earlier reviews [132]. A comprehensive review of the limitations associated with the applications of such models to Fe0 systems is the subject of another paper, hence is beyond the scope of the current study. Briefly, the extension of such models to Fe0 materials is problematic for the following reasons: first, unlike adsorbents such as activated carbon whose surface area is readily available for adsorption at the beginning of the experiment (time t0 = 0; Figure 1), Fe0 systems are highly dynamic and equilibrium is rarely achieved in such systems (i.e., rust never rests), while the total reactive sites contributing to contaminant removal is not exactly known (Figure 1). Thus, most assumptions for such models are not valid for Fe0 materials. Second, as discussed earlier, contaminant removal in Fe<sup>0</sup> systems does not occur solely via adsorption. In view of this, the application of adsorption models to Fe<sup>0</sup> systems constitute another fundamental error in Fe0 literature. Yet despite a number of reviews highlight the mistakes, misuse and inconsistencies in the use of adsorption kinetic and isotherm models [53,54,67–71,131,132], the problem persists. Thus there is need to develop models for contaminant removal in Fe0 systems that takes into account the iron corrosion phenomena and the fundamental mechanisms responsible for contaminant removal as highlighted in the current review.

Reference [157] severely questioned the validity of specific rate constants (kSA) in Fe0/H2O systems. These rate constants are based on the stoichiometry of contaminant reduction by electrons from the metal body (Fe0) and are not considering the interactions of contaminants within the oxide scale [167]. As discussed in Section 2.2, the Fe<sup>0</sup> surface is not accessible to all contaminants. Even kobs values (which are normalized to the surface area to obtain kSA) currently used in the literature are erroneous as they are rooted on the wrong reaction. However, the most severe mistake has been to use the adsorption capacity for Fe<sup>0</sup> which is considered a reducing agent. Moreover, an adsorption capacity is determined in a system where the reactive agent is not completely depleted.

#### **4. Selection and Characterization of Fe0 Materials**

As discussed above, the scientific reason for using all Fe0 materials (e.g., granular iron and bimetallics, iron filings, iron nails, iron wire, nano-Fe0, scrap iron, steel wool) in water treatment is the electrode potential of the redox electrode FeII/Fe0: E0 = <sup>−</sup>0.44 V. Clearly, all reactive Fe<sup>0</sup> materials have the same redox potential. The intrinsic reactivity of individual Fe<sup>0</sup> specimens depends on a myriad of factors from which some are not readily accessible to the researcher. Relevant influencing factors include: alloying elements, Fe<sup>0</sup> form, manufacturing processes, metallography, Fe0 grain size, surface area, and surface oxidation state. Accordingly, each Fe<sup>0</sup> material has its own intrinsic reactivity which should be characterized in order to better understand how it is influenced by operational conditions to induce the intended remediation goal.

The long history of using Fe<sup>0</sup> for technical chemical applications, including water treatment reveals that there has always been efforts to select appropriate materials for individual applications. For example, the porous "spongy iron" (sponge iron or direct reduced iron) was proven more suitable in filtration systems than dense materials [12,13,168]. Similarly, multi-metallic systems and nanoscale materials were recently developed to address (recalcitrant) contaminants that were less sensitive to treatment with granular materials [169]. For completeness, it should be stated that there is no Fe<sup>0</sup> material for all situations such that one should rationally select appropriate materials for each specific application. For example, while treating water in fluidized systems (Anderson Process), dense materials were better than sponge iron [13,168]. Similarly, Hildebrant et al. [140] reported that Fe<sup>0</sup> materials of low reactivity according to the EDTA test exhibited a better efficiency for fluoride removal. The question arises how to select the right Fe<sup>0</sup> specimen for a given application? This calls for the development of standardized protocols for selection and characterization of Fe0 materials for various applications.

It is unfortunate that material selection has not received the due attention given its central role for the technology. Lufingo et al. [11] recently gave an overview of the available tools for the characterization of the intrinsic reactivity of Fe0 specimens. These authors based their work on an excellent review article by Li et al. [162] and insisted on the quantification of Fe2<sup>+</sup> from Equation (1). According to Lufingo et al. [11], quantifying Fe2<sup>+</sup> or H2 evolution from Equation (1) are the best tools to characterize the intrinsic reactivity. However, quantifying both Fe2<sup>+</sup> and H2 in natural systems suffers from the high reactivity of those primary corrosion products within the system, including their adsorption onto FeCPs and their action as own reducing agents. Clearly, measured amounts of Fe2<sup>+</sup> and H2 represent an excess quantity and cannot be strictly used to quantify iron corrosion (at pH values >4.5). These considerations clearly show that the H2 evolution method [163,164] is an approximation, while the EDTA method [170] is disturbed by dissolved O2. All other methods are contaminant-specific, and thus of low value [133,160,161]. Lufingo et al. [11] then proposed iron dissolution in a dilute (2 mM) 1,10-Phenanthroline solution (Phen test) as a facile method free from the inherent shortcomings of all available methods.

In terms of affordability and applicability, the Phen test is currently the best available method to characterize the intrinsic reactivity of Fe<sup>0</sup> specimens. The test lasts for less than 36 hours and characterizes the initial kinetics of Fe<sup>0</sup> dissolution. It is suggested as a candidate for a standard method and is immediately useful for material selection (screening) and quality control [11]. Each Fe0 specimen is characterized by its kPhen value which reflects its initial dissolution at the pH value of natural waters but without the interaction of the oxide scale. The kPhen values for nine steel wool specimens (Fe<sup>0</sup> SW) presented by Lufingo et al. [11] were such that 0.07 <sup>≤</sup>kPhen (μg h−1) <sup>≤</sup>1.30. This shows a ratio of reactivity of 18.5 for Fe<sup>0</sup> SW specimens which are often tested as a uniform class of materials compared to granular materials for example. Hildebrandt et al. [10,140] presented the kEDTA values for 13 reactive Fe0 SW and one granular Fe0 as 3.7 <sup>≤</sup> kEDTA (μg h−1) <sup>≤</sup>130.8. The lowest kEDTA value corresponds to granular Fe0, suggesting a reactivity ratio of up to 36 between Fe0 SW and granular Fe0. No data comparing kPhen values for Fe0 SW and granular Fe<sup>0</sup> are yet available. However, it is certain that the Phen test provides a confidence Fe<sup>0</sup> screening tool and is non-contaminant-specific. Therefore, if a significant body of kPhen data and data for characterized Fe<sup>0</sup> specimens for the removal of selected representative contaminants (or contaminant groups) are made available, then site-specific treatability studies would then be required only to fine-tune design criteria for the optimal performance of remediation Fe0/H2O systems.

