Jan Dolinaa,
Zuzanna Gončukováa,
Marek Bobákb and
Lukáš Dvořák*a
aInstitute for Nanomaterials, Advanced Technologies and Innovation, Technical University of Liberec, Studentská 2, 461 17 Liberec, Czech Republic. E-mail: lukas.dvorak@tul.cz; Tel: +420 485 353 805
bMembrain s.r.o., Pod Vinicí 87, 471 27 Stráž pod Ralskem, Czech Republic
First published on 18th April 2018
Biofouling represents a serious problem limiting the widespread application of membrane technology. Therefore, the aim of this study was to develop and verify a new modification method based on the in situ formation of silver nanoparticles and their incorporation into a membrane polymer to prevent biofouling. The modification method consisted of soaking a commercial hollow-fibre polyethersulfone membrane in a solution of silver ions, diffusion of ions into the membrane polymer, and their reduction using ascorbic acid. Such a modified membrane displayed a lower tendency towards biofouling, exhibiting an about 15% higher permeability compared to an unmodified membrane when filtering actual wastewater treatment plant effluent. The modification also led to the formation of stable silver nanoparticles (mostly in the range of 25–50 nm) homogenously distributed on the surface of the hollow-fibres. This resulted in higher surface hydrophilicity (the water contact angle decreased from 91° to 86°) contributing to the biofouling prevention. The modified membrane also showed high stability, as only 2.1% of the total silver leached after 8 h of filtration. Moreover, no changes in the original membrane cross-section structure or separation properties were observed. Besides the improved antibiofouling properties of the modified membrane, the main advantage of the developed method is its simplicity, short reaction time, absence of high energy-consuming initiation, and the possibility to apply it on site, thus even with commercial membrane modules. It will increase the application potential of membranes in the field of wastewater treatment.
To date, several methods of minimising membrane biofouling have been developed. These include changes in operational conditions (such as an increase in sludge retention time, increase in hydraulic retention time, decrease in the food-to-microorganism ratio, or wastewater characteristics), all of which affect the biomass characteristics or production of extracellular polymeric substances and, subsequently, interactions with the membrane surface and biofouling.5–7 The ratio of filtration time and flux to time of backpulse/backwashing, including relaxation, also affects membrane biofouling.1 However, while such methods have had a positive impact on biofouling, they offer a partial improvement only. Hence, other more effective approaches are needed, including modifications to the original membrane surface or development and preparation of new membrane materials. Such approaches usually result in an increase in the antimicrobial and hydrophilic characteristics of the membrane surface, resulting in a reduction to mutual hydrophobic interactions between the membrane and microorganisms, thereby mitigating biofouling.8
Polyethersulfone (PES), a polymer material commonly used for membrane preparation, exhibits high temperature, chemical and dimensional stability. As PES is a polar material, it can easily be modified. Several methods for surface modification of PES membranes have been developed.9,10 Since PES displays hydrophobic characteristics with low surface energy, the main goal of such modifications is to obtain a more polar surface, thus reducing surface hydrophobicity. In general, current methods of membrane surface modification can be divided into six main groups. These include coating methods (formation of thin layers subsequently fixed by noncovalent bonding to the original membrane surface, e.g. ref. 11); blending methods (two or more polymers mixed to obtain the required properties, e.g. ref. 12 and 13); composite methods (a membrane prepared from two or more materials differing in physical/chemical properties, e.g. ref. 14); chemical methods (new functional groups directly formed on the original membrane surface, e.g. ref. 15); grafting methods (covalent bonding of suitable monomers or functional groups to the membrane following initiation reactions, e.g. ref. 16 and 17) and combined methods.
Examples of combined methods include the exposure of a commercial PES membrane to a sulfonation reaction, resulting in the formation of new functional groups, followed by immersion in a titanium dioxide solution,18 and attachment of graphene oxide after grafting allylamines to the surface of a PES membrane by UV light.19
In most cases, the surface modification of PES membranes makes use of organic compounds and differing process conditions, often with energy-intensive initiation procedures. Often, however, the processes result in partial blockage of the membrane pores, resulting in lower hydraulic performance together with difficulties in controlling the modification reactions. As such, most processes for modifying PES membranes tend to use inorganic substances, including titanium dioxide, silicon particles, aluminium oxides or silver. As both silver ions and silver nanoparticles have proven antibacterial properties, modification of membranes using silver nanoparticles offers a promising option for mitigating membrane biofouling.20 The antimicrobial properties limit mutual interactions between the microorganisms and the membrane surface while still achieving long-term membrane permeability. Indeed, according to Meng et al.,21 nanotechnology in general has great potential for the modification and development of new membranes.
