Optimization of sulfate removal by sulfate reducing bacteria using response surface methodology and heavy metal removal in a sulfidogenic UASB reactor

Accepted Manuscript Title: Optimization of sulfate removal by sulfate reducing bacteria using response surface methodology and heavy metal removal in a sulfidogenic UASB reactor Authors: Tahereh Najib, Mostafa Solgi, Abbas Farazmand, Seyed Mohammad Heydarian, Bahram Nasernejad PII: DOI: Reference:

S2213-3437(17)30266-X http://dx.doi.org/doi:10.1016/j.jece.2017.06.016 JECE 1678

To appear in: Received date: Revised date: Accepted date:

23-1-2017 6-6-2017 10-6-2017

Please cite this article as: Tahereh Najib, Mostafa Solgi, Abbas Farazmand, Seyed Mohammad Heydarian, Bahram Nasernejad, Optimization of sulfate removal by sulfate reducing bacteria using response surface methodology and heavy metal removal in a sulfidogenic UASB reactor, Journal of Environmental Chemical Engineeringhttp://dx.doi.org/10.1016/j.jece.2017.06.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Optimization of sulfate removal by sulfate reducing bacteria using response surface methodology and heavy metal removal in a sulfidogenic UASB reactor Tahereh Najiba, Mostafa Solgib, Abbas Farazmandc, Seyed Mohammad Heydarianc, Bahram Nasernejada1 a

Chemical Engineering Department, Amirkabir University of Technology (Tehran Polytechnic), P.O.

Box 15875-4413, Tehran, Iran b

Chemical Engineering Department, Babol Noshirvani University of Technology, P.O. Box 484,

4714871167 Babol, Iran c

Biotechnology Department, Iranian Research Organization for Science and Technology (IROST), P.O.

Box 15815-3538, Tehran, Iran

1

Corresponding author at: Chemical Engineering Department, Amirkabir University of

Technology (Tehran Polytechnic), P.O. Box 15875-4413, Tehran, Iran E-mail address: [email protected] 1

Abstract

Biological treatment using sulfate reducing bacteria (SRB) has been found out to have the potential for treatment of wastewaters containing sulfate and heavy metals. In the present study, in addition to the investigation of carbon source effect on the growth and sulfate removal, a statistical model using the design of experiments methodology has been developed to optimize the removal of sulfate by the obtained consortium from the anaerobic digested. All parameters were selected in three levels. Investigating the data collected from experiments using response surface methodology proved that the interactions between parameters were insignificant and could be neglected. The optimum removal conditions were achieved at pH of 7.19, initial sulfate concentration of 2153.15 mg/L, COD (Chemical Oxygen Demand) to initial sulfate concentration ratio (COD/SO42-) of 2.72 and the COD related to ethanol to total COD ratio (CODethanol/CODtotal) of 1 which led to sulfate removal of 98%. In five different batch experiments done in UASB reactor, the optimum conditions were applied to evaluate SRB performance in real and synthetic wastewater treatment mostly in all experiments. The differences of experiments were in real or synthetic wastewater, having an extra stage of treatment by H2S, using organic wastewater as carbon source and presence of heavy metal in bacteria medium. The maximum obtained removal percentage for Zn and Ni were 99.996 and 96.87, respectively. Keywords: sulfate reducing bacteria, sulfate removal, wastewater treatment, UASB reactor, Response surface methodology 1. Introduction

Mining effluents containing high amounts of heavy metals and sulfate ions could be considered as one of the major pollutants of the natural water. The presence of sulfate ions in mining effluents is due to bacterial activity on sulfidic ore, and the high concentration of heavy metals derived from mine rocks is a result of the wastewater acidity which increases their solubility [1–3]. Some industries also release these pollutants into the environment. For instance, oil refineries discharge wastewater containing high concentrations of sulfate and effluents containing heavy metal are released into the environment by industries such as metallurgical, 2

electronics, electroplating, paint and pigment manufacturing, stainless steel production, leather tanning, textile, wood preservation etc [4–6]. Heavy metals can replace the essential ions in cells if they enter the human body and thus can disturb their normal function. Some might even have carcinogenic and mutagenic effects. On the other hand, drinking water containing high sulfate concentration could lead to diarrhea. Aside from the health risks they cause, sulfate ions in water can be reduce to hydrogen sulfide which could cause erosion in water delivery metal parts and concrete buildings [4,7–9]. Although chemical precipitation has been used commonly for treating Mining effluents; nevertheless, the high costs and inadequate efficiency results have made its large-scale application uneconomic [10]. Besides, high volume of generated sludge in chemical treatment can be problematic [11]. The most important advantage of biological treatment over the chemical one is the less resolubility of sulfide precipitation than hydroxide precipitation in a wider range of pH [12]. Bioremediation technology using sulfate reducing bacteria (SRB) has been considered by many researchers due to its effectiveness in removing sulfate ions and heavy metals simultaneously from industrial wastewaters and freshwater resources [13]. The anaerobic process is based on the fact that sulfate is considered as the electron acceptor while a source of organic compound substrate acts as the electron donor which leads to the formation of sulfide and bicarbonate ions. The produced sulfide results in precipitation of metal ions which have very low solubility and the produced bicarbonate increases the pH of wastewater [10,13,14]. The process is briefly shown as follows [16]: Electron donor + SO42- → HS- + HCO3-

(1)

Me2+ + HS- → MeS↓ +H+

(2)

HCO3- + H+ → H2O + CO2

(3)

The restriction faced in using SRB is due to bacteria super sensitivity towards heavy metals, thus application of the method in wastewater treatments would be limited to the cases with low concentrations of heavy metals. Moreover, the acidic nature of mining effluents can be a deterrent parameter for SRB activity. As a result, the most practical method is to grow the bacteria separately and then subsequently use the produced biogenic sulfide solution in the treatment. Therefore, the higher heavy metal removal efficiency, in the case that bacteria have been grown in a medium with no heavy metals present, can be achieved when sulfate removal is 3