Our research group has long recognized the need for systematic characterization of Fe0 materials in terms of their intrinsic reactivity and efficiency. The idea followed by this research group as summarized herein is to characterize the Fe<sup>0</sup> reactivity and the efficiency of Fe0/H2O systems in a pollutant-independent-manner. The alternative is to agree on probing pollutants while using standard protocols. A reference Fe<sup>0</sup> material would also be necessary. The use of EDTA and Phen tests to characterize the intrinsic reactivity has already been discussed. The efficiency of the Fe0/H2O system is characterized using the methylene blue (MB) discoloration method (MB method) [60]. Herein, the low affinity of MB for FeCPs is used to trace the abundance of in-situ generated iron oxides in comparatively long-term experiments [62,101,159,171]. The Phen test provides a reliable guidance in selecting Fe0 from the large catalog of available materials and to control the quality of newly manufactured ones. On the other hand, the MB method provides a reliable guidance for the characterization of the efficiency of Fe0/H2O systems. Selected probing agents enable the discussion of the results, and orange II, methyl orange and reactive red 120 were positively tested in this regard [61,62]. Recent results by Hildebrandt et al. [140] suggest that fluoride could be the next candidate for affordable probing reagent. In fact, it was found that fluoride removal is more efficient by low reactive Fe0 specimens.

#### **5. Future Perspectives and Potential Applications**

#### *5.1. Investigating the Fe0*/*H2O System*

The findings on which future research should be rooted must be based on the evidence that the oxide scale on iron is a diffusion barrier and shall never been disturbed in ways that are not reproduced under field situations [22,23,53,54,56,124,126,127,171–175]. This has been the motivation for adopting quiescent bath experiments as a more suitable approach to investigate processes occurring in Fe<sup>0</sup> permeable reactive barriers some two decades ago [20]. Then and now, most research groups consider the shaking intensity as a relevant operational parameter to be investigated almost in all instances without a quiescent system as a reference. However, there is no given convincing reason for the chosen mixing intensities tested in these experiments [124,125].

In 2005, Devlin and Allin [173] designed a glass-encased magnet batch reactor (GEM reactor) to investigate the impacts of selected anions on the efficiency of granular Fe0 in removing aqueous contaminants using 4-chloronitrobenzene as a probe molecule. This design aimed to achieve a better comparability between results of different experiments by fixing the stirring method and the stirring rates [169,173]. Using this approach, all experiments should be conducted in the GEM reactor to ensure that the granular iron remains stationary while the solution is stirred. Noubactep et al. [124,127] later demonstrated that, while using a rotary shaker, the shaking intensity should never be larger than 100 rpm. Given that a stirring device (including the GEM reactor) can be difficult to acquire by low-income laboratories, quiescent experiments can be adopted as a rule [171].

Quiescent batch experiments have also been adopted for the determination of the initial corrosion rate of Fe<sup>0</sup> in EDTA (kEDTA) and Phen (kPhen). Herein, the experiments are stopped before solution saturation, meaning that Fe<sup>0</sup> specimens are characterized under conditions where no oxide scale is available. Contrary to the prevalent approach [152–164], Fe<sup>0</sup> specimens are not characterized for any contaminant removal efficiency [176,177]. Solution saturation corresponds to [Fe] values equivalent to the stoichiometry of iron complexation (e.g.,[Fe] = 112 mg L−<sup>1</sup> for 2 mM EDTA). Lufingo et al. [11,120,121] recently compared the EDTA and the Phen test and established the superiority of the Phen test, which is additionally more affordable.

Another key feature of the remediation Fe0/H2O system is related to the oxide scale. The omnipresent oxide scale on Fe<sup>0</sup> is positively charged under natural conditions (pH > 5.0) [62,100,101,178]. Thus, the Fe0/H2O system is an ion-selective one and preferentially removes negatively charged species like bacteria [142,146]. One original idea has been to characterize the discoloration of methylene blue (MB) by Fe0/sand/H2O systems (MB method). The MB method is grounded on an historical work by Mitchell et al. [178], who observed that sand adsorbs less MB when it is coated with iron oxide. Thus, mixing the same mass of different Fe0 specimens with the given mass of a sand specimen, and allowing them to equilibrate for the long-time in a MB solution enable the differentiation of the Fe0 reactivity of various materials [62,101,176,177]. As a rule, the most reactive material produces the largest amount of iron hydroxides, which, in turn, in-situ coat sand and thus discolors MB the least. MB is thus not a model contaminant, but an operational tracer [171,177]. To ease the interpretation of results, methylene orange (MO) [61,176] or Orange II [62,101] can be used as anionic dyes since they have molecular sizes similar to that of MB. Phukan et al. [62,101,176]

additionally used reactive red 120 which is also an anionic dye, but is much larger in size that MO and Orange II.

One key advantage of the MB method is that its enables a visual observation of preferential flow within a reasonable time scale (some weeks) [62]. In designing an Fe0-based filtration systems, a decrease in hydraulic conductivity (permeability loss) is expected, and an early contaminant breakthrough will be observed due to preferential flow [179–183]. Mineral precipitation is particularly intense in the entrance zone of the filter [181], but the created preferential pathways are extended throughout the whole water column [179,181]. The investigation of the process of creation and extension of preferential pathways has been analytically very challenging [181–183]. On the basis of observations using the MB method [176,177], changes generated in the entrance zone can be better followed and considered in modelling efforts. This last aspect is crucial for the development of the technology as it has been convincingly demonstrated that models currently predicting the service life of Fe<sup>0</sup> filters are rooted on a wrong premise [114,115].

#### *5.2. Designing the Next Generation Filter*

The determination of the amount of a given Fe<sup>0</sup> which long-term corrosion kinetics would allow a designed Fe0/aggregate system to satisfactorily treat a given water for a certain time frame (e.g., 12 months) can be regarded as a routine work. This routine engineering application is however complicated by the complexity of the iron corrosion process and mistakes in past research as outlined herein. One major thinking mistake has been to consider that admixing Fe0 with sand (the most used aggregate) would alter the decontamination kinetics [90,183,184]. Thus, admixing with sand has been mainly considered as an economic tool to safe Fe<sup>0</sup> costs and the negative effects on the resulting system discussed [185]. It has been recently demonstrated that only hybrid Fe0 systems are sustainable [105–107].