The above compounds are usually incorporated into the basic polymer (or mixture) prior to membrane preparation. Although the addition of such compounds can result in the required properties on the final membrane surface, the solution used to modify the base polymer matrix may negatively affect the ability to produce hollow-fibres when preparing the PES membrane. For this reason, most studies have focused on the modification or preparation of flat sheet membranes.
Hollow-fibres are the most frequently used membrane type in membrane bioreactors. Their main advantage lies in their very high packing density (surface-to-volume ratio).1 The high surface area allows the application of lower pressures to achieve the required hydraulic performance, which also results in a lower tendency towards biofouling and cake layer formation. A further advantage of hollow-fibre membranes over flat sheet membranes is the possibility of using intensive backpulse or backflush cleaning processes, thereby allowing the cake layer to be removed on a regular basis.22
As most current modification methods are time or/and energy-consuming, the main goal of this study was to develop a new modification method simple in use. The developed method consisted of using silver nanoparticles formed in situ and their incorporation into the membrane polymer. Silver nanoparticles were chosen in this study as they exhibit a strong antimicrobial effect towards a broad spectrum of different microorganisms (Gram-negative and Gram-positive). Moreover, silver nanoparticles are more efficient than other nanoparticles of transition metals and/or their oxides (e.g. CuO, TiO2 or ZnO). To date, most studies dealing with membrane modification have also been conducted on flat sheet membranes; therefore, the developed method was used to modify a commercially available hollow-fibre membrane to verify its efficiency towards the mitigation of membrane biofouling. Following modification, changes in the filtration performance with demineralized water and wastewater treatment plant effluent were evaluated. Silver leaching was also accomplished during the filtration test. Surface and cross-section morphology was characterized by a scanning electron microscope, and surface roughness was assessed by atomic force microscopy. Moreover, the amount of silver on the surface of the modified membrane using an energy dispersive X-ray detector, and the water contact angle to determine membrane hydrophilicity were also assessed.
The hollow-fibre membrane packed in a tube module supplied by Membrane Solution LLC (USA) was first soaked in a solution of 3.5% (wt) silver nitrate (AgNO3, Sigma-Aldrich), which was circulated through the module for 4 h at room temperature. The membrane was then intensively rinsed three times with demineralised water. Reduction of silver ions, i.e. in situ formation of silver nanoparticles, was conducted by soaking the membrane in a 2% (wt) solution of ascorbic acid (C6H8O6; Sigma-Aldrich) for 2 h. Following silver reduction, the membrane was again rinsed intensively with demineralised water and then heated for 2 h at 70 °C while recirculating the warm demineralized water through the module. The modified membrane was then immediately cooled with fresh demineralised water to laboratory temperature (22–25 °C). The modified membrane was subsequently stored in demineralised water for later use.
The membrane used to verify the efficiency of the developed modification method had an active membrane surface area of 0.7 m2. The hollow-fibres themselves were single capillary with an internal diameter of 0.7 mm, an outer diameter of 1.3 mm, and an out–in flow direction. The membrane had a molecular weight cut-off of 100 kDa.
To evaluate potential changes in the removal of the organic compounds caused by modification of the membrane, chemical oxygen demand (COD) was determined in the membrane module feed (actual WWTP effluent) and permeate. The COD concentration was measured in duplicate using the Hach-Lange (Germany) cuvette test (LCK 414).
The concentration of released silver was measured by inductively coupled plasma atomic emission spectroscopy (ICP-OES) using an Optima 2100 DV spectrophotometer (Perkin Elmer, USA).
The surface morphology of the hollow-fibres was evaluated followed by water evaporation and placing onto a standard pin stub mount provided by Tescan (Czech Republic).
A Carl Zeiss ULTRA Plus scanning electron microscope (SEM) (Carl Zeiss Microscopy GmbH, Germany) was employed to visualise the membrane surface and cross-section morphology at different magnifications. Before the SEM analysis, the membrane samples were first covered with a 2 nm layer of platinum. The SEM analysis was undertaken under an acceleration voltage of 2 kV. The SEM was equipped with a secondary electron detector in order to analyse the chemical composition of the membrane surface. The amount of silver on the membrane surface was determined using a built-in energy dispersive X-ray (EDX) detector (Oxford instruments, UK) with no coating of samples by platinum to avoid the reduction of chemical contrast.
