Abstract

This study aims to assess the efficiency of two natural-based coagulants, namely calcium lactate and tannic acid, and compare them with conventional coagulants, including polyaluminium chloride (PACl) and ferric chloride. Jar test experiments were performed on the raw inlet water of the Isfahan water treatment plant (IWTP) in Iran. Response surface methodology was implemented to design and optimize the experiments. The factors considered in the design were coagulant dose, pH, initial turbidity, and temperature. Results showed the acceptable efficiency of natural coagulants in turbidity reduction, so that they meet the potable standard levels. The final water turbidity in the optimum condition for calcium lactate, tannic acid, PACl, and ferric chloride were 0.58, 0.63, 0.56, and 0.76 NTU, respectively. The comparison between the performances of the coagulants showed no significant difference in turbidity removal. However, the sludge volume produced as well as the impact on pH alteration after coagulation–flocculation were lower when using natural coagulants than with conventional coagulants. Also, the residual aluminum for PACl measured was higher than the desired limit according to Iran's drinking water standard. Finally, the simple additive weighting method was used to rank the four coagulants based on the selected criteria. The results showed that the natural coagulants could be preferable to the conventional coagulants if the concerns regarding disinfection by-product formation due to their residual organics were resolved. Since this issue was fixed in the IWTP due to the ozonation process, calcium lactate was proposed as an efficient alternative to PACl.

INTRODUCTION

Among the natural pollutants in surface waters, there are colloidal impurities. These materials cause turbidity and color in water; they can also be shelters for pathogens against disinfectants. Excessive turbidity is a serious challenge for water treatment. In water treatment, coagulation–flocculation is the process used for the removal of suspended colloids (Hayder & Ab Rahim 2015).

Generally, in water treatment plants, inorganic coagulants, such as aluminum sulfate (alum), ferric chloride, and synthetic organic polymers (polyaluminium chloride (PACl)), are common coagulants. However, water treatment using metal salts results in high sludge volumes, which leads to disposal problems, as well as heavy metal accumulation in the environment (Divakaran & Pillai 2002). Furthermore, some studies have indicated that the residual aluminum in water from alum and PACl applications may induce Alzheimer's disease (Flaten 2001; Muthuraman & Sasikala 2014). Additionally, iron salts can alter the water pH, which necessitates the addition of alkaline chemicals, such as lime, leading to increased sludge volume. Moreover, metal salts are less effective in water temperatures below 10 °C, because the low temperature affects the coagulation kinetics and exerts a significantly negative effect on the floc aggregation rate (Lou et al. 2012).

The above-mentioned concerns and problems in applying mineral coagulants have drawn more attention to natural organic coagulants. These coagulants could be a preferable alternative method for the removal of turbidity from drinking water. They have the advantages of human health safety due to not forming any residual of heavy metals, biodegradability, reduced dose requirements, and less sludge production (Kumar et al. 2017). However, natural flocculants have the risk of residual organic matter which can serve as a precursor matter for disinfection by-products (DBPs).

Rizzo et al. (2008) compared the efficiency of chitosan and conventional coagulants (alum and ferric chloride) in terms of turbidity and natural organic matter removal. The removal efficiencies of alum, ferric chloride, and chitosan for turbidity were 81%, 90%, and 73%, and for total organic carbon (TOC) were 41%, 50%, and 4%, respectively. Also, they assessed the toxicity of the chlorinated coagulated waters in terms of Daphnia magna immobilization percentage which showed more toxicity for the chitosan sample. Finally, the authors concluded that the use of chitosan in drinking water treatment plants may result in more toxic finished water compared to conventional coagulants depending on surface water characteristics and chemicals.

Yarahmadi et al. (2016) compared the turbidity removal efficiency of Descurainia sophia seed extract as a natural coagulant with ferric chloride. The results showed an efficiency of 90% for ferric chloride compared with 43% for Descurainia sophia. The TOC of the treated water using the seed extract showed an increased concentration of 1.0 mg/L. These results were in accordance with the findings of Bina et al. (2009) that studied the application of chitosan as a coagulant and reported an increase of 0.8 mg/L in TOC. Both studies concluded that due to the low increase of TOC, the risk of DBP formation was low.

Camacho et al. (2017) evaluated the impact of Moringa oleifera (MO) seeds as a coagulant in turbidity removal from surface waters. The authors reported that turbidity removal was up to 85% for high turbidity and 60% for low turbidity waters, with a dose of 50 mg/L. It was also demonstrated that MO was capable of removing the aromatic organic matter, up to the range of 40–50%. However, dissolved organic matter results showed an increase of one to three times, depending on the raw water turbidity and the dose of MO.

In order to remove turbidity from water, Choy et al. (2016) tested different types of starches and concluded that rice starch could achieve a turbidity removal of 50% at pH 4 and a dose of 120 mg/L. They also showed that adding PACl with a 0.6 mg/L dose could enhance the removal efficiency by 80%. Oladoja et al. (2017) tested the fruit seed extract as a natural coagulant. The optimum coagulant dosage of 10 mL/L at pH 6 gave a turbidity removal efficiency of above 90%. The produced sludge showed an improved settling characteristic with the sludge volume index of 31 mg/g.