Several hybrid systems have been successfully tested for water treatment including Fe0/activated carbon [184], Fe0/Fe3O4 [185], Fe0/MnO2 [186], Fe0/pyrite [99] and Fe0/sand [187]. While inert sand alone has clearly improved the efficiency of Fe0 systems [188], the efficiency of other tested aggregates was attributed to the specific materials. The Fe0/sand is regarded as in-situ sand coating [141]. It is certain that other materials will be similarly covered by FeCPs. Thus, it is questionable whether the systems really operate as described.

Rephrasing Notter [189], the success of an Fe0 filter depends on four main factors: (i) the quantity and quality of the water to be produced (e.g., daily), (ii) the intrinsic reactivity of used Fe0, (iii) the nature of the contaminant(s), and (iv) the availability of Fe0 material (to renew exhausted systems). This key principle remains unchanged, one century later. All is needed are systematic investigations take each Fe<sup>0</sup> and each water source as a stand-alone design parameter.

The laboratory procedures used to characterize and evaluate Fe0-based systems in batch and column experiments vary considerably among studies. For example, some studies use agitated batch experiments to investigate contaminant removal by Fe<sup>0</sup> [126,127], while others used batch experiments operated in quiescent mode, which is closer to field conditions [61]. This makes direct comparison of results among studies problematic, and could lead to misleading conclusions about the performance of Fe<sup>0</sup> materials. Therefore, there is need to develop standardized protocols for the evaluation of Fe0 material in both batch and column experiments. In the case of batch experiments, such protocols should include specifying the particle/grain size, tests for determination of material reactivity (e.g., EDTA, Phen tests), liquid/solid ratio, sampling frequency and duration of experiment. For column experiments, such protocols should specify filter depth, duration of experiment, grain size of filter material and chemical properties of test solution to be used, while accounting for potential interference among solutes. Besides developing dedicated pristine Fe0 materials for the water treatment industry, scope also exists to use Fe<sup>0</sup> material generated as wastes from other industries. Similarly, iron oxide-rich spent sludge from Fe0-based water treatment systems can be used as raw materials in other industries such as the production of pigments (e.g., iron oxide red) [190], and even as filter material in the construction

industry. This cyclic flow of iron materials between the water treatment industry and other industrial processes could form part of a circular economy. Filter wastes may also be regenerated and recycled to new Fe0 material, and then used in other industry when they contain toxic contaminants such as As or U. The recycling of filter wastes in other industries may require detailed environmental risk assessments, including the evaluation of contaminant leaching and potential ecotoxicological effects.

#### *5.3. Field Applications of Fe0-Based Systems*

Fe0-based systems present unprecedented opportunities for wastewater treatment and safe drinking water provision especially in low-income countries, including those in Africa (Table 4) [144,157,191–201]. In fact, the use of Fe0-based systems (e.g., the Bishof Process) for clean water provision has a long history dating back to the 19th century [13,14,202]. The history of Fe0-based drinking water treatment systems is discussed in detail in earlier review papers [17,18,122]. Recently, our research group has proposed the integration of Fe0-based systems in rainwater harvesting systems as a low-cost technology for decentralized drinking water provision, in what is known as the Kilimanjaro Concept [192–194].


**Table 4.** Potential field applications of Fe0-based systems in drinking water and wastewater treatment. The given reference refers to the oldest known application.

Fe0-based systems also have potential applications in domestic and industrial wastewater treatment systems (Table 4). For example, the Harza Process [27,203] has been used to remove Se from agricultural drainage water. Rahman et al. [200] and Fronczyk et al. [201] proposed the amendment of treatment media for runoff infiltration trenches/pits with granular Fe0. In addition, Fe0-based systems have been used to treat both domestic [204] and industrial wastewaters [205,206], including acid mine drainage [24,207]. Available studies suggest that Fe<sup>0</sup> can be added as filter media in constructed wetlands designed to treat urban stormwater and industrial wastewaters [199]. However, a lot remains to be done to further develop and disseminate Fe0-based technologies for wastewater treatment and decentralized safe water provision in developing countries, which such low-cost technologies are most needed. Considering that the bulk of studies are limited to laboratory scale applications, there is need to optimize the Fe0-based systems and evaluate them under field conditions.

#### **6. Summary and Conclusions**

The corrosion of iron in remediation Fe0/H2O systems is an electrochemical process, coupling Fe0 oxidative dissolution to the reduction of water (protons) and to no other available oxidizing agent, including dissolved O2. This is because the universal oxide scale on Fe0 acts as diffusion barrier to dissolved species and a conduction barrier to electrons from the metal body. In other words, water is the sole chemical which can remove electrons from the Fe<sup>0</sup> surface. Fe<sup>0</sup> oxidation and water reduction must not necessarily occur at the same locality. The spatial separation of oxidative (anodic) and reductive (cathodic) reactions is possible as the metal body allows the free flow of electrons from anodic to cathodic sites. The tendency of Fe0 to give off electrons (Equation (1)) is the same for all Fe0-based materials (E0 = –0.44 V). This makes material selection and characterization critical in designing sustainable Fe0/H2O systems.

The need to characterize Fe0 materials in terms of intrinsic reactivity and efficiency is critical for the design and operation Fe0/H2O systems, a subject that has been addressed by our research group. Specifically, EDTA and Phen tests were used to characterize the intrinsic reactivity of Fe0 materials. In this regard, the Phen test is considered an affordable and appropriate method that provides a reliable guidance in selecting Fe0 from the large catalog of available Fe<sup>0</sup> materials and to control the quality of newly manufactured ones. The efficiency of the Fe0/H2O system is characterized using the methylene blue (MB) discoloration method, while other probing agents investigated include orange II, methyl orange and reactive red 120.

The most characteristic issue of remediation Fe<sup>0</sup> is the small size (<5 mm) of used materials. Assuming uniform corrosion, the corrosion rates for progressive Fe0 oxidation should be normalized to the individual particles. In other words, expression like mmol year <sup>−</sup><sup>1</sup> should be expressed as mmol year−<sup>1</sup> particle−<sup>1</sup> or mmol year−<sup>1</sup> grain−1. The next important issue will be to consider the non-linear kinetics of the corrosion rate such that the service life of a designed system can be deduced knowing the size of used particles and the long-term corrosion rate. Once this is known, considering the expansive nature of iron corrosion would help to design sustainable systems. The choice of the admixing aggregates (e.g., gravel, MnO2, pumice, sand) and the mixing ratios are to be investigated on a case-by-case basis.

A better understanding of the long-term corrosion of relevant Fe0 materials under site-specific conditions is envisioned to ultimately aid in the design of affordable, applicable and efficient remediation Fe0/H2O systems. Applications of Fe0-based systems include; (i) a large variety of water treatment systems, (ii) household and small-scale water treatment plants, including rainwater harvesting systems for drinking water supply, (iii) decentralized domestic wastewater treatment, (iv) urban stormwater, agricultural and industrial wastewater treatment, and (v) as filter media in constructed wetlands. Addressing the key knowledge gaps highlighted here, and extending Fe0-based systems to other application domains such as wastewater treatment for agriculture are focal research areas in our group, which brings together collaborators from various countries.