at its optimum point [17]. The carbon-energy substrate due to its significant effect on the bacterial growth rate, culture composition and the potential impact on the economics of an industrial process is considered a significantly predominant variable [18,19]. Various studies have been performed using sodium lactate as the source of carbon [20]. Although using lactate as the carbon source for these types of treatments and in industrial scale seems uneconomic. Therefore, less expensive carbon sources such as ethanol or waste containing organic substrate could be a great option in reducing the expenses of treatment [21]. Ethanol has been widely applied as the electron source in biological production of sulfide in various types of reactors [19]. Treatment of sulfate/heavy metals-rich wastewater have been reported using different bioreactor configurations such as fluidized bed reactor (FBR), membrane bioreactor (MBR), up-flow anaerobic sludge blanket bioreactor (UASB), anaerobic hybrid reactor (AHR), packed bed reactor (PBR) and continuous stirred tank reactor (CSTR) [10,22,23]. Response surface methodology (RSM) consists of a group of mathematical and statistical techniques based on the fit of a polynomial equation to the experimental data in order to predict the behavior of a system. It is appropriately applicable when one or more response of interest is influenced by several variables. RSM has the advantages of reducing the number of runs necessary to evaluate the multiple variables in experimentation, simultaneous studying of their interactions, higher percentage yield, reduced process variability, a closer substantiation of the output response to the target and less time with minimum cost. Among the RSM designs, the most frequently used second-order designs are the 3k factorial, central composite, and the Box– Behnken designs. The central composite design (CCD) is the most popular method due to its simple structure and good efficiency [24–26]. To the best of our knowledge, no studies have been reported on concurrent comparison of effect of ethanol and lactate individually as well as half and half mixture of them based on COD on sulfate removal by SRB and bacterial growth rate. Thus, in this study three experiments at the same condition by different carbon sources of ethanol, lactate and the mixture of equivalent COD of lactate and ethanol were done. In another stage of this study, the purpose is to achieve a better understanding of the relationship between the variables and also to determine the optimum condition in sulfate removal. Therefore, by applying RSM four parameters of pH, initial concentration of sulfate, COD to initial sulfate concentration ratio (COD/SO42-), and the COD 4

related to ethanol to total COD ratio (CODethanol/CODtotal) were investigated in order to determine their effects on the dependent parameter which is sulfate removal percentage. Consequently, the UASB reactor in the obtained optimum conditions was coupled by a two-phase (gas/liquid) reactor in which the sulfidogenic UASB reactor treated effluent of the previous day was extra treated by the output gas of bioreactor for 24 hr. The aim was investigation of the considerable advancement in the efficiency of sulfidogenic UASB reactor in the capability of heavy metal treatment by the novel 24-hour batch system of gas treatment. 2. Materials and methods 2.1. Bacterial Inoculum

In order to carry out the experiments, the anaerobic digested sludge was collected from a sewage treatment plant in the south of Tehran. To activate the sulfate reducing consortium, the Postgate isolation selective medium was used so that the population of SRB would be the dominant microbial population [27]. The first incubation period with the Postgate isolation selective medium was performed for one week at a temperature of 35 oC. After that and before each run the sludge was washed using distilled water and then centrifuge twice so that no sulfate from the medium would remain in the bacterial environment and so the bacteria would be ready to be used in different testing conditions. The growth medium and all the experiment containers were autoclaved for 15 minutes at a temperature of 121oC. The mixed liquor volatile suspended solids (MLVSS) of the bacteria in the bacteria solution was set to be 12,650 mg/L and the volume-volume percentage of the bacteria in the Postgate growth medium was chosen to be 10 percent, so the bacteria concentration in the experiment condition was 126.5 mg/L. For the optimization experiments, after the inoculation of the bacteria in 100 milliliter serum bottles, each bottle was flushed with the nitrogen gas for 5 minutes in order to make sure of having an anaerobic condition. After that the bottles of each series of runs were put in the incubator shaker. 2.2. Carbon Source Optimization Tests

To optimize the carbon source in Postgate medium, CODethanol/CODtotal variable was chosen in three different levels: level zero (The entire COD is due to sodium lactate), level 0.5 (Half of COD is due to ethanol and other half is due to sodium lactate) level 1 (The entire COD is due to ethanol). The other parameters of the experiments including temperature, pH, the initial concentration of sulfate and COD/SO42- were all kept in their optimum level in literatures (Table 5

1), also the time of experiments was 50 hours. In the selection of carbon sources for feeding an anaerobic consortium, two parameters which got involved are the ability of SRB in use of it and competition between SRB and Methanogenic Archaea (MA) in use of it. SRB are not generally successful in the competition with MA for using acetate. The reason related to low yield of sulfate removal by a mixed consortium in the condition that acetate is the carbon source is that SRB are unable to completely oxidize acetate even in the presence of excess levels of sulfate [28,29]. Due to this inability of SRB, carbon compounds are incompletely oxidized to acetate. Generally, it has been discovered that sulfate reducers can be categorized into two groups: one which can degrade organic compounds incompletely to acetate and the others which can degrade organic compounds completely to carbon dioxide. The second group commonly also uses acetate as a growth substrate throughout two different pathways of acetate oxidation, a modified citric acid cycle and the acetyl-CoA pathway [30]. In the case of competition, methanogens can use only some sources for growth such as hydrogen, carbon dioxide and acetate. It has not been identified any methanogens which can use organic acids, like lactate, propionate and butyrate, which are the appropriate substrates for SRBs. Only in the presence of a high amount of sulfate, SRB compete with methanogens to use of hydrogen or acetate [30]. Thus, acetate was not the selected carbon source in this study. Laanbroek et al. found that in the condition of sulfate inadequacy, SRB use hydrogen, lactate and ethanol instead of propionate and acetate [31]. 2.3. Designing the Experiments and the RSM

To optimize the process and study the interactions between the parameters at the growth medium, a four-factor-three-level face central composite design (CCD) was applied using the RSM. Four parameters of pH, Initial concentration of sulfate, COD/SO42-, and CODethanol/CODtotal were considered as independent parameters, and sulfate removal percentage after 65 hours from the beginning of the experiment at a temperature of 35 °C was considered as the dependent parameter. The range and levels of the independent parameters of Postgate medium can be seen in the Table 2. To study the mutual effects of four independent parameters on the response function, 30 experiments (including 6 replicates experiments in the center point to calculate the repeatability of the method) were carried out to estimate the experimental error [26]. The Design Expert (version 7.0.0, Stat Ease Inc., USA) was used in order to design the 6

experiments and analyze the data. The designed experiments were performed and the experimental data achieved in the lab was given to the software in order to determine the predicted values. The complete matrix of the designed experiments and the results can be seen in the Table 3. In order to propose and select a valid model, the actual data was fit into different model presented by CCD which the quadratic model seemed to be the best candidate [32].

The experimental data was fitted to Polynomial quadratic model to obtain regression coefficients.

Y  0   i xi   ii xi2   ij xi x j

(4)

Where, Y is the predicted response, x is the independent variable, β0 is the constant term, βi are the linear coefficients, βii is the squared coefficients, and the βij is the interaction coefficient. Investigating all factors using the ANOVA variance analysis will help to find out whether the model would be valid, the effect of factors and their interactions. These factors include the Fvalue, p-value, lack of fit, adjusted R-squared (R2Adj), coefficient of determination R-squared (R2), and predicted R-squared (R2Pred). The R2Adj and R2Pred are used to measure the fluctuation around the mean and new data. F-value is the statistical factor which defines the data fluctuation around the mean, while p-value provides data on the level of the significance of the variables. A valid model is capable of predicting the interactions occurred between the variables during the process. 2.4. Sulfidogenic upflow anaerobic sludge blanket (UASB) reactor A pilot-scale bioreactor (4 liters) was designed and made of stainless steel, so that the culture through the silicon tube was pumped and entered the bioreactor at the bottom section and then passed through the sludge. At the top of reactor, a biomass screen was installed to not allow the biomass to leave the reactor in circulation process. After circulation in the specified hydraulic retention time, the solution containing sulfide exited from the top of reactor. The reactor was placed in a water bath to set the temperature at 35±2ºC. All the joints of the reactor were sealed with Teflon tape to have the anaerobic condition in the bioreactor. The gas outlet was connected to a two-phase (gas/liquid) reactor to have the next step of treatment by produced H2S. 7

