Introduction

Access to clean and safe water is one of the fundamental principles of global prosperity. However, produced water—a by-product of oil and gas extraction—poses significant environmental challenges due to its high levels of salts, mineral ions, hydrocarbons, organic acids, and polycyclic aromatic hydrocarbons (PAHs), threatening ecosystems and biodiversity. The quality of produced water (the amount of various contaminants in it) depends on the geographical location, time, and operational conditions of oil and gas extraction. Globally, the volume of produced water is considerable. The water-to-oil ratio worldwide is estimated to be around 2:1 to 3:1, meaning that for every barrel of oil extracted, 2 to 3 barrels of water are produced. According to a report (in 2020), the amount of produced water has reached 250 million barrels per day1,2,3,4,5. Considering that Iran is one of the oil-producing countries, it can be concluded that the volume of produced water in Iran is very significant. This high volume of produced water, coupled with Iran’s existing water stresses, requires effective management and treatment strategies to reduce environmental impacts and potentially reuse water in oilfield operations.

Conventional physical and chemical separation methods for treating produced water are often energy-intensive, inefficient, along with the production of secondary waste, space-intensive, and costly. The membrane process in wastewater treatment offers several advantages over other separation processes as follows: lower energy consumption since it avoids phase changes, the ability to remove even trace contaminants, and high efficiency in eliminating organic compounds, salts, heavy metals, and viruses. It also requires less space and equipment, features a simple design, and is scalable for industrial use, with minimal pressure drop and excellent mass transfer6,7,8. Today, polymeric membranes are more widely used in water purification due to their adaptable pore formation, wide range of polymer materials that can be used in membranes, greater flexibility, less space required for installation, and lower operating cost compared to inorganic membranes. These membranes also offer mechanical strength, chemical stability, and flexibility9. However, unmodified polymer membranes have several limitations, including low permeability, susceptibility to fouling, and restricted chemical resistance10,11,12. To address these subjects, mixed matrix membranes (MMMs) incorporating nanoparticles have been developed. In these membranes (MMMs), the nanomaterials are embedded within the polymer matrix as the filler13. Metals and metal oxides are commonly used as fillers because of their high surface area, exceptional adsorption capacity, and hydrophilic properties14,15. Titanium dioxide (TiO2) nanoparticles possess remarkable hydrophilicity, high thermal stability, and excellent chemical and permeability characteristics. These features make TiO2 nanoparticles ideal fillers for the development of novel mixed matrix membranes with enhanced hydrophilicity15,16.

Polysaccharides are plentiful in nature, cost-effective, and biodegradable, making them potential materials for membrane fabrication. Chitosan (CS) is a well-known polysaccharide that features notable properties such as excellent biocompatibility, antibacterial effects, strong biodegradability, non-toxicity, and numerous hydrophilic hydroxyl and amino groups17,18,18. However, CS has some disadvantages, such as low mechanical strength, low porosity (dense structure), and low stability. The dense structure of chitosan membranes originates from localized aggregation of tangled polymer chains forming tightly packed fiber bundles, which create small pores and low porosity. The dense structure of the pure CS membrane results in low filtration flux18,19. The embedding of nanomaterials as filler in the CS matrix is a procedure to address these disadvantages18.

In most studies conducted on the use of CS membranes for wastewater treatment and pollutant removal, commercial CS has been used17,18,18. However, in this study, CS synthesized from shrimp shells was used to synthesize the membrane. Additionally, most research conducted using chitosan membranes has focused on treating synthetic wastewaters17,18,18. However, this study investigated the performance of chitosan-based membranes for treating real wastewater (produced water from the National Iranian South Oil Company (Ahvaz)). In this study, to increase the filtration flux of the chitosan membrane and enhance its performance in successive passes, titanium dioxide (TiO2) nanoparticles were added as fillers to synthesize a chitosan/TiO2 mixed matrix membrane. This membrane was used to treat oil-produced water with a high pollution load, specifically chemical oxygen demand (COD). The effects of process parameters, including operating pressure, volumetric flow rate, and initial COD concentration, on membrane performance were examined using the central composite design method.