Two such natural coagulants are calcium lactate and tannic acid. Calcium lactate is a biodegradable substance that can be obtained from agricultural waste by the fermentation of the substrates with lactic acid bacteria. It is a completely harmless coagulant, as it is commonly used as a food additive. It is also added to sugar-free foods to prevent tooth decay. Moreover, as a medicine, it is commonly used as an antacid and for calcium deficiency treatment (Devesa-Rey et al. 2011).

Tannic acid is a specific form of tannin, a type of polyphenol found in many plants. It refers to tannins that can be dissolved in water, as opposed to other types of condensed tannins. Tannins are found in foods and beverages, including fruit, tea, and wine, as well as edible grasses, such as corn. Tannins appear in drinking water sources usually from harmless natural sources like leaves and tree bark (Mcallister 2016). Tannic acid is generally safe to consume in small amounts, and according to the National Institutes of Health, tannic acid has an antioxidant effect on the human body, which can protect body cells from mutagens and carcinogens (Kerkar 2017; Chung et al. 1998). It has other health benefits due to its antioxidant and antimicrobial properties. Moreover, it is used for several medical purposes, such as in accelerating blood clotting and reducing blood pressure. Though tannic acid may generally be harmless, it can be dangerous when taken in large amounts (Kerkar 2017; Chung et al. 1998).

In the recent literature, there are some studies in which the researchers have used natural coagulants for water treatment and turbidity removal. Sánchez-Martín et al. (2010) tested a tannin-based coagulant agent called Silvafloc for river water clarification. The authors obtained a 90% turbidity removal in neutral pH with a 20 mg/L dose of the coagulant. The coagulant also decreased total coliforms up to 70%. The treated water presented very low polyphenols content (about 0.4 mg/L). Furthermore, organic matter content was not increased but removed by about 30%. It was observed that total organic matter increased linearly after coagulant addition, but no further increase was observed after the flocculation process, and all organic matter added by the coagulant was removed.

In another study, tannic acid was used by Paydari (2011) for turbidity removal. The author achieved 72% turbidity removal. Using the Taguchi method (Aber et al. 2010) for the experimental design, pH was found as the most effective factor in the removal efficiency. Moreover, Thakur & Choubey (2014) achieved a 91% turbidity removal at the optimal dosage of 3.0 mL/L of a tannin-based coagulant. A commercially produced tannin-based coagulant called Tanfloc was used by Hameed et al. (2016) for municipal wastewater treatment and achieved a removal efficiency of 90% for turbidity and 60% for the ratio of biochemical oxygen demand to chemical oxygen demand (BOD/COD). Grenda et al. (2018) applied the tannin-based coagulant with a cationic polyacrylamide (PAM) flocculant for turbidity and color removal for industrial wastewater. The results showed that the turbidity removal could increase from 68% (while using 200 ppm of bio-coagulant alone) to 93% (by adding 5 ppm of cationic PAM with 40% charge). For color removal, the results of similar tests showed an increase from 82% to 89%.

Calcium lactate has also been found as a promising coagulant for turbidity removal. In a study by Devesa-Rey et al. (2011), the authors optimized the operational conditions for efficient turbidity removal using the response surface methodology (RSM) (Montgomery 2012). The factors of calcium lactate dose, pH, and initial sediment concentration were considered in the experimental design. In another study, Devesa-Rey et al. (2012) compared four natural, unconventional coagulants, namely lactic acid, calcium lactate, sodium lactate, and citric acid, with AlCl3. The results showed that the percentage of turbidity reductions of AlCl3 and lactic acid were 95.4% and 97.1%, whereas calcium lactate and sodium lactate resulted in 88.3% and 77.3% of NTU reduction, respectively. The authors recommended calcium lactate as an alternative to aluminum salts since the risk of the overdosage of aluminum in potable water is a concern. They did not report any analyses regarding residual organics of coagulants and the risk of DBP formation.

Calcium lactate also showed the acceptable efficiency as a coagulant for the treatment of different industrial pollutants, such as lignin from agro-based industrial wastewater (Zahrim et al. 2015; Joy et al. 2018) and melanoidin from a palm oil mill (Azreen et al. 2017).

The literature review shows that tannic acid and calcium lactate have been tested individually for turbidity removal, and their efficiencies have been compared with some conventional coagulants. However, there have been no published studies on the comparison and optimization of both of these natural coagulants, compared with conventional ones, such as FeCl3 and PACl, using RSM. Furthermore, the effect of temperature has rarely been studied as an important factor for the removal efficiency. Considering this gap in knowledge, this study aims to compare and evaluate the efficiency of tannic acid, calcium lactate, FeCl3, and PACl in the turbidity removal of raw water in the Isfahan water treatment plant (IWTP) in Iran. Ultimately, the priority of coagulants is determined using the simple additive weighting (SAW) method (Memariani et al. 2009).

