Enzymatic degradation of bisphenol-A with immobilized laccase on TiO2 sol–gel coated PVDF membrane

Journal of Membrane Science 469 (2014) 19–30

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Enzymatic degradation of bisphenol-A with immobilized laccase on TiO2 sol–gel coated PVDF membrane Jingwei Hou, Guangxi Dong, Yun Ye, Vicki Chen n UNESCO Centre for Membrane Science and Technology, School of Chemical Engineering, The University of New South Wales, Sydney, NSW 2052, Australia

art ic l e i nf o

a b s t r a c t

Article history: Received 29 January 2014 Received in revised form 9 May 2014 Accepted 16 June 2014 Available online 21 June 2014

Bio-degradation with laccase immobilized on the TiO2 functionalized membrane offers an attractive option to augment the conventional wastewater treatment processes for the removal of micropollutants such as Bisphenol A (BPA). Immobilization of the enzyme on nanostructured TiO2 coated membranes addresses the common issues such as the poor activity and stability associated with the free laccase, as well as reducing the loss of laccase. Furthermore, the removal of large molecular weight BPA degradation products presents another benefit with this bio-catalytic membrane system. In this work, the TiO2 nanoparticles were coated on the membrane surface via a low temperature hydrothermal sol–gel process, and the laccase was subsequently immobilized on these membranes by chemical coupling. The PVDF membranes with different pore sizes (0.1 and 0.45 mm) and coating cycles (up to 4 cycles) were used to examine the effects of the TiO2 loadings and the nanostructure of the coating layer on the laccase immobilization performance. The optimal apparent activity and activity recovery were achieved on the 0.1 mm membrane with 3 coating cycles based on the 2,20 -azino-bis-(3-ethyl benzothiazoline-6-sulfonic acid) (ABTS) assay, due to the larger BET surface area and BJH pore diameter observed on this membrane. Substantial improvement in BPA removal efficiency and stability under moderate flux conditions were also obtained on the TiO2 functionalized membranes, in good agreement with the ABTS assay results. In addition, the polymer products derived from the BPA bio-degradation process showed negligible fouling impact on the coated 0.1 mm membrane. & 2014 Elsevier B.V. All rights reserved.

Keywords: Laccase immobilization BPA degradation TiO2 sol–gel coated membrane Wastewater treatment

1. Introduction Bisphenol A (BPA) is heavily used in the plastic industry as the plastic monomer and plasticizer in the production of polycarbonate and epoxy products [1]. Trace amount of BPA has been detected in natural environment as well as in human serum, urine, tissue and blood, due to the BPA leaching from a wide range of plastic products containing BPA [2]. The adverse effect of BPA on human health was evidenced even at low dosage, and its estrogenic activity can lead to cancer in the mammary and the prostate [3]. Therefore it is essential to remove the leached BPA from wastewater to minimize its potential harmful impact propagating through the water cycle. However, the complete removal of BPA via the conventional wastewater treatment processes is difficult to achieve as BPA can potentially absorb into membrane matrix and eventually contaminate the permeate [4]. The search for effective BPA removal from wastewater is therefore of high priority as is with many similar micropollutants. n Correspondence to: School of Chemical Engineering, The University of New South Wales, Sydney, NSW 2052, Australia. Tel.: þ 61 2 9385 4813; fax: þ 61 2 9385 5966. E-mail address: [email protected] (V. Chen).

http://dx.doi.org/10.1016/j.memsci.2014.06.027 0376-7388/& 2014 Elsevier B.V. All rights reserved.

The use of white rot fungi enzymes (e.g. laccase, lignin peroxidase, and versatile peroxidase) to eliminate BPA has been extensively studied [5–7]. Among them, laccase is the most preferable choice because of its low cost, good stability under moderate operating conditions, high catalytic efficiency, and broad specificity [8]. Furthermore, the oxidation of the phenolic compounds with laccase will form large insoluble polymers (Fig. 1) which can be easily removed from aqueous phase by filtration or sedimentation [9]. Although numerous studies have been dedicated to the use of laccase for BPA removal [10–12], the use of free laccase is undesired due to its high sensitivity to the industrial operating conditions, as well as its poor stability and reusability. The use of immobilized enzyme therefore has been proposed to address these issues. The laccase has been immobilized onto various supports, such as glass, agarose, silica, TiO2, organic gel, chitosan, and kaolinite [13]. Among them, the TiO2 nanoparticle is an ideal immobilization candidate due to its tunable surface structure, high surface area, high hydrophilicity, and good stability [14–16]. In addition, the presence of hydroxyl groups on the TiO2 surface enables further functionalization to couple enzymes via covalent bonding, which is preferable in practice due to the stronger attachment formed between enzyme and support [13]. TiO2 nanoparticles have been used to immobilize

20

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

Fig. 1. Schematic degradation pathway of bisphenol-A in the presence of laccase.

enzymes such as glucose oxidase, lipase and urease, and good immobilization performances were achieved [17–19]. In our previous study, laccase was covalently immobilized onto the TiO2 nanoparticles and nearly 80% of the original laccase activity was preserved [20]. Whilst excellent performance was observed on the laccase immobilized on the TiO2 nanoparticle, the reuse and recycle of these bio-catalytic nanoparticles still remains a major challenge, especially in the wastewater treatment processes. This technical difficulty can be overcome by using membrane which can be easily integrated with the present wastewater treatment processes instead of a particulate immobilization substrate. In addition, the enhanced mass transfer offered by convective flow through the porous membrane can substantially improve the efficiency of the catalytic reaction. The use of both polymeric and inorganic membranes for laccase immobilization has been explored in the past. The performance of the immobilized laccase on polymeric membrane suffered from low activity recovery and poor reusability mainly due to the weak physical bonding formed between the enzyme and membrane, as well as the low surface area of the membrane [21–24]. While promising immobilization performance was observed on the ceramic membrane due to its fine-tuned, well-ordered surface and pore morphology [9], the difficulty of inorganic membrane preparation and its high cost hindered its wider application. The nano-composite membrane incorporated with inorganic particles possesses advantages from both organic and inorganic components including low cost, high stability and flexibility, larger surface area, and controllable surface morphology [25,26]. Several studies have been carried out on the use of such nano-composite membrane for enzyme immobilization for biosensors [14,27–29]. In our previous work, TiO2 particles were blended into the polyethersulfone (PES) membrane to prepare nano-composite membrane for laccase immobilization. Significant improvement in activity, activity recovery and stability was observed in comparison with the pure polymeric membrane [20]. However, due to the inherent limitations of the blending process, the majority of the TiO2 was embedded within the PES matrix therefore not accessible to the laccase to function as the immobilization support. A potential solution is to strategically locate the TiO2 particles on the surface of the polymeric membrane readily for laccase immobilization. Recently, a low temperature hydrothermal (LTH) sol–gel coating technique was developed in our group to generate a thin, robust layer of TiO2 nano-composite coating on the in-house ultrafiltration PES membrane [30]. A uniform coating layer with high TiO2 content was obtained with good stability and excellent fouling resistance. The membrane with such a unique structure and properties can be potentially used as the laccase immobilization substrate to improve the BPA degradation performance. However, little research has been focused on the laccase immobilization on TiO2 coated membranes. In this study, commercial polyvinylidene fluoride (PVDF) microfiltration membranes (0.1 and 0.45 mm pore size) from Millipore were chosen as coating support. Larger microfiltration membrane pores could ensure the formation the TiO2 coating on both membrane surface and inside the membrane pores, which provided more accessible TiO2 nanoparticles for laccase immobilization. In

addition, larger pores also ensured good filtration performance after laccase immobilization, especially during the BPA degradation process where insoluble polymers were formed as degradation products. The PVDF membranes coated with the TiO2 nanoparticles via the LTH sol–gel coating technique were explored as novel nanocomposite substrates. The bio-catalytic membranes were prepared by covalently immobilizing laccase onto the coated membrane through 3-aminopropyltriethoxysilane (APTES) and glutaraldehyde (GLU) sequential immobilization process. Comprehensive characterization techniques were applied to investigate the effect of the coating layer properties on the immobilization performance. In addition, the filtration conditions were assessed and optimized based on the BPA degradation performance. The longterm performance of the bio-catalytic membranes as well as the membrane fouling behavior during the BPA degradation process was also evaluated.

