Ultrafiltration study of the polysulfone membrane modified with branched polyethyleneimine

A

effective membrane surface area (m²)

A_s

membrane surface area (cm²)

ATR-FTIR

Attenuated Total Reflectance-Fourier Transform Infrared

BSA

bovine serum albumin

C_a

BSA concentration after adsorption (mg/L)

CA

contact angle

C_b

BSA concentration before adsorption (mg/L)

C_f

feed concentration (mg/L)

CF

compaction factor

EWC

equilibrium water content

FESEM

field emission scanning electron microscope

Flux_RR

flux recovery ratio

F_t

total fouling ratio

F_ir

irreversible fouling ratio

F_r

reversible fouling ratio

HPG

hyperbranched polyglycerol

HA

humic acid

J_w1

pure water flux in first run (L/m² h)

J_w2

pure water flux in second run (L/m² h)

J_p

feed flux (L/m² h)

L

membrane thickness (cm)

LLDP

liquid–liquid displacement porosimetry

Ln

total hydraulic permeability coefficient

M_BSA

amount of BSA adsorbed (mg/cm²)

M_dry

dry membrane weight (g)

M_wet

wet membrane weight (g)

NMP

N-methyl-2-pyrrolidone

o/w

oil-in-water

PEG

polyethylene glycol

PSF

polysulfone

PVDF

poly(vinylidene floride)

PVP

polyvinylpyrrolidone

PWF

pure water flux

PEI

polyethyleneimine

R

rejection%

TMP

transmembrane pressure

UF

ultrafiltration

V

volume of permeate (L)

V_BSA

volume of BSA solution (L)

ΔP

operating pressure (kPa)

ΔT

permeation time (h)

Greek letters

σ

surface tension (mN/m)

ε

porosity

ρ_w

density of water (g/cm³)

Rapid societal growth leads to the expansion of industries, urban cities, and population growth, all of which increase the water demand and place enormous responsibility on the waste-water treatment process (Van Loosdrecht & Brdjanovic 2014). The volume of untreated wastewater discharged into the aquatic environment is currently increasing dramatically, and large quantities of hazardous chemicals that are non-biodegradable and highly toxic are contaminating freshwater bodies. Many polymers are used in waste-water treatment and the removal of toxic materials from water through the membrane separation process. The following are the advantages of membrane technology: (i) they are cost-effective, (ii) easy to operate, (iii) large temperature and pH operating range, and (iv) low energy consumption.

Polymeric materials have been thought to be an ideal candidate for membrane separation processes involving high process conditions. Most membranes are made of organic or inorganic materials. As a result, when compared to membranes made of other materials, polymer membranes are more effective. Polymers are the most important class of material for ultrafiltration membranes. The chemical and physical properties derived from structural components such as molecular weight, chain interaction, and chain flexibility are used to select the polymeric material for a specific membrane. The selection of a high-performance membrane material is critical. The materials used by researchers in the treatment of oily wastewater are as follows: polyvinylidene fluoride (PVDF) (Yuliwati & Ismail 2011; Zhang et al. 2013; Liu et al. 2016; Venault et al. 2016), polyethersulfone (PES) (Kusworo et al. 2017), polysulfone (PSF) (Sinha & Purkait 2013), polypropylone (PP) (Song et al. 2017), and cellulose acetate (Han et al. 2015) are all examples of base polymers. These polymeric materials are further modified by adding different types of additives depending on the requirement. Adding other materials can be achieved by various techniques based on polymer properties, solubility in different organic solvents, and weight percent of the additive requirement.

Polymer-based membranes with high pure water flux (PWF), rejection value, and hydrophilic properties are ideal for optimising oil/water separation performance. The need to deal with trade-off relationships between permeance and flux, as well as membrane fouling and membrane-scaling, is one of the critical issues pertaining to the use of conventional polymeric membranes in oily wastewater treatment (Zuo et al. 2018). Polymer-based membranes have low PWF, fouling issues, and degradation due to oil droplets blocking their pores, which causes the membranes to have short operation times.