MATERIALS AND METHODS

Case study and raw water quality

Figure 1 shows the flow diagram of the IWTP with a capacity of 12 m3/s. Inflow is conveyed from the Cham–Aseman reservoir supplied from the Zayandehrood River, located 50 km west of Isfahan city. Water samples were taken from the influent of the coagulation–flocculation unit five times. The ranges of the physical–chemical parameters of the water quality are presented in Table 1.

Table 1

Water quality parameters of raw water from the IWTP (number of samples = 5)

ParameterTurbidity (NTU)pHAlkalinity (mg CaCO3/L)EC (μs/cm)Temperature (°C)Hardness (mg CaCO3/L)
Range 4.5–10.7 7.8–8 114–223 304–366 14–21 198–244 
ParameterTurbidity (NTU)pHAlkalinity (mg CaCO3/L)EC (μs/cm)Temperature (°C)Hardness (mg CaCO3/L)
Range 4.5–10.7 7.8–8 114–223 304–366 14–21 198–244 
Figure 1

Flow diagram of the IWTP.

Figure 1

Flow diagram of the IWTP.

Jar test

A jar test (Figure 2) was conducted based on the method of ASTM D2035 (ASTM International 2008). In each test, three beakers of 1-L volume were filled with 500 mL of the samples, whose pH, temperature, and initial turbidity were previously adjusted. Based on the standard method, after the addition of the coagulant and coagulation aid, the samples were mixed rapidly at a speed of 120 revolutions per minute (RPM) for 1 min and then mixed slowly for 20 min at a speed of 40 RPM. Afterwards, the samples were kept in rest for 30 min for settlement. Finally, samples were taken from the jars at 2 cm below the water surface level.

Figure 2

Jar test set-up for coagulation–flocculation.

Figure 2

Jar test set-up for coagulation–flocculation.

Adjustment of water samples

To assess the effect of initial turbidity on the removal efficiency, kaolin heavy powder (Sigma-Aldrich, CAS # 1332-58-7) was used as a source of synthetic turbidity. To create turbidity, a few grams of kaolin were added to 1 L of distilled water and stirred using a magnetic mixer at a speed of 400 RPM. Then, the suspension was left in a static state for 24 h for complete hydration. Different volumes were taken from the top layer of the suspension in order to evaluate the turbidity variation in the main sample (Hesami et al. 2014; Choy et al. 2016). Turbidity adjustments were made with an error of ±2NTU.

To adjust the pH in the samples, an ammonia solution (3 normal) for pH increase and HCl (3 normal) for pH decrease were used. For small required doses, the solution was diluted with distilled water. This adjustment was done with an error of ±0.05.

Temperature adjustments of the samples were made using a water bath. As shown in Figure 2, a beaker containing a sample was placed in a vessel containing water. The sample temperatures were adjusted by controlling the temperature of the water in the vessel, using hot water and a water–ice mixture. The temperature error value of this adjustment was ±2 °C.

Coagulants

Two conventional coagulants such as FeCl3 and PACl and two natural organics such as calcium lactate (C6H10CaO6) and tannic acid (C76H52O46) were used in this study. Moreover, cationic polyelectrolyte (PAM) with a 0.05 mg/L dose was used as a coagulant aid. All of the materials were supplied by Merck, Germany.

Experimental design and data analysis

The experiments were designed using the RSM which uses statistical and mathematical methods to model the output variable as the response that is influenced by the input variables called factors. An experiment is a series of tests, called runs, in which changes are made in the input variables in order to identify the reasons for changes in the output response. The RSM obtains the best model, which predicts the response as a fitted surface, and specifies the optimum point of the surface based on the goal of the research (Park 2007).

In this study, Design Expert software version 10.0.3 was used for the design of the experiments based on an optimal method (Montgomery 2012). The optimal design allows the statistical model to be estimated with fewer experimental runs; it can accommodate multiple types of factors with different levels, and type optimal-I minimizes the average prediction variance over the design space. Final turbidity was the response of the experiments. Analysis of variance (ANOVA) was used to evaluate the significance of the model and the variables. The quality of the fitted model was assessed by the coefficient of determination R2. Fisher's F-test was implemented to check the statistical significance of the model. The p-value with a 95% confidence level evaluated the model terms (Kim 2016).

To determine the effective factors, levels, and ranges of the input values for the RSM model, preliminary tests were conducted based on the one factor at a time method. Five factors, including coagulants, coagulant aid, initial turbidity, pH, and temperature, were considered in the tests. Final turbidity was considered as the response in all experiments, because it is an efficient criterion in the evaluation and comparison of coagulants.

Simple additive weighting

In order to compare and prioritize the four studied coagulants based on the experiment results and other criteria, a multi-attribute decision-making (MADM) technique is needed. MADM refers to screening, prioritizing, ranking, or selecting a set of alternatives, usually under independent and conflicting attributes (Anupama et al. 2015). The SAW method is the best-known and most widely used MADM method, due to its simplicity and efficiency. The best alternative (A*) can be derived by the following equations (Tzeng & Huang 2011): 
formula
(1)
 
formula
(2)
where ui (x) denotes the utility of the ith alternative, and i = 1, 2, …, n; wj denotes the weight of the jth attribute; and rij (x) is the normalized preferred rating of the ith alternative with respect to the jth criterion, which is calculated as the following:

For beneficial criteria (larger is better), rij (x) = xij/max xij; and for non-beneficial criteria (smaller is better) rij (x) = min xij/xij; where xij is the rating of alternative i with respect to criterion j.