**Author Contributions:** R.H., H.Y., R.T., X.C., M.X., B.K.A., V.C., M.L., N.P.S.-S., A.I.N.-T., N.G.-B., V.R.S.-T., W.G. and C.N. contributed equally to manuscript compilation and revisions. All authors have read and agreed to the published version of the manuscript.

**Funding:** The work is supported by the Ministry of Education of the People's Republic of China through the Program "Research on Mechanism of Groundwater Exploitation and Seawater Intrusion in Coastal Areas" (Project Code 20165037412).

**Acknowledgments:** We acknowledge support by the German Research Foundation and the Open Access Publication Funds of the Göttingen University.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Application of Ionizing Radiation in Wastewater Treatment: An Overview**

#### **Rehab O. Abdel Rahman 1,\* and Yung-Tse Hung <sup>2</sup>**


Received: 28 October 2019; Accepted: 18 December 2019; Published: 19 December 2019

**Abstract:** Technological applications of nuclear science and technology in different sectors have proved their reliabilities and sustainability over decades. These applications have supported various human civilization needs, ranging from power generation to industrial, medical, and environmental applications. Environmental applications of radiation sources are used to support decision making processes in many fields; including the detection and analysis of pollutant transport, water resources management, and treatment of municipal and industrial wastewaters. This work reviewed recent advances in the research and applications of ionizing radiation in treating different wastewater effluents. The main objective of the work is to highlight the role of ionizing radiation technology in the treatment of complex wastewater effluents generated from various human activities and to address its sustainability. Results of both laboratory and industrial scale applications of this treatment technology have been reviewed, and information on operational safety of industrial irradiators, which affect the sustainability of this technology, has been summarized.

**Keywords:** ionizing radiation; agricultural effluents; dye treatment; pharmaceutical effluents; disinfection

#### **1. Introduction**

Providing clean water and sanitation is one of the sustainable development goals that were proposed by the United Nation (UN). One of the problems that affect this goal is the reduction of freshwater quality. This reduction is attributed to the continuous increase in untreated wastewaters volumes and poor management practices, which led to the introduction of hazardous materials into freshwater sources [1]. Wastewater is defined, as indicated in UNEP/UN-Habitat, as a combination of one or more of the following effluents: domestic, commercial, industrial, horticultural, aquaculture, and storm water [2]. Recent advances in wastewater management helped in addressing some of the problems in water supply, pollution control, water recycling, and environmental protection. Now wastewaters are proposed as a resource, where many wastewater treatment plants are operated by the biogas generated from the anaerobic digestion of sludge, and the effluents from these plants could be used after appropriate treatment to meet the industrial, agricultural, and potable water requirements. Currently, wastewater management in developing countries is characterized by the discharge of large quantity of wastewater into surface water bodies without proper treatment. It is also challenged by the difficulties to sustain financing, operating and maintaining infrastructure for wastewater treatment.

Conventional wastewater treatment plants (WWTP) aims at reducing the contamination levels to acceptable limits required by the national regulatory agencies and at complying with international guidelines, which allow its safe discharge or reuse. Multi-stages of treatment processes are used, where pre-treatment stage is applied to remove coarse and large solids from the waste stream, using physical treatment technologies such as screens and grit chambers [3]. Primary treatment methods are then applied to remove suspended solids that could be settled using gravity sedimentation with

or without coagulation and flocculation [3,4]. The effluent from the primary treatment is directed to the secondary treatment stage to remove residual suspended solids and organic materials by using biological treatment processes [3]. The effluents from secondary stage contain some heavy metals, synthetic bio-refractory organic pollutants, and soluble microbial products derived during biological treatment [5,6]. The synthetic bio-refractory organic pollutants may include emerging micro-pollutants and disinfection by-products. In the tertiary stage of treatment, the effluents from the secondary stage are polished by removing persistence organic containments and heavy metals using advanced wastewater treatment technologies [3,4,7]. These technologies include filtration, sorption, gas stripping, ion-exchange, advanced oxidation processes (AOP), and distillation [7]. Finally, disinfection could be applied depending on the potential use of the treated effluents and the effluents characteristics [3].

Regulations on discharge/reuse indicators varied from country to another. In Denmark, Belgium, Spain, Germany, France, and Netherlands, the regulation covers heavy metals concentrations, total suspended solids (TSS), chemical oxygen demand (COD), 5 days biochemical oxygen demand (BOD5), total nitrogen (TN), total phosphorus (TP), and quantity of effluent discharge [8]. On the other hand, toxicity is only considered as reuse/discharge indicator in Germany, France, and United Kingdom [8]. Table 1 lists some international guidelines on the maximum limits for the reuse of effluents containing heavy metals, and U.S. regulation for their reuse and discharge [9–12]. Table 2 presents effluent discharge and reuse quality requirements in USA, Canada, and EU [11,13–15]. It should be noted that three classes of effluents are listed in the Canadian regulation based on the degree of treatment processes used in the treatment plants, namely, A, B and C, that refers to effluent from tertiary and disinfection treatment, tertiary treatment, and secondary treatment, respectively. For reuse regulations in EU, the limit of each class is determined based on the processes used in the treatment plants, irrigation methods, and the type of corps. Class A represents effluents from tertiary and disinfection treatment used any irrigation method and to produce all food corps. Classes B and C are effluent from secondary treatment and disinfection, which are used to produce food and processed food and non-food corps using all irrigation methods, and drip irrigation, respectively. Finally class D in EU regulation represents effluents from secondary treatment and disinfection, where all irrigation methods are allowed to produce industrial, energy, and seed corps [13]. Discharges from un-complied conventional WWTP can lead to the introduction of persistent chemicals and eco-toxic micro-pollutants into the aquatic systems [16]. The incomplete removal of these containments (even in 10−9–10−<sup>6</sup> g/L concentration range) was reported to induce potential long-term detrimental impact on the environment and the human health [17]. Recent research studies reported in the literature supports the application of advanced oxidation processes (AOP) for wastewater treatment to remove these contaminants [16,17].


**Table 1.** Guidelines of maximum limits for discharge and reuse of some inorganic pollutants in treated wastewater.

NA: Not Available.