2. Experimental 2.1. Materials Hydrophilic polyvinylidene fluoride (PVDF) membranes with pore sizes of 0.1 and 0.45 mm were purchased from Millipore Pty. Ltd. Laccase from Trametes versicolour (EC 1.10.3.2) was supplied by Sigma-Aldrich. 2, 20 -Azino-bis-(3-ethyl benzothiazoline-6-sulfonic acid) (ABTS, Sigma-Aldrich) was used as the assay substrate. Titanium isopropoxide (TTIP) (Z97%, Sigma-Aldrich) was used to prepare the sol solution. 3-Aminopropyltriethoxysilane (APTES, Sigma-Aldrich) and glutaraldehyde (GLU, Ajax Finechem) were used for TiO2 coating layer functionalization. Bisphenol A (BPA) was also supplied by Sigma-Aldrich. All the chemicals were of analytical grade and used without any purification.

2.2. Preparation of sol–gel coated PVDF membrane The detailed preparation procedure of TiO2 sol can be found elsewhere [30,31]. Components were mixed with a molar ratio of TTIP:2, 4-pentanedione:HClO4:H2O:ethanol¼1:0.5:0.5:0.45:4.76. The sol solution was stirred for 1 h at room temperature, and a light yellow transparent sol solution was formed at pH of 1.2. Prior to the sol–gel coating, the PVDF membranes were soaked in Milli-Q water and absolute ethanol for 24 h each to remove all the preservatives, followed by drying at room temperature. Each dip-coating cycle was carried out by the following procedures: the membrane was firstly lowered at a speed of 1.0 mm/s into the sol solution. After 8 s soaking time, the membranes were withdrawn from the solution at a speed of 0.2 mm/s. Subsequently, the coated membranes were dried in the oven at 120 1C for 16 h and then placed in a 90 1C water bath for 24 h. Finally, the membranes were rinsed with Milli-Q water 3 times and dried at room temperature. In this work, up to 4 coating cycles were applied on both 0.1 and 0.45 mm PVDF membranes.

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

21

2.3. Characterization of the sol–gel coated membrane The morphology of the TiO2 coating surface was observed by FESEM Hitachi S900 (4 kV operational voltages). The membrane samples were dried under vacuum and coated with a chromium thin layer prior to imaging. The TiO2 loading on the coated membrane was determined by the high resolution modulated thermo-gravimetric analysis (TGA) Q5000 TA instrument. Membrane was heated up to 1000 1C at a rate of 20 1C/min. For each coated membrane, three samples were taken from different regions of the membrane to evaluate the homogeneity of the coating layer, and the average of three residual weights of TiO2 was reported. Nitrogen adsorption–desorption isotherms were measured at 77 K with Micromeritics Tri-Star 3000 Analyzer system in order to characterize the nano-scale textural properties of the membrane surface. Data were analyzed based on the Brunauer, Emmett, and Teller (BET) and Barrett, Joyner and Halenda (BJH) models. The commercial membranes used in this work had narrowly distributed pore sizes of 0.1 and 0.45 mm, the BJH pore model only studied the mesopore diameter within the range of 1.7–80 nm to reflect the mesoporous structural properties of the TiO2 coating layer and the membrane surface. In this work, the terms “BJH pore” and “mesopore” were used to indicate the nano-scale pore structure (1.7–80 nm) on the membrane surface, especially in the TiO2 coating layer, whilst the term “membrane pore” only referred to the pore diameter of the membrane support (0.1 and 0.45 mm). 2.4. Bio-catalytic membrane preparation and activity assay For all the TiO2 coated membranes, sequential modification by APTES and GLU was applied prior to the laccase immobilization. 2 wt% APTES (with respect to the TiO2 loading on the membrane) was applied to silanize the TiO2 coated membrane surface (6.5 cm in diameter) in absolute ethanol. Then 1 ml 25% (V:V) GLU was used to functionalize the membrane surface in 60 ml Sørensen phosphate buffer solution at pH of 7. Detailed conditions can be found in our previous publication [20]. The uncoated pure PVDF membrane (control) was rinsed with Milli-Q water and ethanol then used directly for laccase immobilization without the above modification processes. In terms of the immobilization procedure, both coated and uncoated membranes (6.5 cm in diameter) were soaked in 30 ml of laccase solution with the different concentrations up to 120 h at 4 1C. All the loosely attached laccase was subsequently removed by rinsing the membranes with Milli-Q water 3 times until no laccase activity was detected in the supernatant. Residual laccase concentration in the supernatant was analyzed by Lowry method to calculate the laccase loading on membrane through the deduction method [32]. The activity of the bio-catalytic membrane was screened by monitoring the oxidation rate of ABTS to ABTS þ at 420 nm using Cary 100 UV–visible Spectrophotometer [33]. The bio-catalytic membrane with 2 cm2 area was suspended in 20 ml ABTS solution (0.5 mM ABTS in pH 3 phosphate–citrate buffer solution) with a stirring speed of 150 rpm. The absorbance was measured once per minute. One unit (U) of laccase activity was defined as the amount of laccase forming 1 mmol ABTS þ per min [20].

Fig. 2. Process scheme of the bio-catalytic membrane reactor for BPA degradation.

degradation process, the trans-membrane pressure and membrane flux were monitored to assess the fouling behavior of the membrane. Quantitative analysis of BPA was carried out with HPLC installed with a UV–visible detector (Shimadzu LC-20AT) and a C18 HPLC column (Alltima HP, 5 mm, 250 mm  4.6 mm id). 0.25 ml of BPA sample was taken from the supernatant in the dead-end cell and mixed with 0.2 ml of 0.1 M HCl immediately to terminate the BPA degradation. Before the HPLC analysis, the mixture was filtered with a PTFE syringe filter (0.2 mm) to remove the insoluble substance. For the HPLC analysis, 50 ml sample was injected and the detection wavelength was set at 278 nm. The mobile phase of water/acetonitrile (50/50, v/v) was run in an isocratic mode at a flow rate of 1 ml/min, and the BPA retention time was 7.670.1 min (column oven temperature 40 1C). The detection limit of the BPA with this method was 0.1 nM. 2.6. Stability of the bio-catalytic membrane The stability of the bio-catalytic membranes in the BPA degradation process was investigated by monitoring the BPA degradation rate up to 4 degradation cycles (24 h for each cycle). All the operational parameters for each degradation cycle remained the same, which were identical to the BPA degradation test described in Section 2.5. Inductively coupled plasma mass spectrometry (ICP-MS) (PerkinElmer quadrupole NexION) was used to analyze the copper and titanium quantity in the remaining solution in the membrane cell at the end of each BPA degradation cycle. Before testing, the samples were subject to pre-treatment as follows: 1 ml sample was mixed with 3 ml hydrochloric acid before being digested in the microwave reaction system at 200 1C for 15 min. The copper and titanium concentrations were calculated based on the calibration curve of the ICP standard solutions.

2.5. BPA degradation by the bio-catalytic membrane 3. Results and discussion BPA degradation study was performed in a stainless steel deadend membrane cell. The membrane area was 0.00084 m2, and a peristaltic pump was used to re-circulate the permeate solution. 40 ml of 150 mM BPA in 0.1 M acetate buffer (pH 5.5) solution was used (Fig. 2). For each BPA degradation cycle, the experiment was conducted for 24 h at room temperature. During the BPA

3.1. Characterization of the coated membrane 3.1.1. Surface morphology The surface morphology of both 0.1 and 0.45 mm PVDF membranes coated with TiO2 was examined by the SEM and the results

22

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

Pure PVDF

1 coating cycle

2 coating cycle

3 coating cycle

4 coating cycle

Pure PVDF

1 coating cycle

2 coating cycle

3 coating cycle

4 coating cycle

Fig. 3. SEM images of the surface of (a) 0.1 mm and (b) 0.45 mm PVDF membrane coated with TiO2 via the LTH sol–gel process.

of the upper surface images are presented in Fig. 3 (the insets are 50 K magnification images showing the nano-scale morphology of the coating layer). In this work, symmetric membranes were applied, and during the dip coating process the membrane was fully submerged in the sol solution for 8 s. Therefore, the TiO2 coating was formed on both outer and inner surface of the membrane, which was confirmed by the EDX image (result not shown). As a result, the upper surface SEM images were shown as representative. In the case of 0.1 mm PVDF membranes with 1 and 4 coating cycles (Fig. 3a), an evenly coated TiO2 layer was observed consisting of TiO2 nanoparticle clusters with an average size of 25– 30 nm. Similar morphology was also observed on 2 and 3 coating cycles (results were not shown here). Furthermore, no obvious decrease in membrane pore size was observed after 4 coating cycles, indicating the formation of a thin TiO2 coating layer on the 0.1 mm membranes. In comparison, very different surface morphologies were observed on the 0.45 mm membranes where the nanoparticle clusters with the sizes up to 150 nm were formed after 2 coating cycles (Fig. 3b), indicating the severe aggregation of the TiO2 particles took place during the gelation process. In addition, after 4 coating cycles, the blockage of membrane pores was evidenced.