Comparing pairs of treatment means

To assess the significance of the difference between the responses of the four coagulants, Fisher's least significant difference (LSD) method was used (Montgomery 2012). This procedure uses the F-statistic for testing the null hypothesis H0: μi = μj. A summary of the formulation is presented as follows: 
formula
(3)
 
formula
(4)
where LSDA,B is the LSD between two treatments, A and B; MSW denotes the mean square error (within the treatments); is the t-statistic with a 95% confidence level; DFW is the degree of freedom; nA and nB are the numbers of observations; and A and B are the mean values of treatments A and B.

RESULTS AND DISCUSSION

The results of the preliminary tests showed that the coagulant aid dose (polyelectrolyte) had very little effect on the response. Therefore, this parameter was eliminated from the experimental design. However, a 0.05 mg/L coagulation aid was added to all experiments at the beginning of the slow mixing in order to increase the efficiency. Thus, the four factors of coagulant dose, pH, initial turbidity, and temperature were selected in the experimental design, as shown in Table 2.

Table 2

Factor levels presented in terms of actual units of measurement

CoagulantFactorCalcium lactate (CL)Tannic acid (TA)PAClFeCl3
Coagulant dose (mg/L) 5–10–15–20 1–5–10 5–10–15–20 5–10–15–20 
pH 5–7–9 4–6–8 5–7–9 4–6–8 
Temperature (°C) 8–15–30 8–15–30 8–15–30 8–15–30 
Turbidity (NTU) 10–250–500–750–1000 10–250–500–750–1000 10–250–500–750–1000 10–250–500–750–1000 
CoagulantFactorCalcium lactate (CL)Tannic acid (TA)PAClFeCl3
Coagulant dose (mg/L) 5–10–15–20 1–5–10 5–10–15–20 5–10–15–20 
pH 5–7–9 4–6–8 5–7–9 4–6–8 
Temperature (°C) 8–15–30 8–15–30 8–15–30 8–15–30 
Turbidity (NTU) 10–250–500–750–1000 10–250–500–750–1000 10–250–500–750–1000 10–250–500–750–1000 

Experimental design

Tables 8–11 in Appendix 1 in Supplementary Materials show the designed experiments for each of the coagulants based on the optimal method, along with their responses obtained after the tests. The Design Expert software proposed 25 experiments with five replicates. The regression models of the final turbidity for the coagulants are given in Equations (5)–(8): 
formula
(5)
 
formula
(6)
 
formula
(7)
 
formula
(8)

The coefficients of A, B, C, and D refer to four factors, which are coagulant dose, initial turbidity, pH, and temperature, respectively. The summary results of the ANOVA tests on the above models are shown in Table 11 in Appendix 2 in Supplementary Materials. A p-value of less than 5% for each term shows the significance of the term; a p-value of more than 10% indicates the term is not significant; and for the range between 5–10%, the term may be significant, which can be judged based on subject matter knowledge (Montgomery 2012).

As shown in the table, p-values for the models indicate that all fitted models are significant. However, the temperature was significant only for tannic acid. Notable insignificant terms were the coagulant dose for calcium lactate and the initial turbidity for tannic acid. However, pH was only significant for the two natural coagulants, and not for PACl or FeCl3.

According to Table 11, the results of calcium lactate showed that the two parameters of pH and initial turbidity were more effective in removing turbidity, that their sums of squares (SS) were maximum and p-values were minimum. Among these two factors, pH was the most effective, with a 36% impact according to the ratio of the SS value to the total SS summation of all factors in the table.

The results of tannic acid in Table 11 showed that the effective terms in turbidity removal were pH and temperature as individual terms. In addition, the terms with interaction effects were coagulant dose and pH, coagulant dose and initial turbidity, coagulant dose and temperature, and initial turbidity and pH. Among these factors, the interaction effect of the temperature and pH was the most effective.

For PACl, the coagulant dose and temperature, as well as the interaction effect of the initial turbidity and pH, were effective in turbidity removal. Among these factors, the temperature was the most effective, with a 56% impact on turbidity removal.

The ANOVA results for FeCl3 in Table 11 indicated that the coagulant dose and initial turbidity, as well as the interaction effects of the coagulant dose and pH, coagulant dose and initial turbidity, and temperature and pH were effective in turbidity removal, while the initial turbidity was the most effective, with a 41% impact on turbidity removal.

Table 3 shows the statistical values of the RSM experiment models. The R2 values (0.77 for calcium lactate and 0.96 for FeCl3) show satisfactory regression models. Moreover, the R2 value should be in close agreement with the R2adjusted value. The R2adjusted value gives the percentage of variation, explained only by the effective independent variables. When the difference between R2 and R2adjusted is less than 0.4, there is very little chance for insignificant terms to be included in the model (Montgomery 2012). The results showed that these differences for all models were within the acceptable range.