2.5. Treatment of sulfate and metal rich wastewater in the UASB reactor 2.5.1. Real and synthetic wastewaters Some batch experiments were done on the treatment of synthetic solution and the others on the treatment of real wastewater (electroplating wastewater). The real wastewater rich in sulfate and dissolved nickel was collected from an electroplating unit located at the countryside of Tehran, Iran. The pH and COD were 2.27 and 16.7 mg/L respectively. The acidic nature of the wastewater and low level of COD are common this kind of wastewater. The sulfate concentration of the wastewater was 2499.97 mg/L. The dissolved nickel concentration was 72.496 mg/L. Stock solutions of nickel and zinc (1,000 mg/L) were prepared by dissolving nickel (II) sulfate (NiSO4) and zinc sulfate (ZnSO4) in distilled water. The working concentrations of synthetic solutions were provided by diluting the stock solutions with distilled water. All chemicals used in the experiments were of analytical reagent grade and obtained from Merck (Darmstadt, Germany). The pH was adjusted with 0.1 M HCl and 0.1 M NaOH. The characterization of real electroplating wastewater and synthetic solutions was given in Table 4. 2.5.2. Batch experiments for treatment The batch experiments were done in five different modes in UASB reactor at 35°C. Mode I of batch treatment experiment was conducted in the obtained optimum conditions by RSM to treat the synthetic solution in which the Zn2+ and Ni2+ concentrations were both 50 mg/L. Heavy metal treatment was done out of the bioreactor, as the effluent containing sulfide was mixed with the solutions containing heavy metal. After sampling, this mixture was stored in the gas/liquid reactor to be contacted with H2S from the bioreactor for 24 hours. The study in the fed batch mode was carried out for 20 days. Throughout the experiments, pH, COD ,the sulfate and sulfide concentration of the effluent and the metal concentration of the mixture before and after H2S contact were measured every 24 hr. Mode II of experiments was carried out similarly in 20 days, except that the carbon source of medium was from cheese whey wastewater (with high level of COD) and the gas treatment was omitted. Mode III, IV and V of the experiments were done in 38 days to treat the real wastewater of nickel electroplating unit. The pH of real wastewater was adjusted to 7.2 and the concentration of nickel was changed to 49.86 mg/L. The treatment in Mode III was done outside of bioreactor, while in Mode IV (with unadapted bacteria) and mode V (adapted bacteria), the real wastewater 8

containing Ni was treated in the bioreactor. The summary of experiments is given in the Table 5.

2.6. Analytical Methods

To fulfill the same condition in terms of the amount of sludge in all runs, the optical density of the consortium was measured with spectrophotometer at 450 nm and related to the MLVSS method through calibration curve [33]. To measure the concentration of sulfate, prior to analysis each sample taken from bottles was centrifuged first in order to obtain a completely transparent and bacteria-free solution. The concentration of sulfate was measured using the turbidimetric method [19]. For pH measurement the unfiltered samples immediately after collection were analyzed by the probe of pH-meter (Metrohm). The concentration of heavy metals in the centrifuged samples was determined by inductively coupled plasma–optical emission spectrometry (ICP–OES) Model 730ES-Variant. Analysis of COD required a preparation of samples in order to remove dissolved sulfide. First, samples were acidified with concentrated sulfuric acid and then sparged with nitrogen gas for 5 minutes [34]. COD measurements were done according to the closed reflux colorimetric method (D5220-High range) of APHA Standard Methods [35]. For sulfide concentration measurement, in order to prevent the loss of sulfide as H2S, pH of samples was increased to 10 by concentrated NaOH. Subsequently, the CordRuwisch was used [36]. The analyses results were the average of three analyses to reduce the possible errors. 3. Results and discussion 3.1. Studying the Effect of CODethanol/CODtotal on Biomass Growth and Sulfate Removal

In order to study the effects of lactate and ethanol carbon source on biomass growth and sulfate removal percentage, experiments were carried out using different growth mediums with carbon sources of ethanol, lactate and ethanol-lactate mixture. It was found out that the optimum biomass production of SRB happens in lactate growth mediums. Although the maximum amount of sulfate removal is not achieved in this condition (Fig. 1), it happened in the ethanol growth medium. The results achieved in this study covers the ones reported by White and Gadd which suggested that growth mediums of lactate has the maximum biomass growth in comparison to other carbon sources; and ethanol generates more sulfide which means more sulfate was used 9

[37]. Although no explanation have been given for it, the difference in the growth and the activity of the bacteria in sulfate removal process could be due to the different cell metabolism inside the bacteria. In the oxidation process of lactate to acetate one ATP -the energy for growthand one NADH are released, while in the oxidation process of ethanol to acetate two NADH are released [38,39]. As under anaerobic condition only a portion of electron carriers (NADH) are involved, the actual amount of ATP produced in anaerobic respiration is always lower than aerobic respiration [40]. Giving more details, ATP synthesis requires entering 3 protons to make one ATP, but in the low concentration of oxygen (anaerobic condition) only two protons release for each NADH. The stoichiometries range from 2 to 8 H+/NADH, equivalent to 0.67- 2.5 ATP/NADH, so in anaerobic condition with the ratio of 2 H+/NADH each NADH makes 0.67 ATP [41]. Therefore, the number of ATP in the medium containing of lactate is higher than lactate and ethanol and ethanol, respectively. Furthermore, although the growth of SRB is less in a medium containing ethanol, but as it produces two NADH per ethanol (as Fig. 1b), the ethanol feeding will reduce more sulfate, this leads to low concentration of sulfate in the system (Fig. 2a).

3.2. Response Surface Methodological Approach for Optimization of Parameters 3.2. 1. Experimental Design, Model Fitting and Statistical Analysis

In order to determine the significance and effectiveness of the second degree model, the statistical analysis on the data obtaind by conducting 30 experiments was performed using the analysis of variance table. The result can be seen in Table 6. The ANOVA analysis table has defined the quadratic model statistically significant. The full quadratic polynomial model for statistical designing based on coded parameters is as follows: 𝑟𝑒𝑚𝑜𝑣𝑎𝑙 = +0.92 + 0.027 ⨯ 𝐴 − 0.12 ⨯ 𝐵 + 0.047 ⨯ 𝐶 + 0.041 ⨯ 𝐷 − 6.250 ⨯ 10−3 ⨯ 𝐴𝐵 − 6.250 ⨯ 10−3 ⨯ 𝐴𝐶 + 5 ⨯ 10−3 ⨯ 𝐴𝐷 − 2.5 ⨯ 10−3 ⨯ 𝐵𝐶 − 0.014 ⨯ 𝐵𝐷 − 6.250 ⨯ 10−3 ⨯ 𝐶𝐷 − 0.064 ⨯ 𝐴2 − 0.25 ⨯ 𝐵2 − 0.079 ⨯ 𝐶 2 − 0.029 ⨯ 𝐷2