According to the theoretical dip-coating model proposed by Brinker [34], during the sol layer gelation process nearly all the solvent evaporation takes place at the liquid–vapor interface, and longer gelation time is required for a thicker sol layer. As a result, a uniform and well distributed TiO2 coating layer is more likely to be formed when a thin sol solution layer covers the substrate surface prior to the gelation process. Therefore, the formation of a much thinner sol solution layer can be expected in the case of the 0.1 mm membranes than the 0.45 mm membranes due to the smaller membrane pore size, as less sol solution can be held in the pore structure, thus resulting in an evenly distributed thin coating layer as observed in Fig. 3a. Similar behavior was also reported by other researchers [35,36]. For instance, Roberts et al. [37] applied dipcoating technique to prepare the inorganic coating layer on a reticulated vitreous carbon support. A thin sol layer was formed after withdrawn, leading to a uniform coating layer after gelation. In addition, no obvious change in the coating layer morphology or nanoparticle cluster size was observed with more coating cycles. By contrast, the large pore size of the 0.45 mm membranes can potentially trap the sol solution inside the pores, consequently forms a thick sol solution layer covering the polymer surface, which requires longer time for gelation. In addition, during the

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

Table 1 TiO2 loading on the sol–gel coated membrane.

23

Table 2 BET surface area, mean BJH pore diameter and pore volume for the PVDF membrane with different coating cycles.

Support

Coating cycle

TiO2 loading (wt%)

0.1 mm PVDF membrane

1 2 3 4

Cycle Cycles Cycles Cycles

1.9 70.1 3.4 70.2 5.9 70.3 7.1 70.3

0.45 mm PVDF membrane

1 2 3 4

Cycle Cycles Cycles Cycles

1.3 70.2 2.0 70.4 5.0 70.9 4.7 70.7

gelation process, the viscosity of the sol coating layer gradually increases due to the hydrolysis and poly-condensation, which allows for the repulsive particles to re-align and forms a gelation network, thus leading to nanoparticle aggregation. This behavior was evident in this study as shown in Fig. 3b. Similar phenomena were also observed by other researchers [38–40]. 3.1.2. Effect of the coating cycle on TiO2 loading The effect of the coating cycles on the TiO2 loading was studied through thermo-gravimetric analysis (Table 1). A steady increase of the particle loading with coating cycles was observed on the 0.1 mm PVDF membrane. In this work, for each coated membrane three TGA samples were taken from different regions of the membrane. Similar TGA results with low standard deviation were observed for each membrane sample. This observation proved that the coating layer was homogenous in terms of TiO2 loading. In comparison, the TiO2 loading on the 0.45 mm membranes was relatively lower than the 0.1 mm counterparts with the same coating cycles, and higher standard deviations indicated the uneven coating layer on the membrane surface. Additionally, a decrease in TiO2 loading was observed on the 0.45 mm membrane with 4 coating cycles. This observation might be due to the detachment of the large TiO2 aggregates during the rinsing step, consequently leading to a lower particle loading and poor distribution. Furthermore, the commercial PVDF membranes used in this work have been modified by the manufacturer to provide hydroxyl groups on their surface [21], which can potentially benefit the formation of a stable TiO2 coating layer due to the [R–O–Ti] bond formed during the gelation process. However, in the case of the 0.45 mm membranes, the gel pre-formation which occurred in the thick sol layer involved the hydrolysis and poly-condensation reactions, which consumed the hydroxyl groups required for the formation of the [R–O–Ti] bond [41]. Therefore, the resulted TiO2 coating on the 0.45 mm membrane was less stable, thus leading to an uneven coating and the loss of TiO2 content. 3.1.3. BET surface area and BJH pore measurements The textural properties of the TiO2 coating layer and the membrane polymer surface were studied through the nitrogen adsorption–desorption measurements. The BET surface area and BJH pore diameter and pore volume of the membranes with different coating cycles were reported (Table 2). In terms of the BJH pore diameter and pore volume, only the pores in the range of 1.7–80 nm were considered in order to reflect the membrane surface mesoporous structures. Two types of pores were present in the TiO2 coated membrane: the pores from the membrane substrate and the mesopores on the membrane surface (including the TiO2 coating layer and the polymer surface). For the membrane pores, the commercial PVDF membranes used in this study have much larger pores (0.1 70.005 mm and 0.45 70.02 mm ) than the globular dimensions of the laccase (7.0  5.0  5.0 nm3) [42], whilst the mesopores within the TiO2 coating layer exhibited pore

BJH pore volume (pore volume between 1.7 and 80 nm), cm3/g

Support

Coating cycle

BET surface area, m2/g

0.1 mm PVDF membrane

0 1 2 3 4

3.6 7 0.3 5.5 7 0.4 8.17 0.6 8.6 7 0.6 8.0 7 0.6

22.0 7 1.7 23.8 7 2.0 20.2 7 2.0 24.07 2.3 15.17 1.7

0.00617 0.0004 0.0081 7 0.0007 0.0147 0.0011 0.0167 0.0013 0.0177 0.0013

0.45 mm PVDF membrane

0 1 2 3 4

3.9 7 0.5 3.9 7 0.5 4.4 7 0.6 6.5 7 0.9 5.2 7 1.1

6.3 7 0.4 13.5 7 1.3 11.17 1.3 8.737 1.7 18.28 7 3.4

0.0056 7 0.0005 0.00917 0.001 0.0081 7 0.001 0.01267 0.0013 0.0086 7 0.0018

Mean BJH pore diameter (pore size between 1.7 and 80 nm), nm

size in the same magnitude of the size of laccase. Previous studies have demonstrated that the mesoporous structure with pore size similar or slightly larger than the dimension of the enzymes has a major contribution to the enzyme immobilization [16,43,44]. In this regard, the mesopores present in the coating layer are likely to determine the laccase immobilization performance [45]. Therefore, the BJH pore diameter and pore volume were only calculated from the pore sizes ranging between 1.7 and 80 nm to reveal the surface mesoporous structure. As shown in Table 2, for both 0.1 and 0.45 mm membranes, increasing coating cycles led to an increase in the BET surface area until the 4th coating cycle. A drop in BET surface area was observed on both 0.1 and 0.45 mm membranes after the 4th coating cycle, but more pronounced on the 0.45 mm membrane. In the case of 0.1 mm membrane, this was due to the compression and densification of the coating layer as a result of the repeated heat treatment [30,46]. Whereas for the 0.45 mm membrane, such a reduction in BET surface area was mainly caused by the detachment of the large TiO2 aggregates, which was evidenced by the loss of TiO2 loading as shown in Table 1. In terms of the BJH pore volume, similar trend to the BET surface area was observed for both 0.1 and 0.45 mm membranes, except a slightly increase of the BJH pore volume after 4th coating cycle for the 0.1 mm membrane. In terms of the BJH pore diameter, no obvious change was observed on the 0.1 mm membranes with 1–3 coating cycles, whereas the reduction of BJH pore diameter after the 4th coating cycle could be attributed to the coating layer compression and densification, which was in agreement with the BET surface area measurement. By contrast, the BJH pore diameters of the 0.45 mm PVDF membranes were generally smaller than the 0.1 mm counterparts due to the formation of nanoparticle aggregations. The dramatic increase in the BJH pore size after the 4th coating cycle could be attributed to the voids left by the detachment of the large TiO2 aggregates. The relatively larger BJH pore diameters observed on the 0.1 mm membranes than the 0.45 mm membranes indicated the formation of a looser coating layer structure in the case of the 0.1 mm membrane, which also contributed to the more pronounced compression and densification behavior observed on the 0.1 mm membranes. 3.2. Preparation of the bio-catalytic membrane 3.2.1. Effect of the immobilization parameters The membranes with different coating cycles were immersed in the laccase solution with different enzyme concentrations (50–600 mg/ml) to assess the optimal laccase concentration for

200

0.3

150

0.2

100

0.1

50

0.0

0

100 200 300 400 500 600 Initial laccase concentration (μg/ml)

0 700

240 200

3.2.2. Effect of the coating cycles and the membrane pore size As shown in Fig. 5a, the uncoated 0.1 mm membranes exhibited the lowest laccase loading (23 mg cm  2), which was mainly due to the weak van der Waals' force between the laccase and supports. In comparison, the coated 0.1 mm membrane exhibited improved laccase loading along with more coating cycles, whilst the laccase loading reached plateau at the 4th coating cycle. The mesopore structure formed during the coating process increased the surface area available for the laccase immobilization, which was evidenced by the BET surface area measurement (Table 2), leading to more anchor points for laccase immobilization. Similar observations were reported on other mesoporous supports including silica and carbon supports [49–51]. In addition, the covalent bonding formed between the laccase and TiO2 coating layer minimized the detachment of laccase, which also contributed to the higher loading compared with the uncoated membranes. The plateau stage at the 4th coating cycle was a result of the reduction of the surface area caused by the compression and densification of the coating layer as previously mentioned. In terms of the laccase activity on the 0.1 mm bio-catalytic membrane, a substantial increase in activity from 0 to 3 coating cycles was observed. Such a significant improvement was mainly

Laccase loading on membrane (μg/cm2)

0.4

160 0.3 120 0.2 80 0.1

40 0

Fig. 4. Effect of the initial laccase concentration on the laccase loading and activity (0.1 mm membrane with 3 coating cycles, ABTS as assay substrate).