Materials and methods

Materials

Shrimp shells (sourced from Bushehr province) were used for chitosan synthesis. Other chemicals used in this study include sodium hydroxide (NaOH from Merck, ≥ 99.0%), hydrochloric acid (HCl from Merck, 37%), ethanol (C2H5OH from Merck, ≥ 99.9%.), acetic acid (CH3COOH from Merck, ≥ 99.8%), titanium dioxide (TiO2 NPs from AERDSIL EVONIK), and bovine serum albumin (BSA from Sigma-Aldrich). The membrane feed was prepared from effluent from oil extraction operations (produced water) from the National Iranian South Oil Company (Ahvaz). Double-distilled water was used in the stages of chitosan synthesis, membrane fabrication (including chitosan and mixed matrix membranes), and dilution of the effluent to prepare feeds with different contamination levels.

Material preparation

Synthesis of chitosan from shrimp shells

Chitosan was prepared from shrimp shells according to the method presented by Khadijeh Amirsadat et al.20 with the assistance of ultrasonic waves. By using ultrasonic waves, the kinetic energy efficiency and process rate are increased due to the release of energy at the microscopic level. This increases the mass transfer rate and reduces the process time. The steps of chitosan synthesis are presented in Fig. 1. Using the acid-base titration method, the degree of deacetylation of the synthesized chitosan was found to be 90.65%.

Fig. 1
figure 1

Chitosan synthesis from shrimp shells.

Full size image

Preparation of chitosan membrane and chitosan/TiO2 NPs mixed matrix membrane

To investigate the effect of TiO2 nanoparticles as a filler and the difference between chitosan and mixed matrix membranes, a chitosan sheet membrane was prepared. 700 mg of chitosan (particle size: 210–250 μm) in 50 ml of acetic acid solution (2% v: v) at 60 °C was stirred for one hour using a shaker incubator equipped with a temperature controller until completely dissolved, and a pale yellow gel-like (diluted) solution was obtained. To remove any solid impurity particles, the solution was passed through filter paper and poured into a laboratory Petri dish with an internal diameter of 10 cm. The Petri dish containing the solution was kept in an oven at 60 °C for one day to evaporate the acetic acid completely. After the acetic acid was completely evaporated, the Petri dish containing the chitosan membrane film was placed in the refrigerator for 15 min (to expose it to cold air). To separate the prepared membrane film from the bottom of the glass Petri dish, the Petri dish containing the membrane film was immersed in a 1 M NaOH solution. In the last step, the prepared membrane was washed with double-distilled water until the pH of the washing water reached neutral pH and then dried in an oven at 60 °C21.

To prepare the matrix membrane of the chitosan/TiO2 NP, after preparing a gel-like (diluted) solution by dissolving 700 mg of chitosan in acetic acid solution and filtering it, a certain amount of TiO2 nanoparticles was added to the solution as per Table 1. The resulting mixture was stirred for 20 min using ultrasonic waves (frequency 40 kHz and power 180 W). The resulting mixture was poured into a Petri dish with a diameter of 10 cm and placed in an oven at 60 °C for one night to evaporate the acetic acid. The remaining steps, such as exposure to cold air (for 15 min) and then immersing in sodium hydroxide solution and washing with distilled water, and drying in an oven at 60 °C, were performed as in the previous step. Figure 2 shows the steps for preparing the chitosan membrane and the mixed matrix membrane of chitosan/TiO2. The composition of prepared membranes is presented in Table 1.

Table 1 Composition of prepared membranes.
Full size table
Fig. 2
figure 2

Preparation steps of chitosan and 3 M chitosan/TiO2 membranes.

Full size image

Membrane characterization

FTIR analyses for detecting the surface chemistry (Spectroum GX, manufacturer: PerkinElmer), water contact angle analysis for detecting the hydrophilicity (JIKAN, CAG-20 SE model), SEM and FE-SEM analyses from the surface and cross-sectional area of the membrane (TESCAN, TESCAN-Vega3, ZEISS, SIGMA 500) were used to identify the chemical structure and morphology of the synthesized membrane with maximum operation. The mechanical properties (the stress-strain curves) of M1 and M4 samples (50 mm*6 mm) were analyzed using a Tensile Strength Tester at P = 50 N. The roughness of chitosan and Chitosan/TiO2 surfaces was measured using the AFM technique (WITec, Germany). To determine the pore size distribution of the synthesized membranes, BET (nitrogen gas adsorption and desorption) analysis was used using a BELSORP mini II, Japan. To perform this analysis, the membrane was cut into small pieces (0.5 mm × 0.5 mm) and used for analysis.