Table 3

Statistical values of the RSM experiment models

CoagulantMean of responseR2R2adjustedPredicted R2AP
Calcium lactate 1.148 0.77 0.65 0.49 9.7 
Tannic acid 1.5312 0.92 0.82 0.67 14.07 
PACl 0.974 0.87 0.81 0.75 12 
FeCl3 1.3796 0.96 0.93 0.86 16.03 
CoagulantMean of responseR2R2adjustedPredicted R2AP
Calcium lactate 1.148 0.77 0.65 0.49 9.7 
Tannic acid 1.5312 0.92 0.82 0.67 14.07 
PACl 0.974 0.87 0.81 0.75 12 
FeCl3 1.3796 0.96 0.93 0.86 16.03 

The predicted R2 indicates how well a regression model predicts responses for new observations. The difference in the predicted and adjusted R2 values was less than 0.2 for all responses, indicating that all four models had sufficient capability to predict the responses. Adequate precision (AP) is a measure of signal-to-noise (S/N) ratio. It gives a factor by which the adequacy of the model to predict the response in the design space can be judged. A value greater than 4 is desirable, which was satisfactory for all the responses.

Comparing the coagulants

Table 4 presents the results of the LSD test to evaluate the significance of the difference between the mean of final turbidity for each pair of coagulants according to Equation (4). The comparison between the LSD value and the absolute mean difference (Equation (5)) in each row shows that there was no significant difference between the means of the final turbidity of the coagulants.

Table 4

Comparing the mean of pairs by the LSD method

ABLSDA,B
CL TA 0.693 0.383 
PACl CL 0.420 0.174 
PACl TA 0.650 0.557 
FeCl3 CL 0.550 0.232 
FeCl3 TA 0.741 0.152 
PACl FeCl3 0.494 0.406 
ABLSDA,B
CL TA 0.693 0.383 
PACl CL 0.420 0.174 
PACl TA 0.650 0.557 
FeCl3 CL 0.550 0.232 
FeCl3 TA 0.741 0.152 
PACl FeCl3 0.494 0.406 

In the next step, the coagulants were compared based on their optimum conditions. In the Design Expert software, the optimization was done based on the goals of minimizing both the final turbidity and the coagulant dose. Table 5 presents the optimum conditions of the four coagulants which were tested, and the results are presented as measured final turbidity. A comparison between the predicted optimized results from the model and the experiments shows that the optimization results were satisfactory. Also, the removal efficiencies based on the predicted final turbidity were calculated and are presented in the table.

Table 5

Optimal points in terms of the operating variables and the final turbidity

CoagulantTemperature (°C)pHInitial turbidity (NTU)Coagulant dose (mg/L)Predicted final turbidity (NTU)Measured final turbidity (NTU)Removal efficiency based on the predicted value (%)
Calcium lactate 12.18 6.92 126.31 0.63 0.58 99.5 
Tannic acid 12.85 5.4 39.69 1.49 0.59 0.32 98.5 
PACl 19 10 0.56 0.47 94.4 
FeCl3 25 7.9 585.34 6.25 0.76 0.68 99.8 
CoagulantTemperature (°C)pHInitial turbidity (NTU)Coagulant dose (mg/L)Predicted final turbidity (NTU)Measured final turbidity (NTU)Removal efficiency based on the predicted value (%)
Calcium lactate 12.18 6.92 126.31 0.63 0.58 99.5 
Tannic acid 12.85 5.4 39.69 1.49 0.59 0.32 98.5 
PACl 19 10 0.56 0.47 94.4 
FeCl3 25 7.9 585.34 6.25 0.76 0.68 99.8 

Based on the optimal values, a more complete comparison of the four coagulants is presented in Table 6. The comparisons include the optimum final turbidity, depth of sludge deposited after 30 min, residual concentrations of iron and aluminum, decrease in pH, change in alkalinity, and cost of consuming each substance based on the optimal dosage.

Table 6

Comparison of the performance of coagulants

CoagulantFinal turbidity (NTU)Depth of sludge deposited after 30 min (cm)Aluminum residuala (mg/L)Iron residualb (mg/L)Decrease in pHChange in alkalinity (mg/L as CaCO3)Cost (Rials/m3)
Calcium lactate 0.58 2.5 0.036 0.064. 0.2 5 decrease 400 
Tannic acid 0.32 2.3 0.065 0.011 0.2 5 decrease 170 
PACl 0.47 3.0 0.152 0.095. 0.4 5 increase 75 
FeCl3 0.68 3.5 0.018 0.042 0.8 35 decrease 100 
CoagulantFinal turbidity (NTU)Depth of sludge deposited after 30 min (cm)Aluminum residuala (mg/L)Iron residualb (mg/L)Decrease in pHChange in alkalinity (mg/L as CaCO3)Cost (Rials/m3)
Calcium lactate 0.58 2.5 0.036 0.064. 0.2 5 decrease 400 
Tannic acid 0.32 2.3 0.065 0.011 0.2 5 decrease 170 
PACl 0.47 3.0 0.152 0.095. 0.4 5 increase 75 
FeCl3 0.68 3.5 0.018 0.042 0.8 35 decrease 100 

aAluminum in raw water: 0.055 mg/L.

bIron in raw water: 0.071 mg/L.