(7)

The insignificant terms –the terms with p-value larger than 0.1- can be omitted from the model; therefore the equation of reduced model is as follows: 𝑟𝑒𝑚𝑜𝑣𝑎𝑙 = −3.56388 + 1.03513 ⨯ 𝐴 + 4.91837 ⨯ 10−4 ⨯ 𝐵 + 0.15449 ⨯ 𝐶 + 0.082222 ⨯ 10

𝐷 − 0.071993 ⨯ 𝐴2 − 1.14219 ⨯ 10−7 ⨯ 𝐵2 − 0.028406 ⨯ 𝐶 2

(8)

The F-value of the model was found to be 54.20 and the value of probability of the model was less than 0.0001 which proves that the model is significant and there is only 0.01% probability of the differences occurring purely by chance. R2 is the statistical parameter to measure how close the data and the fitted model are, in other words it is useful to determine the accuracy of the fitted model on the data. As the values become closer to 1, the fitted model becomes more capable of providing the results closer to the actual values, as a function of independent variables. A model with an R2 value higher than 0.8 can be considered well fitted. The Predicted R2 determines the accuracy of the model in predicting the results for new observations. Thus, its higher values represent better prediction of the results. The adjusted R2 is a developed version of R2 which is always smaller than it and its value only increases if the improvement due to adding a new term is more than what is expected by chance. In order for the model make satisfactory predictions, the adjusted and predicted R2 values should be close [42]. In the present research the value of R2 was 0.9806 which is very close to 1. Also the values of R2adj and R2pred were 0.9625 and 0.9169 respectively. Good agreement of these values represents the suitable fitting of the model using the software. Another term named residual can also be seen in the ANOVA table which is the sum of two other terms: lack of fit and pure error. The crucial term here is the F-value of the lack of fit and its low value shows that the lack of fit is insignificant compared to the pure error. An insignificant lack of fit is desirable, because the objective is to fit the model on the data. In the recent design the F-value of lack of fit was estimated 4.06 which prove its insignificance. The presented model is acceptable if only the terms of pure error (the error caused by repeating the experiment) and lack of fit (sum of the squared factors omitted from the model) are insignificant. The p-value of the lack of fit was found to be 0.0678 in which the lack of fit value is considered insignificant compare to the pure error, and therefore the model is acceptable. In the diagnostic section of the design experiment software some provided diagrams could be applied to investigate the validity of the model. The section includes the analysis of normality of the experiment data, showing the residual for the predicted results and the level of closeness of the predicted results and the actual results. The normal probability diagram can be seen in Fig. 11

3.a, in which the plotted points close to the straight line proves the normal distribution of errors with an average value of zero. Another way to determine the sufficiency of the model is to study the residual which is shown in Fig. 3.b, random scatters evenly distributed above and below the horizontal axis proves the sufficiency of the model. Also in Fig. 3.c the values which were obtained from the presented model is compared with the values which were achieved from the actual results. The clustering of the points around the diagonal line indicates a satisfactory correlation between the experimental and predicted data, confirming the robustness of the model [14,43]. 3.2.2. Using Surface Response Diagrams to Determining the Optimized Condition

In the perturbation diagram presented in Fig. 4, the effects of the four parameters on sulfate removal are shown at the same time. This diagram represents a general view of the effect of changing all parameters on the result. The slope variation of the line in D (CODethanol/CODtotal) shows that although this parameter was not as much effective as parameter B in sulfate removal, nevertheless, the cost of operation makes this parameter noteworthy, because ethanol is much cheaper than lactate and also the result of sulfate removal in the condition of ethanol is better than using both ethanol and lactate simultaneously and using lactate individually. This fact was the reason of choosing CODethanol/CODtotal=1 as an experimental condition. The influence of A (pH) and C (COD/SO42-) are almost the same in their higher values, but in lower values the influence of C is more impressive. The optimum value for the B (SO42-) is more obvious than other parameters, which means by making small changes in its value, the percentage removal of sulfate changes noticeably. The contour and surface response plots are used in order to determine the optimized value for the variables in which the result would be maximized. In these diagrams the parameters influence and the interaction between two independent factors is illustrated. Fig. 5 and Fig. 6 show the samples of the surface respond and the contour plot for pairs of pH and initial sulfate concentration. Each contour and surface response plot provides data on the interaction between two parameters while the other two parameters have been kept constant at the center point. For example, Fig. 6 shows the two-dimensional response surfaces of sulfate removal percentage and the relationship between pH and initial sulfate concentration at the center point. It presents a simultaneous effect of pH and initial sulfate concentration on the sulfate removal percentage. It illustrates that nearly the pH=7 and almost the center of initial sulfate concentration range have 12

positive effect on the removal. Looking more closely to the figure, the maximum sulfate removal (95.806%) was observed for pH of 7.19 and initial sulfate concentration of 2153.15 mg/L. According to this figure the removal of sulfate of >88% was observed in the initial sulfate concentration between 1250 mg/L and 2875 mg/L and pH more than 6.125 when the value of other independent variables were kept at the center point. From this figure, sulfate removal will be decreased by increasing the initial sulfate concentration from 3750 mg/L when the value of other independent variables was kept at the center point. 3.2.3. Effect of Various Individual Parameters

Effect of various individual parameters on the sulfate removal is shown in Fig. 7. According to the Fig. 7.a by increasing the pH in any constant concentration of sulfate, the sulfate removal rate is increased and at a pH level of around 7.2 the removal rate is maximized. However the bacteria showed better activity in basic pH range than in acidic pH range. But as it can be seen in the figure in acid pH range up to pH=6 the decrease in bacterial activity is not significant and therefore treatment of acidic wastewaters seems possible using these bacteria. However in order to increase the activity of the bacteria and also to have a chemical pretreatment, using NaOH could be effective. According to Fig. 7.b the increase in sulfate concentration up to a specific amount will result in an increase to its removal because the sulfate itself acts as a substrate in the medium for SRB and lack of it could decrease the activity level of the bacteria, so increasing the initial concentration of sulfate up to optimum point will increase the level of the bacterial activity, while more increase will result in a negative effect on bacterial activity and also sulfate removal. Higher concentrations of sulfate reverse in the activity of the sulfate reducing bacteria [4]. This phenomenon is due to inhibition of metabolism and as a result inhibition of SRB growth at higher sulfate concentration [44]. Moreover, in the condition of high sulfate concentration the amount of dissolved sulfide increases, therefore, the redox potential increases which causes inhibitory effect on the sulfate reducing bacteria [45]. The effect of COD/SO42- on the sulfate removal is shown in Fig. 7.c. The importance of the COD/SO42- term is that the reaction and the process could be limited by one of the two substrates (sulfate or carbon source) and also in the cost of preparing the pure carbon source. The sulfate reducing bacteria do their activity as oxidation-reduction reactions, in which the carbon source is 13

oxidized and the released electron from this reaction is used to reduce sulfate. Therefore the number of electron which is transferred between these two reactions determines the optimized ratio of the two substrates. In the other words, the COD/SO42- is important controlling parameter of the electron flow in the oxidation-reduction reactions [46]. It can be seen in the following reaction that in order to reduce one mole of sulfate to one mole of sulfide, 8 electrons are required:

SO42  8e   4H 2 O  S 2  8OH 

(9)

Depending on the carbon source chosen and the number of electrons being released by oxidation process, the mole to mole ratio of carbon source over the sulfate is determined. Furthermore, considering the calculation form which the mentioned mole to mole ratio was estimated, the ratio of the necessary oxygen required for carbon oxidation (COD) over sulfate concentration could be determined. It is noteworthy that as Pavlina Kousi et al. have mentioned in their report, the minimum organic supplements required for reducing the given sulfate must be a little more than its stoichiometric ratio, because a specific amount of the energy released during the decomposition of organic compounds will be used in bacterial growth and maintenance processes [19]. The energy released during the bacterial growth is also dependent to carbon source and as it was mentioned earlier for example, bacteria in a medium with lactate lose more energy during the growth than in a medium with ethanol, so COD/SO42- in an environment with lactate carbon source is better to be more than in an environment with ethanol carbon source. At first, by increasing the mentioned ratio (keeping the initial concentration of sulfate constant and increasing the carbon source) the electrons required for reducing the sulfate is provided, but after a certain point due to the extra amount of COD and as a result lack of sulfate the oxidation process becomes difficult. It has also been reported by Gomez et al. that the reduction in sulfate removal process in much higher COD/SO42- due to the incomplete oxidation of carbon source, release of H+ and as a result some decrease in the pH level is considered to be a negative factor in bacterial activity [47]. In the reports presented by Choi and Rim it is suggested that if the COD/SO42- is more than 2.7, it will have a negative effect on the activity of SRB because the methanogens will compete over hydrogen and acetate with SRB. The optimized value calculated in the recent work using the software was 2.72 which has an adequate approximation to the value reported by Choi and Rim [48]. Nagpal et al. also found out that the optimized value for this ratio was between 1.5 and 2.25 which was enough to create the 14

maximum amount of sulfide [22]. In addition, Velasco et al. have reported that the complete oxidation of ethanol to carbon dioxide creates three times more hydrogen sulfide than the partial oxidation, and their maximum sulfide concentration was obtained at the COD/SO42- of 2.5 [49]. 3.2.4. Determining the Optimized Removal Conditions Using the Software

Finally the results which had the maximum desirability were given by the ANOVA analysis considering the effect of different parameters and the interactions between them. In the present work one of the desired results were chosen by software as the optimized condition which is shown in Table 7.

3.3. Treatment of sulfate and metal rich wastewater in the bioreactor The performance of sulfidogenic UASB rector regarding the removal of sulfate and heavy metal was assessed through determination of the bioreactor effluent parameters. The trend of changes in pH and sulfate in experiments of I, II, III, IV and V, also COD and sulfide in experiments of I and II were shown in Fig. 8 and the results of heavy metal removal were illustrated in Fig. 9. The pH of all modes was adjusted to 7.19, the obtained optimal condition. The sulfate concentration in mode I, II and III were almost 2153.15 (optimum concentration), but in mode IV and V, it was the concentration of sulfate in the real wastewater. In the last two modes, all the nutrients of Postgate medium and carbon source were solved in the real wastewater. The carbon source in all modes was ethanol, because of the highest removal of sulfate among ethanol, lactate and mixture of ethanol-lactate. As a result of SRB activity, the pH of effluent increased in all experiments except for II. In experiment II, the high level of COD resulted in incomplete oxidation of carbon source to acetate. As the incomplete oxidation of carbon source to acetate generated protons, the pH level decreased to 5.25 [47].

Using organic wastes or organic wastewaters with high level of COD has two main problems. The capability of SRB in COD reduction as their substrate is limited, so environmental issue related to abundance of organic carbon released from organic substrate and incomplete oxidation makes the second stage of treatment inescapable. As it can be seen in the Fig. 8.c. the COD of effluent in experiment II reached to nearly 7000 at the end of experiment. The other problem is the low efficiency of SRB performance, due to the high value of COD/SO42-. The limited sulfate removal and sulfide production in Fig. 8.b. and Fig. 8.d. also less heavy metal removal in Figure 15

9 compared to experiment I showed the problem of COD/SO42- high value. Results from other studies have convinced this fact. Bratkova et al. showed the efficiency of COD removal in high organic loading bioreactor was 12–19% (COD reached 8825–9080 mg/L), because of the incomplete oxidation of organic compounds [50]. Moreover, Zhang and Wang used Chicken manure, dairy manure and sawdust as the organic substrates for treatment of synthetic Acid mine drainage by SRB. The effluent from their sulfidogenic UASB reactor was characterized with high values of COD (3100–13,800 mg/L). They demonstrated that the high percentage of heavy metal removal attributed to the sorption, since the sulfate removal was limited [51]. The amount of heavy metal removal efficiency can be deeply affected by the respective solubility and stability of any metal sulfide [52]. For example, the solubility product constant of nickel sulfide salt and Zn sulfide salt are 3 × 10−19 and 2 × 10−25, respectively [52]. In Fig. 9.a and b, it can be seen that the removal percentage of Zn is more than Ni. The substantial percentage of Zn removal can be explained due to its low solubility product with sulfide, while the product constant of nickel sulfide salt is much more than Zn. Moreover, Nickel sulfide precipitation has been found very pH-dependent [53]. The other reason for high percentage of Zn removal is due to the feature of Zn for being bounded by extracellular polymeric substances (EPS) extracted from sulfate reducing bacteria. Charge density, attractive interaction and conformation types of the polymer are the parameters which determine the affinity of different metal ions [54]. However, no studies have been found on the ability of Ni2+. Nevertheless, an extra treatment stage done by H2S from bioreactor can increase the removal of Ni. As it can be seen in Fig. 9.a., the red lines show the differences in treatment with and without the extra treatment stage. During the last days, this stage was more effective, since as Fig. 8.d. shows the dissolved S2- decreased, thus H2S was more in these days. In addition of the produced bicarbonate by SRB activity, hydrogen sulfide gas output increased the pH of wastewater. In Fig. 8.f., the best result in sulfate removal was related to mode III, because nickel removal was done out of bioreactor and bacteria activity did not hinder by nickel toxicity. Moreover, bacteria had better performance in mode V in comparison with mode IV, because in mode V the bacteria had been adapted to Ni before experiment. In Fig. 9.c. nickel removal was more efficient in mode III, because mixing of bioreactor effluent and real wastewater made the volume twice. 16

In mode IV and V a portion of removal was related to biosorption of Ni by biomass. Nevertheless as Ni removal was done inside the bioreactor in mode IV and V and SRB had been adapted only for experiment V, the result in experiment V was more efficient. 4. Conclusion