Produced water treatment

Produced water analyze

The level of produced water contamination was expressed in terms of COD, TSS, and TDS. To determine the COD (Chemical Oxygen Demand) level, standard kits from Pars Chemical Research Company (Beh-Azma), a COD reactor, and a spectrophotometer (Jasco730, Japan) were used. To assess the Total suspended solids (TSS) in wastewater, standard D-5907 was used22. Initially, the filter paper was placed in an oven at 105 °C for one hour to remove moisture. After cooling the filter paper in a desiccator for 30 min, its initial mass (minitial) was measured using a highly accurate balance. Next, 100 mL of wastewater was passed through the filter paper, which was then dried at 105 °C for one hour. Once cooled in a desiccator, the filter paper’s weight was measured again (mfinal), and the TSS value (mg/L) was calculated using Eq. (1)22. In this equation, V represents the volume of the solution (100 mL = 0.1 L)

$$\:TSS=\frac{{m}_{final}-{m}_{initail}}{V}$$
(1)

The amount of Total Dissolved Solids in the produced water (TDS) was also measured according to standard D-590722. First, to completely remove moisture from the Petri dish, it was placed in an oven at 180 °C for one hour. After cooling in a desiccator, its initial mass (md, initial) was recorded using a four-decimal balance. Then, 50 mL of the wastewater sample was poured into the Petri dish, which was placed on a heater stirrer to evaporate the water thoroughly through heating. Next, the Petri dish was again placed in an oven at 180 °C for one hour. Finally, after cooling in a desiccator, the final mass (md, final) was measured, and the TDS concentration (mg/L) was calculated using Eq. (2)22.

$$\:TDS=\frac{{m}_{d,\:final}-{m}_{d,\:initial}}{V}$$
(2)

Also, the pH of the produced water was determined. Table 2 shows the characterization of the produced water sample used in this study.

Table 2 Characterization of produced water sample.
Full size table

The membrane system

To perform the wastewater treatment process using chitosan and chitosan/TiO2 membranes, a membrane system consisting of a membrane holding module, a water pump, a tank for the input wastewater (feed), a tank for the treated wastewater, two pressure gauges, and pipes and fittings was used, the schematic and photo of which are presented in Fig. 3. The holder used is a disc type and made of Teflon, and the desired membrane film with a diameter of 4 cm is placed inside it. After sealing the system, the tests were conducted by examining the effect of various parameters. The general procedure of the tests was that a certain amount of the produced water with a certain contamination (COD) was considered as the input for the membrane process and was introduced into the system as feed by a pump. The pressure was adjusted to the desired range using a pump. Dial gauges (one on the flow entering the membrane module and the other on the retentate flow) were used to adjust and record the pressure. The correct amount of pressure applied to the membrane film surface was considered equal to the average of the numbers shown by these two gauges. Considering that the permeate flow is discharged to the tank at atmospheric pressure and the gauge shows the relative pressure, the pressure gauge was not installed on the permeate flow. After ensuring the membrane system is sealed (no leaks), the pressure can be adjusted to the desired number by gradually closing the retentate flow valve.

Evaluation of permeability, fouling, and separation properties of polymer membranes

The pure water flux (PWF) test is one of the common methods for evaluating the performance of membranes for wastewater treatment. This test was performed for both chitosan and chitosan/TiO2 mixed matrix membranes. To perform this test, the membrane system in question was first thoroughly rinsed, and the system’s sealing was checked. In the next step, the membrane was placed in the membrane holding module. All membrane tests were performed at ambient temperature. Distilled water was poured into the feed tank, and the pump was turned on to allow the pure water to flow past the membrane for several minutes without pressure. After ensuring the sealing and absence of leakage of the system, the pressure was slowly increased to 3 bar by gradually closing the retentate flow valve, and when the pressure reached 3 bar, the measurement of the permeate flow rate began. The permeate volume was measured at 5-minute intervals until a constant volume was reached. Equation (3) was used to measure the pure water flux (L/(m2.h))23,24.

$$PWF = \frac{V}{{A \times t}}$$
(3)

In the above equation, A is the effective membrane area, V is the permeate volume collected at a given time, and t is the sampling time interval.

Fig. 3
figure 3

The membrane system.