Zhao et al. added hyperbranched polyglycerol (HPG) to a PVDF membrane as an additive. This experiment revealed that PVDF with HPG has higher PWF, porosity, and permeability than plain PVDF membranes (Zhao et al. 2007). Kim and Lee created a PSF polymer membrane with PEG 600 as an additive in 1998, and it was discovered that porosity and PWF were increased when compared to a plain PSF membrane (Kim & Lee 1998). Chakrabarty et al. investigated the effect of additive PEG with a mixture of different molecular weight poly(vinyl-pyrrolidone) (PVP) on the PSF (base polymer) membrane. The results were intriguing because the PWF of the modified membrane initially increased with the molecular weight of PVP up to 24,000 Da, but then the blended membrane began to become dense, resulting in decreases in PWF. Similarly, PEG with a molecular weight of 6,000 Da was found to have the highest BSA rejection. This experiment concluded that PEG with a molecular weight of 6,000 Da is the best additive or pore-forming agent for asymmetric PSF membranes (Chakrabarty et al. 2008). Saljioughi and Mohammadi investigated the effect of additive PVP blended in cellulose acetate at various weight percent levels ranging from 0 to 9 wt%. The findings showed conflicting results when using a 1.5 wt% additive. It was observed that the PWF of the membrane increased as the weight percent of PVP went from 0 to 1.5, due to the creation of macrovoids in the membrane's sublayer. However, increasing the weight percent of PVP above 1.5 resulted in a decrease in PWF, as macrovoid suppression occurred (Saljoughi & Mohammadi 2009).

Table 1 displays various hydrophilic polymers used in the literature as an additive for membrane modification. It is noted that this hydrophilic polymer used as an additive plays an important role in the mitigation of fouling tendency, and change in morphology and structure of polymeric membrane. Table 1 gives information about the base polymer used, different additives, organic solvent, PWF, and foulant rejection.

This research aims to evaluate the effectiveness of adding branched polyethyleneimine (PEI) to polysulfone (PSF) membranes for ultrafiltration to remove humic acid and oil–water emulsion. After that, the PSF membrane-modified with PEI was created and cast. Using Fourier Transform Infrared Attenuated Total Reflection Spectroscopy (FTIR-ATR) and Field Emission Scanning Electron Microscopy (FESEM), synthetic membranes were examined. PWF, HA rejection, BSA adsorption, permeability, and contact angle (CA) of membrane surface have been studied. The modified PSF membranes were found to have improved permeability, reduced fouling, and better selectivity compared to unmodified PSF membranes. The optimal PEI concentration for modifying the PSF membrane was determined to be 3 wt%, and the modified PSF membranes were effective in removing humic acid and oil–water emulsion from wastewater.

PSF (average M_w = 35,000 gmol) was used as the basic polymer in the membrane casting solution, which was purchased from Sigma-Aldrich Co., USA. N-methyl pyrrolidone (NMP) of analytical reagent-grade has been provided by LOBA Chemie, India. Sigma-Aldrich Co., USA and Otto Chemie Private Limited in India supplied bovine serum albumin (BSA). Without additional purification, all compounds were utilised. Throughout the studies, deionised (DI) water cleaned by the Millipore system (Millipore, France) was employed.

PSF and PEI polymer were mixed in organic solvent NMP in which base polymer wt% was kept at 15 and 12% for two different sets. The PEG of molecular weight 6,000 Da was kept 5 wt% of the total weight of the solution, the combination was kept under constant heating at 45–50 °C under magnetic stirring. After cooling to room temperature, the solution was kept for degasification for 12–15 h and then the solution was fabricated with the help of a casting knife. The membrane was prepared by the phase inversion method by dipping the glass containing the solution into a DI water bath and was allowed to solidify and convert into a white colour sheet. The fabricated membranes were kept in replaced DI water bath for 12–15 h to remove unreacted polymer and eliminate the solvent remaining in the membrane.

The presence of PEI and PSF of the produced membrane was confirmed by FTIR analysis. An FTIR spectrometer was used to get the FTIR spectra of the membrane, which was ground with KBr (IRAffinity-1, Shimadzu, Japan). The weight percentage of the additive is mentioned in Table 2.

FESEM (ZEISS LSM 510 Meta) was used to capture top surface and cross-sectional images of the PEI- branched PSF membranes. A sample of all membranes was mounted on a thin coating of a carbon tape on an FESEM stub of a suitable size. All of the samples were gold-coated to give electrical conductivity for non-conducting polymer membranes, and micrograph pictures were taken in an extreme vacuum with an acceleration voltage of 10–25 kV. At different magnifications, a number of SEM top and cross-section photos were captured. These photographs show the top layer's visual information as well as the membranes' cross-sectional structure.