Concluding, the experiments of carbon sources effect showed that the best growth happened in the presence of the carbon source consisted of sodium lactate while the best sulfate removal occurred in the medium with ethanol. The influences of pH, initial concentration of sulfate, COD/SO42- and CODethanol/CODtotal and their interactions were investigated by a RSM and ANOVA. The desirability function optimization showed that the optimum removal (97.87%) was obtained at a pH, initial concentration of sulfate, COD/SO42- and CODethanol/CODtotal of 7.19, 2153.15 mg/L, 2.72 and 1, respectively. Finally, Results from batch experiments proved that bioprecipitation by SRB is a promising method; nevertheless, conditions affected the result. For example, using organic wastewater with high level of COD reduced the efficiency of sulfate and heavy metal removal. In the condition that heavy metal removal was not significant, gas from bioreactor had much potential to enhance removal. In the condition which bacteria grew in the media containing heavy metal, a portion of removal attributed to sorption on biomass. In order to minimize the toxic effect of metals on bacteria, adaptation could result better activity of bacteria. Acknowledgements

This work was supported by Amirkabir University of Technology (Tehran Polytechnic) and Iranian Research Organization for Science and Technology (IROST).

17

REFESENCES [1]

S.S. Mohanty, T. Das, S.P. Mishra, G.R. Chaudhury, Kinetics of SO4−2 reduction under different growth media by sulfate reducing bacteria, BioMetals. 13 (2000) 73–76.

[2]

J. Lu, T. Chen, J. Wu, P.C. Wilson, X. Hao, J. Qian, Acid tolerance of an acid mine drainage bioremediation system based on biological sulfate reduction., Bioresour. Technol. 102 (2011) 10401–10406.

[3]

C. Candeias, P.F. Ávila, E. Ferreira da Silva, A. Ferreira, A.R. Salgueiro, J.P. Teixeira, et al., Acid mine drainage from the Panasqueira mine and its influence on Zêzere river (Central Portugal), J. African Earth Sci. 99 (2013) 705–712.

[4]

S. Al Zuhair, S. Al Zuhair, M. H El-Naas, H. Al Hassani, Sulfate inhibition effect on sulfate reducing bacteria, J. Biochem. Technol. 1 (2008) 39–44.

[5]

H.T.Q. Kieu, E. Müller, H. Horn, Heavy metal removal in anaerobic semi-continuous stirred tank reactors by a consortium of sulfate-reducing bacteria, Water Res. 45 (2011) 3863–3870.

[6]

H. Demiral, I. Demiral, F. Tümsek, B. Karabacakoǧlu, Adsorption of chromium(VI) from aqueous solution by activated carbon derived from olive bagasse and applicability of different adsorption models, Chem. Eng. J. 144 (2008) 188–196.

[7]

H. Massara, C.N.C. Mulligan, J. Hadjinicolaou, Hexavalent chromium removal by viable, granular anaerobic biomass., Bioresour. Technol. 99 (2008) 8637–8642.

[8]

T.J.K. Visser, S.J. Modise, H.M. Krieg, K. Keizer, The removal of acid sulphate pollution by nanofiltration, Desalination. 140 (2001) 79–86.

[9]

R.K. Sani, B.M. Peyton, L.T. Brown, Copper-induced inhibition of growth of Desulfovibrio desulfuricans G20: assessment of its toxicity and correlation with those of zinc and lead, Appl. Environ. Microbiol. 67 (2001) 4765–4772.

[10]

M. Zhang, H. Wang, Preparation of immobilized sulfate reducing bacteria (SRB) granules for effective bioremediation of acid mine drainage and bacterial community analysis, Miner. Eng. 92 (2016) 63–71.

[11]

Ihsanullah, A. Abbas, A.M. Al-Amer, T. Laoui, M.J. Al-Marri, M.S. Nasser, et al., Heavy metal removal from aqueous solution by advanced carbon nanotubes: Critical review of adsorption applications, Sep. Purif. Technol. 157 (2016) 141–161.

[12]

K. Jalali, S.A.B. M, S.A. Baldwin, The role of sulphate reducing bacteria in copper removal from aqueous sulphate solutions, Water Res. 34 (2000) 797–806.

[13]

B. Brahmacharimayum, P.K. Ghosh, Sulfate bioreduction and elemental sulfur formation in a packed bed reactor, J. Environ. Chem. Eng. 2 (2014) 1287–1293.

[14]

R. Ahmadi, A. Rezaee, M. Anvari, H. Hossini, S.O. Rastegar, Optimization of Cr(VI) removal by sulfate-reducing bacteria using response surface methodology, Desalin. Water Treat. 57 (2016) 11096–11102.

[15]

A M. Jiménez-Rodríguez, M.M. Durán-Barrantes, R. Borja, E. Sánchez, M.F. Colmenarejo, F. Raposo, Heavy metals removal from acid mine drainage water using biogenic hydrogen sulphide and effluent from anaerobic treatment: effect of pH., J. Hazard. Mater. 165 (2009) 759–65.

[16]

R. Singh, A. Kumar, A. Kirrolia, R. Kumar, N. Yadav, N.R. Bishnoi, et al., Removal of sulphate, 18

COD and Cr(VI) in simulated and real wastewater by sulphate reducing bacteria enrichment in small bioreactor and FTIR study., Bioresour. Technol. 102 (2011) 677–682. [17]

T. Jong, D.L. Parry, Removal of sulfate and heavy metals by sulfate reducing bacteria in shortterm bench scale upflow anaerobic packed bed reactor runs, Water Res. 37 (2003) 3379–3389.

[18]

M. Zhang, H. Wang, X. Han, Preparation of metal-resistant immobilized sulfate reducing bacteria beads for acid mine drainage treatment, Chemosphere. 154 (2016) 215–223.

[19]

P. Kousi, E. Remoundaki, A. Hatzikioseyian, F. Battaglia-Brunet, C. Joulian, V. Kousteni, et al., Metal precipitation in an ethanol-fed, fixed-bed sulphate-reducing bioreactor., J. Hazard. Mater. 189 (2011) 677–684.

[20]

P. Elliott, S. Ragusa, D. Catcheside, Growth of sulfate-reducing bacteria under acidic conditions in an upflow anaerobic bioreactor as a treatment system for acid mine drainage, Water Res. 32 (1998) 3724–3730.

[21]

M.M.M. Gonçalves, A C. A Da Costa, S.G.F. Leite, G.L. Sant’Anna, Heavy metal removal from synthetic wastewaters in an anaerobic bioreactor using stillage from ethanol distilleries as a carbon source., Chemosphere. 69 (2007) 1815–1820.

[22]

S. Nagpal, S. Chuichulcherm, L. Peeva, A. Livingston, Microbial sulfate reduction in a liquid– solid fluidized bed reactor, Biotechnol. Bioeng. 70 (2000) 370–380.

[23]

C. Mack, J.E. Burgess, J.R. Duncan, A sulphide-extraction membrane bioreactor for the recovery of metals from industrial process wastewater : ph optimization, Event (London). (2004) 618–623.