0

1

2 Coating cycle

3

4

0.0

0.5

240 200

0.4

160 0.3 120 0.2 80 0.1

40 0

0

1

2 Coating cycle

3

4

0.0

100

80

Activity recovery (%)

immobilization. Similar trends were observed on all the membranes, and the result from the 0.1 mm PVDF membrane with 3 coating cycles is shown in Fig. 4 as being representative. The laccase loading increased along with initial laccase concentration. In terms of the bio-catalytic membrane activity, an increase in activity was observed when increasing the laccase concentrations from 50 to 400 mg/ml. Further increase in laccase concentration did not improve the activity, such an observation indicated that all the accessible surface area was occupied by the laccase at this concentration level [47], and further increase in enzyme concentration led to the undesired lateral interaction between the overcrowded laccase, thus causing the loss in activity. Therefore, the plateau observed on the laccase activity at higher concentration was the trade-off between the increased laccase loading and the loss of activity due to the laccase over-crowding [48,49]. The effect of the immobilization time on the laccase loading was also investigated by monitoring the laccase concentration in the supernatant during the immobilization process. For all the membranes, the laccase concentration in the supernatant was stabilized after 70 h, indicating saturation of binding sites has been achieved at this point and further increase in the immobilization time would not improve the laccase loading on membrane. Therefore, 400 mg/ml initial laccase concentration and 70 h immobilization time were used for all the subsequent immobilization experiments.

0.5

Bio-catalytic membrane activity (U/cm2)

0.4

Bio-catalytic membrane activity (U/cm2)

250

Laccase loading on membrane (μg/cm2)

0.5

Laccase loading on membrane (μg/cm2)

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

Bio-catalytic membrane activity (U/cm2)

24

0.1 μm membranes 0.45 μm membranes

60

40

20

0

1 coating cycle

2 coating cycles

3 coating cycles

4 coating cycles

Fig. 5. Laccase loading and activity on (a) 0.1 mm and (b) 0.45 mm membrane with different coating cycles, and (c) the comparison of the activity recovery for the 0.1 and 0.45 mm membranes with different coating cycles (ABTS as assay substrate).

due to the increase in BET surface area. Larger surface area prevented the undesired lateral interaction between the immobilized laccase and ensured the good contact between the laccase and immobilization support thus leading to a higher apparent activity [49,52]. However, a nearly 50% drop in activity was observed on the membrane with 4 coating cycles, even though it had relatively unchanged laccase loading and only marginal decreases in BET surface area when compared with the membrane with 3 coating cycles. Such a loss in activity could be attributed to the much smaller mesopores (15.1 nm for the membrane with 4 coating cycles compared to 24.0 nm for the membrane with 3 coating cycles) which significantly increased the mass transfer

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

resistance between the immobilized laccase and the assay substrate (ABTS). Additionally, the small mesopores also induced the undesired lateral interaction between the crowded laccase molecules which also reduced the apparent activity. For the 0.45 mm bio-catalytic membranes, similar laccase loading and activity trends over coating cycles were observed in comparison with the 0.1 mm membranes (Fig. 5b). The highest apparent activity obtained on the 0.45 mm membrane with 3 coating cycles was 0.41 70.4 U/cm2, which was very close to the 0.1 mm membrane with 3 coating cycles (0.42 70.4 U/cm2). The laccase loading on all the 0.45 mm membranes were generally higher than the 0.1 mm counterparts regardless the coating cycles, which can be attributed to the smaller surface mesopores in the 0.45 mm membranes where laccase can be easily trapped inside thus increasing the loading. As shown in Fig. 5c, the activity recoveries of the 0.45 mm membranes were lower than the 0.1 mm membranes for all the coating cycles, which were mainly caused by the smaller mesopores on the 0.45 mm membranes. The smaller mesopores had three adverse effects on the immobilized laccase activity: (i) inducing the undesired lateral interaction between the over-crowded laccase molecules, (ii) causing multipoint attachment between the laccase and support, and (iii) lowering the ABTS accessibility to immobilized laccase. The performance of the immobilized enzymes on mesoporous supports has been investigated by other researchers [50,53,54]. A proper choice of the mesopore diameter which better facilitates the enzymatic performance is partially dictated by the size and shape of the enzymes, with a larger mesopore generally preferred as a rule of thumb. In the case of laccase, Wang et al. [55] immobilized laccase onto the mesoporous silica nanoparticles with average pore diameter of 18 nm (in the same range of the current work), which resulted in a good activity recovery. In another work, Zhu et al. reported that the silica spheres with larger mesopores displayed higher laccase activity compared with the support with smaller pores [56]. Another research carried out by Liu et al. demonstrated that the mesoporous silica support with average pore size of 10 nm showed much higher laccase loading but lower activity recovery than the supports with larger pores, which was in line with our observation [49]. In the current work, the coating layer on the 0.1 mm membrane had mesopore size around 20 nm, which was 2–4 times larger than the laccase globular dimensions (7.0  5.0  5.0 nm3) [42], while the mesopores on the 0.45 mm membrane coating layer was around 10 nm, only slightly larger than the globular dimensions of the laccase. In this study, 20 nm mesoporous structure was considered as the optimal pore size for the laccase immobilization due to the higher activity and activity recovery observed on the 0.1 mm membranes. The results from the current work were also compared against our previous work on the laccase immobilization on the TiO2 nanoparticles blended membrane, substantial improvement in terms of the laccase loading, apparent activity and activity recovery rate was achieved (as shown in Table 3) [20]. Such a significant

25

improvement was the result of the vastly different membrane architecture. In the blended membrane, the majority of the TiO2 nanoparticles were embedded within the polymer matrix thus inaccessible to the laccase. Therefore, considerable amount of laccase was attached to polymer surface which led to enzyme denature due to the unfavorable interaction between hydrophobic polymer and adsorbed enzymes. In the current work, the TiO2 nanoparticles were presented on both membrane surface as well as membrane pores, which provided good accessibility for laccase immobilization. Such a nanoparticle mesoporous coating structure also offered preferable micro-environments (high curvature, high hydrophilicity) which could better preserve enzyme's natural conformation and ensure good affinity between the immobilized enzyme and substrates. As a result, the TiO2 coated membrane exhibited higher activity recovery compared with blended membranes. Furthermore, the results (laccase loading, activity and activity recovery) obtained from current work were also comparable or higher than other works was observed (Table 3) [9,21,45,48,57], proving the effectiveness of TiO2 coating layer in laccase immobilization. 3.3. BPA degradation by the bio-catalytic membrane 3.3.1. Determination of the optimal operational flux The removal of BPA by the bio-catalytic membranes was investigated in this work by recirculating the effluent stream through the membrane mounted in a dead-end cell, and Fig. 6 shows the effect of the operational flux on the BPA removal performance for both 0.1 and 0.45 mm bio-catalytic membranes. The bio-catalytic membrane with 1 coating cycle was studied as a representative to determine the optimal operational flux. As shown in Fig. 6, the background adsorption tests were carried out on the 0.1 and 0.45 mm membranes with 1 coating cycle (no immobilized laccase) under the flux of 20 L/m2 h, and marginal adsorption of BPA (less than 8% after 24 h) was observed. This is due to the hydrophilic nature of the TiO2 coated membrane used in this work, which has low affinity towards hydrophobic BPA [30,58]. In terms of the effect of the operational flux, for the 0.1 mm membrane with 1 coating cycle, the highest BPA removal efficiency was achieved under 20 L/m2 h (Fig. 6a), further increase in the operational flux to 30 L/m2 h slightly reduced the BPA removal efficiency. Similarly, for the 0.45 mm membrane, operational flux of 20 L/m2 h provided the highest BPA removal efficiency (Fig. 6b). However, it was observed that further increase in the operational flux to 30 L/m2 h significantly reduced the BPA removal efficiency: only 30% of BPA was removed under this condition after 24 h. The BPA removal efficiency by the bio-catalytic membranes was mainly determined by the activity of the immobilized laccase and the mass transfer efficiency. Higher operational flux enhanced the mass transfer efficiency between the immobilized laccase and BPA, thus contributing to a better BPA removal performance. However,