Atomic force microscopy was used to assess membrane surface roughness. First, the samples were fixed on a glass plate and placed into the holder of a Nanowizard 3 microscope (JPK Instruments, Germany). The samples were measured in contact mode using a cantilever with a 1–2 nm sharp tip of a silicone probe (SHOCONA-SS, Applied NanoStructures, Inc., USA) to scan a 10 × 10 μm portion of the membrane surface. Following analysis, JPK data processing software provided by JPK Instruments (Germany) was employed to calculate roughness parameters. To minimise the influence of curvature, the surface area analysed was set at 2 × 2 μm. A three-dimensional structural profile of the membrane surface was also formed using this JPK data processing software.
Unlike other methods currently applied for membrane modification, modification through the in situ formation of silver nanoparticles, as developed in this study, did not lead to any change in flow characteristics (Fig. 1B). The average flux at different TMPs before and after modification was basically the same (Fig. 1B). In comparison, Rahimpour et al.,23 who synthesised two types of PES membrane modified with titanium dioxide nanoparticles (one prepared through entrapment and the other by coating and UV irradiation), observed a lower initial flux in the titanium dioxide entrapped membranes compared with the initial flux of an untreated PES membrane. In further tests, however, both membranes exhibited enhanced antibiofouling properties and long-term flux stability compared to the neat PES membrane.23
In the second filtration test, actual effluent from a municipal WWTP was used as a feed to demonstrate the efficiency of the developed method to mitigate biofouling. While the absolute permeability values for the reference membrane were slightly higher than the modified membrane (Fig. 2A), the normalised permeability (K/K0) was higher for the modified membrane (Fig. 2B). This disproportion could be explained by a slightly different concentration of suspended solids in the feed used on both membranes, as the feed comprised actual WWTP effluent. In fact, the concentration of suspended solids throughout the test with the reference membrane was 10.7 mg L−1, compared to a concentration of 15.0 mg L−1 during the test with the modified membrane.
Fig. 2 Permeability over time: (A) absolute values (K) for the modified and reference membranes, (B) normalised permeability (K/K0), after eliminating the influence of different feed characteristics. |
To better compare the permeability between these two membranes, normalised permeability values (permeability [K/K0] obtained 60 minutes after the stabilising process) against the initial permeability of each membrane were plotted (Fig. 2B). Following a sharp initial decrease in permeability caused by unstable process conditions and the formation of a filtration layer on the membrane surface, the permeability increased in the modified membrane over time (Fig. 2). This suggests that (a) the compounds responsible for the fouling were not fixed to the surface, and (b) that the modified membrane showed a lowered tendency towards biofouling compared to the reference (unmodified) membrane. Moreover, after membrane cleaning, i.e. following an intensive reversal flow of permeate in the opposite direction to filtration in the module performed after 24 h of testing, the modified membrane displayed an increased recovery of permeability.
A rapid increase in permeability difference (dK/Kref) was also shown by the modified membrane when comparing the permeability difference to a reference membrane plotted over time (Fig. 3). At the start of the experiment, the modified membrane showed 5% lower permeability than the reference membrane; however, after approximately 6 h, the difference in permeability (dK/Kref) began to decrease rapidly and, after 12 h, the permeability values were comparable. At this point, the permeability values for the modified membrane were always higher than those for the reference membrane. The greatest difference in dK/Kref values (approximately 20%) was recorded after 24 h, i.e. just before cleaning the membrane module. During further phases of the test, dK/Kref values for the modified membrane were approximately 15% higher than those for the reference membrane.
Fig. 3 Difference in permeability (dK/Kref; in %) between the modified membrane and a reference (unmodified) membrane during a filtration test with WWTP effluent. |
These results suggest that the modification of the membrane through the developed in situ formation of silver nanoparticles had no negative effect on membrane filtration performance. On the contrary, both the permeability of the modified membrane and its anti-biofouling properties improved due to the modification procedure.
The separation properties of the membranes were evaluated based on the retention of organic substances (expressed as COD). The reference and modified membranes exhibited comparable separation properties, with COD retention for the reference membrane reaching 27.2% while the modified membrane reached 30.5%. Retention of COD was caused by the capture of suspended solids and soluble substances with a molecular weight greater than 100000 daltons (molecular weight cut-off guaranteed by the supplier) on the membrane surface.