**Table 2.** Effluent reuse and discharge limits in different regulations [11,13–15].

AOP use in-situ generated hydroxyl or sulfate radicals for organic pollutants degradation and heavy metals toxicity reduction [5,18,19]. AOP convert synthetic organic pollutants and soluble microbial products into simple biodegradable and harmless products, which lead to the reduction of COD and BOD in the treated effluents. AOP technologies employ various activation methods, where induced oxidation is achieved via exposure to photochemical or ionizing radiations, or chemicals [18,20,21]. Several studies were conducted to assess the feasibility of using ionizing radiation, in the form of gamma (γ) rays or electrons (e−), to remove persistence contaminants and to disinfect treated water and sludge. The results of these studies indicated that the ionizing radiation treatment is technically and economically promising. The efficiency of this technology was also proved in the reduction of persistence heavy metals under different treatment conditions. Fifteen pilot plants and several full scale irradiation treatment facilities that employ electron accelerator or γ irradiators were established with varying capacities. International Atomic Energy Agency (IAEA) reported that there is a need to continue research in this area to increase the general awareness of these processes in the environmental engineering community [22]. This awareness could be achieved by presenting integrated reviews on the advances in ionizing radiation applications in industrial effluents treatment from technical, operational safety, and economical aspects. Some review papers were published that reviewed the role of ionizing radiation in the degradation of azo dyes [23], summarized the results of the IAEA coordination research project [24], and have reported the optimized doses and procedures [25]. There is a lack in review papers that summarize recent advances in this field and that provide integrated insights into the role of this technology in eliminating hazardous biodegradation and disinfection products. In this work, the effort is directed to summarize the current understanding of the decomposition and removal mechanisms for organic and inorganic pollutants, respectively. Principles and advances in investigating the scientific basis of the applicability of this technique in the treatment and disinfection of agriculture, dyes, pharmaceutical, and petrochemical effluents will be presented. Operational safety and economical factors that affect the sustainability of this technology will be summarized. Finally, knowledge gaps will be identified and research areas that need to be addressed will be highlighted.

#### **2. Decomposition and Removal Mechanisms**

Generally any emitted radiation is characterized by its ability to deposit some of its energy in the surrounding media. This energy excites the media atoms by striping their electrons or break the chemical bonds between its molecules [26]. This particular characteristic leads to the application of ionizing radiation in treating wastewater, where the radiation imparts some of its energy in the radiolysis of water molecules. The effectiveness of the irradiation process is evaluated by calculating the radiation chemical yield (G-value, μmol·J −) that quantifies the number of formed species due to the absorption of 100 eV and is given by:

$$\mathbf{G} = \frac{6.023 \times 10^{23} \mathbf{C}}{\mathbf{D} \times 6.24 \times 10^{16}} \tag{1}$$

where C is the formed species concentration (mol/L) and D is the absorbed dose (Gy). Figure 1 illustrates the two stages radiolysis process for water molecules and its corresponding time scale, radiolysis products (primary intermediates), and their G-values. Primary intermediate (PI) then reacts with the pollutants (A) leading to the formation of secondary intermediates (SI) that are less persistence in a bi-molecular reaction to produce the degradation products (C + D), as follows [25,27,28]:

$$\text{A} + \text{PI} \underset{\rightleftharpoons}{\text{SI}} \text{SI} \rightarrow \text{C} + \text{D} \tag{2}$$

**Figure 1.** Two stages radiolysis process.

In actual wastewater, there is a competition between the pollutants and the anions in the solution, i.e., Cl<sup>−</sup>, CO3 <sup>2</sup><sup>−</sup>, HCO3 <sup>−</sup>, SO4 <sup>2</sup>−, for the reaction with the primary intermediates. This competition can affect the efficiency of the overall treatment process [27]. Gehringer studied the competition kinetics of two pollutants (A, B) in wastewater. The reaction rate was attributed to the bi-molecular rate constant (kA) between the solute concentration ([A] or [B]) and the primary intermediates concentration [OH] as follow:

$$-\frac{\text{dA}}{\text{dt}} = \text{k}\_{\text{A}}[\text{A}][\text{OH}] \tag{3}$$

Table 3 lists the bi-molecular reaction rate constants for some pollutants [22,25,27,29]. The amount of the specific radicals available for interaction with certain solute could be calculated using the reaction probability (PA) according to the following equation:

$$P\_{\mathcal{A}} = \frac{\mathbf{k}\_{\mathcal{A}}[\mathbf{A}]}{\mathbf{k}\_{\mathcal{A}}[\mathbf{A}] + \mathbf{k}\_{\mathcal{B}}[\mathbf{B}]} \tag{4}$$

Figure 2 illustrates the reactions between the primary intermediates and the available anions in the wastewater effluents. Wojnarovits and Takacs [30] indicated that intermediate reaction with chloride ions is dominant at pH < 5, whereas the reaction with carbonate and bicarbonate ions are dominant at neutral or slightly alkaline pH. The produced free radicals will subsequently react with dissolved organic matters in the wastewater via direct electron transfer (outer sphere electron), addition to double bonds (inner sphere electron) or abstraction of H-atoms from C-H bonds. The last reaction takes place between oxidizing radicals and saturated molecules. It should be noted that in some cases, it is very hard to determine if the reaction is electron transfer or radical addition/elimination.


**Table 3.** Bimolecular reaction rate constants for some pollutants [22,25,27,29].

**Figure 2.** Schematic diagram of intermediate reaction with free radicals.

Ionizing radiation can reduce various forms of mercury via reaction with e−aq (Equation (5)) and H (Equation (6)) to form unstable compound that dimerized (Equation (7)) to produce insoluble form [31,32].

$$\text{HgCl}\_2 + \text{e}\_{\text{aq}}^- \rightarrow \text{HgCl} + \text{Cl}^- \tag{5}$$

$$\rm H\rm HgCl\_2 + H \to H\rm gCl + Cl^- + H^+ \tag{6}$$

$$2\text{HgCl}\_2 \rightarrow \text{Hg}\_2\text{Cl}\_2\tag{7}$$

The presence of hydroxyl radicals inhibited the production of insoluble mercury (Hg2Cl2) through re-oxidation of HgCl [31]. To enhance the removal performance of this technology the use of organic radical to act as hydroxyl scavenger was proposed. For example, ethanol may be used (Equations (8) and (9)) with the following reactions [9]:

$$\text{CH}\_3\text{CH}\_2\text{OH} + \text{OH} \rightarrow \text{CH}\_3\text{CHOH} + \text{H}\_2\text{O} \tag{8}$$

$$\mathrm{HgCl}\_2 + \mathrm{CH\_3CHOH} \to \mathrm{HgCl} + \mathrm{Cl^-} + \mathrm{CH\_3CHO} + \mathrm{H^+} \tag{9}$$