Table 3 Performance of the immobilized laccase using various supports. Support type

Modification process

Laccase loading (mg/cm2)

Bio-catalytic membrane activity (U/cm2)

Activity recovery (%)

Reference

0.1 mm PVDF membrane 0.45 mm PVDF membrane In house prepared PES 0.45 mM PVDF membrane Polypropylene membrane p(HEMA-g-GMA)–NH2 Ceramic membrane Poly HEMA film

3 Coating cycles 3 Coating cycles TiO2 nanoparticle blended in PES Hydrazine hydrate Chromic acid/GLU 2-bromo-2-methylpropionyl bromide GLU crosslinking Physical adsorption

119 710 1617 15 52 7 5 7.5 36 Up to 139 148 139

0.42 7 0.04 0.417 0.04 0.131 70.013 0.36 – Up to 0.099 – 0.099

397 3.5 287 4.1 7.4 7 0.6 20 – – – 3.55

Current work Current work [20] [21] [58] [49] [9] [46]

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

180

180

150

150

BPA Concentration / μM

BPA Concentration / μM

26

120 90 60 30 0

0

5

10

15 Time (h)

20

90 60 30

0

5 10 BPA Solution Recirculation Cycle

15

180

150

BPA Concentration / μM

BPA Concentration (μM)

180

120 90 60 30 0

120

0

25

0 L/m2 h 5 L/m2 h 10 L/m2 h 20 L/m2 h 30 L/m2 h

0

5

10 15 Time (h)

20

25

0 L/m2 h 5 L/m2 h 10 L/m2 h 20 L/m2 h 30 L/m2 h

150 120 90 60 30 0

0

5 10 BPA Solution Recirculation Cycle

15

Fig. 6. Effect of the operational flux on the BPA removal efficiency by (a, c) 0.1 mm and (b, d) 0.45 mm bio-catalytic membranes with 1 coating cycle (laccase loading: 87 mg/cm2 for 0.1 mm and 135 mg/cm2 for 0.45 mm membrane).

it also shortened the residence time of BPA on the bio-catalytic membranes, leading to a lack of reaction between BPA and immobilized laccase on the membranes. Furthermore, higher flux also increased the hydraulic shear force near the membrane surface, which could potentially lead to the detachment of the TiO2 coating layer and immobilized laccase from the membrane surface, or the conformational change of the immobilized enzyme. Therefore an optimal operational flux existed because of the tradeoff between these factors. Similarly, an optimal operational flux was also found in the bio-catalytic fluidized bed reactors [59–62]. On the other hand, the polymer products formed from the BPA degradation process can accumulate on the membrane surface and bury the immobilized laccase, which increase the mass transfer resistance between the laccase and BPA especially at a higher flux. This may also contributed to the reduced BPA removal efficiency observed in this study under the flux of 30 L/m2 h. In order to better understand the effect of the operational flux, the residual BPA concentration vs. number of recirculation cycles was presented in Fig. 6c and d, where the number of recirculation cycles was calculated by dividing the reaction time by the hydraulic retention time. Very similar results were observed with both sized membranes. Under static hydraulic condition (0 LMH), both 0.1 and 0.45 mm bio-catalytic membranes could degradation BPA over time, which indicated that the molecular diffusion could still provide the BPA concentration gradient in the bulk solution and near the membrane surface which facilitated the degradation of BPA under the static conditions. When the operation flux of 5–20 LMH was applied, improved BPA degradation rate was observed compared with 0 LMH, indicating the mass transfer

efficiency was improved with the BPA solution flowing through the bio-catalytic membrane. Additionally, similar amount of BPA was degraded within each recirculation cycle, which could be explained by the interaction between increased mass transfer efficiency and reduced BPA residence time as discussed above. However, further increase the operational flux to 30 LMH reduced the BPA degradation efficiency for each cycle of liquid recirculation especially for 0.45 mm membrane. The reason could be attributed to the increased hydraulic shear forces that led to immobilized enzyme denature as discussed above. In this work, the 0.1 and 0.45 mm membranes with 2–4 coating cycles also exhibited similar BPA removal profiles under different operational flux (results not shown), therefore 20 L/m2 h was regarded as the optimal operational flux for the bio-catalytic membrane system, and used in the subsequent studies.

3.3.2. Effects of the coating cycles and the membrane pore size As shown in Fig. 6a and b, the BPA removal profile consisted of a rapid BPA concentration decrease during the first 4 h followed by a much slower gradual reduction during the remaining period (4–24 h). In order to better compare the BPA removal efficiency by the bio-catalytic membranes with different coating cycles, the BPA removal rate at 5 h was used as the indicator and the results are presented in Fig. 7. In the case of the 0.1 mm membranes, the membrane with 3 coating cycles exhibited the highest BPA removal rate, which was in agreement with the previous membrane activity measured using ABTS as substrate. In addition, the trends of the BPA

100

240 200

BPA removal (%)

80

160 60 120 40 80 20 0

40

0

1

2

3

4

0

Laccase loading on membrane (μg/cm2)

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

100

240

80

200 160

60 120 40 80 20 0

40

0

1

2 3 Coating cycle

4

0

Laccase loading on membrane (μg/cm2)

BPA removal (%)

Coating cycle

Fig. 7. BPA removal rate after 5 h for (a) 0.1 mm and (b) 0.45 mm bio-catalytic membranes with different coating cycles, background BPA adsorption was deducted.

removal rate for the membranes with different coating cycles were in line with the laccase loading on the 0.1 mm bio-catalytic membranes. In comparison, the 0.45 mm bio-catalytic membranes exhibited slightly higher BPA removal rate than the 0.1 mm counterparts. Furthermore, for the 0.45 mm membranes, the highest BPA removal rate was observed with the membrane with 4 coating cycles, which was different from the previous activity results where the membrane with 3 coating cycles exhibited the highest activity using ABTS as assay substrate (Fig. 5b). Nevertheless, the laccase loadings on the 0.45 mm membranes were in line with the BPA removal rates. As previously discussed, the smaller mesopores present in the 0.45 mm membranes led to a lower activity recovery partly due to the lower accessibility of ABTS. However, better BPA removal performances were obtained with the 0.45 mm membranes than the 0.1 mm counterparts. This observation could be attributed to the smaller BPA molecular dimension (around 1.12  0.33  0.58 nm3) [63] compared with ABTS (around 1.7  1  1 nm3) [64]. Therefore the immobilized laccase within the mesopores were still accessible by BPA, considering the mesopores on the 0.45 mm membranes were around 10 nm. Consequently, the higher BPA removal rate achieved with the 0.45 mm membranes over the 0.1 mm membranes was mainly due to the higher laccase loading. However, it should be noted that 90% BPA removal rate was obtained with the 0.1 mm membrane with about 120 mg/cm2 laccase loading (3 coating cycles) while the similar BPA removal rate was achieved by 0.45 mm membrane with much higher laccase loading (4 coating cycles with 190 mg/cm2). This observation could be attributed to the laccase deformation caused by the multi-point attachment on the 0.45 mm membranes or conformational change caused by the over-crowded laccase.