Fig. 4 Cross-section morphology of (A) the modified membrane, and (B) the reference membrane; 500× magnification. |
Under higher magnification, the presence of silver nanoparticles on the surface and in the active surface layer of the modified membrane was clearly visible (Fig. 5A). Furthermore, compared with the reference membrane (Fig. 5B) there was clear evidence of the presence of silver nanoparticles inside the polymer matrix (from the surface up to the macrovoid structure) of the modified membrane (Fig. 5A).
Fig. 5 Cross-section morphology of (A) the modified membrane, with silver nanoparticles clearly visible, and (B) the reference membrane; 10000× magnification. |
The performed modification resulted in the incorporation of silver nanoparticles into the active membrane layer as well as into the dense inner pore structure. As silver nanoparticles exhibit hydrophilic properties, they can be beneficial for the mitigation of biofouling also on the surface of membrane pores.
Fig. 6 Scanning electron microscopy images of the surface of (A) the modified membrane, and (B) the reference membrane; 10000× magnification. |
Comparison of EDX spectra for the reference and modified membranes also clearly indicates the presence of silver on the modified membrane (Fig. 7A; ratios of individual elements [mass and molar ratios] shown in Table 1). The EDX analysis confirmed that silver nanoparticles were homogeneously distributed (Fig. 7B), with a mass ratio of silver of 3.6% on the surface of the modified membrane.
Elements | Reference membrane | Modified membrane | ||
---|---|---|---|---|
Mass ratio [%] | Molar ratio [%] | Mass ratio [%] | Molar ratio [%] | |
Carbon (C) | 69.74 | 77.57 | 67.77 | 77.84 |
Oxygen (O) | 23.48 | 19.61 | 21.73 | 18.74 |
Sulphur (S) | 6.78 | 2.82 | 6.87 | 2.95 |
Silver (Ag) | — | — | 3.63 | 0.46 |
The size of the silver nanoparticles present on the surface of the modified membrane was also assessed as it has been shown that the size of nanoparticles affects their antimicrobial activity.28 The mean diameter of the silver nanoparticles detected on the modified membrane was 43 ± 18 nm (n = 20). According to Panáček et al.,28 smaller silver nanoparticles of around 25 nm exhibit the strongest antimicrobial activity, while large nanoparticles of around 50 nm have a reduced effect. This is due to the higher surface/volume ratio of smaller silver nanoparticles facilitating the release of silver ions and/or production of reactive oxygen species.29 The particle size distribution analysis performed in this study indicated that approximately 10% of the silver nanoparticles had a diameter of <25 nm, and only 15% of were larger than 50 nm. As most of the nanoparticles ranged from 25–50 nm, the membrane could be expected to display strong antimicrobial properties. This also corresponds with the results of the filtration tests, where a lowered tendency toward biofouling was observed in the modified membrane.
In addition to the size of the silver nanoparticles, their antimicrobial activity may also be influenced by their shape.30 In this study, the silver nanoparticles present on the surface of the modified membrane were clearly spherical (Fig. 8). According to Pal et al.,30 spherical silver nanoparticles exhibit average antimicrobial activity compared to truncated triangular nanoparticles, which show strong activity, and rod-shaped nanoparticles, which display an inferior antimicrobial effect.
Fig. 8 Size of the silver nanoparticles present on the surface of the modified membrane; 87000× magnification. |
Parameters | Modified membrane | Reference membrane |
---|---|---|
Average roughness Ra [nm] | 24 | 4 |
Peak-to-valley roughness Rt [nm] | 150 | 29 |
Max. height [nm] | 79 | 11 |
Max. depth [nm] | −71 | −18 |
Membrane | Average water contact angle [°] | Standard deviation [°] |
---|---|---|
Modified | 86.1 | 1.6 |
Reference | 91.6 | 3.3 |
The developed modification method resulted in stable silver nanoparticles firmly anchored to the membrane surface and in the active layer (confirmed by SEM and EDX), leading to a membrane with improved antibiofouling properties. SEM analysis also confirmed no negative impact of the in situ formation of silver nanoparticles on membrane cross-section morphology. A decrease in the water contact angle indicated that the modification method resulted in a more hydrophilic character of the modified membrane surface than the reference membrane. As a result, the modified membrane displayed an approximately 15% higher permeability when filtering actual effluent from a municipal wastewater treatment plant.
The simplicity of the developed method, the short reaction time, and the ability to be performed on site resulting in a lower tendency toward biofouling represent its great potential for real application.
This journal is © The Royal Society of Chemistry 2018 |