The typical FTIR peaks of the PSF membrane were compared to the modified PSF membrane, indicating the presence of PEI in the membrane. The FTIR spectra of each modified membrane were obtained using a new attachment (ATR-8200 HA). The membrane samples were examined by attaching them to the ATR clamp.

Pore counts per unit surface area with a radius between r_i and r_k are represented by N_i,k. The dynamic viscosity of the alcohol-rich phase is denoted by d, which is the pore length as well as the thickness of the membrane layer.

Here, and are the weights of the membrane after and before soaking, respectively. L is the thickness of the membrane obtained from SEM. A is the effective area of membrane soaked in water, while dW is the water density (1 gm/cm³).

The liquid–liquid displacement Porosimetry (LLDP) is a method used to perform an in-line porosimetry analysis of the membrane cartridges, getting their pore size distributions (PSDs), and mean pore diameters (d_avg). This will enable us to get the pore size and pore density of each pore in the membrane. Here, a mixture of water, isobutanol, and methanol in a ratio of 25:15:7 v/v was used in surface tension between alcohol- and water-rich phases. A mixture of water and alcohol was shaken in a separating funnel and was left to stand for 4 h to ensure complete separation between the two liquids. The water-rich phase settled at the bottom of the funnel and the alcohol-rich phase remained at the top. The bottom product was used as a wetting liquid for the membrane and the top product for displacing liquids as discussed in the section “Methods and experiments”.

F_r, F_t, and F_ir ratios were examined and are provided to understand how humic acid ultrafiltration causes membrane fouling fully. Tables 4 provides more data. As shown in Figures 14 and 15, irreversible fouling was reduced after initially rising in the instance of 15M0 when PEI was present at a concentration of 3 wt%. Less foulant build-up or deposition in the membrane's pores or on its surface is indicated by lower F_ir levels, and vice versa. The concentration of PEI may increase reversible fouling by up to 3 wt%. The rise in F_r demonstrated that reversible fouling of the membrane takes the place of irreversible fouling during the ultrafiltration process. In other words, incorporating PEI on the membrane surface prevents irreversible fouling. The detail explains these findings that the hydration layer generated on the membrane surface can reduce the contact between the foulants and the surface of the membrane by the incorporation of the amine group, preventing the irreversible foulant binding. Similar to F_ir, the trend in F_t was seen. It was determined from Figures 14 and 15 that in ultrafiltration studies with all types of feeds (HA and o/w emulsion), the bare membrane has the maximum value of F_t and F_ir. However, the 15M3 membrane has the least value of F_t and F_ir, making it less responsive to the deposition of irreversible foulants. The Flux_RR of unmodified and modified membranes is shown in Figures 14 and 15. Flux_RR improves for modified membranes as F_ir decreases because Flux_RR and irreversible fouling are directly related to each other. The Flux_RR enhanced from 0.526 (for plain PSF membrane) to 0.954, and 0.85 (for 12M3 membrane) in the humic acid and oil–water emulsion ultrafiltration process, respectively. Compared to HA ultrafiltration, the value of Flux_RR in oil–water emulsion ultrafiltration was lower. This might be because HA molecules are easily removed by simply hydraulic cleaning and require higher Flux_RR. Due to the oil–water emulsion's insolubility in water, oil droplets persist within the pores or on the membrane surface even after simple hydraulic washing, making it challenging to clean the membranes. Therefore, chemical cleaning can eliminate these types of foulants from membranes.

PEI used as an additive in PSF base polymer membrane has shown positive results such as the decrease in fouling ratio and increase in the rejection value of HA along with the increase in the weight percentage of PEI in composition from 0 to 3%. The permeability of the membrane 12M3 has increased from 0.19 to 0.78 L/m²hr Kpa with respect to the membrane 15M0. The rejection value of HA increased from 9.32 to 61.088% for 0–3% of PEI. Overall, the difficulty in the direct blending of PEI with PSF is a major issue due to the agglomeration of PEI particles on the membrane surface. For a higher weight percentage of PEI in PSF base polymer membrane, the co-polymerisation process can be utilised for better fouling ratio results and rejection percentage.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.