[24]

S. Dubey, S. Nath, Y. Chandra, Optimization of removal of Cr by γ-alumina nano-adsorbent using response surface methodology, Ecol. Eng. 97 (2016) 272–283.

[25]

M. Dastkhoon, M. Ghaedi, A. Asfaram, A. Goudarzi, Ultrasonics Sonochemistry Improved adsorption performance of nanostructured composite by ultrasonic wave : Optimization through response surface methodology , isotherm and kinetic studies, Ultrason. - Sonochem. 37 (2017) 94– 105.

[26]

U. Guyo, T. Makawa, M. Moyo, T. Nharingo, B.C. Nyamunda, T. Mugadza, Application of response surface methodology for Cd(II) adsorption on maize tassel-magnetite nanohybrid adsorbent, J. Environ. Chem. Eng. 3 (2015) 2472–2483.

[27]

J.R. Postgate, The Sulphate-reducing Bacteria, CUP Archive, 1979.

[28]

P. Lens, M. Vallerol, G. Esposito, M. Zandvoort, Perspectives of sulfate reducing bioreactors in environmental biotechnology, Rev. Environ. Sci. Biotechnol. 1 (2002) 311–325.

[29]

B. Bharati, G.P. Kumar, A Study on the efficiency of five different carbon sources on sulfate reduction, J. Environ. Res. Dev. 7 (2012) 416–420.

[30]

G. Muyzer, A.J.M. Stams, The ecology and biotechnology of sulphate-reducing bacteria, Nat. Rev. - Microbiol. 6 (2008) 441–454.

[31]

H.J. Laanbroek, H.J. Geerligs, L. Sijtsma, H. Veldkamp, Competition for sulfate and ethanol among Desulfobacter, Desulfobulbus, and Desulfovibrio species isolated from intertidal sediments, Appl. Environ. Microbiol. 47 (1984) 329–334.

[32]

A. Shokri, K. Mahanpoor, D. Soodbar, Evaluation of a modified TiO2 (GO-B-TiO2) photo catalyst for degradation of 4-nitrophenol in petrochemical wastewater by response surface methodology based on the central composite design, J. Environ. Chem. Eng. 4 (2016) 585–598.

[33]

P.-E. Jørgensen, T. Eriksen, B.K. Jensen, P.E. Jorgensen, T. Emksen, Estimation of viable biomass in wastewater and activated sludge by determination of ATP, oxygen utilization rate and FDA 19

hydrolysis, Water Res. 26 (1992) 1495–1501. [34]

M. Altun, E. Sahinkaya, I. Durukan, S. Bektas, K. Komnitsas, Arsenic removal in a sulfidogenic fixed-bed column bioreactor, J. Hazard. Mater. 269 (2014) 31–37.

[35]

American Public Health Association, & American Water Works Association. (1965). Standard methods for the examination of water and wastewater. American Public Health Association. (1999) 733.

[36]

R. Cord-Ruwisch, A quick method for the determination of dissolved and precipitated sulfides in cultures of sulfate-reducing bacteria, J. Microbiol. Methods. 4 (1985) 33–36.

[37]

C. White, G.M. Gadd, A comparison of carbon/energy and complex nitrogen sources for bacterial sulphate-reduction: potential applications to bioprecipitation of toxic metals as sulphides, J. Ind. Microbiol. 17 (1996) 116–123.

[38]

R. Deitrich, S. Zimatkin, S. Pronko, Oxidation of ethanol in the brain and its consequences, Alcohol Res. Heal. 29 (2006) 266.

[39]

M.T. Ferreira, A.S. Manso, P. Gaspar, M.G. Pinho, A.R. Neves, Effect of oxygen on glucose metabolism: utilization of lactate in Staphylococcus aureus as revealed by in vivo NMR studies, PLoS One. 8 (2013) e58277.

[40]

P.J. Russell, Biology: The Dynamic Science, Volume 1 w/PAC, Cengage Learning.

[41]

M. Wheelis, Principles of modern microbiology, Jones & Bartlett Publishers, 2011.

[42]

L. Cited, T.Y. Wei, C.C. Wan, Heterogeneous photocatalytic oxidation of phenol with titanium dioxide powders, Ind. Eng. Chem. Res. 30 (1991) 1293–1300.

[43]

F. Ay, E.C. Catalkaya, F. Kargi, A statistical experiment design approach for advanced oxidation of Direct Red azo-dye by photo-Fenton treatment., J. Hazard. Mater. 162 (2009) 230–236.

[44]

O.O. Oyekola, R.P. van Hille, S.T.L. Harrison, Kinetic analysis of biological sulphate reduction using lactate as carbon source and electron donor: Effect of sulphate concentration, Chem. Eng. Sci. 65 (2010) 4771–4781.

[45]

S. Dev, S. Roy, D. Das, J. Bhattacharya, Improvement of biological sulfate reduction by supplementation of nitrogen rich extract prepared from organic marine wastes, Int. Biodeterior. Biodegradation. 104 (2015) 264–273.

[46]

S.V. Mohan, N.C. Rao, K.K. Prasad, P.N. Sarma, Bioaugmentation of an anaerobic sequencing batch biofilm reactor (AnSBBR) with immobilized sulphate reducing bacteria (SRB) for the treatment of sulphate bearing chemical wastewater, Process Biochem. 40 (2005) 2849–2857.

[47]

D. Villa-Gomez, H. Ababneh, S. Papirio, D.P.L. Rousseau, P.N.L. Lens, Effect of sulfide concentration on the location of the metal precipitates in inversed fluidized bed reactors., J. Hazard. Mater. 192 (2011) 200–207.

[48]

E. Choi, J.M. Rim, Competition and inhibition of sulfate reducers and methane producers in anaerobic treatment, Water Sci. Technol. 23 (1991) 1259–1264.

[49]

A. Velasco, M. Ramírez, T. Volke-Sepúlveda, A. González-Sánchez, S. Revah, Evaluation of feed COD/sulfate ratio as a control criterion for the biological hydrogen sulfide production and lead precipitation., J. Hazard. Mater. 151 (2008) 407–413.

[50]

S. Bratkova, B. Koumanova, V. Beschkov, Biological treatment of mining wastewaters by fixedbed bioreactors at high organic loading, Bioresour. Technol. 137 (2013) 409–413.

[51]

M. Zhang, H. Wang, Organic wastes as carbon sources to promote sulfate reducing bacterial 20

activity for biological remediation of acid mine drainage, Miner. Eng. 69 (2014) 81–90. [52]

G.K. Mothe, K. Pakshirajan, G. Das, Heavy metal removal from multicomponent system by sulfate reducing bacteria: mechanism and cell surface characterization, J. Hazard. Mater. 324 (2016) 62-70.

[53]

F.D. Reis, A.M. Silva, E.C. Cunha, V.A. Leão, Application of sodium- and biogenic sulfide to the precipitation of nickel in a continuous reactor, Sep. Purif. Technol. 120 (2013) 346–353.