Full size image

The flux recovery ratio (FRR) was calculated for all synthesized membranes using the following method: First, pure water was filtered, and the initial pure water flux (PWF1) was measured. Then, an aqueous solution containing bovine serum albumin (BSA) at a concentration of 10 g/L was passed through the membrane for one hour, and the permeate flux (JBSA) was measured for one hour. After this, the membrane was washed with distilled water to remove the cake layer formed on its surface, which consists of proteins that were weakly bound during filtration. Following the wash, distilled water was passed through the membrane again at an operating pressure of 3 bar. The pure water flux (PWF2) was measured until a stable value was reached. The percentage of flux recovery (FRR, %), total fouling ratio (TFR, %), reversible fouling ratio (RFR, %), and irreversible fouling ratio (IRFR, %) were determined using Eq. (4)23.

$$\:FRR\:\left(\%\right)=\frac{PWF2}{PWF1}\times\:100;\:TFR\:\left(\%\right)=\frac{PWF1-{J}_{BSA}}{PWF1}\times\:100\:;\:\:RFR\:\left(\%\right)=\frac{PWF2-{J}_{BSA}}{PWF1}\times\:100;\:\:IRFR\left(\%\right)=TFR-RFR$$
(4)

After identifying the membrane with the highest pure water flux and the best pure water flux recovery ratio, wastewater treatment experiments were conducted using the chitosan/TiO2 mixed matrix membrane (M4), which demonstrated superior performance. The performance of the membrane was evaluated based on the percentage of Chemical Oxygen Demand (COD) reduction. To examine the influence of three parameters—initial COD concentration, volumetric flow rate, and operating pressure—experiments were designed using the central composite design method. Table 3 outlines the levels for the investigated parameters. A total of 20 experiments were conducted, as detailed in Table 4.

Table 3 Levels of parameters under investigation.
Full size table
Table 4 Results of experimental design for membrane-based water purification.
Full size table

For the experiments, a feed solution with the desired COD concentration was prepared by diluting the initial wastewater (C0). This feed, which has a specific pollutant concentration (C0), was pumped through the system under the operating conditions detailed in Table 4. The permeate stream was then collected, and the COD of the permeate (Cf) was measured. The percentage reduction in COD was calculated using Eq. (5). All experiments were carried out at ambient temperature.

$$COD{\text{ }}reduction{\text{ }}\% = \frac{{\left( {C_{0} - C_{f} } \right) \times 100}}{{C_{0} }}$$
(5)

Results and discussion

Characterization of chitosan and chitosan/TiO2 3 M membranes

In the FTIR spectrum of the chitosan membrane (Fig. 4a), there is a broad peak at 3380 cm⁻¹, which is attributed to intermolecular hydrogen bonding. This bonding results from the overlap of the stretching vibrations of the hydroxyl (-OH) and amino (-NH₂) groups25. Additionally, a peak observed at 2921 cm⁻¹ corresponds to the symmetric and asymmetric vibrations of the C-H bond. Another peak at 1656 cm⁻¹ is related to the vibration of the -NH₂ group. The spectrum also features a peak at 1513 cm⁻¹, which can be assigned to the secondary amide (amide II)26. Lastly, the peaks appearing between 1055 and 1110 cm⁻¹ are associated with the asymmetric stretching vibrations of the C-O-C bond. In the FTIR spectrum of the chitosan/TiO2 (M4) mixed matrix membrane (Fig. 4a), the peak at 3406 cm−1 can be assigned to the bonding between Ti+4 of TiO2 nanoparticles and the amino and hydroxyl groups of chitosan (26 cm−1 shift for the peak at 3380 cm−1 in the FTIR spectrum of chitosan)27. There is a sharp peak at 1028 cm−1, which can be assigned to Ti-O-C28. Also, there are new peaks at 580–610 cm−1, which can be attributed to Ti-O bond stretching and Ti-O-Ti bond bridging29,30. These findings indicate the interactions between TiO2 and he functional groups of chitosan and thus, the successful synthesis of the chitosan/TiO2 mixed matrix membrane. Due to the hydrophilic nature of TiO2 nanoparticles, interactions between Ti-O and the amino and hydroxyl groups of chitosan increase the overall membrane surface’s affinity for water, in other words, increasing hydrophilicity.

According to Fig. 4b, the maximum tensile stress of the chitosan membrane increased from 18.3 MPa to 33.7 MPa when TiO2 nanoparticles were embedded in the chitosan matrix (M4 membrane). The presence of nanoparticles also affected the strain (elongation at break, %), reducing the elongation at break from 7% for the chitosan membrane to 4% for the chitosan/TiO2 (M4) membrane. This enhancement in mechanical properties results from the interaction between TiO2 nanoparticles and functional groups of chitosan (hydrogen bonding and electrostatic interactions), as well as the effective dispersion of these nanoparticles within the chitosan matrix31.