According to Iran's drinking water quality standard for the outflow turbidity limit of 1 NTU (ISIRI 2009), it can be seen that the two coagulants of natural origin, in comparison with conventional coagulants, showed satisfactory results. The four coagulants were tested on samples with initial turbidities of 900 NTU in order to compare the sludges produced. It was observed that after 30 min, the depths of the deposited sludge produced by the two natural coagulants were less than the other two mineral coagulants. Moreover, FeCl3 produced the highest amount of sludge (3.5 cm). Therefore, it can be concluded that the use of natural coagulants, which have biodegradability and produce less sludge, eliminates most of the health concerns associated with conventional and mineral coagulants.

In the case of heavy metals, the amount of iron and aluminum and the residual values of the samples were measured by the atomic absorption method. As expected, the concentration of residual aluminum in the water after the coagulation and flocculation process by PACl was more than with the other materials. The limit for aluminum, according to Iran's drinking water standard, is 0.1 mg/L for large water treatment plants (ISIRI 2009), which for PACl was slightly more than the acceptable value. The main concern regarding the use of PACl in water treatment is the high amount of residual aluminum in water after coagulation and flocculation. According to the standard, the limit for iron remaining in drinking water is 0.3 mg/L. It was observed that the residual iron in the treated water was below the limit for all four coagulants.

The cost (last column in Table 6) is calculated based on the optimal dosage of the coagulant and the cost per kilo of each of the chemicals, for 1 m3 of treated water. Considering that the cost of PACl is lower than other coagulants, it can be said that under these conditions, the use of natural coagulants is not economically feasible.

The amount of pH reduction after the addition of the calcium lactate and tannic acid coagulants was 0.2; and the values for PACl and FeCl3 were 0.4 and 0.8, respectively. When using FeCl3, the reduction in pH was more significant, which necessitates the use of lime for pH adjustment. Moreover, changes in alkalinity in terms of mg/L as CaCO3 were 5 units decrease for both calcium lactate and tannic acid; 5 units increase for PACl, and 35 units decrease for FeCl3. FeCl3 acidifies the water; therefore, alkalinity had to be added to the water.

Implementation of coagulants for a raw water sample

In this section, a comparison based on the raw water in addition to the comparisons made based on the optimum conditions was provided. Pretests were done to determine the best coagulant dose for maximum turbidity removal. In order to evaluate the potential of producing organic matter as the precursors of DBPs, TOC was measured. The final results are shown in Table 7.

Table 7

Results of the jar test on the raw water sample

CoagulantDose (mg/L)Poly-El (mg/L)Turbidity (NTU)pHTOC (mg/L)Removal efficiency (%)
Raw water T = 21 °C – – 11.8 7.9 1.1 – 
Calcium lactate 15 0.05 2.92 7.8 3.9 75 
Tannic acid 0.05 4.3 8.0 2.4 64 
PAC 20 0.05 0.75 7.4 1.1 94 
FeCl3 0.05 1.8 6.6 1.5 85 
CoagulantDose (mg/L)Poly-El (mg/L)Turbidity (NTU)pHTOC (mg/L)Removal efficiency (%)
Raw water T = 21 °C – – 11.8 7.9 1.1 – 
Calcium lactate 15 0.05 2.92 7.8 3.9 75 
Tannic acid 0.05 4.3 8.0 2.4 64 
PAC 20 0.05 0.75 7.4 1.1 94 
FeCl3 0.05 1.8 6.6 1.5 85 

According to Table 7, PACl had the highest removal efficiency, while it had the highest dose based on the pretest results. The comparison between TOC values in the raw water and after settlement shows an increase in organic matter, after implementing the organic coagulants. It increases the risk of DBP formation.

Based on the recommendation by USEPA (2010) for DBP control, regarding the residual TOC less than 4 mg/L, and the alkalinity greater than 120 mg/L as CaCO3 (185 mg/L as CaCO3), 15% removal of TOC is necessary. This issue has to be considered in the operation of the treatment plant.

As illustrated in Figure 1, the ozonation unit is present after sedimentation in the IWTP which decreases TOC before filtration and chlorination units. Therefore, the potential of DBP formation due to the residuals of organic coagulants will greatly decline. Thus, the criterion for increasing organic residual in this study was not considered in the selection of the best coagulant.

Selecting the best coagulant

According to the SAW method, and taking into account different economic, technical, operational, and environmental measures, different coagulants were evaluated. The decision matrix for selecting the best coagulant is given in Table 8. In this matrix, at first, the criteria for choosing the best coagulant had to be determined. Afterwards, considering the importance of each criterion in terms of performance, environmental impact, cost, etc., weights were given; then, the criteria became dimensionless. For each material, according to the laboratory observations and results of the experiments, for each criterion, scores of 1–10 were assigned. For example, for criterion 1 (effectiveness in turbidity removal), all scores were considered the same. It is worth mentioning that the criterion 4 (effects on pH reduction) is applicable only for the case study and water resources with similar ranges of pH (7.8–8.0).