Successive reduction of Cr species from (VI) state to (III) state is achieved via reaction with H [32]. On the other hand, cadmium (II) and lead (II) are reduced via reaction with H or/and e−aq (Equation (10)) to produce Cd(I) that undergoes disprotonation (Equation (11)) or oxidization through a reaction with OH<sup>−</sup> or H2O2 (Equation (12)), or react with hydrogen to produce unstable MH<sup>+</sup> species that decay (Equation (13)) [18].

$$\text{M}^{2+} + \text{e}\_{\text{aq}}^{-} \rightarrow \text{M}^{+} \tag{10}$$

$$\text{CH}^+ \rightarrow \text{M} + \text{M}^{2+} \tag{11}$$

$$\text{M}^+ + \text{OH} \rightarrow \text{M}^{2+} + \text{OH}^- \tag{12}$$

$$2\text{ H} + 2\text{M}^+ \to 2\text{MH}^+ \to \text{H}\_2 + \text{M}^{2+} + \text{M} \tag{13}$$

To reduce the oxidation effect and subsequently to enhance cadmium and lead precipitation, an organic OH− scavenger could be used or the process could be operated in the absence of oxygen [31–33].

#### **3. Advances in Treating Agricultural Wastewaters**

Despite chemical pesticides, herbicides, and fungicides are applied according to national agricultural guidelines to enhance the agricultural production efficiency. Residues of these persistence pollutants and their toxic byproducts exist in agricultural wastewater, and could migrate to surface-and ground-waters. Ionizing radiation treatment technology proved its effectiveness in decomposing a varied number of these pollutants using tertiary treatment method. Most of the researches conducted in this area focused on optimizing the irradiation conditions, investigating the effect of other waste components on the degradation process, and the possibility of using combined treatment technologies to enhance the overall efficiency of the treatment process. Table 4 summarizes the performance of ionizing radiation in removing some pollutants in agricultural wastewaters.


**Table 4.** Performance of agricultural wastewater pollutants degradation using irradiation.

NA: Not Available.

The optimal irradiation conditions for six commercial pesticides, i.e., diazinon, dimethoate, procloraz, metiocarb, imidacloprid, and carbofuran, using electron-beam facility, were determined for different wastewater compositions and at different operational conditions [36]. The initial pesticide concentrations (Ci) varied in the range 40 < Ci < 400 ppm, the initial pH in the range (4.5 < pH < 8) and the irradiation doses were varied from 2–10 KGy. The application of 5 KGy was found to be the most efficient for the treatment of procloraz, dimethoate, imidacloprid, and carbofuran (99% degradation), but not efficient for the treatment of metiocarb. Low irradiation doses (<1 KGy) led to the formation of hydroxylated intermediates that build-up in the solution with the progress of the irradiation process. By increasing the irradiation period, a complete removal of the pollutants is achieved as a result of hydroxylation of the accumulated intermediates. The effect of the combined process of aeration and irradiation treatment was evaluated. It was found that the use of aeration can improve the metiocarb removal by 18%. In general, the hydroxylation of the aromatic ring to produce hydroxyl-cyclohexadienyl radicals is reversible [37]. Using aeration to inhibit the formation of these reversible radicals by peroxidation will lead to more efficient degradation [37]. The optimum irradiation conditions for the degradation of three herbicide and one fungicide, i.e., 2,4-dichlorophenoxyacetic acid (2,4-D), 3,6-dichloro-2-methoxy-benzoic acid (dicamba), 4-chloro-2-methylphenoxyacetic acid (MCPA), and carbendazim were determined. It was found that complete radiolytic degradation of dicamba (Ci = 110 ppm) is achieved using an irradiation dose of 5 KGy and the process is insensitive to the presence of other waste constituents, i.e., NO3 −. 2,4-D-degradation was found to be affected by the presence of other waste constituents [35]. Another study addressed the use of ionizing radiation in the treatment of 2,4-D, and MCPA at lower concentration levels (<50 ppm), in the presence Cl−, Br− and NO3 −. The results confirmed the role of oxidative radicals (OH−) in the degradation of these pollutants and the sensitivity of the irradiation process to the presence of the anions, where the secondary intermediate formed at low doses (<1 KGy) are more toxic than the original pollutants [38,39].

The feasibility of using ionizing radiation in the treatment of agricultural effluents containing chlorinated organic pesticide, i.e., (4-chloro phenoxyacetic acid (4-CPA), 2.4-dichlorophenoxyacetic acid (2,4-D), 2.4-dichlorophenoxyacetic propionic acid (2,4-DP), and 2.4-dichlorophenoxyacetic butanoic acid (2,4-DB) was studied. The results indicated good efficiency of the irradiation process at 1 KGy in decomposing these pollutants. It was observed that chlorine was released as a result of their degradation. To reduce the disinfection byproducts in the treated wastewater prior to its use in chicken and fish livestock, an irradiation of the wastewater stream at 16.2 KGy dose was proposed, and 27 KGy was proposed for sludge treatment [40].

Swine wastewaters, which are alkaline agricultural wastewaters, that contain different pollutants, i.e., carbohydrates, proteins, lipids nitrates, nitrite, phosphate, and ammonia, were treated using a combined process of irradiation and ion-exchange biological treatment method. The use of electron-beam at 75 KGy achieved 85.1% removal efficiency of chemical oxygen demand (COD) at organic loading rate of 1.41 kg/m3.day and achieved 75% removal efficiency of total nitrogen. The nitrogen removal was found to be sensitive to variation in current density [41].

#### **4. Advance in Dyes Treatments**

The presence of organic dyes in industrial effluents can lead to serious health and environmental problems. Ionizing radiation treatment was proposed to remove these contaminants. Research efforts were directed to study model compounds, single polluted solution, simulated wastewater, and real effluents. Table 5 summarizes the kinetic reactions of model compounds with their primary intermediates, their corresponding rate coefficients and their secondary intermediates [23]. Most of these compounds, except Azobenzene, decompose as a result of hydroxylation reaction, where the rate coefficients are slightly varied depending on the molecular structure of the contaminant. On the other hand, reactions with e− and H have small contribution to the decomposition process [23]. G value for phenol degradation was found to be inversely proportional to the irradiation dose. Degradation reaction is first order and favors neutral pH. Phenol degradation is enhanced with the addition of

oxidants, i.e., O3, or S2O8 <sup>2</sup>−. S2O8 <sup>2</sup><sup>−</sup> was more efficient due to the selectivity of SO4 <sup>2</sup><sup>−</sup> radical to the formed by-products (carboxylic acids) [42].