[54]

J. Wang, Q. Li, M.-M. Li, T.-H. Chen, Y.-F. Zhou, Z.-B. Yue, Competitive adsorption of heavy metal by extracellular polymeric substances (EPS) extracted from sulfate reducing bacteria., Bioresour. Technol. 163 (2014) 374–376.

21

Fig. 1. Oxidation pathways of (a) lactate to acetate and (b) ethanol to acetate

Sulfate Cnoncentration (mg/L)

3000

a

2500 2000 1500 1000

Lactate 500

Ethanol Lactate & Ethanol

0 0

5

10

15

20

25

30

35

40

Time (hr)

22

45

50

55

Biomass Concentration (mg/L)

3500

b

3000 2500 2000 1500

Lactate

1000

Ethanol 500

Lactate & Ethanol

0 0

5

10

15

20

25 30 35 Time (hr)

40

45

50

55

Fig. 2. The changes of (a) sulfate concentration and (b) biomass concentration through time, using sulfate reducing consortium; pH=7, SO42-=2500 mg/L, COD/SO42-=2, and T=35°C

23

24

Fig. 3. (a) Normal probability of residuals, (b) residuals vs. predicted and (c) correlation of actual and predicted values by the model

Fig. 4. The perturbation diagram for the significant parameters in percentage of sulfate removal in function of coded values, the coded values in the base point (code=0) are: pH=7, SO42-=2500 mg/L, COD/SO42-=2.25, CODethanol/CODtotal=0.5

Fig. 5. Surface response for sulfate removal, in function of pH, SO42; T= 35°C, COD/SO42-=2.25, CODethanol/CODtotal=0.5.

25

Fig. 6. Contour plot for sulfate removal, in function of pH, SO42; T= 35°C, COD/SO42-=2.25, CODethanol/CODtotal=0.5.

26

Fig. 7. Effect of various individual parameters: pH (a), SO42- (b), COD/SO42- (c), on the sulfate removal. One parameter is varied while the others are kept constant at their center points.

27

Fig. 8. Trend of (a) pH, (b) Sulfate, (c) COD and (d) Sulfide in experiments I, II and (e) pH and (f) sulfate in experiments III, IV and V.

Fig. 9. Trend of heavy metal removal in experiments I (a), II (b) and III, IV and V (c).

28

Table 1 The optimum conditions for sulfate removal by SRB obtain from literature Parameter

Optimize value of literatures o

Temperature pH Initial SO42COD/SO42-

35 C 7 mg/L 2500 2-2.5

Reference [4] [4] [4] [21,22]

Table 2 Experimental range and levels of independent variables Range and levels (coded) Factors

-1

0

+1

pH

A

6

7

8

SO42- (mg/L)

B

1000

2500

4000

COD/SO42-

C

0.5

2.25

4

CODethanol/CODtotal

D

0

0.5

1

Table 3 Sequence of runs for the CCD Run

Variable properties

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

pH 7 8 8 6 7 7 8 6 6 6 8 7 8 7 8 7 7 6 8 7 7 7

SO424000 4000 4000 4000 1000 2500 1000 4000 4000 2500 1000 2500 1000 2500 1000 2500 2500 1000 4000 2500 2500 2500

Sulfate removal% COD/SO422.25 0.5 4 4 2.25 2.25 0.5 0.5 4 2.25 0.5 2.25 4 0.5 4 2.25 4 4 4 2.25 2.25 2.25

CODethanol/CODtotal 0.5 1 1 1 0.5 0.5 0 1 0 0.5 1 0 1 0.5 0 0.5 0.5 0 0 0.5 0.5 0.5

29

Actual 0.51 0.39 0.45 0.41 0.79 0.96 0.55 0.36 0.4 0.77 0.68 0.76 0.72 0.75 0.63 0.90 0.89 0.60 0.44 0.95 0.95 0.94

Predicted 0.54 0.40 0.49 0.44 0.78 0.91 0.56 0.35 0.36 0.82 0.64 0.87 0.73 0.78 0.65 0.91 0.87 0.60 0.41 0.91 0.91 0.91

23 24 25 26 27 28 29 30

1000 2500 1000 4000 2500 4000 1000 2500

6 7 6 8 7 6 6 8

4 2.25 0.5 0.5 2.25 0.5 0.5 2.25

1 0.5 1 0 1 0 0 0.5

0.70 0.91 0.55 0.32 0.98 0.30 0.50 0.90

0.68 0.91 0.59 0.32 0.96 0.27 0.50 0.87

Table 4 Characterization of the electroplating wastewater Parameter

Real wastewater

Synthetic solution containing of Ni

Synthetic solution containing of Zn

pH

2.27

7.2

7.2

16.7

-

-

SO4 (mg/L)

2499.97

81.84

73.48

Ni (mg/L)

72.469

50

-

Zn (mg/L)

ND

-

50

COD (mg/L) 2-

Table 5 Information on the experiments. Experiment

Wastewater

Heavy metal Treatment

Gas Treatment

Adaptation

Be able to be compared with

I

Synthetic

Outside of Bioreactor

YES

NO

II

II

Synthetic

Outside of Bioreactor

NO

NO

I,III

III

Real

Outside of Bioreactor

NO

NO

II,IV

IV

Real

Inside of Bioreactor

NO

NO

III,V

V

Real

Inside of Bioreactor

NO

YES

IV

Table 6 ANOVA for response surface quadratic model for the response variables Source of variation Model A-pH

Sum of squares

Degree of freedom

Mean Square

F-Value

1.4 0.013

14 1

0.10 0.013

54.20 7.22

p-value probability <0.0001a 0.0169a

B-SO42-

0.25

1

0.25

137.68

<0.0001a

C-COD/ SO42-

0.039

1

0.0039

21.21

0.0003a

DCODethanol/CODtotal

0.030

1

0.030

16.46

0.0010a

AB

6.250E-004

1

6.250E-004

0.34

0.5695b

AC

6.250E-004

1

6.250E-004

0.34

0.5695b

30

AD

4.000E-004

1

4.000E-004

0.22

0.6484b

BC

1.000E-004

1

1.000E-004

0.054

0.8192b

BD

3.025E-003

1

3.025E-003

1.64

0.2202b

CD

6.250E-004

1

6.250E-004

0.34

0.5695b

A2

0.010

1

0.010

5.66

0.0310b

B2

0.16

1

0.16

86.62

<0.0001a

C2

0.016

1

0.016

8.65

0.0101a

D2

2.113E-003

1

2.113E-003

1.14

0.3019b

Residual

0.028

15

1.848E-003

Lack of Fit

0.025

10

2.468E-003

4.06

0.0678b

Pure Error

3.043E-003

5

6.085E-004

Correction Total

1.43

29

2

R

0.9806 2

Adj. R

0.9625

Values of ‘p-value’ <0.0500 indicate model terms are significant. a

Significant.

b

Not significant.

Table 7 Determinate optimized run by CCD method

Number

pH

SO42-

COD/SO42-

1

7.19

2153.15 2.72

31

CODethanol/CODtotal removal

Desirability

1

0.99805

0.978674