Table 8

SAW analysis results for selecting the appropriate coagulant

No.CriteriaWeightPAClCalcium lactateTannic acidFeCl3
Effectiveness in removing turbidity 0.11 
Costs 0.14 10 
Amount of sludge deposited 0.15 
Effects on pH reduction 0.14 
Presence of heavy metals in the water 0.18 
Efficiency in high turbidity conditions 0.12 
Dependence on operating conditions 0.16 10 
 0.76 0.84 0.77 0.62 
No.CriteriaWeightPAClCalcium lactateTannic acidFeCl3
Effectiveness in removing turbidity 0.11 
Costs 0.14 10 
Amount of sludge deposited 0.15 
Effects on pH reduction 0.14 
Presence of heavy metals in the water 0.18 
Efficiency in high turbidity conditions 0.12 
Dependence on operating conditions 0.16 10 
 0.76 0.84 0.77 0.62 

Bold denotes the highest utility.

For criterion 5 (the presence of heavy metals), for PACl, due to the presence of the residual Al, the lowest score was given. The sum of the product of the criteria weights as well as the normalized values of the scores are presented in the last row. Based on Equation (2), the coagulant with the highest utility () was selected as the best. According to Table 8, calcium lactate was the best option.

CONCLUSIONS

In this study, the four coagulants of calcium lactate, tannic acid, FeCl3, and PACl were compared for water treatment. The experiments were designed by the RSM and the ANOVA showed the validity of the regression models of the effective factors in the final turbidity response. Optimum conditions of the models were checked through experiments, and the following results were obtained: 0.63 NTU (the removal efficiency of 99.5%) after the application of calcium lactate; 0.59 NTU (the removal efficiency of 98.51%) after the application of tannic acid; 0.76 NTU (the removal efficiency of 99.74%) after the application of FeCl3; and 0.56 NTU (the removal efficiency of 94.4%) after the application of PACl. The results of the experiments show that natural coagulants can reduce the turbidity of water up to the standard level. Furthermore, the comparison between the four coagulants showed that the changes in pH and alkalinity, as well as sludge production, were lower using natural coagulants when compared with the tested chemical coagulants. The tests on a raw water sample showed an increase in TOC after sedimentation. However, due to the presence of the ozonation process in the IWTP, the concerns regarding DBP formation due to their residual organics were resolved. In general, a multi-criteria analysis for choosing the preferred water treatment option shows that natural coagulants, such as calcium lactate and tannic acid, despite their higher costs, excel over conventional coagulants, and ultimately calcium lactate was chosen as the best option. The results showed that natural coagulants can be considered as a preferable option to the conventional coagulants, provided that their residual organics are removed and the risk of DBP formation is reduced. Since this concern was resolved in the IWTP due to the presence of the ozonation process, calcium lactate was proposed as a competent alternative to the conventional coagulants.

ACKNOWLEDGEMENTS

The authors thank Mr. Mojtaba Ghobadian and his colleagues at the Isfahan water and wastewater company for their support in providing raw water samples.

SUPPLEMENTARY MATERIAL

The Supplementary Material for this paper is available online at https://dx.doi.org/10.2166/aqua.2019.075.