**Table 5.** Degradationkinetics of model compounds [23].

65% de-coloration of aqueous solution containing Alizarin Yellow GG (AY-GG, Ci = 100 ppm) was achieved using 9 KGy dose. Irradiation pre-biodegradation was proposed to improve the efficiency of treatment by 30% due to the formation of heterocyclic aromatic amines and cyanides. Increasing the irradiation dose can enhance the biodegradation due to the elimination of toxic secondary intermediates [43]. The aqueous solutions of Reactive Blue 15 (RB15) and Reactive Black 5 (RB5) dyes were irradiated with doses 0.1–15 KGy and at 2.87 and 0.14 KGy/h dose rates. Complete de-coloration was observed at 1 and 15 KGy doses for RB5 and RB15, respectively [25]. The de-coloration mechanisms for Apollofix Red (AR) and RB5 in aqueous solutions were investigated using γ-rays irradiation [44]. It was found that the de-coloration that was due to reactions with e− and H, and increased linearly with increasing the irradiation dose, whereas de-coloration that was due to reaction with OH− radical increase logarithmically. De-coloration was attributed to the destruction of the color bearing part of the dye as a consequence of OH− addition to the aromatic ring [25]. Aqueous solutions containing AR and Apollofix Yellow (AY) were irradiated with doses of 1.0–8.0 KGy and at 0.14 KGy/h dose rate. The complete de-coloration was observed using 3.0 and 1.0 KGy doses for AR and AY, respectively [25].

The use of ionizing radiation to decompose wastewater effluents of varying pH (1.6 < pH < 11.5) and COD concentration (650 < Ci < 2210 ppm) from dye manufacturing factory was evaluated [45]. The use of 15 KGy decreased the optical density by 95% and removed COD by 72%. Use of aeration and H2O2 enhanced the degradation. The treatment of simulated wastewater contains four dyes, i.e., direct and reactive azo and anthraquinone dyes, using electron beam and γ radiation at different dyes concentration and irradiation doses in the presence and absence of H2O2 was studied [46]. The highest de-coloration performance (≈100%) was obtained via electron beam irradiation using 7 KGy for direct and reactive azo dye (Ci = 1000 ppm). A combined process of electron-beam and biological treatment was used in a pilot plant study in treating two waste streams. The first stream was generated during the operation of dying process and the second stream was from polyester fiber production enriched with ethylene glycol and terephthalic acid in 1998. An industrial facility was later commissioned and operated based on the same technology in 2005 [24,25]. The use of electron-beam technology to de-color and detoxify three effluents that represent chemical, final, and standard textile effluents was conducted. The effluents were de-colored (96, 55, and 90%) using 40, 2.5, and 2.5 KGy, respectively. The use of irradiation pre-biological treatment was found to reduce the toxic effect to the subsequent biological treatment process [47]. The potential use of combined treatment for alkaline and

nearly neutral textile effluent of COD (632 < COD < 127 ppm), BOD (311 < BOD < 490 ppm), and turbidity (75–77%) at irradiation doses (<3 KGy) was investigated. It was found that the application of coagulation prior to irradiation enhanced the de-coloration performance [48]. Electron beam irradiation of simulated effluents that contains reactive dye (reactive yellow 15), size (starch), synthetic size (PVA), alkali, color, and pigment (pigment red 139) was studied [49]. Application of 1 KGy irradiation for post-biodegradation treatment enhanced the quality of the treated effluents.

#### **5. Advances in Wastewater and Sludge Disinfection**

Disinfection of treated wastewaters using ionizing radiation has been studied extensively. The application of 4 KGy reduced the BOD5, COD, and total organic carbon (TOC) to acceptable limits [50]. In another study, anaerobic digested sludge irradiation at 1 KGy eliminated 98% of the total and fecal coliform, whereas BOD5 was reduced to an acceptable limit at 4 KGy and COD was not affected by irradiation up to 20 KGy [44]. The disinfection of acidic industrial effluents and sludge (2 < pH < 3) of high concentration of BOD5 and COD (7093, and 32,664 ppm, respectively) using γ irradiation was investigated [17]. Sludge and treated water disinfection could be achieved using 7 and 4 KGy, respectively, whereas the reduction of the BOD5 and COD concentration to acceptable limits was obtained at 18 KGy. A comparative study was conducted to evaluate the use of different advanced oxidation methods in the disinfection of the municipal wastewater, re-growth control, and the associated operating costs [51]. The results revealed that UV efficiency was affected by the seasonal variation in the wastewater composition, whereas ionizing radiation efficiency was respectively unaffected by this factor. Ionizing radiation provided high stable disinfection efficiency (95%) for the total colony count and total coliform at radiation doses >0.25 KGy and inhibited the re-growth. From economical point of view, the electric power consumption for UV and ozone is three orders of magnitude higher than that required for ionizing radiation. Electron beam irradiation used to disinfect a municipal wastewater at irradiation doses (<3 KGy) and removed 90% of coliforms [51].

#### **6. Advances in Pharmaceutical and Petrochemical Wastewater Treatments**

The possibility of treating pharmaceutical effluents was addressed by studying the effect of ionizing radiation on the biodegradability and toxicity of individual drugs. Changes in biodegradability and toxicity induced in aqueous solutions containing sulfamethoxazole (SMX, Ci = 0.1 mmol/L) using ionizing radiation treatment revealed that SMX biodegradability was improved by applying 0.4 KGy dose. At 2.5 KGy dose, SMX conversion to biologically treatable substances was noted [52]. The degradation of carbamazepine (CBZ) by ionizing radiation was enhanced by the application of oxidant, 10 mM H2O2 [20]. The decomposition of mutagenic and carcinogenic secondary intermediates, i.e., acridine (ACIN), was enhanced in the presence of H2O2. The ionization of aqueous solutions containing ciprofloxacin (CIR) and norfloxacin (NOR) (Ci = 10−<sup>4</sup> M) was investigated. The degradation reaction proceeds via OH<sup>−</sup> and e<sup>−</sup> reactions with comparable rate constant (≈10<sup>9</sup> mol−1·L·s<sup>−</sup>1). At low irradiation doses, the antibacterial activities of the secondary intermediates vanished. Pollutants hydroxylation during γ irradiations proceeds on the hydroxylated molecules and desethylene derivatives and during pulse radiolysis is attributed to absorbance of hydroxyl-cyclohexadienyl radicals. In hydrated electron reactions, electron adduct is formed then it underwent protonation yielding cyclohexadienyl type radical [53]. The treatment of real pharmaceutical effluent using combined process of coagulation, biological treatment, and γ irradiation was investigated [54]. Two tested neutral effluents, namely, low organic strength (LSW; BOD < 6730 ppm, COD < 12,715 ppm) and high organic strength (HSW; BOD < 27,242, COD < 51,223 ppm). The use of irradiation led to maximum reduction in COD of 45% in acidic media at 50 KGy (LSW) and 30% in acidic media at 100 KGy (HSW). The application of coagulation pre-treatment was found to affect the efficiency, where 55% and 50% could be achieved using 100 KGy, for LSW and HSW, respectively. The use of H2O2 led to enhanced COD and TOC removal efficiency, when compared to S2O8 <sup>2</sup>−. The combined treatment led to overall 92.7% <sup>±</sup> 2.3% and 90.2% ± 2.9% removal of COD from LSW and HSW, respectively. In a separate study, combined