The images obtained from the surface morphology and the cross-sectional area of the chitosan and the chitosan/TiO2 mixed matrix membranes (M4) are presented in Fig. 5. The images presented in Fig. 5a, which show the surface of the chitosan membrane, indicate the dense structure of the pure chitosan membrane, and its pore volume is small. In the images related to the surface of the mixed matrix membrane (Fig. 5b, c), pores are observed, which means that the presence of TiO2 nanoparticles as a filler in the polymer matrix creates pores in its structure. Increasing the pore volume helps the permeability of the synthesized membrane and improves its performance for wastewater treatment. Also, comparing the cross-sectional areas of the chitosan membranes (Fig. 5d) and the chitosan/TiO2 mixed matrix membrane (Fig. 5f) indicates an increase in porosity in the cross-sectional area of the chitosan membrane with the presence of nanoparticles. To determine the porosity of the synthesized membranes, the membranes were soaked in deionized water for 24 h and then weighed. Next, the membrane samples were dried in an oven at 60 °C for 24 h and weighed again. The porosity (ε) of the membranes was estimated using Eq. (6)32. In this equation, Ww and Wd are the masses (g) of the wet and dry membranes, respectively. ρW is the density of water (cm³/g), and L is the membrane thickness (cm), measured from the FE-SEM image of the cross-sectional area.

$$\:\epsilon\:=\frac{({W}_{w}-{W}_{d})}{A\times\:L\times\:{\rho\:}_{W}}100\%$$
(6)

Table 5 shows the increase in porosity of the chitosan membrane with the addition of TiO2 NPs as a filler and the rise in membrane porosity as the filler weight% increases. However, higher filler concentration (0.4 wt%) leads to an increase in solution viscosity and blockage of the pores (resulting from the agglomeration of nanoparticles)32,33. Increasing the concentration of filler (nanoparticles) in the polymer matrix reduces the mobility and fluidity of the polymer chain and increases viscosity. Also, at high concentrations of nanoparticles (fillers), nanoparticles tend to aggregate and cluster due to insufficient polymer coverage and space stabilization, which can physically block the membrane pores and reduce membrane porosity.

Fig. 4
figure 4

(a): FTIR spectra of Chitosan (M1) and Chitosan/TiO2 3 M membranes (M4), (b): Mechanical stability of M1 and M4 membranes.

Full size image
Table 5 Porosity of prepared membranes.
Full size table
Fig. 5
figure 5

The surface morphologies of (a): Chitosan, (b, c): Chitosan/TiO2 3 M membranes (M4), The cross-sectional area of (d): Chitosan, (f): Chitosan/TiO2 3 M membranes (M4).

Full size image

Figure 6 shows the water contact angle results for the chitosan and chitosan/TiO2 3 M membranes. The data presented in this figure indicate that the contact angle for the pure chitosan membrane decreased from 82.5° to 61° (M4) for the chitosan/TiO2 mixed matrix membranes, indicating an increase in the hydrophilicity of the chitosan membrane due to the presence of hydrophilic titanium dioxide nanoparticles.

Fig. 6
figure 6

Water contact angle for chitosan (M1) and chitosan/TiO2 3 M membranes.

Full size image

The AFM images presented in Fig. 7a and Fig. 7b depict the surface roughness of membranes M1 and M4, respectively. The height increased from 342 nm for the chitosan membrane to 494 nm for the mixed matrix membrane M4, reflecting an increase in the surface roughness of the mixed matrix membrane due to the presence of nanoparticles (fillers). As surface roughness increases because of the embedding of TiO2 nanoparticles in the polymer matrix, the water contact angle decreases, indicating an increase in the hydrophilicity of the membrane18. These results are consistent with the results of water contact angle analysis (Fig. 6). Figure 7c, d shows nitrogen adsorption and desorption isotherms for chitosan and mixed matrix membrane (M4). The isotherms reveal a significant increase in specific surface area from 11.01 m²/g for chitosan to 76.56 m²/g for mixed matrix membrane (M4), attributed to the incorporation of TiO₂ nanoparticles between chitosan polymer chains, creating additional space. The mean pore diameters for chitosan and chitosan/TiO2 membranes are 24.447 nm and 28.782 nm, respectively. The BJH pore size distribution of chitosan (Fig. 7e) and chitosan/TiO2 (Fig. 7f) membranes indicates the meso-macroporous structure of these membranes.