27

3.4. Stability of the bio-catalytic membrane Four consecutive BPA degradation cycles (24 h for each cycle, 40 ml 150 mM BPA, room temperature, pH 5.5 and operational flux of 20 L/m2 h) were performed on the bio-catalytic membrane to examine the stability of the immobilized laccase. At the end of each run, all the remaining BPA solution in the membrane cell was replaced with 40 ml of freshly prepared BPA solution. Furthermore, the remaining BPA solution after each degradation cycle was analyzed by ICP-MS to monitor the copper and titanium content in the solution. The BPA degradation speed was calculated based on the BPA concentration reduction profile in the first hour of each degradation cycle, where a linear profile was observed as a function of time. The BPA removal rate at 5 h was used as the indicator to reflect the bio-catalytic membrane performance in each degradation cycle, and the results are presented in Fig. 8. Comparing results from the 0.1 mm bio-catalytic membranes, the uncoated biocatalytic membrane lost all BPA removal efficiency after 4 cycles. As expected, the relatively weak physical adsorption via van der Waals' force could not provide sufficient laccase attachment on support, as a result, the immobilized enzyme could easily detach from the support after repeated use. In comparison, the coated 0.1 mm bio-catalytic membranes exhibited improved stability due to the strong covalent bonding formed between the laccase and TiO2. In this work, the optimal stability was obtained on the 0.1 mm bio-catalytic membrane with 1 coating cycle, where the BPA removal efficiency at the 4th degradation cycle was nearly the same as at the 1st degradation cycle. However, because the 0.1 mm membrane with 3 coating cycles exhibited the highest initial BPA removal rate, it still retained the highest BPA removal rate after the 4th degradation cycle. The stability of immobilized laccase was also investigated in previous studies. Results in current work showed that up to 91.7% original BPA degradation speed was preserved after 96 h of operation (4 degradation cycles), which are comparable or better than stability results from other researchers. For instance, when laccase was immobilized on polyacrylonitrile beads, after 30 successive degradation cycles (90 min for each cycle), 85% of original BPA degradation speed was preserved [65]. In the study by Songulashvili et al. [66], porous silica beads were applied as the laccase immobilization support, and further application of the resulting bio-catalyst in BPA degradation revealed no loss of degradation speed after three consecutive BPA degradation cycles (60 min for each cycle). In addition, the silica based bio-catalytic nanoparticles [67] and cross-linked laccase aggregates [68] prepared by Cabana et al. were applied in a packed bed reactor for continuous BPA elimination, and both exhibited good stability for up to 7 days. GLU was applied in all these works as the crosslinker, indicating the effectiveness of covalent bonding in the preservation of laccase stability. However, it should be noted that all these stability tests were carried out in a buffer solution, which would be different from practical wastewater treatment environments. There is currently no research about the immobilized laccase under wastewater conditions using BPA as the targeted substrate. The investigation in this line of research is currently undergoing. Analyzing the results from the 0.45 mm bio-catalytic membranes, all the membrane experienced remarkable loss in BPA removal rate after repeated use regardless the coating cycles. Several factors can contribute to the loss of bio-catalytic efficiency: laccase deformation, laccase detachment, and increase in mass transfer resistance. Previous research has demonstrated that the copper atoms (the active core of laccase) are the backbones of the laccase conformation, and the copper ions will enter the aqueous phase once the immobilized laccase is deformed [69]. Therefore,

28

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

100

BPA removal (%)

80 60

Uncoated 1 coating cycle 2 coating cycles 3 coating cycles 4 coating cycles

40 20 0

1

2

3

4

Degradation run 100

BPA removal (%)

80 60

3.5. Fouling behavior on the bio-catalytic membrane Uncoated 1 coating cycle 2 coating cycles 3 coating cycles 4 coating cycles

40 20 0

1

2

3

4

Degradation cycle

100

100

80

80

60

60

40

40

20

0

20

0.1 μm membrane 0.45 μm membrane

1

2 3 Degradation cycle

4

0

Remaing active laccase on membrane (%)

Fig. 8. BPA removal rate after 5 h for (a) 0.1 mm and (b) 0.45 mm bio-catalytic membranes in repeated degradation cycle (laccase loading: 0.1 mm membranes: 23 mg/cm2 (uncoated); 87 mg/cm2 (1 coating cycle); 95 mg/cm2 (2 coating cycle); 119 mg/cm2 (3 coating cycle); 113 mg/cm2 (4 coating cycle); and 0.45 mm membranes: 7.5 mg/cm2 (uncoated); 135 mg/cm2 (1 coating cycle); 150 mg/cm2 (2 coating cycle); 161 mg/cm2 (3 coating cycle); 188 mg/cm2 (4 coating cycle)).

Relative BPA degradation speed (%)

the 0.45 mm membrane with 3 coating cycles, it only preserved approximately 30% of the original BPA degradation speed at the 4th degradation cycle. This observation indicated that a large proportion of the active laccase on the 0.45 mm membrane with 3 coating cycles cannot be accessed by BPA, which can be partially attributed to the smaller mesopores. Even though BPA could enter these small mesopores as aforementioned, the polymer formed after the BPA degradation can easily block the mesopores and make the immobilized laccase inaccessible by other BPA molecules, thus leading to a significant loss in BPA degradation efficiency after repeated use. In comparison, the remaining active laccase and the relative BPA degradation speed displayed good agreement for the 0.1 mm membrane during the stability tests, indicating limited impact on the mass transfer resistance caused by the polymer products from the degradation process. This observation could be attributed to the larger mesopores in the 0.1 mm bio-catalytic membranes. Furthermore, the titanium content in the remaining solution at the end of each degradation cycle was also measured by ICP-MS to investigate the stability of the TiO2 coating layer. Less than 1% loss in TiO2 was observed on both 0.1 and 0.45 mm membranes, indicating excellent TiO2 coating layer stability under the current BPA degradation operating conditions.

Fig. 9. Relative BPA degradation speed and the remaining active laccase on the biocatalytic membrane during the stability tests (membrane with 3 coating cycles, 20 L/m2 h, laccase loading 119 mg/cm2 for 0.1 mm and 161 mg/cm2 for 0.45 mm membrane).

the detached and deformed laccase can be quantified by analyzing the copper quantity in the remaining solution at the end of each cycle. Subsequently the remaining active laccase on the biocatalytic membrane can be calculated by deducing the original laccase loading on the membrane by the detached and deformed laccase in the remaining solution. The results in Fig. 9 demonstrated that even though over 60% of active laccase remained on

As aforementioned, the polymer products formed during the BPA degradation process can potentially accumulate on the membrane surface and clog the membrane pores, consequently causing membrane fouling [9]. In this study, the trans-membrane pressure (TMP) was monitored during the BPA degradation process for up to 96 h (4 degradation cycles) to evaluate the degree of membrane fouling. The 0.1 mm membrane with 3 coating cycles was examined due to its good initial BPA removal efficiency and long term stability. As shown in Fig. 10, no obvious increase in TMP was observed after 96 h, indicating membrane fouling during the BPA degradation process was negligible. Such a good anti-fouling property of this membrane can be attributed to the hydrophilic nature of the TiO2 coating layer which enhanced the membrane antifouling performances [30].

4. Conclusions In this work, laccase was successfully immobilized on the TiO2 sol–gel coated 0.1 and 0.45 mm PVDF membranes. The results revealed that both the pore size of the membrane support and the coating cycles had significant impact on the bio-catalytic membrane performance in terms of apparent activity, laccase loading, activity recovery, BPA removal efficiency, and laccase stability. Analyzing the membranes with different pore sizes, under current coating parameters, the 0.1 mm membranes appeared to be a better support for laccase immobilization as they exhibited higher activity recovery and better stability in the BPA degradation. By contrast, even though the 0.45 mm membranes gave higher laccase loading and better initial BPA degradation efficiency, they displayed lower activity recovery and poor stability in the BPA degradation due to the smaller coating layer mesopores leading to the laccase deformation and higher mass transfer resistance during the BPA degradation process. This study also demonstrated that the coating cycles had strong influence on the bio-catalytic membrane performance. Compared with the uncoated bio-catalytic membranes, the coated 0.1 and 0.45 mm bio-catalytic membranes exhibited substantial improvement in terms of apparent activity, laccase loading and BPA

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

3.0

150 2nd cycle

3rd cycle

4th cycle [4]

2.5

120

2.0 90 1.5 60 1.0 30

TMP (kPa)

BPA concentration (μM)

1st cycle

[5]

[6]

[7]

0.5

[8]

0 0

20

40

60

80

0.0 100

[9]

Time (h) Fig. 10. TMP Profile during the BPA degradation process (0.1 mm membrane with 3 coating cycles, 20 L/m2 h, laccase loading 119 mg/cm2 membrane).

[10]

[11]

degradation performance. Both 0.1 and 0.45 mm bio-catalytic membranes with 3 coating cycles displayed the highest apparent activity with ABTS as assay substrate. In the BPA degradation test, even the 0.45 mm bio-catalytic membrane with 4 coating cycles displayed the highest initial degradation efficiency, the 0.1 mm membrane with 3 coating cycles offered the optimized BPA removal performance due to its high initial degradation efficiency and good stability. The fouling behavior of the bio-catalytic membrane was also examined in this study; only marginal increase in TMP was observed on the 0.1 mm membrane with 3 coating cycles after 4 consecutive BPA degradation cycles, indicating good fouling resistance of this particular membrane. Furthermore, superior stability was achieved on the 0.1 mm membranes with 3 coating cycles, where less than 10% loss of BPA removal efficiency was observed after 96 h of operation (4 degradation cycles). The result in this work also showed that the absolute TiO2 loading and the structure of the coating layer were both very important to the biocatalytic membrane performance in terms of apparent activity, activity recovery and membrane stability in BPA degradation. Whilst the 0.1 mm coated membranes in this work exhibited good performance in BPA degradation, the bio-catalytic membrane performance under real industrial conditions where a wide range of physico-chemical conditions exist (e.g. various pH, the existence contaminants such as humic acid and potential enzyme inhibitors) is worthy of investigation to examine the feasibility of utilizing such a bio-catalytic membrane system in the wastewater treatment processes. Renewing the enzyme coating and long-term stability remain significant technical challenges to be addressed; however, nanostructured TiO2 coated membranes provide a potential alternative to immobilize a wide range of enzymes for water treatment and chemical reactors.