process of coagulation, electron beam treatment, and biological treatment was performed. An overall reduction in COD of 94% and 89% was achieved LSW and HSW, respectively [55]. The radiation doses were varied from 25–100 KGy at different pH that represent acidic, neutral and alkaline media. The slightly improved performance of the electron irradiation process was related to the reaction of e−aq with H through parallel reaction with the organic contaminants to generate H+, which subsequently inhabits the recombination of e−aq and OH−. The consideration of the total cost of electron beam irradiation facility (20 MeV, 100 kW), its shielding and maintenance, and capital costs was estimated and the cost of the treatment was estimated to be 0.6 USD/m<sup>−</sup>3.

The decomposition of naphthalene (Ci = 5–32 ppm) in aqueous solution was studied using γ irradiation combined with both H2O2 and TiO2nano-particles. The application of 3 KGy dose led to high naphthalene removal performance (>98%) and TOC reduction (28–31%) due to hydroxylation reaction. This performance is enhanced by 35% due to the presence of 40 ppm of H2O2 and 48% due to the presence of 0.8 g/L TiO2 [56].

#### **7. Sustainability of the Technology**

In general, the sustainability of the nuclear industry is governed by its technical competitive performance, economical feasibility, and safe operational practice. As presented in the previous sections, the technical performance of the application of ionizing radiation technology in wastewater treatment is effective in disinfection and reduces the bio-refractory nature of several persistence organic pollutants. Several studies proved the economical feasibility of the e-beam and γ irradiation technologies [21,32,41,48,51,54–63]. For e-beam technology, the capital costs include the accelerator price, building shield, conveyer, cooling and ventilation systems, and monitoring system [57]. The capital cost of the e-beam accelerator is dependent on the power (P, kW) consumed to produce optimum dose (D, KGy) for specified plant flow rate (Q, m3/h) taking into account the utilization factor (ϕ) [48,60]:

$$\mathbf{p} = \frac{\mathbf{D}\mathbf{Q}}{\mathbf{q}}\tag{14}$$

The cost of the accelerator installation (Ki, k\$) is determined using the applied electron energy (E, MeV), power, accelerator type and manufacturer (b, d), and installation coefficient (a) as follows [48,57]:

$$\mathbf{K}\_{\mathbf{i}} = \mathbf{a} \cdot \mathbf{b} (1 \pm \mathbf{d}) \mathbf{E} \sqrt{\mathbf{p}} \tag{15}$$

The values of the coefficients b and d varied with time due to the evolution and advances in the manufacturing process [48]. On the other hand, the costs of γ irradiator are not accompanied by power consumption [54], where γ source irradiate spontaneously. The average cost (ATC, k\$/m3) of this irradiation technique is determined based on the price of the used radioactive source (R, k\$/Ci), required activity (I, Ci), irradiation time (t, h) and half life (t0.5, h), and the irradiation chamber volume (v, m3)

$$\mathbf{A} \left( \mathbf{C}\_{i} \right) = \mathbf{R} \times \mathbf{I} \times \mathbf{t} / \left( \mathbf{t} \mathbf{0}.5 \times \mathbf{v} \right) \tag{16}$$

To facilitate the cost comparison between the use of e-beam technology and other treatment technologies the relative treatment costs estimated at different time is presented in Figure 3 [48,64]. In terms of cost, only disinfection process, which uses chlorine, is better than the use of e-beam technology. Similar data are not available for γ irradiation. To compare the cost of this technique alone and cost of this technology combined with coagulation, the data reported by Changotra et al. [54] were plotted for two types of industrial effluents (Figure 4).

**Figure 3.** Relative treatment costs for wastewater using e-beam and other disinfection and conventional treatment methods [48,64].

**Figure 4.** Treatment costs for γ irradiation and combine coagulation-γ irradiation [54].

There are some technical and non-technical issues that need to be resolved to ensure the sustainability of using this technology. These issues are listed below [22]


From safety point of view, radiological accidents associated with different application of industrial irradiator, either γ irradiators or accelerators are limited, and could be classified as level 4 on the international nuclear and radiological event scale (INES). This level is used to describe accident with local consequence; Figure 5 summarizes the radiological exposures in these historical accidents [65]. Except the 1967 accelerator accident, one worker was exposed to radiation in each accident with a

probability 0.092 and 0.12 a−<sup>1</sup> for accelerators and γ irradiators, respectively. It should be noted that due to the more stringent regulatory safety requirements that were issued later, similar accidents were not reported over three decades ago. These regulatory requirements are not only related to safety aspects but also to the security aspects to ensure the application of 3S concept (Safety, Security, and Safeguards) [66,67]. To reduce these accidents, and based on the national regulation, the following measures should be considered [68–70]:


**Figure 5.** Summary of the historical industrial irradiator accidents; (**a**) accelerators, (**b**) γ irradiator.

Currently, the control strategy in these facilities is based on the combination of physical protection means, i.e., shields, barriers, and interlocks, and operational procedures [65,71]. This strategy ensured an accident probability for industrial irradiator in the order of 2 <sup>×</sup> <sup>10</sup>−<sup>4</sup> <sup>a</sup><sup>−</sup>1.

#### **8. Conclusions**

The review of ionizing radiation technology in decomposing bio-refractory organic contaminants and disinfecting different wastewater effluents were presented in this study. Factors that affect the sustainability of this technology were summarized. From this review the following conclusions could be drawn.


**Author Contributions:** Conceptualization, R.O.A.R. and Y.-T.H.; methodology, R.O.A.R.; resources, R.O.A.R. and Y.-T.H.; writing—original draft preparation, R.O.A.R.; writing—review and editing, R.O.A.R.; supervision, Y.-T.H.; funding acquisition, Y.-T.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Water* Editorial Office E-mail: water@mdpi.com www.mdpi.com/journal/water

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18