Fig. 7
figure 7

AFM images of (a): chitosan, (b): chitosan/TiO2 3 M (M4); BET isotherm/desorption for (c): chitosan, (d): chitosan/TiO2 3 M (M4); BJH diagrams for (e): chitosan, (f): chitosan/TiO2 3 M (M4).

Full size image

The Molecular weight cut-off (MWCO) of chitosan/TiO2 3 M membrane (M4) was determined as follows: the aqueous solution of Polyethylene glycol (PEG) with various molecular weights (0.4, 0.8, 4, 10, 20, 35, 100 kDa, supplied from Sigma-Aldrich Co.) was passed through the membrane system. The concentration (Cf) of PEG in the feed aqueous solution was 10 mg/L, and the PEG concentration in the permeate (Cp (mg/L)) was detected using a Shimadzu 5000 TOC analyzer. The rejection percentage (%) of PEG for each molecular weight was calculated according to Eq. (7). The molecular weight with 90% rejection was defined as the MWCO of the membrane34. The pressure and flow rate of these experiments are as follows: 3.86 bar and 157.98 mL/min. Results indicate that the MWCO of chitosan/TiO2 (M4) membrane is 83.225 kDa.

$$\text{Re} jection{\text{ }}\% = \frac{{\left( {C_{f} - C_{p} } \right) \times 100}}{{C_{f} }}$$
(7)

Membrane performance

The pure water flux for chitosan and chitosan/TiO2 membranes is presented in Fig. 8a, which indicates an increase in the permeability of the synthesized mixed matrix membranes with an increase in the amount of hydrophilic TiO2 filler. However, at high filler concentrations, as stated in the study of the porosity of the synthesized membranes, the pure water flux decreased due to the decrease in membrane porosity. It can be seen that with the increase in the hydrophilicity of the membrane due to the presence of hydrophilic titanium dioxide nanoparticles and also the increase in its porosity, its pure water flux increased, which is confirmed by the analysis of the water contact angle and the comparison of the porosity of the synthesized membranes. The flux recovery ratio (FRR) for the synthesized membranes is presented in Fig. 8b. The addition of hydrophilic TiO2 nanoparticles enhances the hydrophilicity of the synthesized mixed matrix membrane. This enhanced hydrophilicity leads to better adsorption of water molecules and the formation of a water layer on the membrane’s surface. This water layer decreases the BSA adsorption. By reducing BSA adsorption, the amount of membrane fouling decreases, leading to an increase in FRR. The fouling parameters of M1 and M4 membranes are shown in Fig. 8c. As illustrated, embedding TiO2 nanoparticles into the natural chitosan matrix and increasing its hydrophilicity created a water layer on the surface of the mixed matrix membrane (M4). This water layer prevented the adsorption and deposition of contaminant molecules on the membrane surface, resulting in a lower Total Fouling Ratio (TFR) for membrane M4 compared to the polymer membrane M1. In fact, the fouling resistance of the mixed matrix membrane M4 is higher compared to the polymer membrane M1, which is due to the increased hydrophilicity of the membrane13,23. Additionally, the value of the irreversible fouling ratio (IRFR) parameter decreases by about 65% due to the presence of titanium dioxide nanoparticles. This indicates an improved ability of the mixed matrix membrane surface to remove the formed cake through cleaning with rinsing with water. It also suggests a reduced tendency for pollutant molecules/ions to interact with the functional groups of the mixed matrix membrane surface23.

Fig. 8
figure 8

(a): The pure water flux (PWF), (b): Flux recovery ratio (FRR) for synthesis membranes, (c): FRR, TFR, IFR, and IRFR for M1 and M4 membranes.