[12]

[13]

[14] [15]

[16]

[17]

[18]

[19]

[20] [21]

[22]

[23]

[24]

[25]

Acknowledgments This research was supported under Australian Research Council's Discovery Projects funding scheme (DP1095930). Jingwei Hou gratefully acknowledges the financial support of the China Scholarship Council (CSC) and the University of New South Wales (UNSW).

[26] [27]

[28]

[29]

References

[30] [1] E. Burridge, Bisphenol A: product profile, Eur. Chem. News 17 (2003) [2] L.N. Vandenberg, R. Hauser, M. Marcus, N. Olea, W.V. Welshons, exposure to bisphenol A (BPA), Reprod. Toxicol. 24 (2007) 139–177. [3] C.A. Richter, L.S. Birnbaum, F. Farabollini, R.R. Newbold, B.S. C.E. Talsness, J.G. Vandenbergh, D.R. Walser-Kuntz, F.S. vom Saal,

14–20. Human Rubin, in vivo

[31]

29

effects of bisphenol A in laboratory rodent studies, Reprod. Toxicol. 24 (2007) 199–224. T.A. Ternes, M. Stumpf, J. Mueller, K. Haberer, R.D. Wilken, M. Servos, Behavior and occurrence of estrogens in municipal sewage treatment plants – I. Investigations in Germany, Canada and Brazil, Sci. Total Environ. 225 (1999) 81–90. H. Cabana, J.P. Jones, S.N. Agathos, Elimination of endocrine disrupting chemicals using white rot fungi and their lignin modifying enzymes: a review, Eng. Life Sci. 7 (2007) 429–456. L. Lloret, G. Eibes, G. Feijoo, M.T. Moreira, J.M. Lema, F. Hollmann, Immobilization of laccase by encapsulation in a sol–gel matrix and its characterization and use for the removal of estrogens, Biotechnol. Prog. 27 (2011) 1570–1579. Y.-J. Kim, J.A. Nicell, Impact of reaction conditions on the laccase-catalyzed conversion of bisphenol A, Bioresour. Technol. 97 (2006) 1431–1442. S. Kurniawati, J.A. Nicell, Characterization of trametes versicolor laccase for the transformation of aqueous phenol, Bioresour. Technol. 99 (2008) 7825–7834. V. Chea, D. Paolucci-Jeanjean, M.P. Belleville, J. Sanchez, Optimization and characterization of an enzymatic membrane for the degradation of phenolic compounds, Catal. Today 193 (2012) 49–56. S.-y. Okazaki, J. Michizoe, M. Goto, S. Furusaki, H. Wariishi, H. Tanaka, Oxidation of bisphenol A catalyzed by laccase hosted in reversed micelles in organic media, Enzym. Microb. Technol. 31 (2002) 227–232. H. Cabana, J.P. Jones, S.N. Agathos, Preparation and characterization of crosslinked laccase aggregates and their application to the elimination of endocrine disrupting chemicals, J. Biotechnol. 132 (2007) 23–31. T. Fukuda, H. Uchida, M. Suzuki, H. Miyamoto, H. Morinaga, H. Nawata, T. Uwajima, Transformation products of bisphenol A by a recombinant trametes villosa laccase and their estrogenic activity, J. Chem. Technol. Biotechnol. 79 (2004) 1212–1218. N. Durán, M.A. Rosa, A. D'Annibale, L. Gianfreda, Applications of laccases and tyrosinases (phenoloxidases) immobilized on different supports: a review, Enzym. Microb. Technol. 31 (2002) 907–931. X. Chen, S. Dong, Sol–gel-derived titanium oxide/copolymer composite based glucose biosensor, Biosens. Bioelectron. 18 (2003) 999–1004. H. Lu, J. Yang, J.F. Rusling, N. Hu, Vapor-surface sol–gel deposition of titania alternated with protein adsorption for assembly of electroactive, enzymeactive films, Electroanalysis 18 (2006) 379–390. Z. Dai, M. Fang, J. Bao, H. Wang, T. Lu, An amperometric glucose biosensor constructed by immobilizing glucose oxidase on titanium-containing mesoporous composite material of no. 41 modified screen-printed electrodes, Anal. Chim. Acta 591 (2007) 195–199. Q. Li, G. Luo, J. Feng, Q. Zhou, L. Zhang, Y. Zhu, Amperometric detection of glucose with glucose oxidase absorbed on porous nanocrystalline TiO2 film, Electroanalysis 13 (2001) 413–416. Y. Ikeda, Y. Kurokawa, Hydrolysis of 1,2-diacetoxypropane by immobilized lipase on cellulose acetate-TiO2 gel fiber derived from the sol–gel method, J. Sol-Gel Sci. Technol. 21 (2001) 221–226. H. Hatayama, T. Swabe, Y. Kurokawa, Immobilization of urease on composite fibre by using a gel formation of cellulose acetate and titanium iso-propoxide, J. Sol-Gel Sci. Technol. 7 (1996) 13–17. J. Hou, G. Dong, Y. Ye, V. Chen, Laccase immobilization on titania nanoparticles and titania-functionalized membranes, J. Membr. Sci. 452 (2014) 229–240. C. Jolivalt, S. Brenon, E. Caminade, C. Mougin, M. Pontié, Immobilization of laccase from trametes versicolor on a modified PVDF microfiltration membrane: characterization of the grafted support and application in removing a phenylurea pesticide in wastewater, J. Membr. Sci. 180 (2000) 103–113. S. Georgieva, T. Godjevargova, M. Portaccio, M. Lepore, D.G. Mita, Advantages in using non-isothermal bioreactors in bioremediation of water polluted by phenol by means of immobilized laccase from rhus vernicifera, J. Mol. Catal. B: Enzym. 55 (2008) 177–184. K.P. Katuri, S. Venkata Mohan, S. Sridhar, B.R. Pati, P.N. Sarma, Laccasemembrane reactors for decolorization of an acid azo dye in aqueous phase: process optimization, Water Res. 43 (2009) 3647–3658. K. Rasera, J. Ferla, A.J.P. Dillon, R. Riveiros, M. Zeni, Immobilization of laccase from pleurotus sajor-caju in polyamide membranes, Desalination 245 (2009) 657–661. X. Zhao, L. Lv, B. Pan, W. Zhang, S. Zhang, Q. Zhang, Polymer-supported nanocomposites for environmental application: a review, Chem. Eng. J. 170 (2011) 381–394. G. Dong, H. Li, V. Chen, Challenges and opportunities for mixed-matrix membranes for gas separation, J. Mater. Chem. A 1 (2013) 4610–4630. Y. Zhang, P. He, N. Hu, Horseradish peroxidase immobilized in TiO2 nanoparticle films on pyrolytic graphite electrodes: direct electrochemistry and bioelectrocatalysis, Electrochim. Acta 49 (2004) 1981–1988. W. Li, R. Yuan, Y. Chai, L. Zhou, S. Chen, N. Li, Immobilization of horseradish peroxidase on chitosan/silica sol–gel hybrid membranes for the preparation of hydrogen peroxide biosensor, J. Biochem. Biophys. Methods 70 (2008) 830–837. Y. Kurokawa, T. Sano, H. Ohta, Y. Nakagawa, Immobilization of enzyme onto cellulose-titanium oxide composite fiber, Biotechnol. Bioeng. 42 (1993) 394–397. A. Razmjou, J. Mansouri, V. Chen, M. Lim, R. Amal, Titania nanocomposite polyethersulfone ultrafiltration membranes fabricated using a low temperature hydrothermal coating process, J. Membr. Sci. 380 (2011) 98–113. S. Meng, J. Mansouri, Y. Ye, V. Chen, Effect of templating agents on the properties and membrane distillation performance of TiO2-coated PVDF membranes, J. Membr. Sci. 450 (2014) 48–59.