REFERENCES

REFERENCES
Anupama
K. S. S.
Gowri
S. S.
Rao
B. P.
Rajesh
P.
2015
Application of MADM algorithms to network selection
.
International Journal of Innovative Research in Electrical, Electronics, Instrumentation and Control Engineering
3
(
6
),
64
67
.
ASTM International
2008
ASTM D2035-08: Standard Practice for Coagulation-Flocculation Jar Test of Water
.
ASTM International
,
West Conshohocken, PA. Available from: https://www.astm.org.
Azreen
I.
Zahrim
A. Y.
Chong
S. H.
Ng
S. W.
2017
Estimation of melanoidin concentration in palm oil mill effluent ponding system and its treatment using calcium lactate
. In:
IOP Conference Series: Materials Science and Engineering
, Vol.
206
.
IOP
,
012078
.
Bina
B.
Mehdinejad
M.
Nikaeen
M.
Attar
H. M.
2009
Effectiveness of chitosan as a natural coagulant aid in treating turbid waters
.
Journal of Environmental Health Science & Engineering
6
(
4
),
247
252
.
Camacho
F. P.
Sousa
V. S.
Bergamasco
R.
Teixeira
M. R.
2017
The use of Moringa oleifera as a natural coagulant in surface water treatment
.
Chemical Engineering Journal
313
,
226
237
.
Choy
S. Y.
Prasad
K. N.
Wu
T. Y.
Raghunandan
M. E.
Ramanan
R. N.
2016
Performance of conventional starches as natural coagulants for turbidity removal
.
Ecological Engineering
94
,
352
364
.
Chung
K.
Wei
C.
Johnson
M.
1998
Are tannins a double-edged sword in biology and health
.
Trends in Food Science & Technology
9
,
168
175
.
Devesa-Rey
R.
Bustos
G.
Cruz
J. M.
Moldes
A. B.
2012
Evaluation of non-conventional coagulants to remove turbidity from water
.
Water, Air, & Soil Pollution
223
(
2
),
591
598
.
Divakaran
R.
Pillai
V. S.
2002
Flocculation of river silt using chitosan
.
Water Research
36
(
9
),
2414
2418
.
Grenda
K.
Arnold
J.
Gamelas
J. A.
Rasteiro
M. G.
2018
Up-scaling of tannin-based coagulants for wastewater treatment: performance in a water treatment plant
.
Environmental Science and Pollution Research
1
12
.
Hayder
G.
Ab Rahim
A.
2015
Effect of mixing natural coagulant with alum on water treatment
. In:
Presented at the 3rd National Graduate Conference (NatGrad 2015)
.
Universiti Tenaga Nasional, Putrajaya Campus
,
Kajang, Selangor
, pp.
206
209
.
Hesami
F.
Bina
B.
Ebrahimi
A.
2014
The effectiveness of chitosan as coagulant aid in turbidity removal from water
.
International Journal of Environmental Health Engineering
3
(
1
),
8
.
Institute of Standards and Industrial Research of Iran (ISIRI)
.
2009
Drinking Water Physical and Chemical Specifications
.
ISIRI No. 1053, 5th Revision (in Persian)
.
Kerkar
P.
2017
Is Tannic Acid in Water Bad for You?
Pain Assist Inc.
.
Kumar
V.
Othman
N.
Asharuddin
S.
2017
Applications of natural coagulants to treat wastewater – a review
. In:
MATEC Web of Conferences
, Vol.
103
.
EDP Sciences
,
06016
.
Les Ulis, France, paper no. 06016, pp.1–9, eds. M. J. Zainorizuan, L. Yee Yong, L. Alvin John Meng Siang, O. Mohamad Hanifi, R. Siti Nazahiyah & A. Mohd Shalahuddin
.
Mcallister
J.
2016
Is Tannic Acid in Water Harmful?
.
Memariani
A.
Amini
A.
Alinezhad
A.
2009
Sensitivity analysis of simple additive weighting method (SAW): the results of change in the weight of one attribute on the final ranking of alternatives
.
Journal of Optimization in Industrial Engineering
2
(
4
),
13
18
.
Montgomery
D. C.
2012
Design and Analysis of Experiments
, 8th edn.
Wiley & Sons
,
New York
,
USA
.
Muthuraman
G.
Sasikala
S.
2014
Removal of turbidity from drinking water using natural coagulants
.
Journal of Industrial and Engineering Chemistry
20
(
4
),
1727
1731
.
Oladoja
N. A.
Saliu
T. D.
Ololade
I. A.
Anthony
E. T.
Bello
G. A.
2017
A new indigenous green option for turbidity removal from aqueous system
.
Separation and Purification Technology
186
,
166
174
.
Park
G.-J.
2007
Analytic Methods for Design Practice
.
Springer Science & Business Media
, Berlin,
Germany
.
Paydari
P.
2011
Evaluation of the Oak Seed as Coagulant and Coagulant Aid in Water Treatment
.
Master's Thesis
,
Civil Engineering Department, Isfahan University of Technology
,
Iran
.
Rizzo
L.
Di Gennaro
A.
Gallo
M.
Belgiorno
V.
2008
Coagulation/chlorination of surface water: a comparison between chitosan and metal salts
.
Separation and Purification Technology
62
(
1
),
79
85
.
Sánchez-Martín
J.
González-Velasco
M.
Beltrán-Heredia
J.
2010
Surface water treatment with tannin-based coagulants from Quebracho (Schinopsis balansae)
.
Chemical Engineering Journal
165
(
3
),
851
858
.
Thakur
S. S.
Choubey
S.
2014
Use of tannin based natural coagulants for water treatment: an alternative to inorganic chemicals
.
International Journal of ChemTech Research
6
,
3628
3634
.
Tzeng
G. H.
Huang
J. J.
2011
Multiple Attribute Decision Making Methods and Applications
.
Chapman and Hall/CRC
, pp.
55
67
.
US EPA
2010
Comprehensive Disinfectants and Disinfection Byproducts Rules (Stage 1 and Stage 2): Quick Reference Guide
.
Yarahmadi
T.
Peyda
M.
Mohammadian Fazli
M.
Rezaeian
R.
Soleimani
N.
2016
Comparison of water turbidity removal efficiencies of Descurainia sophia seed extract and ferric chloride
.
Journal of Human, Environment and Health Promotion
1
(
2
),
118
124
.
Zahrim
A. Y.
Nasimah
A.
2018
Calcium lactate as a promising coagulant for the pretreatment of lignin-containing wastewater
. In:
Engineering Technologies for Renewable and Recyclable Materials
.
Apple Academic Press
,
Palm Bay, FL, eds. Joy, J., Jaroszewski, M., Praveen, K. M., Thomas, S. & Haghi, R.
, pp.
289
314
.

Supplementary data