Full size image

Effect of operation parameters

Membrane tests for the treatment of produced water using a chitosan/TiO2 mixed matrix membrane (M4) were performed according to Table 4, and the desired response, which is the percentage of COD reduction, was calculated (presented in Table 4). Statistical methods were applied to analyze the data obtained from the experimental design using Design Expert 10.0.4 software. Various mathematical models, including linear, quadratic (second-order), two-factor interaction, and cubic models, were applied by the software to suggest a suitable mathematical relationship (appropriate mathematical model) between the response and the parameters under investigation. This allows for the prediction of the performance of the membrane process applied in the wastewater treatment. Table 6 presents the statistical parameters used to evaluate the accuracy of the desired mathematical models. The values of the regression coefficient parameter (R2), Adjusted-R2, and Predicted R2 for the second-order model (quadratic) are 0.994, 0.9582, and 0.9885, respectively. These regression coefficient parameters are very close to 1. Also, the p-value parameter of the second-order model is equal to > 0.0001 (less than 0.05 at the 0.95 confidence level), and the p-value for the lack of fit of this model is higher than 0.05. The set of these cases indicates the suitability and appropriate accuracy of this mathematical model for describing and predicting the process response value (percentage of pollutant load reduction) under different operating conditions (pressure, initial concentration, and volumetric flow rate). The coefficient of variation (C.V%), calculated by dividing the standard deviation by the mean of the response, indicates the amount of dispersion in the mean unit. Given the low value of both C.V% and standard deviation (Std. Dev), it can be assumed that the data have relatively little dispersion from each other. Additionally, the Adeq Precision value greater than 4 is also desirable.

Table 6 Evaluation of the accuracy of models.
Full size table

The ANOVA (Analysis of Variance) table for reducing COD of effluent with chitosan/TiO2 matrix membrane is presented in Table 7. The value of the statistical parameter F-value of the proposed model (quadratic) for the synthesized mixed matrix membrane is 182.63, indicating that the proposed model is significant. The p-value values less than 0.05 (> 0.0001) for the proposed model indicate that the model terms are statistically significant for the desired process. Terms with a p-value greater than 0.05 do not affect the model. The p-value parameter value for the lack of fit term is equal to 0.0689 for the desired mixed matrix membrane, indicating a high and acceptable accuracy of the proposed model for describing and predicting the membrane performance under different operating conditions.

Table 7 Analysis of variance (ANOVA) for membrane treatment of produced water.
Full size table

Based on the p-values for the parameters studied, initial concentration, volumetric flow-rate, and pressure all significantly influence the final response (COD reduction). The interactions among pressure, flow rate, and initial concentration also affect the final response (Fig. 9). The initial concentration and volumetric flow rate parameters have a parabolic effect on the chitosan/TiO2 membrane performance, and the pressure parameter has a linear impact on it. Equation (8) shows the quadratic model proposed by Design Expert 10.0.4 software to predict the amount of COD reduction in terms of the desired operating parameters, including pressure, feed concentration, and volumetric flow rate. The pollution load (COD) reduction, calculated using Eq. (8), is presented in Table 4 and is very close to the experimental data. Figure 10a compares the experimental and predicted COD reduction values by the second-order model (Eq. 8). This comparison demonstrates the high accuracy of the proposed second-order model in describing and forecasting the performance of the membrane process using a chitosan/TiO2 3 M membrane. The perturbation diagram presented in Fig. 10b shows the linear effect of the pressure parameter and the nonlinear effect of the initial pollutant concentration and volumetric flow rate parameters on the performance of the membrane process.

$$\begin{aligned} \text{Re} duction\% & = 2.3949 + 23.973P + 0.019C_{0} \\ & + 0.0687Q + 0.0024*P*C_{0} - 0.029*P*Q \\ & - 0.000023*C_{0} *Q - 1.54016P^{2} \\ & - 0.0000064C_{0}^{2} + 0.0000833Q^{2} \\ \end{aligned}$$
(8)

The three-dimensional plots illustrating the simultaneous effect of pressure, initial concentration, and volumetric flow rate on the performance of the synthesized mixed matrix membrane are presented in Fig. 11. The data show that as pressure increases, the membrane performance for COD reduction improves. As the initial concentration increases, the membrane performance decreases due to the possibility of fouling and cake formation behind the membrane, and with increasing volumetric flow rate, the membrane performance decreases due to the decrease in contact time and retention time. As shown in Fig. 11a, b, the membrane performance in reducing COD of the effluent increases with increasing pressure, indicating a positive effect of pressure on membrane performance due to the increase in driving force of the membrane process. Increasing pressure increases the flux. However, the membrane performance is likely to decrease at high pressure due to the rise of membrane fouling. Research has shown that increasing pressure increases the fouling rate and the rate of fouling of the membrane pores35. As the data shown in Fig. 11a, c indicate, with increasing initial concentration, the hydrophilicity of the membrane causes contaminants in the effluent to accumulate behind it, leading to cake formation, concentration polarization, and blockage of the membrane pores. This decreases membrane performance and the effectiveness of separating and removing contaminants as the initial effluent concentration increases36,37. Also, as shown in Fig. 11b, c, because the contact time between the effluent and the membrane decreases with increasing volumetric flow rate (reduction in residence time), the membrane performance declines, and the COD removal rate also drops, which is reported in other research reports38.