30

J. Hou et al. / Journal of Membrane Science 469 (2014) 19–30

[32] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with the folin phenol reagent, J. Biol. Chem. 193 (1951) 265–275. [33] R. Bourbonnais, M.G. Paice, Oxidation of non-phenolic substrates: an expanded role for laccase in lignin biodegradation, FEBS Lett. 267 (1990) 99–102. [34] C.J. Brinker, G.C. Frye, A.J. Hurd, C.S. Ashley, Fundamentals of sol–gel dip coating, Thin Solid Films 201 (1991) 97–108. [35] Y. Sheng, L. Liang, Y. Xu, D. Wu, Y. Sun, Low-temperature deposition of the high-performance anatase-titania optical films via a modified sol–gel route, Opt. Mater. 30 (2008) 1310–1315. [36] D. Zhang, T. Yoshida, H. Minoura, Low-temperature fabrication of efficient porous titania photoelectrodes by hydrothermal crystallization at the solid/ gas interface, Adv. Mater. 15 (2003) 814–817. [37] M. Roberts, A.F. Huang, P. Johns, J. Owen, Dip–spin coating of reticulated vitreous carbon with composite materials to act as an electrode for 3D microstructured lithium ion batteries, J. Power Sources 224 (2013) 250–259. [38] C.M. Malengreaux, A. Timmermans, S.L. Pirard, S.D. Lambert, J.-P. Pirard, D. Poelman, B. Heinrichs, Optimized deposition of TiO2 thin films produced by a non-aqueous sol–gel method and quantification of their photocatalytic activity, Chem. Eng. J. 195–196 (2012) 347–358. [39] S.M Attia, J. Wang, G. Wu, J. Shen, J. Ma, Nanostructure study of TiO2 films prepared by dip coating process, J. Mater. Sci. Technol. 18 (2002) 31–33. [40] E.V. Rebrov, J.C. Schouten, Limiting withdrawal rate and maximum film thickness during dip-coating of titania sols onto a Si substrate, Chem. Eng. Process. 50 (2011) 1063–1068. [41] J.-W. Lee, C.-W. Won, B.-S. Chun, H. Sohn, Dip coating of alumina films by the sol–gel method, J. Mater. Res. 8 (1993) 3151–3157. [42] T. Bertrand, C. Jolivalt, P. Briozzo, E. Caminade, N. Joly, C. Madzak, C. Mougin, Crystal structure of a four-copper laccase complexed with an arylamine: insights into substrate recognition and correlation with kinetics, Biochemistry 41 (2002) 7325–7333. [43] N. Carlsson, H. Gustafsson, C. Thörn, L. Olsson, K. Holmberg, B. Åkerman, Enzymes immobilized in mesoporous silica: a physical–chemical perspective, Adv. Colloid Interface Sci. 205 (2014) 339–360. [44] S.-i. Matsuura, R. Ishii, T. Itoh, S. Hamakawa, T. Tsunoda, T. Hanaoka, F. Mizukami, Immobilization of enzyme-encapsulated nanoporous material in a microreactor and reaction analysis, Chem. Eng. J. 167 (2011) 744–749. [45] J.N. Talbert, J.M. Goddard, Enzymes on material surfaces, Colloids Surf. B: Biointerfaces 93 (2012) 8–19. [46] J.-H. Yang, Y.-S. Han, J.-H. Choy, TiO2 thin-films on polymer substrates and their photocatalytic activity, Thin Solid Films 495 (2006) 266–271. [47] A. Rekuć, J. Bryjak, K. Szymańska, A.B. Jarzębski, Very stable silica-gel-bound laccase biocatalysts for the selective oxidation in continuous systems, Bioresour. Technol. 101 (2010) 2076–2083. [48] G. Bayramoğlu, M. Yakup Arıca, Immobilization of laccase onto poly(glycidylmethacrylate) brush grafted poly(hydroxyethylmethacrylate) films: enzymatic oxidation of phenolic compounds, Mater. Sci. Eng. C 29 (2009) 1990–1997. [49] Y. Liu, Z. Zeng, G. Zeng, L. Tang, Y. Pang, Z. Li, C. Liu, X. Lei, M. Wu, P. Ren, Z. Liu, M. Chen, G. Xie, Immobilization of laccase on magnetic bimodal mesoporous carbon and the application in the removal of phenolic compounds, Bioresour. Technol. 115 (2012) 21–26. [50] N. Ehlert, P.P. Müller, M. Stieve, P. Behrens, Immobilization of alkaline phosphatase on modified silica coatings, Microporous Mesoporous Mater. 131 (2010) 51–57. [51] L.F. Bautista, G. Morales, R. Sanz, Immobilization strategies for laccase from trametes versicolor on mesostructured silica materials and the application to the degradation of naphthalene, Bioresour. Technol. 101 (2010) 8541–8548. [52] Z. Wang, X. Liu, M. Lv, J. Meng, Simple synthesis of magnetic mesoporous FeNi/carbon composites with a large capacity for the immobilization of biomolecules, Carbon 48 (2010) 3182–3189.

[53] Y. Wang, F. Caruso, Mesoporous silica spheres as supports for enzyme immobilization and encapsulation, Chem. Mater. 17 (2005) 953–961. [54] P.H. Pandya, R.V. Jasra, B.L. Newalkar, P.N. Bhatt, Studies on the activity and stability of immobilized α-amylase in ordered mesoporous silicas, Microporous Mesoporous Mater. 77 (2005) 67–77. [55] F. Wang, C. Guo, L. Yang, C. Liu, Magnetic mesoporous silica nanoparticles: fabrication and their laccase immobilization performance, Bioresour. Technol. 101 (2010) 8931–8935. [56] Y. Zhu, S. Kaskel, J. Shi, T. Wage, K.-H. van Pée, Immobilization of trametes versicolor laccase on magnetically separable mesoporous silica spheres, Chem. Mater. 19 (2007) 6408–6413. [57] S. Georgieva, T. Godjevargova, D.G. Mita, N. Diano, C. Menale, C. Nicolucci, C.R. Carratelli, L. Mita, E. Golovinsky, Non-isothermal bioremediation of waters polluted by phenol and some of its derivatives by laccase covalently immobilized on polypropylene membranes, J. Mol. Catal. B: Enzym. 66 (2010) 210–218. [58] S. Yüksel, N. Kabay, M. Yüksel, Removal of bisphenol A (BPA) from water by various nanofiltration (NF) and reverse osmosis (RO) membranes, J. Hazard. Mater. 263 (2013) 307–310. [59] M.N. Kathiravan, R. Karthiga Rani, R. Karthick, K. Muthukumar, Mass transfer studies on the reduction of Cr(VI) using calcium alginate immobilized bacillus sp. in packed bed reactor, Bioresour. Technol. 101 (2010) 853–858. [60] Y. Ding, Y. Sun, Small-sized dense magnetic pellicular support for magnetically stabilized fluidized bed adsorption of protein, Chem. Eng. Sci. 60 (2005) 917–924. [61] F. Wang, Y. Hu, C. Guo, W. Huang, C. Liu, Enhanced phenol degradation in coking wastewater by immobilized laccase on magnetic mesoporous silica nanoparticles in a magnetically stabilized fluidized bed, Bioresour. Technol. 110 (2012) 120–124. [62] C. Menale, C. Nicolucci, M. Catapane, S. Rossi, U. Bencivenga, D.G. Mita, N. Diano, Optimization of operational conditions for biodegradation of chlorophenols by laccase-polyacrilonitrile beads system, J. Mol. Catal. B: Enzym. 78 (2012) 38–44. [63] K.O. Agenson, J.-I. Oh, T. Urase, Retention of a wide variety of organic pollutants by different nanofiltration/reverse osmosis membranes: controlling parameters of process, J. Membr. Sci. 225 (2003) 91–103. [64] Y. Chen, V. Lykourinou, T. Hoang, L.-J. Ming, S. Ma, Size-selective biocatalysis of myoglobin immobilized into a mesoporous metal–organic framework with hierarchical pore sizes, Inorg. Chem. 51 (2012) 9156–9158. [65] C. Nicolucci, S. Rossi, C. Menale, T. Godjevargova, Y. Ivanov, M. Bianco, L. Mita, U. Bencivenga, D.G. Mita, N. Diano, Biodegradation of bisphenols with immobilized laccase or tyrosinase on polyacrylonitrile beads, Biodegradation 22 (2011) 673–683. [66] G. Songulashvili, G.A. Jimenéz-Tobón, C. Jaspers, M.J. Penninckx, Immobilized laccase of Cerrena unicolor for elimination of endocrine disruptor micropollutants, Fungal Biol. 116 (2012) 883–889. [67] H. Cabana, C. Alexandre, S.N. Agathos, J.P. Jones, Immobilization of laccase from the white rot fungus Coriolopsis polyzona and use of the immobilized biocatalyst for the continuous elimination of endocrine disrupting chemicals, Bioresour. Technol. 100 (2009) 3447–3458. [68] H. Cabana, J.P. Jones, S.N. Agathos, Utilization of cross-linked laccase aggregates in a perfusion basket reactor for the continuous elimination of endocrine-disrupting chemicals, Biotechnol. Bioeng. 102 (2009) 1582–1592. [69] X. Li, P. Yu, L. Yang, F. Wang, L. Mao, An electrochemical method for investigation of conformational flexibility of active sites of trametes versicolor laccase based on sensitive determination of copper ion with cysteine-modified electrodes, Anal. Chem. 84 (2012) 9416–9421.