Fig. 9
figure 9

The interactions of parameters (a): Pressure-initial concentration, (b): Pressure-flow-rate, (c): Initial concentration-flow-rate.

Full size image
Fig. 10
figure 10

(a): The experimental values versus predicted values by Eq. (8); (b): The perturbation diagram.

Full size image
Fig. 11
figure 11

The 3D plots for the simultaneous effect of (a): pressure and initial concentration, (b): pressure and volumetric flow-rate, (c): initial concentration and volumetric flow-rate.

Full size image

Another aspect of the data analysis using experimental design software is to determine the optimal operating conditions for achieving the highest performance efficiency in a process. According to the experimental design carried out with the software, the optimum points for the purification of produced water using a membrane system with a synthesized mixed matrix membrane are as follows: pressure: 3.86 bar, initial concentration 1975.1 mg/L, and volumetric flow rate 157.98 ml/min. Under these conditions, the software predicts a COD reduction value of 91.4157%. The wastewater treatment experiments performed under these optimal conditions demonstrated an experimental COD reduction of 90.05%. Also, the chitosan/TiO2 mixed matrix membrane under these conditions can reduce TSS and TDS by 84% and 86%, respectively.

Investigating membrane regeneration and reuse

To evaluate the reusability and stability of the mixed matrix membrane, the synthesized membrane (M4) was tested over 10 cycles. After wastewater treatment under optimal conditions, the membrane film was soaked in a 0.5 M sodium hydroxide solution for 5 h, then rinsed with distilled water and dried at 60 °C in an oven. The dried membrane was reused in the purification process under the same optimal conditions. The results showing pollution reduction per cycle are presented in Fig. 12a. The data reveal a 22% decrease in wastewater treatment efficiency after 10 cycles. Despite this, the membrane still demonstrates acceptable performance for repeated use. This performance decline after 10 cycles may be due to slight changes in membrane morphology, as shown in Fig. 12b. To further evaluate its stability and strength, tests were extended to 21 cycles. It was observed that after 21 cycles, a very small hole appeared in the membrane.

Fig. 12
figure 12

(a): The COD reduction using chitosan/TiO2 (M4) at 10 cycles, (b): The SEM image of chitosan/TiO2 membrane after 10 cycles.

Full size image

Comparison of research results

To assess the performance of the membrane synthesized in this study compared to those reported in other articles for COD reduction and pollutant removal, the top performances of these membranes are shown in Table 8. Comparing the data in this table reveals that the synthesized membrane has a strong ability to treat wastewater and lower pollutant loads.

Given the high volume of produced water in oil and gas processes, the lack of sour water resources in Iran and elsewhere in the world, and the significant performance of the synthesized bio-based membrane (Chitosan/TiO2), it can be said that this membrane can be considered as a candidate for industrial application for the treatment of effluents from other industries.

Table 8 Some membranes applied to wastewater treatment.
Full size table

Conclusions

This study aims to purify produced water using a chitosan-TiO2 mixed matrix membrane, where the chitosan, serving as the polymer matrix, was synthesized from shrimp shells. The impact of the filler nanoparticle (TiO2 NP) weight% on pure water flux (PWF) and flux recovery ratio (FRR) showed that increasing the filler content to 3 wt% improved both PWF and FRR in the mixed matrix membrane. However, higher filler contents led to decreases in water flux and recovery ratio due to increased solution viscosity and pore blockage. In the second stage, the performance of the chitosan/TiO2 (M4) mixed matrix membrane for treating high COD wastewater was evaluated under different conditions using the central composite experimental design. Results indicated that the 3 M membrane, under optimal conditions (operating pressure: 3.86 bar, initial COD: 1975.1 mg/L, flow rate: 157.98 mL/min), could reduce COD, TDS, and TSS in the influent wastewater by 90.05%, 86%, and 84%, respectively. Testing the chitosan/TiO2 membrane over 10 consecutive cycles demonstrated its remarkable and acceptable potential for semi-industrial and industrial applications. The synthesized bio-based membrane (Chitosan/TiO2) demonstrated strong performance in treating industrial wastewater with high pollutant loads (COD: 3833 mg/L, TSS: 4880 mg/L, TDS: 214940 mg/L), making it a promising candidate for effluent treatment in other industries.